Spatiotemporal Co‐Delivery of Hydrogen and Magnesium via Microneedle Patches for Neuroinflammation Modulation After Spinal Cord Injury: A Multi‐Modal In Vivo Study

ABSTRACT

Spinal cord injury (SCI) treatment is hindered by a “triple barrier:” the physical dura, a chemical storm of reactive oxygen species (ROS), and an immune barrier mediated by pro‐inflammatory microglia. To address these, we developed a “Hydrogen/Magnesium dual‐engine microneedle” (MN‐Mg) system by embedding magnesium (Mg) microparticles within a methacrylated hyaluronic acid (HAMA) hydrogel matrix via micromolding. Possessing high mechanical strength, Mg‐MNs penetrate the dura mater. Upon implantation, the system triggers a controllable magnesium‐water reaction, initiating a spatiotemporally synergistic dual‐engine therapeutic mode. The “hydrogen engine” rapidly releases high‐concentration hydrogen gas (H₂) during the acute SCI phase. It efficiently scavenges ROS storm (reducing levels by 55%) by inhibiting the MAPK pathway and downregulating AP‐1 transcription, creating an antioxidant window for neural repair. Subsequently, the “magnesium engine” provides sustained Mg^{2 +} release during the subacute phase, exerting a dual restorative effect: induces microglia polarization toward the pro‐reparative M2 phenotype (4.8‐fold increase) and promotes axonal regeneration (2.9‐fold increase). This synergy leads to locomotor recovery in a rat SCI model, with scores improving from 5.5 ± 1.05 (Controls) to 14.8 ± 1.17 at 8 weeks. This “Penetration‐Confinement‐Dual‐Engine Modulation” paradigm enables spatiotemporal synergistic H2/Mg therapy supported by an elucidated ROS‐MAPK/AP‐1 regulatory axis, advancing SCI treatment from symptomatic relief to targeted neural microenvironment reconstruction.

A spatiotemporal co‐delivery microneedle system (MN‐Mg) is developed for spinal cord injury repair. By generating hydrogen gas and magnesium ions in situ, this dual‐engine platform effectively scavenges ROS and modulates microglial polarization via the MAPK/AP‐1 signaling axis. This synergistic therapy promotes neurovascular reconstruction and significant locomotor recovery, offering a novel paradigm for targeted neural microenvironment reconstruction.

1. Introduction

Spinal cord injury (SCI), a debilitating condition of the central nervous system (CNS) resulting from trauma, ischemia, or compression, continues to exhibit a rising global incidence [1]. Due to the inherently limited regenerative capacity of mature neurons, patients often suffer from permanent functional deficits, imposing a substantial socioeconomic burden [2]. The pathophysiological progression of SCI comprises two stages: primary and secondary injury [3]. Primary injury involves direct structural damage to the spinal cord caused by mechanical forces. Secondary injury, initiated during the acute phase, is characterized by a robust inflammatory response and oxidative stress, which together propagate a vicious cycle involving reactive oxygen species (ROS) storms, neuroinflammatory cascades, and excitotoxicity, thereby exacerbating the zone of neurological deficit [4, 5, 6].

The predominant mechanism underlying secondary injury‐induced functional loss is ischemia‐induced mitochondrial dysfunction, which precipitates a pronounced ROS surge within 72 h post‐injury [7]. This excessive ROS burst explosively activates key inflammatory signaling pathways, notably the mitogen‐activated protein kinase (MAPK) and nuclear factor kappa‐light‐chain‐enhancer of activated B cells (NF‐κB) axes, leading to the release of pro‐inflammatory cytokines such as tumor necrosis factor‐alpha (TNF‐α) and interleukin‐1 beta (IL‐1β) [8, 9]. These cytokines facilitate the recruitment of a substantial influx of immune effector cells to the injury site. This acute‐phase inflammatory storm directly causes an imbalance in immune homeostasis by skewing microglial polarization toward the pro‐inflammatory M1 phenotype and inhibiting the differentiation of reparative M2 cells, ultimately leading to neuronal apoptosis and glial scar formation. Animal models have demonstrated a significant inverse correlation between the magnitude of the inflammatory response within the critical 24–72 h post‐injury window and long‐term neurofunctional prognosis [10, 11, 12, 13]. Therefore, developing targeted strategies that can simultaneously scavenge ROS and reverse inflammatory polarization during the acute phase has emerged as a pivotal therapeutic objective for breaking the vicious cycle of secondary injury.

Clinical intervention for secondary injury following SCI has long been hampered by the severe adverse effects associated with corticosteroid pulse therapy. Among neuroprotective strategies, hydrogen gas (H₂) has emerged as a promising candidate for antioxidant therapy due to its small molecular size and ability to neutralize hydroxyl radicals selectively. This property allows H₂ to overcome the low penetration efficiency and narrow free‐radical‐scavenging spectrum of commonly used antioxidants such as ganglioside GM‐1 and vitamin E (α‐tocopherol) [14, 15, 16]. However, the intrinsic physical properties of H₂ present a triple delivery barrier: its low saturation concentration at atmospheric pressure necessitates reliance on hazardous high‐concentration gas supplies; its high diffusivity results in a short effective duration; and the blood–spinal cord barrier significantly impairs its delivery efficiency to the target site. These barriers prevent H₂ from establishing a sustained and therapeutically effective concentration during the acute phase of SCI, failing to align with the temporal window of the ROS burst (6–72 h) and also impede its penetration through the physiological barrier of the dura mater to reach the core lesion site. Crucially, when ROS remains unscavenged, they drive microglial polarization toward a pro‐inflammatory phenotype by activating the NF‐κB pathway. The subsequent release of TNF‐α continuously activates the NADPH oxidase 2 (NOX2) complex, triggering a new round of ROS storms. This self‐sustaining vicious cycle of “oxidative stress → immune activation → ROS regeneration” reveals the core mechanism behind the failure of clinical trials involving conventional antioxidants: their single‐pathway intervention, although capable of transiently reducing peak free radical levels, is powerless to block the persistent oxidative damage mediated by microglia. This insight highlights the key to breaking the impasse: a spatiotemporal synergistic strategy is required to precisely eliminate the ROS storm during the acute phase and concurrently reverse the pro‐inflammatory transformation of microglia.

