3D-MSCs apoptotic bodies-integrated conductive hydrogel mitigates spinal cord injury via immunoregulation and alleviating neuronal pyroptosis

Abstract

Spinal cord injury (SCI) is a severe condition that affects the central nervous system. Current clinical interventions are insufficient to address the profound consequences of SCI, highlighting the great need for alternative treatment options. In recent years, the use of hydrogels to locally deliver extracellular vesicles (EVs) has become a potential method for SCI therapy, conductive hydrogels can also promote SCI repair by establishing functional connections between neurons. However, increasing the bioactivity of EVs and optimizing hydrogel-based therapies remain significant challenges. In this study, MXene nanosheets endow composite hydrogels with excellent conductivity. Bone marrow mesenchymal stem cells (BMSCs) were grown in 3D suspensions to form spheroids, after which apoptosis was induced to isolate apoptotic bodies (ABs). Then 3D-ABs were incorporated into a Gelatine methacryloyl (GelMA) hydrogel containing MXene nanosheets for their sustained release in vivo. Furthermore, the composite hydrogel provides mechanical support and mimics the electrical transmission properties of neurons. After local injection into SCI mice, the composite hydrogel effectively filled the lesion cavity, promoted the reconstruction of functional neural connections, suppressed neuroinflammation and alleviated neuronal pyroptosis owing to its conductive components and apoptotic vesicles. This novel injectable composite hydrogel represents a promising therapeutic option for SCI repair.

Highlights

• •Pioneering integration of 3D MSCs-derived ABs into hydrogel for spinal cord injury (SCI) repair.
• •MXene nanosheets promote the reconstruction of functional neural connections.
• •AMG hydrogel fosters functional restoration after traumatic SCI.
• •AMG hydrogel suppresses the inflammatory microenvironment and inhibits pyroptosis in the acute phase after SCI.
• •AMG hydrogel promotes axonal penetration and remyelination in the chronic phase after SCI.

1. Introduction

Spinal cord injury (SCI) is a condition of the central nervous system with a low likelihood of full recovery. After being injured, neurons cannot divide and accumulate harmful metabolic byproducts, making the loss of neural function almost irreversible [1]. Notably, in 2019, SCI affected 20.6 million people worldwide, with an incidence rate of 0.9 million and 6.2 million years lived with disability, reflecting a sharp increase since 1990 [2]. Complex pathological and biochemical events occur after SCI, with the inflammatory microenvironment and secondary neuronal death being key factors in determining prognosis [3,4].

The inflammatory response following SCI is not only a consequence of injury but also a major driving force in its progression [5]. Inflammation involves not only neurons but also glial cells, including microglia, astrocytes, and oligodendrocytes. Microglia, in particular, play a dual role in neuroinflammation [6,7]. During the progression of SCI, microglia become activated, and signalling molecules, such as glutamate, high-mobility group box 1 protein (HMGB1), and nucleotides, are secreted from damaged neurons. Once activated, microglia tend to polarize towards the M1 phenotype to release proinflammatory cytokines, reactive oxygen species (ROS), and other neuroinflammatory mediators. However, when microglia shift towards the classical M2 phenotype, they secrete anti-inflammatory cytokines, including tumor necrosis factor (TNF), IL-10, and IL-4 [[8], [9], [10], [11]]. Therefore, promoting the polarization of microglia from the M1 phenotype to the M2 phenotype while suppressing excessive neuroinflammation is considered a key strategy for treating SCI. Among the various factors that influence the inflammatory microenvironment, the excessive production of ROS overwhelms the body's natural antioxidant defences, leading to oxidative stress [12,13]. As a result of this imbalance, the spinal cord becomes highly sensitive to oxidative damage, and reducing oxidative stress can further improve the postinjury inflammatory microenvironment. Exacerbated inflammation after injury intensifies neuronal death and structural damage, leading to progressive neurological dysfunction. As a recently identified mode of inflammatory cell death, pyroptosis has been detected in multiple tissues, such as the spinal cord and brain [14]. The classic pyroptosis pathway involves the cleavage of Gasdermin D (GSDMD) by activated caspase-1, which produces the active fragment N-terminal GSDMD (GSDMD-N). This fragment causes membrane rupture, forms the NOD-like receptor thermal protein domain associated protein 3 (NLRP3) inflammasome, and drives inflammatory cytokine release, further exacerbating inflammation [[15], [16], [17]]. One major cause of neurological impairment after SCI is neuronal pyroptosis; therefore, controlling this process can significantly improve therapeutic outcomes.

The current standard clinical treatment for SCI includes decompressive surgery to reduce spinal cord pressure and restore spinal stability combined with high-dose intravenous administration of methylprednisolone sodium succinate (MPSS) during the acute phase [18]. However, surgical intervention can only prevent further secondary damage by alleviating mechanical compression, and corticosteroid therapy is associated with serious complications [19]. To date, there is still no clinically proven method that effectively promotes neurological recovery. Although cell transplantation has shown potential in supporting neural regeneration, challenges such as immune rejection and short-term efficacy remain [20]. Given these limitations, developing effective biomaterials holds significant promise for advancing SCI treatment. Current biomaterial options face specific limitations. While nanoparticles and liposomes are incapable of providing the structural support necessary for guided axonal regeneration, pre-formed polymer scaffolds [21], on the other hand, pose challenges due to their invasive implantation, poor conformability to irregular injury cavities, and excessively high mechanical stiffness [22]. These drawbacks highlight the need for an alternative biomaterial strategy. Hydrogels can simulate the physiological environment of the extracellular matrix, exhibit excellent biocompatibility, and serve as structural materials to fill cavities, thereby promoting nerve regeneration and functional recovery. Gelatine methacryloyl (GelMA) is a biocompatible and easily shaped material. The properties of its key ingredient, gelatine, mimic those of the extracellular matrix (ECM), and the stiffness of GelMA hydrogels can be tuned for various applications by adjusting the component ratios [23,24]. GelMA has a 3D porous network microstructure, which not only facilitates localized drug delivery but releases drugs in a more gradual manner [25]. The conversion and transmission of physiological information between cells is accomplished primarily through electrophysiological signals [26,27]. Conductive hydrogels can mimic the physiological environment of electrically active tissues, regulating cell adhesion and growth. MXenes, which are 2D transition metal carbides or nitrides, offer excellent electrical conductivity [28,29]. Additionally, MXenes also exhibit good biodegradability, allowing them to be broken down in and eliminated from the body [30]. Therefore, we designed a conductive hydrogel, MXene@GelMA, for tissue repair in the injured spinal cord.

Bone marrow stem cell (BMSC) therapy has recently gained attention for its strong immunomodulatory effects and has been widely used in disease models such as SCI and focal cerebral ischaemia [31]. Therefore, we chose BMSCs as the primary focus of our research. Although BMSC transplantation is effective for treating SCI, the inability of stem cells to remain at the injury site for an extended period and their limited differentiation capabilities necessitate the identification of better ways to harness the therapeutic potential of BMSCs [32]. Research indicates that transplanted stem cells rapidly undergo apoptosis after injection and form apoptotic bodies (ABs), the primary entities responsible for the therapeutic effects of these cells [[33], [34], [35]]. The ABs produced during apoptosis range from 0.5 to 5 μm in size and are involved in the exchange of bioactive molecules between cells [36,37]. Currently, ABs derived from BMSCs have been extensively employed in therapeutic interventions for a variety of diseases [38]. Since their contents largely mirror the those of their parent cells [39], modulating the cellular microenvironment can alter the protein and miRNA contents of ABs.

Many types of cells can form 3D spheroids in suspension or nonadhesive environments. Compared with traditional cell culture methods, 3D culture techniques offer unique advantages in terms of structure and cell‒cell interactions [[40], [41], [42], [43]]. In a 3D culture environment, changes in the mechanical and physical structures alter stem cell proliferation, differentiation, and paracrine signalling, resulting in enhanced bioactivity, which may be due to changes in cellular components such as miRNAs [44,45]. Recent studies have shown that BMSCs exhibit more potent bioactivity when cultured in a 3D environment [[46], [47], [48]]. Studies have shown that 3D suspension cultures can alter the therapeutic efficacy of ABs against diseases by influencing the composition of miRNAs within them [44]. Therefore, we utilized 3D suspension cultures of BMSCs to increase the biological activity of ABs and promote SCI repair.

In this study, we designed GelMA hydrogels loaded with MXene and 3D stem cell-derived ABs for injured spinal cord repair (Fig. 1), and further validate its therapeutic mechanism for SCI through transcriptome sequencing. Compared with traditional invasive treatments (scar removal and local drug injection), this approach not only reduces unnecessary trauma but also allows for controlled drug release to exert therapeutic effects. The hydrogel improved neurological function after SCI by alleviating oxidative stress, promoting macrophage M2 polarization, and inhibiting neuronal pyroptosis, offering a novel strategy for SCI therapy.

Fig. 1.

The diagram of the design and fabrication of delivery system loading BMSCs derived 3D-ABs and Mxene nanosheets for the repair of SCI. This composite hydrogel alleviated the inflammatory microenvironment in the acute phase and promoted axon regeneration in the chronic phase.

2. Results

2.1. Preparation and characterization of the AMG hydrogel

In this study, suspension culture was used to produce 3D-BMSC spheroids (Fig. 2A). Flow cytometry were used showing that BMSCs expressed high levels of the surface markers such as CD29, CD44, CD73, CD90, and CD105 (Fig. S2A-E, Supporting Information), and the BMSC-specific markers CD29 and CD44 were used to stain both 2D-BMSC and 3D-BMSC spheroids to confirm the formation of spheroids with good biological activity [49,50] (Fig. 2B). After apoptosis was induced with STS, 3D-ABs were isolated from 3D-BMSC spheroids using standard ultracentrifugation techniques. Dynamic light scattering (DLS) analysis showed that the 3D-ABs had an average size of approximately 700 nm (Fig. 2E), while scanning electron microscopy (SEM) revealed that the apoptotic vesicles exhibited a characteristic morphology (Fig. 2C). Immunofluorescence staining of 3D-ABs using the apoptotic vesicle marker C1QC confirmed their identity (Fig. 2D). WB analysis was performed to characterize the apoptotic vesicles before and after incorporation into the composite hydrogel (Fig. 2F), and the apoptotic vesicle markers histone H3 (H3), histone 2B (H2B), C1QC, and C3B were detected. Key factors differing between 2D-ABs and 3D-ABs also analysed, highlighting the superior therapeutic potential of 3D-ABs (Fig. S3A-B, Supporting Information). Thus, the 3D-ABs were confirmed to have been internalized within the cellular structures. The spinal cord contains various cell types, among which microglia and neurons are the most closely related to SCI. Microglia are important in the inflammatory response, whereas neurons determine prognosis after injury. Therefore, it is critical that ABs be efficiently internalized by microglia and neurons to exert therapeutic effects. We used PC12 cells as a neuronal model and BV2 cells as a microglial model to investigate ABs internalization. To validate this finding, we extended our investigation to primary neurons (Fig. S4A, Supporting Information) and macrophages. The results confirmed the successful phagocytosis of Dil-stained 3D-ABs in these primary cells (Fig. S4B-C, Supporting Information). In vitro, Dil-labelled 3D-ABs (20 μg/mL) were added to PC12 and BV2 cell cultures, and cellular uptake was assessed 24 h later using flow cytometry (FCM). Approximately 97.8% of the PC12 cells and 95.4% of the BV2 cells exhibited positive uptake of 3D-ABs (Fig. 2G). Furthermore, 3D-ABs uptake was assessed in vitro and in vivo using immunofluorescence staining. In vitro, F-actin was applied to label PC12 cells and CD68 was applied to label BV2 cells. We observed apparent internalization of the Dil-labelled 3D-ABs via fluorescence microscopy. In vivo, 3D-ABs were injected into the injured spinal cord, and frozen sections of spinal cord tissue was obtained three days postinjury. Under a microscope, numerous Dil-positive 3D-ABs were internalized by neurons and microglia. Moreover, we confirmed that the 3D-ABs released from the AMG composite hydrogel exhibited the same cellular uptake properties (Fig. 2H, I). The AMG-ABs retained biological effects comparable to those of free the 3D-ABs both in vitro and in vivo.

Fig. 2.

Preparation and characterization of 3D-ABs and Mxene nanosheets. (A) Graphical illustration of the 3D-ABs isolation procedure. (B) 2D-BMSCs and 3D-BMSCs stained by CD29 and CD44 (scale bar, 40 μm). (C) SEM images of 3D-ABs (scale bar, 1 μm). (D) 3D-ABs stained by CIQC (scale bar, 10 μm). (E) DLS analysis for the 3D-ABs. (F) Expression of biomarkers of 3D-ABs and Hydrogel-3D-ABs including C1QC, C3B, H2B, and H3, β-actin was utilized as a loading control. (G) FCM analysis of the percentage of Dil-positive PC12 and BV2 cells after treatment with DiI-labelled 3D-ABs. (H) Uptake of Dil-labelled 3D-ABs and Hydrogel-3D-ABs by PC12 and BV2 cells (scale bar, 20 μm). (I) Frozen sections of DiI-labelled 3D-ABs and Hydrogel-3D-ABs treated spinal cord were stained for Neun and CD68 (scale bars, 20 μm). (J) SEM and EDS elemental mapping images of the Mxene nanosheets (scale bar, 20 μm). (K) Live/dead cell staining images for PC12 and BV2 cells after different treatments (scale bar, 200 μm).

As widely studied 2D nanomaterials, MXenes have garnered significant attention in regenerative medicine and tissue engineering because of their exceptional electrical conductivity and outstanding biocompatibility. Following our team's previously established synthesis method, we obtained MXenes by etching the aluminium layer from the Ti₃AlC₂ powder [51]. The morphology and structure of the MXene were observed using SEM, revealing its thin, transparent 2D nanosheet-like structure with lateral dimensions reaching several hundred nanometres and a thickness ranging from a monolayer to a few layers (Fig. 2J). Then, energy-dispersive spectroscopy (EDS) elemental mapping confirmed the uniform distributions of C, F, O, and Ti on the characteristic layered structure surface (Fig. 2J). Furthermore, subsequent coculture experiments demonstrated its excellent biocompatibility (Fig. 2K). According to previous studies, MXene most effectively promotes injury recovery at a concentration of 200 μg/mL [30].

GelMA is a synthetically modified composite material with excellent biocompatibility. Gelatine, the main component of GelMA, has biological properties similar to those of the extracellular matrix (ECM). Under specific conditions, methacrylic anhydride (MA) reacts with gelatine, substituting methacryloyl groups for amino and hydroxyl groups, increasing the reactivity of the product and enabling the formation of new linkages between gelatine chains [52]. As the molecular weight increases, these crosslinks facilitate the transition of the material into a gel state. In this study, we chose GelMA as the drug delivery platform. Before crosslinking, MXene nanosheets were added to the GelMA precursor solution at a concentration of 200 μg/ml to prepare the MG hydrogel. Next, the optimal concentration of 3D-ABs in the composite hydrogel was determined to be 1.25 mg/ml by monitoring the effects of varying concentrations of 3D-ABs on motor function recovery in SCI mice (Fig. S1A, Supporting Information). 3D-ABs were incorporated into GelMA at a concentration of 1.25 mg/ml to generate the AG hydrogel, and both the vesicles and nanosheets were incorporated into the GelMA solution to obtain the AMG composite hydrogel (Fig. 3A). The final hydrogels were prepared for subsequent experiments after UV light-mediated crosslinking. The three composite hydrogels were observed using SEM, revealing their well-structured porous morphology (Fig. 3B). The elemental composition of the AMG hydrogel, which was analysed by energy dispersive X-ray spectroscopy (EDS), included the elements C, N, O, and Ti, indicating that the MXene nanosheets were dispersed within the hydrogel (Fig. 3C). To examine the condition of the vesicles contained within the hydrogel, a visualization technique was employed. 3D-ABs were stained with Dil and added to the hydrogel, and 3D fluorescence imaging revealed the uniform dispersion of vesicles within the hydrogel (Fig. 3D). Owing to the excellent conductivity of MXene nanosheets, we further investigated the electrical properties of the three composite hydrogels. Electrochemical characterization was performed using a potentiostat with hydrogel-coated ITO glass as the working electrode. I‒V curves were measured from −0.5 V to 0.5 V using a dual-probe Keithley 2400 source meter (Fig. 3E). The conductivities of the AG, MG, and AMG hydrogels were 1.10033 S/m, 1.54767 S/m, and 1.57061 S/m, respectively (Fig. 3F). MXene nanosheet addition significantly increased the conductivity of the hydrogel, and the inclusion of vesicles did not alter the conductivity. Rheological shear rate scanning tests were conducted at 24–26 °C and shear rates between 1 and 100 RAD/s to assess the physical characteristics of the composite hydrogels. For all the hydrogels, the storage modulus (elastic modulus, G′) exceeded the loss modulus (viscous modulus, G″) over the shear rate range of 1–100 RAD/s, indicating good stability (Fig. 3H). Notably, the storage modulus of the hydrogels decreased after the addition of MXene nanosheets. For compression testing, the hydrogels were crosslinked under 405 nm UV light for 60 s and compressed at a rate of 1 mm/min. Compared with hydrogels containing vesicles only, the addition of MXene decreased the compression modulus (Fig. 3G). On the basis of the swelling curves of the different hydrogels, the addition of MXene nanosheets led to an increase in the swelling ratio (Fig. 3I). These changes in physical properties could be attributed to the opacity of the MXene nanosheets, which may have affected the light-induced crosslinking of GelMA. An investigation was conducted to assess the rate of vesicle release from the composite hydrogels. The amount of ABs released from the materials was measured daily (μg). The release experiment was performed for six days, and more than 90% of the ABs were released by the end of the sixth day (Fig. 3J). The degradation rates of the different hydrogels varied, and the addition of MXene slightly increased the degradation rate (Fig. 3K). Hydrogel conductivity declines with progressive degradation (Fig. S5C, Supporting Information). Mercury intrusion porosimetry measurements confirmed favorable porosity and pore dimensions in the hydrogels, providing adequate space for axonal ingrowth (Fig. S5A-B, Supporting Information).