Based on the pathological nature of this vicious cycle, magnesium ions (Mg^{2 +}) emerge as a critical factor capable of disrupting it due to their dual modulatory roles in neural repair. Mg^{2 +} promotes microglial polarization toward the M2 phenotype by activating the phosphoinositide 3‐kinase/protein kinase B (PI3K/Akt) pathway, effectively inhibiting TNF‐α release and downregulating NOX2 expression, thereby interrupting the ROS regeneration pathway at its source. Contrastingly, Mg^{2 +} competitively blocks N‐methyl‐D‐aspartate receptors (NMDARs) and voltage‐gated calcium channels, reducing intracellular calcium influx and thus mitigating excitotoxic cell death, and promoting neuronal axon regeneration—a distinct advantage not provided by other immunomodulators [17, 18, 19, 20]. This dual repair mechanism of “immunomodulation + neuroregeneration” enables Mg^{2 +} to simultaneously address the two central aspects of secondary injury: suppressing the amplification of the inflammatory cascade and restoring the foundation for neural function. However, existing delivery systems exhibit fundamental limitations. Uncontrolled degradation of magnesium implants leads to localized alkalosis‐induced cell death, whereas injectable hydrogels, although improving sustained release, fail to penetrate the dura mater due to insufficient Young's modulus, resulting in subtherapeutic Mg^{2 +} concentrations at the target site. This delivery failure causes a significant delay in immunomodulation; once pro‐inflammatory microglia have established a self‐sustaining damage loop via the TNF‐α/NOX2 axis, a single‐pathway intervention cannot reverse the inflammatory cascade or initiate neural repair [21, 22].

This study was conducted to develop a spatiotemporally coupled solution: employing H₂ to target and eliminate the ROS storm during the acute phase, thereby creating a therapeutic window for immune repair, and concurrently utilizing Mg^{2 +} to remodel the immune microenvironment through M2 polarization and directly promote axonal regeneration during the subacute phase. To this end, we propose a magnesium‐loaded hyaluronic acid microneedle system (MN‐Mg; Figure 1) to achieve a breakthrough therapeutic strategy through a synergistic material‐structure‐function design, addressing three major clinical challenges in SCI treatment: dura mater penetration, uncontrolled hydrogen diffusion, and immune microenvironment imbalance.

FIGURE 1.

Schematic illustration of the fabrication of the magnesium‐loaded microneedle system and its therapeutic mechanism for promoting spinal cord injury repair.

We hypothesize that this synergistic strategy of “hydrogen‐blocking oxidative stress initiation + magnesium‐driven dual repair” will directly address the core pathological process following SCI. It involves penetrating the dura mater to release H₂ for free radical scavenging at the lesion site, while the sustained release of Mg^{2 +} will facilitate continuous immune‐neural Modulation. Their spatiotemporal complementarity will enable Modulation of the neural microenvironment, offering a novel approach to overcome the limitations of conventional therapeutic paradigms.

Following minimally invasive penetration of the dura mater by the rigid needle body, the hyaluronic acid‐based micro‐confinement system precisely restricts the diffusion path of hydrogen generated from the magnesium‐water reaction. This in situ hydrogen supply mechanism overcomes the spatiotemporal control bottlenecks of conventional gas therapies and rapidly eliminates the acute‐phase ROS storm, thereby suppressing the activation of the downstream mitogen‐activated protein kinase/activator protein 1 (MAPK/AP‐1) signaling axis. Concurrently, the released Mg^{2 +} establishes a sustained modulatory microenvironment at the injury site, inducing microglial polarization toward a reparative phenotype, downregulating pro‐inflammatory mediators such as TNF‐α, and promoting neuroregeneration. This metal‐gas spatiotemporal coupling mechanism proposes a redefined therapeutic paradigm across three dimensions: physical delivery, chemical modulation, and immune remodeling. It avoids the off‐target effects associated with systemic administration and achieves microenvironmental homeostasis by disrupting the “oxidative stress‐immune activation” vicious cycle. This technology transcends the limitations of traditional single‐pathway interventions, offering an integrated solution for SCI that combines precise delivery with synergistic modulation. Its “penetration‐confinement‐modulation” trinity design concept introduces a new avenue for the treatment of central nervous system diseases.

This investigation comprises three integrated aims: (i) development and characterization of methacrylated hyaluronic acid (HAMA)‐based microneedles embedding Mg particles for dura mater penetration and confined hydrogen diffusion; (ii) in vitro assessment of microglial polarization, ROS scavenging, neural stem cells (NSCs) differentiation, and neuronal markers; and (iii) in vivo evaluation in a rat SCI model assessing locomotor function, histopathology, inflammatory milieu, neural regeneration, and angiogenesis.

2. Results

2.1. Preparation and Characterization of HAMA‐Based Microneedles

Initially, we fabricated microneedle patches using HAMA (Figure 2A). Scanning electron microscopy (SEM) revealed that each patch consisted of a 10 × 10 array (100 needles in total). The individual needles were conical in shape, with a base diameter of 300 ± 15 µm and a height of 600 ± 25 µm (Figure 2B), dimensions compatible with the thickness range of the rat dura mater (300–500 µm). For the magnesium‐containing microneedles (c), numerous dark spherical particles were observed on the surfaces of the tips and bases (Figure 2C). The diameter of these spherical protrusions (40 ± 8 µm) observed under SEM matched the size of the pristine magnesium particles used (40 µm) (Figure 2D).

FIGURE 2.

Characterization of the Mg‐loaded MNs. (A,B) Optical and SEM images of blank MNs. (C,D) Optical and SEM images of Mg‐loaded MNs. (E,F) H₂ and Mg^{2 +} release profiles. (G) Elemental analysis (EDS) of MNs and pure Mg after 0, 3, and 7 days of immersion. (H) XPS spectra of MNs and pure Mg. (I) pH changes after 7 days of immersion. (J) pH‐dependent H₂ generation from the MNs.

The Mg particles embedded in the Mg‐MNs can sustainably release Mg^{2 +} and H₂ through a confined hydrolysis reaction, and the 3D network structure of HAMA can retain both Mg^{2 +} and H₂. To evaluate the release kinetics of hydrogen and Mg^{2 +} from the Mg‐MNs, we immersed both Mg‐MNs and an equivalent mass of free Mg particles in phosphate‐buffered saline (PBS; pH 7.4, 37°C) and monitored them dynamically for 7 days. By day 7, the Mg‐MNs had cumulatively released 84.6 ± 1.04 µmol of hydrogen, significantly higher than the free Mg particle group (p < 0.05) (Figure 2E). Similarly, the cumulative Mg^{2 +} release from Mg‐MNs reached 11.60 ± 0.04 mg, which was also significantly higher than that from pure Mg particles (3.10 ± 0.22 mg; p < 0.05) (Figure 2F). The weight loss profile of MN‐Mg patches showed a rapid initial degradation within the first 2 days, followed by a slower, sustained phase, consistent with the cumulative release of Mg^{2 +} and H₂ (Figure S1D). To elucidate the mechanism underlying the enhanced reactivity and sustained release of the MN‐Mg system, we performed a multi‐modal surface characterization using SEM‐EDS and XPS to verify the “spatial confinement” hypothesis. First, SEM‐EDS mapping provided visual evidence of the surface passivation states. After immersion in deionized water for 3 and 7 days, the free Mg particles (green) exhibited a dense accumulation of elemental Oxygen (red) on their surfaces, signifying the rapid formation of a thick Mg(OH)₂ passivation shell. In stark contrast, the Mg particles within the MN‐Mg group showed only minimal Oxygen signals at both time points (Figure 2G), suggesting that the HAMA hydrogel network effectively restricts the access of water and the accumulation of hydroxide precipitates, thereby inhibiting shell formation. To quantitatively validate this anti‐passivation effect, we conducted XPS analysis to determine the surface valence states of Magnesium after 7 days of degradation. In the pure Mg group, the surface underwent severe oxidation: the metallic Mg⁰ peak (1304.14 eV) accounted for only 48.11% of the total Mg signal, while the oxidized Mg‐O peak (1306.26 eV) dominated at 51.89%. Consistent with this, the O 1s spectrum showed a high proportion of lattice oxygen (Mg‐O, 531.82 eV).Conversely, the MN‐Mg group demonstrated superior preservation of metallic magnesium. The metallic Mg⁰ content (1302.8 eV) was maintained at 69.14%, with the oxidized Mg‐O fraction (1303.8 eV) significantly reduced to 30.86%. Additionally, the O 1s spectrum of MN‐Mg showed a distinct distribution with a lower proportion of lattice oxygen compared to the pure Mg group (Figure 2H). Collectively, the ∼21% difference in metallic Mg retention (69.14% vs. 48.11%) provides definitive quantitative evidence that the HAMA network spatially confines the hydrolysis reaction, preventing the formation of a dense passivation layer and ensuring the continuous release of H₂ and Mg^{2 +}. We found that the pH of the HAMA‐MN degradation solution decreased from 7 to 6.2 ± 0.1 within 72 h, whereas the pH of the solution containing pure Mg increased to 10.3 ± 0.17. The pH of the Mg‐MNs group was 8.0–8.3 (Figure 2I). These findings suggest that the acidic degradation products of HAMA, specifically glucuronic acid (HA⁻), neutralize Mg(OH)₂ on the particle surface via the reaction:

$$\mathrm{Mg}{(\mathrm{OH})}_2+2{\mathrm{H}\mathrm{A}}^{-}\to {\mathrm{Mg}}^{2+}+2{\mathrm{H}}_2\mathrm{O}+2{\mathrm{A}}^{-}$$

This reaction lowers the solution pH and reduces the passivation layer of Mg(OH)₂ on the Mg surface (Figure S1A). Consequently, although H₂ release from Mg‐MNs was reduced in high‐pH environments, they still produced a substantial amount of H₂ (22.6 ± 1.12 µL) under neutral pH 7 conditions (Figure 2J).

Given the unique structural characteristics of the biological spine, which must endure compressive forces from daily activities and tensile forces from surrounding tissues, the mechanical properties of the biological scaffold are critical for maintaining stability and fixation at the injury site following implantation [23]. We characterized the mechanical properties of the microneedles using a universal testing machine. The hydrogel microneedles exhibited a Young's modulus of approximately 400 MPa, indicating good mechanical strength (Figure S1B). To evaluate the mechanical adaptability of the microneedle patches under physiological conditions, we measured the Young's modulus of MN‐HAMA, MN‐MgCl₂, and MN‐Mg after immersion in PBS for 1 h. As shown in Figure S1B, all groups exhibited a significant reduction in modulus upon hydration, decreasing from approximately 400 MPa (dry state) to around 40 kPa (wet state). There were no significant differences among the three groups in the hydrated state. This modulus closely matches that of native spinal cord tissue, suggesting that the microneedles can minimize mechanical mismatch and reduce the risk of secondary injury after implantation. These findings demonstrate the excellent mechanical compliance and safety profile of our hydrogel‐based microneedle system. To evaluate dura mater penetration efficiency, we applied the microneedle patches to ex vivo rat dura mater (thickness: 420 ± 35 µm) under a pressure of 5 N/cm² for 30 seconds. Subsequent methylene blue staining demonstrated that the fabricated microneedles possessed a high specific surface area and excellent piercing capability, as evidenced by complete penetration of the dura mater and distinct indentation marks (Figure S1C).

2.2. Biocompatibility Assessment of HAMA‐Based Microneedles

To assess the biocompatibility of the various microneedle patch groups, we first evaluated the in vitro viability of cells cultured on them. Live/Dead cell staining combined with semi‐quantitative fluorescence analysis revealed that fluorescence intensity of cells on all hydrogel groups was lower than that of the Control group (p < 0.05) (Figure S2A,B). Additionally, results from the Cell Counting Kit‐8 (CCK‐8) assay showed no significant differences in cell proliferation activity among the groups on days 1 and 3 of co‐culture (Figure S2C). By day 5, the proliferation rate in the Control group was significantly higher than that in the magnesium‐containing hydrogel groups. This difference between the microneedle groups and the Control group was likely attributable to the more favorable mechanical stress microenvironment of the cell culture plate surface. Nevertheless, no statistical differences were observed among the MN‐HAMA, MN‐MgCl₂, and MN‐Mg groups, indicating that the incorporation of magnesium did not compromise the biocompatibility of the HAMA matrix.