Fig. 3.

Characterization of the AMG hydrogel. (A) Photographs displaying the sol-gel transformation of the hydrogel. (B) SEM images of AG, MG, and AMG hydrogel (scale bar, 200 μm). (C) EDS elemental mapping images of the AMG hydrogel (scale bar, 100 μm). (D) Fluorescence images of 3D-ABs uniformly dispersed in the hydrogel. (E) I-V Characteristic Curves of different hydrogels. (F) Comparison of the electrical conductivity of the different hydrogels (n = 3). (G) Compressive stress-strain curves and (H) rheological test results for the different hydrogels. (I) Swelling ratio of the different hydrogels (n = 3). (J) Cumulative release of 3D-ABs of composite hydrogel over 6 days and daily release curve of composite hydrogel (n = 3). (K) Degradation of the different hydrogel in vitro (n = 3). (L) Biocompatibility of the composite hydrogel in vitro evaluated by live/dead staining (scale bar, 200 μm).

2.2. Biocompatibility of the AMG hydrogel

Biocompatibility is critical, especially for materials that enter the delicate tissue of the spinal cord. To evaluate the in vitro cytotoxicity of the hydrogel, we cocultured it with PC12 and BV2 cells, both of which maintained good viability (Fig. 3L). The hydrogel was then co-cultured with primary neurons and macrophages, with live/dead staining confirming its favorable biocompatibility (Fig. S4D-E, Supporting Information). Additionally, we injected the hydrogel subcutaneously into the dorsal regions of mice, composite hydrogel groups showed no significant differences in body weight or hepatic function parameters compared to control groups (Fig. S1B, D, E), and histological analysis confirmed that there were no significant pathological changes in the major organs (heart, liver, lungs, kidneys, or spleen) of the mice in all the groups (Fig. S1C).

2.3. AMG hydrogel effectively improves histological and functional performance in mice after SCI

Functional assessments and histological changes were evaluated 28 days after SCI to determine the therapeutic effect of the implanted hydrogel. To assess cavitation caused by inflammation and neuronal cell death after injury, HE and Masson staining were performed on spinal cord tissue. The AMG group exhibited minimal cavitation, indicating that the hydrogel containing MXenes and ABs from 3D-BMSCs effectively mitigated secondary damage (Fig. 4A). Additionally, to further explore the role of the hydrogels in motor function recovery, footprint analysis and Basso Mouse Scale (BMS) scoring were conducted (Fig. 4D). On Days 21 and 28, the BMS scores were greater in the hydrogel-treated SCI group than in the SCI group, and the AMG group demonstrated the most notable improvement. Footprint analysis performed on Day 28 revealed significant improvements in hindlimb function in the hydrogel-implanted groups, but no hindlimb motion was noted in the SCI group (Fig. 4B). Footprint analysis of stride length revealed a similar trend. (Fig. 4F). Furthermore, motor-evoked potentials (MEPs) were recorded on Day 28 postsurgery and analysed to examine electrophysiological recovery (Fig. 4E). The results demonstrated that the amplitude was significantly lower in the SCI group than in the control group. Improvements in amplitude were observed in all hydrogel-implanted groups, with the AMG group showing the most significant increase. We then evaluated the extent of right lower limb muscle atrophy by staining the gastrocnemius muscles with laminin and measuring the average muscle fibre cross-sectional area. The results revealed the smallest cross-sectional area in the SCI group. In contrast, the hydrogel-implanted groups presented significantly larger cross-sectional areas than did the SCI group, further demonstrating the protective effects of the hydrogel (Fig. 4C, G). A more detailed VCR motion video analysis revealed that the ankle, knee, and hip joints in the AMG group could move to a certain extent, and the support performance of the hind limbs and the rhythm of the motion significantly improved (Fig. 4H-M).

Fig. 4.

Composite hydrogel promotes post-SCI functional recovery. (A) H&E and Masson staining of the SCI areas of 28 days (scale bar, 1000 μm). (B) On day 28 after SCI, images of mouse footprints were captured. Blue: fore paw print; Red: hind paw print. (C) Representative images of laminin staining on the gastrocnemius muscles to mark the area of muscle fibers (scale bar, 50 μm). (D) Following SCI, the groups' BMS scores were recorded on days 1, 3, 7, 14, 21, and 28. (E) MEP results of mice in each group at 28 dpi. (F) Stride length (mm) analyses of mice at 28 dpi. (G) Quantitative analysis of the cross-sectional area of gastrocnemius muscle fibers in the right hind limbs of mice in each group(n = 9, 3 animals per group with 3 muscle fibers counted for each animal). (H) Schematic diagram of marked joints in the hindlimb. (I) Representative color-coded bar plots of hindlimb movements in each group. (J) Curves depicting angle variations in hip, knee, and ankle joints during hindlimb motion across groups. (K–M) Quantitative analysis of swing frequency, maximum muscle tension, and cycle duration during a single-step cycle in hindlimbs of mice from each group. The data are presented as the means ± SEMs (n = 3); ∗p < 0.05, indicates significant differences; ns, is not significant. Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test.

2.4. Mechanisms of the AMG hydrogel

To explore the underlying mechanism of the AMG hydrogel, transcriptome sequencing of the SCI and AMG groups was carried out. Among the genes in the AMG group and the SCI group, 253 were differentially expressed (175 were upregulated, and 78 were downregulated) (p < 0.05) (Fig. 5A, B). The bar chart of the enriched KEGG pathways revealed that neuroactive ligand–receptor interactions, the PI3K/AKT signalling pathway, the MAPK signalling pathway, and the neurotrophin signalling pathway were enriched in these genes (Fig. 5C). Furthermore, Western blotting indicated that the p-PI3K and p-AKT levels were higher among the AMG group. This result indicated that the mechanism by which the AMG hydrogel regulates tissue repair after SCI is related to activation of the PI3K/AKT pathway (Fig. 5F, G, H). Additionally, according to the GO analysis, the DEGs were involved primarily in oxidative stress, axonogenesis, pyroptosis, neuron projection, and the inflammatory response. The transcriptome sequencing results revealed that the AMG hydrogel may enhance axonogenesis while inhibiting oxidative stress, pyroptosis, and inflammation, thereby promoting functional recovery after SCI (Fig. 5D, E). To validate our transcriptomic findings, a subset of up- and down-regulated genes was verified using qPCR (Fig. S7B, Supporting Information).

Fig. 5.

Validation of the potential mechanism of AMG composite hydrogel in treating SCI. (A) Heatmap of upregulated and downregulated genes in the spinal cords of mice treated with AMG hydrogel for 3 days (n = 4 mice per group). (B) Volcano plot of DEGs in the SCI and SCI + AMG groups. (C) Transcriptome KEGG analysis. (D, E) Transcriptome GO analysis. (F) Representative western blots showing the expression of PI3K, p-PI3K, AKT, and p-AKT, GAPDH were utilized as a loading control. (G, H) The optical density values for p-PI3K and p-AKT were measured and evaluated in all groups. The data are presented as the means ± SEMs (n = 3 mice per group); ∗p < 0.05, indicates significant differences; ns, is not significant. Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test.

2.5. AMG hydrogel inhibiting oxidative stress in vitro and in vivo

Recent research has demonstrated a strong relationship between oxidative stress and the progression of inflammation. Excessive ROS production exacerbates the inflammatory microenvironment, causing cellular and tissue damage. To examine the effect of the composite hydrogel on inflammation associated with oxidative stress, we simulated the postinjury microenvironment by treating PC12 and BV2 cells with TBHP and coculturing them with the composite hydrogel. ROS levels were assessed by staining with DCFH-DA, which revealed a reduction in fluorescence intensity in the hydrogel-treated groups, with the AMG group showing the most significant decrease (Fig. 6A, C, D). Furthermore, spinal cord sections were prepared from samples collected on Day 3 post-SCI and stained with DHE to measure the ROS levels in the tissue. The hydrogel-treated groups presented substantially reduced ROS levels, with the AMG group showing the most pronounced decrease (Fig. 6B, E). These findings demonstrate that the composite hydrogel effectively mitigated the inflammatory microenvironment following SCI both in vitro and in vivo. Additionally, GSEA provided additional evidence that the AMG composite hydrogel alleviated oxidative stress (Fig. 6F). Then we conducted a rescue assay comparing the SCI group, SCI + LY294002 group, SCI + AMG group, and SCI + AMG/LY294002 group to further determine the relationship between inflammation and PI3K/AKT signaling pathway in the SCI model treated with AMG hydrogel. Western blot analysis of p-PI3K and p-AKT levels in spinal cord tissues from different experimental groups confirmed successful inhibition of the PI3K/AKT pathway by the specific inhibitor LY294002 (Fig. S8A-C, Supporting Information). Subsequent DHE staining of spinal cord cryosections at 3 days post-injury revealed that PI3K/AKT pathway inhibition exacerbated the inflammatory microenvironment, as evidenced by comparing the SCI group with the SCI + LY294002 group. This effect was consistently observed when comparing the AMG group with the AMG + LY294002 group (Fig. S8D-E, Supporting Information). These results collectively demonstrate that the composite hydrogel suppresses the inflammatory microenvironment following SCI through activation of the PI3K/AKT pathway. We also evaluated the long-term functional recovery in different groups. First, HE and Masson staining were performed to examine the glial scar areas across the four groups (Fig. S9A, Supporting Information). The results revealed a larger glial scar area in the AMG + LY294002 group compared to the AMG group. Laminin staining and MEP was used to assess the degree of gastrocnemius muscle fiber atrophy in the affected limbs, showing more pronounced muscle atrophy in the AMG + LY294002 group (Fig. S9C, E, G, Supporting Information). Subsequently, footprint analysis and BMS score were employed to evaluate motor functional recovery (Fig. S9B, D, F, Supporting Information). Mice treated with the PI3K inhibitor exhibited poorer motor function recovery, and this finding was further supported by analysis of locomotion videos from the different groups (Fig. S9H-L, Supporting Information).