We further evaluated cell adhesion on the various hydrogel groups using Integrin β1 staining. The results demonstrated that incorporation of Mg led to a marked increase in pseudopodia formation and extended spreading lengths of NSCs on the microneedle patches, with the most pronounced effect observed in the MN‐Mg group (p < 0.05) (Figure S3A,B). As a transmembrane protein receptor, Integrin plays a crucial role in signal transduction between the extracellular matrix and the cytoskeleton [24]. These findings suggest that the synergistic action of Mg^{2 +} and H₂ in the MN‐Mg group promotes the extension of NSC pseudopodia, thereby providing a favorable platform for cell growth and adhesion.

We then evaluated the in vivo biocompatibility of the various microneedle patch groups through subcutaneous implantation in rats. 2 weeks post‐implantation, hematoxylin and eosin (H&E) staining revealed the presence of a small number of mononuclear inflammatory cells at the implantation sites across all microneedle patch and Mg particle groups, with no significant difference compared to the MN‐HAMA group (Figure S4A,B). This indicates that the incorporation of Mg had minimal impact on the biocompatibility of HAMA. Furthermore, histological analysis of major organs (heart, liver, spleen, lungs, and kidneys) from rats in all hydrogel‐implanted groups showed no evident accumulation of degradation products from the microneedle patches and no significant pathological abnormalities compared to the control group (Figure S4C). These results confirm that the incorporation of Mg into the HAMA patches did not induce systemic toxicity and that the Mg‐MNs exhibit excellent biocompatibility.

2.3. Promotion of Microglial Polarization by HAMA‐Based Microneedles

To evaluate the in vitro immunomodulatory capability, we stimulated RAW264.7 microglial cells with lipopolysaccharide (LPS) to mimic an inflammatory microenvironment. We established several experimental groups: untreated microglial cells (Sham group), LPS‐treated cells (Control group), and co‐culture groups of LPS‐treated cells with either MN‐MgCl₂ or MN‐Mg. After 7 days of co‐culture, intracellular ROS levels were measured using a 2’,7’‐dichlorodihydrofluorescein diacetate (DCFH‐DA) probe (Figure 3A,B), and the immunomodulatory capability of the HAMA‐based microneedles was assessed by evaluating microglial polarization (Iba‐1) toward the M1 (CD86) and M2 (CD206) phenotypes (Figure 3C,D). Semi‐quantitative fluorescence analysis revealed that, compared to the Control group, ROS fluorescence intensity was reduced in both the MN‐MgCl₂ and MN‐Mg groups, with the MN‐Mg group exhibiting the lowest intensity. These findings indicate that Mg‐MNs effectively suppress intracellular ROS generation and mitigate the oxidative stress response during the acute inflammatory phase.

FIGURE 3.

Evaluation of ROS scavenging and microglia polarization. (A,B) Intracellular ROS staining with DCFH‐DA and corresponding fluorescence quantification. (C,D) Immunofluorescence staining and quantification of M1 (CD86⁺) and M2 (CD206⁺) polarized microglia (Iba‐1⁺). (E,F) ELISA analysis of TNF‐α and IL‐10 secretion levels. n = 3. Statistical analysis is conducted using one‐way ANOVA and Tukey's post hoc test. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

Microglia can be categorized into the pro‐inflammatory M1 and anti‐inflammatory M2 phenotypes. We first labeled microglia with Iba‐1 and subsequently marked the M1 and M2 phenotypes with CD86 and CD206, respectively. Immunofluorescence (IF) imaging and semi‐quantitative analysis demonstrated that CD86 fluorescence intensity in the MN‐Mg group was significantly lower than that in the other control groups (p < 0.05), whereas CD206 fluorescence intensity was significantly higher than that in all other control groups (p < 0.05). These results confirm that the MN‐Mg group effectively induces microglial polarization toward the M2 phenotype and inhibits polarization toward the M1 phenotype in vitro, indicating robust immunomodulatory properties [25]. Furthermore, we examined the expression of inflammatory cytokines via quantitative reverse transcription polymerase chain reaction (qRT‐PCR) and found that the MN‐Mg group exhibited decreased expression of the pro‐inflammatory cytokine TNF‐α and increased expression of the anti‐inflammatory cytokine IL‐10 (Figure 3E,F). Taken together, these multifaceted results comprehensively demonstrate that the MN‐Mg group possesses significant immunomodulatory functions, capable of markedly reducing ROS levels, promoting the polarization of microglia from the M1 to the M2 phenotype, and suppressing the inflammatory response.

2.4. HAMA‐Based Microneedles Promote Nerve Regeneration In Vitro

The CNS contains a substantial population of NSCs capable of differentiating into neuronal cells, making them a widely recognized ideal substitute for neuronal loss following SCI. However, the immune microenvironment resulting from secondary injury after SCI severely impairs the differentiation of NSCs into neurons, highlighting the importance of modulating this process for effective neural tissue repair. We co‐cultured the various microneedle patch groups with NSCs and assessed their potential to induce NSC differentiation into neurons and promote axonal growth by evaluating three markers: Microtubule‐Associated Protein (Tau), Growth‐Associated Protein 43 (GAP43), and Microtubule‐Associated Protein 2 (MAP2).

Immunofluorescence (IF) microscopy images (Figure 4A) revealed that NSCs in the Control and MN‐HAMA groups exhibited poor differentiation capacity and limited axonal growth, with cells remaining spherical and growing in clusters (forming neurospheres). In contrast, NSCs in the MN‐MgCl₂ and MN‐Mg groups gradually adhered, dispersed, and differentiated into neurons to varying degrees, displaying significant axonal growth and interconnectivity. Semi‐quantitative fluorescence analysis of the staining for Tau, GAP43, and MAP2 (Figure 4B–D) showed that fluorescence intensities in both the MN‐MgCl₂ and MN‐Mg groups were significantly higher than those in the unloaded MN‐HAMA group (p < 0.05). These findings suggest that incorporation of Mg particles into the microneedle patches can sustainably promote NSC neuronal differentiation through the slow release of Mg^{2 +}, thereby facilitating repair following nerve injury.

FIGURE 4.

Promotion of NSC neuronal differentiation by the microneedle systems. (A) Immunofluorescence staining of neuronal markers (GAP‐43, Tau, and MAP2) in NSCs cultured on different MN substrates. (B‐D) Quantification of the fluorescence intensity of GAP‐43 (B), Tau (C), and MAP2 (D) in the differentiated NSCs. n = 3. Statistical analysis is conducted using one‐way ANOVA and Tukey's post hoc test. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

2.5. HAMA‐Based Microneedles Improve Motor Function and Pathology After SCI in Rats

We established a T9 spinal cord compression injury model in rats and gently applied microneedle patches from each group onto the dural surface at the injury site. Rats were divided into a Sham group, a Control group (injury only), and groups receiving post‐injury implantation of MN‐HAMA, MN‐MgCl₂, or MN‐Mg. Tissues were harvested at 1, 4, and 8 weeks post‐operation, as depicted in the experimental timeline (Figure S5A). Motor function recovery in each group was assessed using the Basso, Beattie, Bresnahan (BBB) locomotor rating scale (Figure S5B) and the Inclined Plane Test (IPT) score (Figure S5C).

At 8 weeks post‐operation, BBB scores for the Control and MN‐HAMA groups were 5.5 ± 1.05 and 5.8 ± 1.32, respectively, with no significant difference between them. This indicates that spontaneous neural repair capacity in rats following SCI is limited, and motor function recovery is difficult without intervention. In contrast, the MN‐Mg group exhibited a BBB score of 6.0 ± 1.10 by week 3 and an IPT score of 46.67 ± 5.16 by week 4, both significantly different from those in the other groups. By week 8, the BBB score of the MN‐Mg group reached 14.8 ± 1.17, and the IPT score increased to 59.17 ± 3.76, indicating that rats achieved weight‐bearing and coordinated forelimb–hindlimb movement. Although the MN‐MgCl₂ group demonstrated partial weight‐bearing capacity, gait coordination remained significantly inferior to that of the MN‐Mg group (Figure S5D). Hindlimb motor function and footprint analysis corroborated the BBB and IPT scores (Figure S5E). At 8 weeks post‐operation, rats in the MN‐MgCl₂ group exhibited intermittent plantar support without significant gait restoration. In contrast, rats in the MN‐Mg group displayed a uniform gait, with markedly reduced toe drag marks and a significant improvement in the rotation angles of their forepaws and hindpaws. Rats in the Control and MN‐HAMA groups continued to drag their hindlimbs. These findings suggest that the MN‐Mg group significantly improves motor coordination in rats following SCI through the synergistic action of H₂ and Mg^{2 +}, potentially by inhibiting secondary spinal cord injury and promoting axonal remodeling.

We further evaluated post‐SCI pathology in each group using H&E staining of tissue sections. As shown in Figure S6, tissue destruction and congestion were evident at the compression site in the spinal cords of all injured groups, confirming successful model establishment. Quantitative analysis of the cavity area on sagittal spinal cord sections using ImageJ software revealed that cavity areas in the Control and MN‐HAMA groups did not change significantly between 4 and 8 weeks post‐operation (p > 0.05), confirming limited spontaneous neural repair. The cavity area in the MN‐MgCl₂ group was smaller than that in the Control group (p < 0.05), indicating that Mg^{2 +} can partially promote neural repair. However, the MN‐Mg group demonstrated the most pronounced therapeutic effect, with its cavity area at 8 weeks post‐operation significantly smaller than that of all other groups (p < 0.05). Moreover, newly formed capillaries (neovascularization) and nerve fiber bundles were visible at the lesion border. These findings further support that Mg‐MNs exert synergistic immunomodulatory and neurorestorative effects through sustained release of H₂ and Mg^{2 +}, thereby creating a favorable microenvironment for neural repair and effectively reducing the post‐injury spinal cord cavity size. Furthermore, to validate the rationale of our microneedle design (600 µm length), we measured the thickness of the rat spinal dura mater using histological sections. As shown in Figure S6D, the average thickness of the dura mater at the injury site is 150 ± 42 µm (n = 3). This indicates that our 600 µm microneedles can effectively penetrate the protective dura barrier while maintaining a safe insertion depth (∼450 µm) into the dorsal spinal cord, avoiding damage to deep neural circuits.

2.6. HAMA‐Based Microneedles Modulate the Inflammatory Response After SCI in Rats

Following SCI, activated inflammatory cells release pro‐inflammatory cytokines, triggering a cascade of secondary injury events that exacerbate tissue damage and hinder neural repair. Therefore, evaluating the early anti‐inflammatory effects of microneedle patches in vivo is of paramount importance. To verify the antioxidative efficacy of MN‐Mg during the peak of the secondary injury cascade, we evaluated ROS levels in the injured spinal cord at 3 days post‐injury using DHE staining. As shown in Figure S7, the SCI and HAMA groups exhibited intense red fluorescence at the lesion site, indicating a massive accumulation of superoxide anions characteristic of the acute oxidative burst. Quantitative analysis showed that the relative fluorescence intensity of the Control group was significantly higher than that of the Sham group (p < 0.05). In sharp contrast, the MN‐Mg treatment significantly attenuated this acute oxidative surge. Compared with the Control group, the fluorescence intensity of DHE in the MN‐Mg group was significantly reduced (p < 0.05), approaching the level of the Sham group. Furthermore, consistent with our observations at 7 day (Figure 5A,B), this early intervention effectively prevented the transition to persistent oxidative stress. Collectively, these results demonstrate that MN‐Mg acts as a potent antioxidant system that not only suppresses the initial oxidative storm (Day 3) but also maintains a low‐oxidative microenvironment into the sub‐acute phase (Day 7), thereby protecting neural tissue from secondary degeneration. To further assess microglial polarization at the injury site, we labeled microglia with Iba‐1 and specifically marked the M1 and M2 phenotypes using inducible nitric oxide synthase (iNOS) and arginase‐1 (Arg‐1), respectively (Figure 5A). Compared to the Sham group, increased microglial accumulation was observed at the injury site in the Control and all treatment groups. IF imaging and semi‐quantitative analysis showed that although the MN‐MgCl₂ group exhibited moderately reduced iNOS fluorescence intensity and elevated Arg‐1 fluorescence compared to the Control and MN‐HAMA groups, the MN‐Mg group displayed significantly lower iNOS fluorescence intensity and significantly higher Arg‐1 fluorescence intensity than all other control groups (both p < 0.05) (Figure 5C,D). These findings were consistent with the DHE fluorescence results, indicating that MN‐Mg possesses the most potent anti‐inflammatory capability.