Fig. 6.

Composite hydrogel inhibits oxidative damage in vitro and in vivo. (A) DCFH-DA staining of PC12 and BV2 cells treated with different hydrogel under TBHP treatment (scale bar, 200 μm). (B) Frozen sections of different hydrogel-treated spinal cords were stained for DHE (scale bar: 50 μm). (C, D) DCFH-DA integrated intensity analysis of PC12 and BV2 cells (n = 3). (E) Analysis and quantification of DHE integrated intensity in each group (n = 3). (F) Transcriptome GSEA analysis. The data are presented as the means ± SEMs (n = 3); ∗p < 0.05, indicates significant differences; ns, is not significant. Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test.

2.6. AMG hydrogel promotes macrophage polarization and reduces neuronal pyroptosis In Vitro

The present study investigated the mechanisms by which the composite hydrogel influences macrophage polarization in vitro. BV2 cells were cocultured with the hydrogel under inflammatory conditions via stimulation with TBHP. WB analysis confirmed that the composite hydrogel effectively reduced TBHP-induced iNOS upregulation while increasing Arg-1 expression (Fig. 7A, B). To assess M1 and M2 macrophage polarization, immunofluorescence staining was used. Following TBHP stimulation, the expression of inducible nitric oxide synthase (iNOS), a marker of M1 macrophages, significantly increased. Compared with TBHP treatment, hydrogel treatment led to decreased iNOS expression, most notably in the AMG group (Fig. 7C, E). Following TBHP stimulation, the level of Arginase-1 (Arg-1), a marker of M2 macrophages, remained high and further increased in the hydrogel-treated groups, with the AMG group showing the most pronounced increase (Fig. 7D, F). Under identical conditions, primary macrophages were treated to assess the therapeutic effects of the hydrogel via western blot and immunofluorescence staining. The results demonstrated that the hydrogel treatment group significantly promoted macrophage polarization toward the M2 phenotype, with the AMG group exhibiting a more pronounced effect (Fig. S6A-D, H-I, Supporting Information).

Fig. 7.

Composite hydrogel promotes the polarization of BV2 cells to M2 types and alleviates PC12 cell pyroptosis in vitro inflammatory environment. (A) Representative western blots showing protein expression of iNOS and Arg-1 in each group, β-actin was utilized as a loading control. (B) Quantitative analysis of relative expression of iNOS and Arg-1. (C) Representative immunofluorescence images of CD68 positive and iNOS positive BV2 cells (scale bar: 20 μm). (D) Representative immunofluorescence images of CD68 positive and Arg-1 positive BV2 cells (scale bar: 20 μm). (E, F) Quantitative analysis of relative fluorescence intensity of iNOS and Arg-1. (G) Representative western blots showing the expression of NLRP3, Caspase-1, IL-1β, ASC, GSDMD-N, and IL-18 protein associated with pyroptosis, β-actin was utilized as a loading control. (H) PI staining of PC12 cells in each group (scale bar, 100 μm). (I) Quantitative analysis of PI staining of PC12 cells (n = 3). (J) Quantitative analysis of relative expression of NLRP3, Caspase-1, IL-1β, ASC, GSDMD-N and IL-18 (n = 3). The data are presented as the means ± SEMs (n = 3); ∗p < 0.05, indicates significant differences; ns, is not significant. Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test.

Propidium iodide (PI) staining and WB analysis were used to assess whether the composite hydrogel alleviated neuronal pyroptosis. PC12 cells were treated with glucose and oxygen deprivation (OGD) to simulate the inflammatory conditions of SCI and were then cocultured with the composite hydrogel for 24 h. PI staining was used to quantify the degree of cell death. In the control group, cell death was minimal, and some areas lacked PI-positive cells. In contrast, OGD treatment led to a marked increase in cell death. Hydrogel treatment effectively reduced the degree of cell death, with the AMG group showing the lowest death rate (Fig. 7H, I). Additionally, WB analysis revealed reduced expression of pyroptosis-related proteins, including NLRP3, Caspase-1, GSDMD-N, ASC, IL-18, and IL-1β, in the hydrogel-treated groups, with the AMG group demonstrating the most pronounced changes (Fig. 7G, J). To further validate the mitigating effect of the composite hydrogel on neural death within an inflammatory context, we repeated the experiment using primary neurons. The results demonstrated that the hydrogel effectively alleviated neural pyroptosis, whereas the AMG group exhibited a significantly superior therapeutic outcome (Fig. S6E-F, G, J, Supporting Information).

2.7. AMG hydrogel inhibits inflammatory response and promotes macrophage polarization after SCI in vivo

Spinal cord sections were prepared from samples collected on Day 3 after injury to examine the inhibitory effect of the composite hydrogel on acute inflammation, which increases following SCI. To evaluate inflammatory cell infiltration during the acute phase of SCI, ELISA analyses further revealed that treatment with composite hydrogel groups led to noteworthy reductions in proinflammatory cytokines (IL-1βand IL-18), with effects more pronounced in the AMG group (Fig. S7A, Supporting Information). CD68 immunolabelling was performed to identify the macrophages/microglia in each group. Immunofluorescence imaging revealed fewer CD68⁺ cells in the hydrogel-implanted groups than in the SCI group, with the most obvious decrease observed in the AMG group (Fig. 8A, D). To distinguish the polarization states of the macrophages/microglia, iNOS and CD86 were used as markers for the M1 phenotypes, Arg1 and CD206 were used as markers for the M2 phenotypes. IBA-1 used to label all macrophage/microglia populations. Compared with the SCI group, all the hydrogel-treated groups presented decreased iNOS and CD86 expression, as indicated by the immunofluorescence images, with the AMG group displaying the most notable reduction (Fig. 8B, E) (Fig. S11A, E, Supporting Information). Conversely, Arg-1 and CD206 was upregulated in all the hydrogel-treated groups compared with that in the SCI group, with the AMG group exhibiting the highest expression level (Fig. 8C, F) (Fig. S11B, F, Supporting Information). WB analysis confirmed these findings, demonstrating that iNOS and CD86 expression was significantly reduced while Arg-1 and CD206 expression was elevated in the hydrogel-treated groups, with the AMG group exhibiting the most substantial changes (Fig. 8G, H) (Fig. S11C-D, Supporting Information). We subsequently employed the PI3K inhibitor LY294002 to investigate the relationship between the PI3K/AKT pathway and macrophage polarization during the treatment of spinal cord injury with the composite hydrogel. Four experimental groups were established: SCI, SCI + LY294002, AMG, and AMG + LY294002. Immunofluorescence colocalization analysis comparing the AMG and AMG + LY294002 groups revealed that PI3K inhibition increased the density of iNOS and CD86 in macrophages, while reducing the density of Arg-1 and CD206 (Fig. S10C-D, G-H, Fig. S11G-H, K-L, Supporting Information). WB analysis further confirmed elevated expression of iNOS and CD86, along with decreased expression of Arg-1 and CD206 in the AMG + LY294002 group (Fig. S10E, K, Fig. S11I, J, Supporting Information). Therefore, the AMG composite hydrogel promotes macrophage polarization toward the M2 phenotype through activation of the PI3K/AKT pathway.

Fig. 8.

Composite hydrogel regulates post-SCI inflammation. (A) Immunofluorescence images of inflammatory cell infiltration at the injured site 3 days after SCI (scale bar, 100 μm). (B) Immunofluorescence image of iNOS expression of microglia 3 days after SCI (scale bar, 40 μm). (C) Immunofluorescence image of Arg-1 expression of microglia 3 days after SCI (scale bar, 40 μm). (D) Quantitative analysis of the number of CD68 positive cells at the injured site 3 days after SCI. (E, F) Quantitative analysis of relative fluorescence intensity of iNOS and Arg-1 protein expression. (G) Representative western blots showing the expression of iNOS and Arg-1 protein 3 days after SCI, GAPDH was utilized as a loading control. (H) Quantitative analysis of relative expression of iNOS and Arg-1. The data are presented as the means ± SEMs (n = 3); ∗p < 0.05, indicates significant differences; ns, is not significant. Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test.