FIGURE 5.

Immunomodulatory function of the MNs in vivo at day 7. (A) Immunofluorescence staining for ROS (DHE), M1 marker (iNOS), and M2 marker (Arg‐1) in spinal cord sections. (B–D) Quantification of the fluorescence intensity for DHE (B), iNOS (C), and Arg‐1 (D). (E–G) Flow cytometry analysis and quantification of the percentage of M1 (F) and M2 (G) microglia. n = 3. Statistical analysis is conducted using one‐way ANOVA and Tukey's post hoc test. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

We further employed flow cytometry to evaluate the modulation of microglial subtypes by the microneedle patch groups at 7 days post‐SCI in vivo. CD11b was used as a general marker for microglia, and the M1 and M2 phenotypes were labeled with CD86 and CD206, respectively (Figure 5E). Compared to the Control group, the proportions of CD11b⁺/CD86⁺ (M1) and CD11b⁺/CD206⁺ (M2) double‐positive microglia did not change significantly in the MN‐HAMA group (p > 0.05) (Figure 5F,G). In contrast, the MN‐MgCl₂ group, due to the introduction of Mg^{2 +}, exhibited a reduced proportion of M1 microglia and an increased proportion of M2 microglia [26, 27, 28]. Our engineered Mg‐loaded microneedle system (MN‐Mg), which simultaneously releases H₂ and Mg^{2 +}, significantly induced microglial polarization toward the M2 phenotype at the injury site, with a corresponding significant reduction in the proportion of M1 microglia. This effect was significantly different from that of all other groups (p < 0.05). Collectively, these findings demonstrate that during the acute phase of SCI, the Mg‐loaded MN‐Mg system can effectively modulate immune cell subtypes.

2.7. HAMA‐Based Microneedles Promote Neural Regeneration and Inhibit Scar Formation After SCI in Rats

We evaluated the effects of the Mg‐loaded microneedle system on neural tissue repair and regeneration after SCI in vivo by assessing three key aspects: neuronal regeneration and axonal sprouting, astrocyte activation and glial scar formation, and angiogenesis. Given the abundance of NSCs in the CNS, we first assessed the capacity of the MN‐Mg microneedle patches to recruit NSCs. At 4 and 8 weeks post‐injury, we observed the distribution of Nestin‐positive stem cells at the injury site in each group (Figure 6A; Figure S8A). Migrated endogenous NSCs were visible in all groups; however, no significant difference was observed between the Control and MN‐HAMA groups (p > 0.05), indicating that NSC recruitment following SCI is weak and indirectly reflects the limited capacity for spontaneous neural repair. In contrast, fluorescence intensity in the MN‐MgCl₂ and MN‐Mg groups progressively increased, and semi‐quantitative analysis showed that the MN‐Mg group exhibited significantly higher intensity than all other groups at all time points (p < 0.05) (Figure 6C). These findings suggest that the MN‐Mg group significantly enhances the homing of endogenous NSCs.

FIGURE 6.

Long‐term therapeutic effects of the MNs, including NSC recruitment, neurogenesis, and glial scar inhibition at 8 weeks. (A) Immunofluorescence staining showing endogenous NSC (Nestin, red) migration to the injury site. (B) Staining for newly formed neurons (Tuj‐1, red) and reactive astrocytes (GFAP, green). (C–E) Quantification of the fluorescence intensity for Nestin (C), Tuj‐1 (D), and GFAP (E). n = 3. Statistical analysis is conducted using one‐way ANOVA and Tukey's post hoc test. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

Following SCI, inflammation‐induced secondary injury is frequently accompanied by chronic encapsulation by glial scar tissue, which impedes the penetration of regenerating neurons and axons, thereby obstructing Modulation of the damaged neural tissue. To investigate the regenerative effects of the Mg‐loaded microneedle system on the injured spinal cord, we further assessed neuronal regeneration and glial scar formation at the lesion site. The distribution of neurons and astrocytes in the lesion area was analyzed using Tuj‐1 and glial fibrillary acidic protein (GFAP) double‐staining, respectively. The results revealed a substantial accumulation of GFAP‐positive astrocytes at the injury border in the Control and MN‐HAMA groups (indicated by intense green fluorescence), forming a dense glial scar that hindered axonal extension and neuronal migration. Meanwhile, Tuj‐1–positive neurons were scarcely detectable. In contrast, the MN‐Mg group exhibited a progressive reduction in GFAP‐positive astrocyte density in and around the injury site, and the scar border formed by activated astrocytes became increasingly indistinct. This was accompanied by greater infiltration of Tuj‐1–positive neurons. Over time, these neurons migrated toward the cavity center to establish connections with distal neurons, restoring a relatively intact spinal cord morphology by 8 weeks. The generation of Tuj‐1–positive cells and suppression of GFAP expression in the MN‐MgCl₂ group were moderately improved than in the Control and MN‐HAMA groups but remained inferior to the MN‐Mg group (Figure 6B,D,E; Figure S8B,D,E).

The extent of axonal regeneration and Modulation of functional connections reflects the degree of neural tissue repair and motor function recovery following SCI. Additionally, angiogenesis in the injured spinal cord region enhances blood supply to damaged neurons, supporting neural regeneration and myelination. We further evaluated the capacity of the Mg‐loaded microneedle system to promote neural regeneration during the middle and late stages of SCI by using CD31 to label angiogenesis, Neurofilament Heavy (NFH) to label axons, Myelin Basic Protein (MBP) to label oligodendrocyte‐derived myelin, and Neuromodulin (GAP‐43) to label axonal sprouting of regenerating nerves (Figure 7A,B; Figure S9A,B). Semi‐quantitative fluorescence analysis showed that at 4 and 8 weeks, the number of CD31‐positive vascular endothelial cells in the MN‐Mg group was significantly higher than that in the other groups (Figure 7F; Figure S9F), indicating robust angiogenesis that provided a favorable platform for neural tissue regeneration. In the Control and MN‐HAMA groups, NFH‐ and MBP‐labeled axons and myelin sheaths were sparse and unevenly distributed within the injury area. In contrast, fluorescence intensities of NFH and MBP in the lesion center of the MN‐Mg group were significantly higher than those in other groups and demonstrated improved continuity. Axons were observed extending from the lesion border toward the cavity center, suggesting that newly formed neuronal axons were penetrating the spinal cord cavity. Under normal physiological conditions, GAP‐43 is expressed at low levels to maintain spinal cord function, but its expression is upregulated during nerve growth and regeneration to facilitate axonal extension and enhance neural plasticity and neurotransmitter transmission. The number of GAP‐43–positive cells was significantly higher in the MN‐Mg group than in all other groups.

FIGURE 7.

Promotion of axonal regrowth and angiogenesis at 8 weeks post‐injury. (A) Immunofluorescence staining for mature axons (NFH), myelin sheaths (MBP), and axonal sprouting (GAP‐43) in the injured spinal cord. (B) Staining for newly formed blood vessels (CD31). (C–F) Quantification of the fluorescence intensity for NFH (C), MBP (D), GAP‐43 (E), and CD31 (F). n = 3. Statistical analysis is conducted using one‐way ANOVA and Tukey's post hoc test. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

In summary, these results collectively demonstrate that the Mg‐loaded microneedle system promotes the migration of endogenous NSCs, reduces glial scar formation, and facilitates neuronal regeneration and angiogenesis, thereby providing a favorable platform for axonal sprouting, extension, and remodeling of the injured neural tissue after SCI.

2.8. Transcriptomic Analysis Reveals the Mechanism of MN‐Mg Microneedles in Modulating Inflammation After SCI

Our in vivo and in vitro results demonstrated that the MN‐Mg group possessed superior anti‐ROS and immunomodulatory capabilities after SCI due to its unique H₂‐generating mechanism. To further investigate the regulatory mechanism by which the Mg‐loaded hydrogel microneedles block oxidative stress via H₂ to facilitate SCI repair, we established a rat T9 spinal cord injury model. The microneedles were implanted at the lesion site, and spinal cord tissues from the injured area were collected at 3 days post‐operation during the acute inflammatory phase for transcriptome sequencing (RNA‐seq).

Analysis of differentially expressed genes (DEGs) revealed 223 upregulated and 347 downregulated genes in the MN‐Mg group compared to the MN‐MgCl₂ group. Principal component analysis (PCA) showed a distinct separation between the two groups, indicating significant differences in their transcriptomic profiles (Figure S10A–C). Gene Ontology (GO) enrichment analysis of the DEGs revealed that, compared to the MN‐MgCl₂ group, the MN‐Mg group exhibited differential expression in terms related to “immune system process” and “immune response” (Figure S10D,E), suggesting that H₂ release influences immunomodulation at the SCI site. Furthermore, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the DEGs indicated that pathways related to “oxygen binding” and “neurotransmitter transporter activity” were upregulated in the MN‐Mg group, while pathways such as the “MAPK pathway,” “NF‐κB signaling pathway,” and “IL‐6 signaling pathway” were downregulated (Figure S10F,G). These findings suggest that the MN‐Mg group may exert an inhibitory effect on the inflammatory response during the early phase of SCI.

The MAPK pathway regulates core biological processes, including cell proliferation, differentiation, apoptosis, and stress responses, and plays a central role in cellular adaptation to microenvironmental changes. We further elucidated the pathway by which MN‐Mg mediates immunomodulation via the MAPK pathway using Western blot (WB) analysis (Figure 8A,B). As shown in the WB results (Figure 8C), no significant difference in ERK phosphorylation was observed between the MN‐Mg group and the MN‐HAMA or MN‐MgCl₂ groups. However, phosphorylation of JNK and p38 was significantly decreased in the MN‐Mg group, indicating inhibition of the JNK and p38 branches of the MAPK pathway. Concurrently, expression of the downstream proteins c‐Jun and c‐Fos was also significantly reduced. Since activator protein‐1 (AP‐1) is a heterodimeric transcription factor complex formed by c‐Fos and c‐Jun via a leucine zipper structure, these results suggest that AP‐1 expression was suppressed in the MN‐Mg group.

FIGURE 8.

Mechanism of MN‐Mg in suppressing the MAPK/AP‐1 inflammatory pathway. (A) Schematic of the proposed H₂‐mediated inhibition of the MAPK/AP‐1 axis. (B,C) Western blot analysis and quantification of MAPK pathway activation (p‐ERK, p‐JNK, p‐p38) in microglia. (D,E) Western blot analysis and quantification of AP‐1 components (c‐Fos, c‐Jun) and downstream cytokines (IL‐6, CXCL1). n = 3. Statistical analysis is conducted using one‐way ANOVA and Tukey's post hoc test. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