2.8. AMG hydrogel reduces neuronal pyroptosis, promotes the survival of neurons following SCI in vivo

To evaluate the therapeutic potential of the composite hydrogel after SCI, we examined neuronal survival and pyroptosis. Spinal cord sections were prepared from samples collected on Day 3 post-SCI and subjected to immunofluorescence staining. Neun was used to label neurons, while GFAP was used to label astrocytes to examine the survival of neurons around the SCI site. Compared with the SCI group, the hydrogel-treated groups presented increased anterior horn neurons, with the AMG group showing the most significant increase. These results indicate that the composite hydrogel promotes neuronal survival following SCI (Fig. 9A). We further examined key proteins involved in the pyroptotic pathway to assess the potential inhibitory effect of the hydrogel on pyroptosis. On the third day after SCI, spinal cord tissues were collected for immunofluorescence staining. The fluorescence intensities of caspase-1 and GSDMD-N were strong in the neurons of the SCI group. In contrast, the hydrogel-treated groups displayed weak fluorescence signals, with the AMG group showing the most substantial reduction (Fig. 9B-E). Western blot analysis was used to evaluate the expression levels of caspase-1, IL-1β, GSDMD-N, ASC, IL-18, and NLRP3. Compared with those in the sham group, pyroptosis-related protein levels were significantly increased in the SCI group. However, hydrogel implantation reduced the expression of these proteins, and the AMG group presented the most pronounced reduction (Fig. 9F, G). These findings suggest that the AMG composite hydrogel effectively suppresses neuronal pyroptosis and promotes neuronal survival following SCI. We next investigated the relationship between the PI3K/AKT pathway and neural pyroptosis during the treatment of spinal cord injury with the composite hydrogel. Immunofluorescence colocalization analysis comparing the AMG and AMG + LY294002 groups demonstrated that inhibition of the PI3K/AKT pathway increased the density of Caspase-1 and GSDMD-N in neurons (Fig. S10A-B, I-J, Supporting Information). WB analysis further confirmed elevated expression levels of Caspase-1 and GSDMD-N in the AMG + LY294002 group (Fig. S10F, L, Supporting Information). These results indicate that the AMG composite hydrogel suppresses neural pyroptosis through activation of the PI3K/AKT pathway.

Fig. 9.

The composite hydrogel inhibits post-SCI pyroptosis in neurons. (A) Immunofluorescence images of residual neurons existing in the anterior horn of the spinal cord (scale bar, 500 μm and 200 μm). (B) Immunofluorescence image of Caspase-1 expression of neurons 3 days after SCI (scale bar, 20 μm). (C) Quantitative analysis of relative fluorescence intensity of Caspase-1 protein expression in neurons within the specified groups (n = 3). (D) Immunofluorescence image of GSDMD-N expression of neurons 3 days after SCI (scale bar, 20 μm). (E) Quantitative analysis of relative fluorescence intensity of GSDMD-N protein expression in neurons within the specified groups (n = 3). (F) Representative western blots showing the expression of NLRP3, Caspase-1, IL-1β, ASC, GSDMD-N, and IL-18 protein 3 days after SCI, GAPDH was utilized as a loading control. (G) Quantitative analysis of relative expression of NLRP3, Caspase-1, IL-1β, ASC, GSDMD-N and IL-18 (n = 3). The data are presented as the means ± SEMs (n = 3); ∗p < 0.05, indicates significant differences; ns, is not significant. Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test.

2.9. Composite hydrogel promotes astrocyte polarization after SCI in vivo

Following CNS injury, astrocytes activate into two phenotypes: the neurotoxic A1 phenotype and the neuroprotective A2 phenotype. The pathological outcome largely depends on the balance between these phenotypes [[53], [54], [55]]. In this study, we examined and compared the polarization states of activated astrocytes during the acute phase after SCI across the SCI group and hydrogel groups. Using GFAP to label astrocytes, with C3 and S100A10 distinguishing A1 and A2 phenotypes respectively, immunofluorescence results demonstrated that the AMG group exhibited the highest S100A10 density and the lowest C3 density, indicating a significant promotion of astrocyte polarization toward the A2 phenotype (Fig. S12A-B, E-F, Supporting Information). Consistent with this, WB analysis revealed substantially decreased C3 expression and increased S100A10 expression in hydrogel-treated groups, with the AMG group showing the most pronounced therapeutic effect (Fig. S12C-D, Supporting Information). To investigate the relationship between the PI3K/AKT pathway and astrocyte polarization in the hydrogel's therapeutic mechanism, we observed that PI3K inhibition reduced S100A10 density and increased C3 density via immunofluorescence, suggesting a shift toward A1 polarization upon pathway suppression (Fig. S12G-H, K-L, Supporting Information). This finding was subsequently validated by WB analysis (Fig. S12I-J, Supporting Information).

2.10. Composite hydrogel enhanced axonal regeneration and remyelination in vivo

Axonal regeneration is a key indicator of functional recovery following SCI. Spinal cord sections were prepared from samples collected on Day 28 post-SCI and stained with NF200 to label neurofilaments, and GFAP was applied to label astrocytes. First, axonal penetration were assessed by quantifying NF-positive axons sheaths present within the lesion cavity. Concurrently, the glial scar, formed by GFAP-positive astrocytes at the lesion periphery, provides a clear definition of the injury boundaries. Compared to the injury group, the hydrogel-treated groups exhibited smaller lesion areas demarcated by GFAP boundaries and increased NF200-positive axons within the lesion site, with the AMG group showing the most pronounced effects (Fig. 10A, C). Myelin integrity is essential for proper neuronal function. MBP was used to label myelin sheaths to assess myelin regeneration, whereas S100β was used to label astrocytes. Compared with that in the SCI group, myelin regeneration was greater in the hydrogel-treated groups, and the AMG group demonstrated the most effective promotion of myelin repair (Fig. 10B, D). GSEA further revealed the role of the composite material in promoting axonal regeneration and neurological functional recovery (Fig. 10E).

Fig. 10.

Composite hydrogel enhances axonal regeneration and remyelination at 28 dpi. (A) Representative immunofluorescence images of NF200 and GFAP (scale bar, 200 and 20 μm). (B) Representative immunofluorescence images of S100β and MBP (scale bar, 200 and 20 μm). (C) Quantitative analysis of axon penetration in each group (n = 3). (D) Quantitative analysis of remyelination in each group (n = 3). (E) Transcriptome GSEA analysis. The data are presented as the means ± SEMs (n = 3); ∗p < 0.05, indicates significant differences; ns, is not significant. Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test.

3. Discussion

Severe inflammation following SCI leads to an imbalance in the microenvironment and neuronal death, making SCI particularly challenging to treat in clinical practice [56,57]. To mitigate the damage caused by the inflammatory microenvironment, we designed a conductive hydrogel incorporating MXene nanosheets and 3D apoptotic bodies (3D-ABs) to achieve both physical and biochemical effects. This composite hydrogel protects surviving neurons, reduces inflammation, improves the local microenvironment, and promotes functional recovery in mice.

The main goal in the treatment of acute SCI is to limit secondary damage, promote neural regeneration, and facilitate functional recovery [56,58]. We developed a safe and effective strategy for injured spinal cord repair. In this approach, vesicles are slowly released from the hydrogel and absorbed by surrounding tissue, wherein their bioactive cargo promotes tissue repair. The MXene nanosheets within the composite hydrogel provide conductive properties, improve the injury microenvironment, and promote connections between regenerating neurons and damaged axons, playing critical roles in preventing pathological changes throughout various injury phases [59,60]. Our findings demonstrate that the composite hydrogel can activate the PI3K/AKT signalling pathway, alleviate oxidative stress, promote M2 macrophage polarization, and inhibit neuronal pyroptosis, contributing to its treatment efficacy.

Bone marrow mesenchymal stem cells (BMSCs), previously known as bone marrow stromal fibroblasts, are adult stem cells derived from the mesoderm with the potential for self-renewal and multipotent differentiation [61]. Their ability to give rise to mesenchymal lineages, such as bone, cartilage, fat, and haematopoietic-supporting stromal cells, highlights their value in regenerative medicine, making them highly promising for therapeutic applications [62]. Studies have shown that extracellular vesicles derived from BMSCs also possess great therapeutic potential [38,63,64]. Compared with the more widely studied exosomes, apoptotic bodies (ABs) are a special class of extracellular vesicles that exhibit less immunosuppressive effects and higher production yields. ABs participate in intercellular communication by delivering proteins, miRNAs, and various nucleic acids to target cells [65]. The molecules contained within ABs are significantly influenced by the physiological or pathological conditions, suggesting that the microenvironment of parent cells can influence the properties of the ABs.

In recent years, an increasing number of studies have shown that cellular behaviour in 3D culture systems differs markedly from that in traditional 2D environments [66]. Cell activity is improved in 3D culture because of the improved interactions with the surrounding matrix. The physical and mechanical properties of the extracellular environment are important for regulating BMSC behaviours through cell–matrix adhesion and feedback mechanisms, including adhesion, growth, migration, and differentiation [67,68]. Differences in cell behaviour between 2D substrates and 3D suspension systems result in changes in cell contact patterns. Suspension culture enables the formation of spheroids, which exhibit more physiologically relevant tissue-like structures. Studies indicate that culturing cells in 3D can influence the therapeutic properties of ABs by altering their cargo.