AP‐1 is a key transcription factor in the pro‐inflammatory response, driving the expression of pro‐inflammatory cytokines (e.g., IL‐6, TNF‐α) and chemokines (e.g., CXCL1, CXCL8), thereby exacerbating inflammation [29, 30, 31]. To validate the mechanism by which MN‐Mg inhibits the MAPK/AP‐1 axis in microglia via H₂ to reduce pro‐inflammatory cytokine release and suppress inflammation, we activated AP‐1 in the MN‐Mg group by adding an AP‐1 agonist, 15(S)‐HPETE. The results showed that protein expression of c‐Fos and c‐Jun, the components of AP‐1, was significantly enhanced, confirming AP‐1 activation (Figure 8H). Compared to the MN‐HAMA and MN‐MgCl₂ groups, expression of the pro‐inflammatory factors IL‐6 and CXCL1 was significantly lower in the MN‐Mg group. Conversely, in the MN‐Mg + AP‐1 agonist group, their expression was markedly elevated (Figure 8D,E). To further clarify the impact of AP‐1 on microglial immunomodulation, we measured intracellular ROS levels using a DCFH‐DA probe. The results showed that AP‐1 activation in the MN‐Mg group significantly attenuated its ROS‐scavenging ability, demonstrating a marked difference compared to that of the MN‐Mg group alone (Figure S11). Flow cytometry further revealed changes in the proportions of M1/M2 microglial subtypes. Compared to the MN‐Mg group, AP‐1 activation led to a marked increase in the proportion of M1 microglia and a significant decrease in the proportion of M2 microglia (Figure S12).

These results suggest that the MN‐Mg group inhibits the MAPK pathway via H₂, which subsequently suppresses AP‐1, thereby downregulating the expression of IL‐6 and CXCL1, blocking oxidative stress in microglia, and suppressing the inflammatory response.

3. Discussion

The pathological cascade following SCI presents a multidimensional complexity, characterized by the interrelated core challenges of acute inflammatory storms, glial scar proliferation, and impaired neural regeneration. As the initiating event in this cascade, the inflammatory response triggers a burst of oxidative stress and establishes an inhibitory microenvironment through the overactivation of microglia and astrocytes, severely hindering the neural repair process [32, 33]. To address this multifaceted regulatory demand, we innovatively developed a magnesium‐loaded microneedle system. The breakthrough design integrates three functional modules: H₂ generation via magnesium hydrolysis, H₂ retention within the hyaluronic acid network, and sustained release of dual active components. Our findings demonstrate that MN‐Mg achieves confined, sustained delivery of H₂ and Mg²⁺, scavenges ROS, and drives microglia toward the reparative M2 phenotype, while also promoting NSC differentiation and axonal regeneration. In vivo, MN‐Mg improved locomotor outcomes, reduced cavity formation and glial scarring, enhanced angiogenesis, and upregulated neuroregenerative markers; transcriptomic analyses corroborate inhibition of MAPK/AP‐1–driven inflammation as a central mechanism, supporting the integrated anti‐oxidant, immunomodulatory, and pro‐regenerative effects of this spatiotemporal therapy. This spatiotemporal synergistic delivery system precisely provides a novel solution for simultaneous modulation of the inflammatory microenvironment and promotion of neural regeneration.

Traditional hydrogen therapy is hampered by limitations such as low solubility, short biological half‐life, and difficulty in achieving localized accumulation, thereby impeding long‐term anti‐inflammatory and neuroprotective efficacy [34, 35]. The microneedle system developed in this study overcomes these limitations through a confined hydrolysis reaction and a 3D network confinement effect. The hyaluronic acid matrix physically adsorbs H₂, forming a diffusion gradient that significantly prolongs the H₂ release period while concurrently enabling sustained Mg^{2 +} release. Microstructural analysis confirmed that the spatial confinement effect of the microneedles effectively inhibits the formation of a passivation layer, thereby enhancing the utilization efficiency of the metal.

The core innovation of this system lies in the incorporation of Mg particles into a HAMA microneedle matrix with geometric parameters tailored to match the thickness of the rat dura mater (300–500 µm). Universal mechanical testing demonstrated that the MN‐Mg system possesses a Young's modulus approaching 1 MPa, indicating excellent mechanical strength. The microneedles can fully penetrate the dura mater, creating distinct indentations and ensuring precise delivery of active components to the injury site. The rationale for selecting a microneedle height of 600 µm, despite the rat thoracic dura mater thickness being approximately 150–200 µm, lies in the mechanical challenge of penetrating soft, viscoelastic tissues. The ‘bed of nails’ effect causes significant tissue deflection during the application of microneedle arrays, reducing the effective insertion depth by up to 40%–50%. Therefore, the 600 µm design ensures sufficient penetration into the subarachnoid space to deliver the therapeutic payload. Importantly, concerns regarding parenchymal damage are mitigated by the material properties of the HAMA matrix. Unlike metal or silicon microneedles, the MN‐Mg patches undergo a rapid modulus transition upon hydration, softening to match the stiffness of the neural tissue. This ‘smart’ softening behavior, combined with the conical geometry that separates rather than transects fibers, minimizes the risk of mechanical disruption to the blood‐spinal cord barrier (BSCB) or dorsal horn neurons, as evidenced by the lack of hemorrhage or cavity formation in our acute‐phase histological analysis. The MN‐Mg system achieves in situ H₂ generation via a confined hydrolysis reaction (Mg + 2H₂O → Mg(OH)₂ + H₂↑) and utilizes the HAMA 3D network to physically adsorb H₂ and establish a diffusion gradient, thereby significantly extending its half‐life. Kinetic experiments revealed that MN‐Mg released a cumulative total of 84.6 ± 1.04 µmol of H₂ after 7 days of immersion in PBS (pH 7.4, 37°C), which was significantly higher than the release from pure Mg particles. The cumulative release of Mg^{2 +} reached 11.60 ± 0.04 mg, also significantly higher than that from pure Mg particles (3.10 ± 0.22 mg; p < 0.05). This enhanced release is attributed to the inhibition of the Mg(OH)₂ passivation layer: SEM‐EDS analysis showed that after 3 or 7 days of immersion in deionized water, the surface of pure Mg particles exhibited oxygen accumulation (O₂, red), indicating the formation of a Mg(OH)₂ shell. In contrast, the O₂ content on the surface of Mg particles within the MN‐Mg system was significantly lower, demonstrating that the microneedle system reduces passivation layer formation through spatial confinement, thereby promoting direct contact between Mg and water. Consequently, MN‐Mg