Electrical conductivity is a crucial characteristic in the design of materials for neural repair [69]. As a new type of 2D nanomaterial, MXene nanosheets have excellent conductivity and biocompatibility, supporting their use in bioactive materials [70]. Such conductivity promotes intercellular communication and improves the microenvironment of the injury site, facilitating tissue regeneration. Through their ability to recapitulate the native cellular microenvironment, conductive materials significantly influence the immune landscape. By affecting ion channels and membrane potentials, they orchestrate intracellular signaling cascades that direct inflammatory responses toward resolution [71,72]. Although the swelling ratio of the composite hydrogel decreased, it remained higher than those of other conductive hydrogels. Another advantage of using hydrogels as the base material lies in their tunable mechanical strength. The storage modulus of our AMG composite is approximately 100 Pa, which is within the favorable range for neural differentiation (≈0.1–1 kPa), supporting tissue repair [73,74].

Despite several key issues in SCI treatment—such as the ability of hydrogels to reduce oxidative stress, improve the inflammatory microenvironment, and alleviate neuronal pyroptosis—our study still has several limitations regarding the underlying mechanisms and drug delivery strategy. For example, we did not investigate the specific roles of the individual components within the 3D-ABs, such as particular proteins or RNAs. We will explore these roles in the future to better understand the therapeutic mechanisms. It should be noted that most 3D-ABs being released within a week. The acute phase of SCI is characterized by inflammatory response, which subsequently triggers macrophage polarization and neural pyroptosis, thereby severely compromising functional recovery. The composite hydrogel exerts its therapeutic effect by mitigating acute-phase inflammation and associated neural death. As SCI repair is inherently long-term and multi-staged, the current hydrogel system has limitations in providing prolonged therapeutic support for chronic repair. We postulate that future improvements could be achieved by engineering hydrogels with slower degradation kinetics, designing multi-stage release systems, or developing intelligent, stimuli-responsive platforms for on-demand vesicle delivery. Such advancements could be highly beneficial for achieving comprehensive neural recovery. Although the electrical conductivity of the hydrogel gradually decreases during the degradation process, it remains sufficiently high during the acute phase of injury to mitigate the inflammatory environment and neuronal death. Future work should focus on optimizing the composite hydrogel formulation to achieve longer in vivo persistence and more reliable electrical properties. Additionally, incorporating engineered ABs into the hydrogel system could be a promising research direction. The therapeutic efficacy of ABs may be enhanced by loading specific proteins or RNAs. Surface modification could also be applied to ABs to achieve targeted delivery, which could synergize with the sustained drug release from the hydrogel for more effective SCI treatment. Furthermore, our study employed spinal cord hemisection model to evaluate the therapeutic effects of the composite hydrogel. This model was selected owing to its high controllability and the clear interpretability of results when assessing axonal regeneration. While the hemisection model offers advantages for assessing specific outcomes like axonal regeneration due to its well-defined lesion, its pathophysiology is distinct from the more clinically prevalent contusion injuries, which involve broader secondary injury processes. Consequently, while useful for initial assessment, the hemisection model has limitations, and further exploration and validation of the composite hydrogel in spinal cord contusion models are warranted. Finally, the immune rejection and long-term effects of hydrogels need to be further evaluated in clinical studies.

4. Conclusion

This study developed a physicochemical bifunctional apoptotic vesicle delivery system containing MXene, which can alleviate the inflammatory microenvironment by promoting microglial M2 polarization, inhibiting neuronal pyroptosis and promoting axon regeneration by activating the PI3K/AKT pathway, to achieve better functional recovery in SCI mice. This bifunctional hydrogel combining conductive materials and bioactive vesicles is effective, providing new treatment ideas for the clinical treatment of injured spinal cord repair.

5. Experimental section

Reagents and Antibodies:Pentobarbital sodium, the Masson staining reagents, and the HE staining reagents were obtained from Solarbio Science & Technology (Beijing, China). Dihydroethidium (DHE; cat. no. GC30025) was obtained from GlpBio in Montclair, California, USA. Primary antibodies against NLRP3 (cat. no. 15101) were obtained from Cell Signaling Technology (Beverly, Massachusetts, USA). The Proteintech Group in Chicago, Illinois, USA, provided antibodies against Caspase-1 (cat. no. 22915-1) and GAPDH (cat. no. 104941). Abcam in Cambridge, UK, supplied goat anti-mouse IgG H&L Alexa Fluor 594 (cat. no. ab150116), Alexa Fluor 488 (cat. no. ab150113), goat anti-rabbit IgG H&L Alexa Fluor 488 (cat. no. ab150077), Alexa Fluor 594 (cat. no. ab150080), DAPI solution (cat. no. ab104139), and ASC (cat. no. ab180799). Mouse monoclonal antibodies against NeuN (cat. no. ab104224), rabbit monoclonal antibodies against NeuN (cat. no. ab177487), rabbit polyclonal antibodies against laminin (cat. no. ab11575), rabbit monoclonal antibodies against neurofilament heavy polypeptide (cat. no. ab207176), rabbit monoclonal antibodies against CD86 (cat. no. ab239075), rabbit monoclonal antibodies against CD206 (cat. no. ab300622), rabbit monoclonal antibodies against C3 (cat. no. ab97462), and β-Actin (cat. no. ab213262) were provided by R&D Systems (Minnesota, USA). Rabbit monoclonal antibodies against S100A10 (cat. no. PA5-95505) were provided by Thermo Fisher Scientific (Rockford, IL, USA). Antibodies against IL-1β (cat. no. A1112) and IL-18 (cat. no. A1115) were provided by ABclonal Technology (Cambridge, MA, USA). Antibodies against GSDMD-N (cat. no. DF12275), CD44 (cat. no. BF9213), and Integrin beta1 (cat. no. AF5379) were obtained from Affinity Biosciences in Ohio, USA. Antibodies against GFAP (cat. no. sc-166458) and Iba-1 (cat. no. sc-32725) were purchased from Santa Cruz Biotechnology in Dallas, Texas, USA. DHE (dihydroethidium; cat. no. S0063), Dil (C1991S), calcein AM (cat. no. C2012), LY294002 (cat. no. S1737) and propidium iodide dye (cat. no. ST1569) were obtained from Beyotime Biotechnology (Jiangsu, China). A BCA kit (cat. no. 23227) was acquired from Thermo Fisher Scientific (Rockford, IL, USA).

Cells Culture: Mouse bone marrow mesenchymal stem cells (MUBMX-01001) were purchased from Ori Cell Bio Co., Ltd. The cells were maintained in complete mBMSC culture medium under standard conditions (Ori Cell Bio, MUXMX-90012). mBMSCs were cultured in a humidified incubator at 37 °C with 5% CO₂ and 95% air. Cells were incubated with the following fluorescently conjugated antibodies: PE anti-mouse CD105 Antibody (cat. no. 12-1051-82, Invitrogen), PE anti-mouse CD90 Antibody (cat. no. E-AB-F1283D, Elabscience), PE anti-mouse CD73 Antibody (cat. no. 12-0731-82, Invitrogen), FITC anti-mouse CD44 Antibody (cat. no. 11-0441-81, Invitrogen), and PE anti-mouse CD29 Antibody (cat. no. 12-0291-81, Invitrogen). Subsequent analysis was performed using a flow cytometer (CytoFLEX, Beckman). BV2 microglia (CL-0493) were obtained from Procell Life Science & Technology Co., Ltd. The cells were cultured in an incubator at 37 °C with 5% CO₂ and 95% air. The cells were maintained in DMEM (Gibco, C11995500BT) supplemented with 10% sterile foetal bovine serum (FBS; Gibco, 10099141C) and 1% penicillin–streptomycin (Gibco, 1719675). PC12 cells were obtained from the Cell Storage Center of Wuhan University (Wuhan, China). PC12 cells were exposed to β-NGF (50 ng/mL, catalogue no. 11050-HNAC, Sino Biological, Beijing, China) for three days to induce their differentiation into neuron-like cells before the experiments were performed. The cells were maintained in an incubator at 37 °C with 5% CO₂ and 95% air in RPMI-1640 medium (Gibco, 11875119) supplemented with 10% heat-inactivated FBS and 1% penicillin–streptomycin. Mouse Bone Marrow-Derived Macrophages (CP-M141) were obtained from Procell Life Science & Technology Co., Ltd. The cells were maintained in Mouse Bone Marrow-Derived Macrophage Complete Medium under standard conditions (cat. no. CM-M141, Procell Life Science & Technology Co., Ltd). Cells were cultured in a humidified incubator at 37 °C with 5% CO₂ and 95% air. Mouse Cortical Neurons (CP-M143) were obtained from Procell Life Science & Technology Co., Ltd. The cells were maintained in Mouse Cortical Neuron Complete Medium under standard conditions (cat. no. CM-M143, Procell Life Science & Technology Co., Ltd). Neurons were cultured in a humidified incubator at 37 °C with 5% CO₂ and 95% air.

A total of 100 μl of the BMSC suspension (100 μl/well, 1000 cells/well) was added to a 96 U-shaped well plate (Engineering for Life, EFLSP101) that had been treated with the antiadhesion coating solution. After 48 h of culture, spherical 3D-BMSCs were observed in the plate.

Isolation and characterization of the ABs: Following treatment with 0.5 μmol/L STS (Med Chem Express, HY-15,141) to induce apoptosis, 3D-BMSCs were incubated at 37 °C in a humidified atmosphere containing 5% CO₂. After 12 h of incubation, the cell supernatants were collected and centrifuged at 300×g for 5 min to remove the remaining cellular debris. Afterwards, the supernatants were centrifuged three times at 2000×g for 30 min each. The resulting ABs derived from the 3D-BMSCs were resuspended in PBS for future use. To evaluate the protein content in the 3D-ABs, a BCA assay was performed. Scanning electron microscopy (SEM) was used to assess the morphology and size of the ABs. The size distribution of the 3D-ABs was measured using dynamic light scattering (DLS; Malvern, UK). Western blotting (WB) was used to confirm the surface markers H3, H2B, C1QC, and C3B in the ABs. β-Actin was used as an internal reference for protein quantification.

Preparation of the hydrogels: GelMA (supplied by EFL) was derived from porcine skin gelatine via alkali treatment. Methacrylic anhydride (MAA) was introduced at a feed ratio of 0.017 mL of MAA to 1 g of gelatine, resulting in a degree of substitution of 30%. In accordance with published methods, the sequential addition reaction was conducted at 50 °C at pH 8.5–9 for 180 min. Unsaturated bonds were introduced into the gelatine structure through a one-step reaction in which methacrylic anhydride reacted with the amino and hydroxyl groups.

MXene nanosheets were synthesized in the laboratory as previously described. Briefly, 1 g of LiF, 5 mL of deionized (DI) water and 15 mL of HCl were added to a Teflon vessel and stirred at room temperature for 10 min. Next, 1 g of Ti3AlC2 powder was gradually added to the mixture at 40 °C over 48 h with constant stirring. The obtained solution was then placed in an oil bath and stirred at 35 °C for 48 h. The resulting crude product was centrifuged (3500 rpm, 5 min) and repeatedly rinsed with DI water until a neutral pH was achieved. Exfoliation was performed via ultrasonication in an ice water bath for 2 h under a nitrogen atmosphere. Finally, the suspension was centrifuged at 3000 rpm for 30 min to yield a dispersion containing MXene nanosheets. Finally, GelMA was dissolved in PBS with 0.25% photoinitiator lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP; Yongqinquan Intelligent Equipment Co., Ltd., Suzhou, China), and the solution was placed in a water bath at 60–70 °C in the dark for 20–30 min. The solution was allowed to cool to ambient temperature, and 1.25 mg of 3D-BMSC-ABs and 200 μg of MXene nanosheets were added to 1 mL of GelMA solution (5%, w/v). The mixture was stirred to obtain the composite pregel solution as a uniform dispersion.

Microstructural Analysis of Hydrogels: To evaluate the internal morphological characteristics, the photochemically cross-linked samples were incubated at −20 °C overnight, subjected to lyophilization, and sectioned perpendicular to the direction of UV exposure. Microstructural images of the gold-coated cross-sections were captured using a scanning electron microscope (SEM; JEOL, JCM-6000Plus, Japan) operating in secondary electron mode at an accelerating voltage of 15 kV. Enclosed pores were manually traced, and the average (n = 30) cross-sectional area of the enclosed pores for each composition was calculated using ImageJ software. To estimate the volumetric porosity, the liquid displacement method was employed. Freeze-dried samples of consistent dimensions were immersed in acetone until saturation occurred. The saturation point was determined by observing the swelling behavior of the samples. After complete saturation, the sample weight (Ws) was measured. The volumetric porosity (Vp) was calculated using the following formula (n = 3):

$${\mathrm{V}}_{\mathrm{p}}=\left({\mathrm{W}}_{\mathrm{s}}-{\mathrm{W}}_{\mathrm{d}}\right)\ast {\mathrm{\rho }}_{\text{ac}}/{\mathrm{V}}_{\mathrm{d}}\text{,}$$

where ρ_ac and V_d denote acetone density and geometric volume of the samples, respectively.

Analysis of shear thinning: Rheological shear rate sweep tests were performed on the 5% GelMA solutions with a rheometer (MCR102, Anton Paar, Austria) at 24–26 °C to evaluate the shear-thinning behaviour in the shear rate range of 1–100 rad/s.

Mechanical testing: The mechanical properties were evaluated by using compressive testing machine (UTM-2203, China). A 200 μL Disk-shaped GelMA sample was photocrosslinked under UV light at 405 nm for 60 s. The tests were performed at rate of 1 mm/min and 25 °C until fracture occurred. The research team repeated each experiment at least three times (n ≥ 3).

Swelling testing: To evaluate the swelling behaviour, the GelMA samples were evaluated gravimetrically by immersion in 2 mL of PBS at 37 °C and weighing 0, 2, 6, 12, and 24 h later. Excess PBS was removed from the sample surface using wax paper after washing. The weights of the swollen samples were subsequently measured. The samples were then lyophilized for 24 h, after which the final dry mass was recorded.

Cumulative ABs release in vitro: To evaluate the sustained release profile of the vesicles, a dialysis method was employed using PBS as the release medium. The concentration of vesicles in the medium was quantitatively determined at predetermined time intervals using a BCA assay kit. The cumulative release percentage was calculated, and the release profile was subsequently generated.

In vitro degradation: The samples from each group, with identical initial volumes, were lyophilized and weighed initially. They were then incubated in PBS and retrieved at pre-determined time points. The collected samples were lyophilized and weighed a second time. The percentage of remaining weight was determined as follows:

$$\text{Remaining}\text{Weight}\%=\mathrm{W}\left(\text{before}\right)/\mathrm{W}\left(\text{after}\right)\ast 100\%$$

Animals: Artificial urination, which reduces the risk of urinary retention after SCI, is more easily achieved in female animals because the female urethra is shorter. Therefore, female animals are often selected for SCI research. This study used healthy adult female C57BL/6J mice that were 6–8 weeks of age and weighed between 20 and 25 g. The Experimental Animal Center (registration number SCXK [ZJ] 2015-0001) affiliated with the Institute in Zhejiang, People's Republic of China, provided the mice. The mice were housed under standard laboratory conditions, including a temperature range of 21–25 °C, a 12-h light/dark cycle, and 50–60% humidity. Food and water were available ad libitum during the entire experimental period.

Groupings: All analyses were conducted with uniform group sizes (n = 3, where “n” denotes biological replicates), except for RNA sequencing (n = 4). 131 C57BL/6J mice were randomly divided into six groups: the sham (n = 18), SCI (n = 19), SCI + MG (n = 18), SCI + AG (n = 18), SCI + AMG (n = 19), SCI + LY294002 (n = 18), SCI + AMG + LY294002 (n = 18) and SCI + ABs (n = 3) groups. An overdose of pentobarbital sodium was used to euthanize the mice on Days 3 and 28 after surgery, after which tissue samples were collected for histological analysis.

In Vivo Spinal Cord Hemisection Model: Each mouse was anaesthetized via intraperitoneal injection of 1% (w/v) pentobarbital sodium (50 mg kg−1) before surgery. A longitudinal incision was made along the midline of the spine through the skin and subcutaneous tissue. Subsequently, the muscles were bluntly dissected to expose the vertebral bodies from T9 to T11. Following a T10 laminectomy performed using microforceps, the spinal cord at the T10 level was exposed. A customized 2 mm-wide curette was used to unilaterally remove spinal cord tissue at the T10 level, with care taken to preserve the central blood vessels. After hydrogel implantation, the wound was closed in layers with 4–0 silk sutures. In the Sham group, the mice underwent laminectomy at T9–T10 but their spinal cords were not damaged. After anaesthesia, the mice were kept warm until their physiological temperature recovered. Additionally, they received manual abdominal expression three times daily to support urinary excretion until normal urinary function resumed. Gentamicin sulfate (30 mg/kg) was injected once per day for three days after surgery.

Evaluation of Functional Behaviour: Functional recovery was assessed using the Basso Mouse Scale (BMS) on Days 0, 1, 3, 7, 14, 21, and 28 after SCI. The BMS ranges from zero (complete paralysis) to nine (normal locomotor function). Behavioural assessments were conducted using an open-field setup. Footprint analyses were recorded 28 days after surgery. Blue dye was used for the forelimbs, and red dye was used for the hindlimbs to generate footprints, which were analysed for stride length. Stride length was measured by calculating the distance between adjacent hindlimb steps. Two blinded investigators independently evaluated the footprints of each mouse.

Analysis of hindlimb recovery behaviour: All of the experimental mice were placed on a track to walk in a straight line, and the motion trajectories of the iliac crest, hip, knee, ankle, and toes were extracted from high-speed video recordings using DeepLabCut GUI 2.3.4. Hindlimb stick figures and the joint angles of the hip, knee, and ankle were plotted using MATLAB 2022a, with swinging and stance phases marked in red and black, respectively. The resulting data, including the height and angular amplitude of five key landmarks, were summarized for each group in bar and radar charts.

Motor-evoked potential (MEP) experiments: Anaesthesia was induced in all the mice by using 1% pentobarbital sodium. Recording electrodes were placed in the gastrocnemius muscle of both legs, the stimulation electrode was placed at the rostral edge of the lesion, and the reference electrode was placed in the subcutaneous abdominal tissue. The spinal cord received electrical pulses (10 mA, 0.1 ms, 1 Hz) through an electrical stimulator. Stimulation was repeated every 15 s, and the amplitude of the evoked potentials was measured after each pulse. Mice were randomly selected from each treatment group, and all the experimenters were blinded to the group allocations.

Dihydroethidium Staining: Spinal cord tissues were gradually dehydrated with 15%, 20%, and 30% sucrose solutions. After dehydration, excess moisture was removed by using clean filter paper. The tissue was then embedded in OCT compound and cooled until solidified. Frozen spinal cord tissues were then cut into 5-μm sections and stained with DHE according to the manufacturer's instructions. Transverse spinal cord sections from 3 mm rostral to the lesion were observed by fluorescence microscopy (Olympus, Japan) to acquire high-resolution images. ImageJ software was used to quantify DHE staining by calculating the integrated optical density (IOD).

Tissue Pretreatment for HE & Masson Staining: Mice were euthanized using an overdose of anaesthesia on the 28th day after surgery, after which the heart was perfused with ice-cold PBS (pH 7.4) and 4% (w/v) PFA. The entire 10 mm spinal cord segment, with the injury site in the centre, was fixed in 4% (w/v) PFA for 24 h, embedded in paraffin, and sectioned longitudinally. Five-micron-thick tissue sections were cut with a microtome and placed on poly-L-lysine-coated slides following established procedures for HE staining. For Masson's trichrome staining, longitudinal sections were first deparaffinized, incubated in a mordant solution of 10% potassium dichromate and 10% trichloroacetic acid, and then stained with haematoxylin to visualize the nuclei. After differentiation in ethanol and hydrochloric acid, the sections were treated with a mild ammonia solution and stained with Masson's trichrome according to established protocols. Images were obtained with an Olympus light microscope (Tokyo, Japan).

Western Blot (WB) Analysis: On Day 3 after SCI, the mice were humanely euthanized. A 1 cm portion of the spinal cord was precisely dissected and homogenized in ice-cold RIPA lysis buffer containing protease and phosphatase inhibitor cocktail III. The spinal cord proteins were extracted using specialized isolation reagents and quantified via BCA assays. Proteins (60 μg) were resolved via 12% (w/v) SDS‒PAGE and transferred to PVDF membranes (Millipore). After blocking in 5% (w/v) skim milk for 2 h, the membranes were incubated overnight at 4 °C with primary antibodies (1:1000 dilution) specific for IL-1β, GSDMD-N, IL-18, Caspase-1, ASC, NLRP3, GAPDH, iNos, Arg1, C3B, H3, H2B, C1QC, β-actin, PI3K, p-PI3K, AKT, and p-AKT. After incubation with HRP-conjugated IgG secondary antibodies at room temperature for 2 h, the band signals were visualized via ECL and imaged with a Bio-Rad ChemiDoc XRS + system.

Immunofluorescence (IF) Analysis: Following standard protocols, transverse sections of the spinal cord were subjected to immunofluorescence staining. The tissues were fixed and subjected to high-pressure antigen retrieval after dewaxing and rehydration. Blocking was performed at 37 °C for 30 min with 5% bovine serum albumin in PBS, after which the sections were incubated with primary antibodies (1:200 dilution) targeting specific proteins, including Caspase-1, GSDMD-N, NeuN, GFAP, NF200, iNos, Arg1, MBP, S100β and Iba-1, at 4 °C overnight. On the following day, the sections were incubated with secondary antibodies at 37 °C for 1 h and then stained with DAPI. An Olympus fluorescence microscope (Japan) was used to capture high-resolution images of the transverse sections prepared from a sample taken 3 mm rostral to the lesion site. Image analysis included 3 randomly selected anterior horn areas from each sample. The integrated density values of Caspase-1 and GSDMD-N in neurons and iNos and Arg1 in microglia were quantified by using ImageJ software. The numbers of CD68-positive microglia (per 0.1 mm²) were determined manually in a double-blind manner.

In Vitro IF Studies: The hydrogel was placed in 6-well plates. PC12, BV2, Neuron and Macrophage were digested into single-cell suspensions, collected by centrifugation at 1000 rpm for 5 min, and seeded at a density of 1 × 10⁵ cells per well into the 6-well plates, either with or without hydrogel. Except for the control group, all other groups were treated with TBHP for 24 h or subjected to OGD (sugar-free medium + 37 °C, 5% CO₂, and 1% O₂) for 2 h. Cells from the different groups were then collected for subsequent assays. Cells were seeded into confocal dishes and fixed with 4% paraformaldehyde for 30 min. The samples were then permeabilized and blocked by incubation in PBS containing 0.2% Triton X-100 and 6% BSA at 37 °C for 1 h. Subsequently, the samples were incubated with corresponding primary antibodies at 4 °C overnight, followed by incubation with corresponding secondary antibodies at 37 °C for 1 h. Between each step, the samples were washed three times with PBS. Images were captured using a confocal microscope. For DCFH-DA staining, after cells were seeded into confocal dishes, the DCFH-DA probe was added and incubated at 37 °C in the dark for 30 min. The unincorporated DCFH-DA probe was removed by washing with serum-free culture medium, and the cells were then imaged under a confocal microscope.

Enzyme-Linked Immunosorbent Assay (ELISA): The protein levels of IL-18 (Servicebio, Wuhan, China) and IL-1β (Servicebio, Wuhan, China) in the spinal cord tissue were detected using ELISA kits according to the manufacturer's protocol. The optical density of each sample was measured at a wavelength of 450 nm using a microplate reader. The expression level of the target protein in the sample was calculated from the standard curve.

Quantitative PCR (qPCR): Total RNA was extracted from cells and spinal cord tissues using TRIzol (Takara, Shiga, Japan). One microgram of RNA was used for reverse transcription to generate cDNA. The RNA levels of genes of interest were detected by qPCR as described previously [18]. The relative expression of each gene was calculated using the 2−ΔΔCT method and normalized to the β-actin level. The sequences of primers used in this study are listed in Table S1 (Supporting Information).

RNA Sequencing and Functional Enrichment Analysis: Tissue samples were collected three days after SCI surgery, and tRNA was extracted with TRIzol reagent following the manufacturer's guidelines. The RNA concentration was measured with a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA), and RNA integrity was assessed with an Agilent 2100 bioanalyzer (Agilent Technology, USA). The TruSeq Stranded mRNA LT Sample Prep Kit (Illumina, San Diego, CA, USA) was used for library preparation according to the manufacturer's protocol. Transcriptomic sequencing and data analysis were performed by OE Biotech Co., Ltd. (Shanghai, China). The Illumina HiSeq X Ten platform was used after library preparation to produce 125 bp or 150 bp paired-end reads. Trimmomatic was used to remove poly-N sequences and low-quality reads. Finally, the clean reads were mapped to the murine genome (GRCm38.p6) with HISAT2. Gene expression was quantified by using the FPKM values from Cufflinks and read counts from HTSeqcount. Differential gene expression was analysed with the DESeq (2012) package in R with a threshold of P < 0.05 and a fold change >2 or < 0.5. Differentially expressed genes (DEGs) were subjected to hierarchical clustering to assess the expression profiles, and GO enrichment analysis of the DEGs was conducted using a hypergeometric test in R.

Statistical Analysis: GraphPad Prism Software version 8.0.1 was used for statistical analysis. The results are presented as the means ± SEMs, and statistical significance was determined according to standard methods. Statistical differences between two independent groups were determined by a two-tailed unpaired t-test. When three or more groups with normally distributed datasets were compared, two-way analysis of variance (ANOVA) and Tukey's multiple comparisons test were used. Nonparametric Mann‒Whitney U tests were used to analyse data that did not reach the assumption of normality. A value of ∗P < 0.05 was considered to indicate statistical significance.

CRediT authorship contribution statement

Jiacheng Zhang: Writing – original draft, Project administration, Formal analysis, Data curation. Linyi Xiang: Software, Investigation, Formal analysis. Xiong Cai: Software, Resources. Yibo Geng: Project administration, Methodology, Investigation. Yuzhe Wu: Validation, Resources, Formal analysis. Jingwei Shi: Visualization, Formal analysis. Haojie Zhang: Methodology, Investigation. Xinli Hu: Visualization, Validation. Xiaoqiong Jiang: Software, Resources. Peijun Zhu: Resources, Investigation. Yun Shu: Investigation. Ruize Miao: Software. Jiayi Zhao: Project administration. Junsheng Zhu: Resources. Zhenglin Li: Software. Xiangyang Wang: Writing – review & editing, Visualization, Resources. Jian Xiao: Writing – review & editing, Software, Conceptualization. Kailiang Zhou: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Liangliang Yang: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Jianjun Qi: Writing – review & editing, Funding acquisition.

Ethics approval and consent to participate

All experiments involving animals were conducted according to ethical policies, and the procedures were approved by the Ethics Committee of Wenzhou Medical University, China (Approval no. wydw 2025‒0145).

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declaration of competing interest

The authors declare no conflict of interest.

Appendix A. Supplementary data

The following are the Supplementary data to this article: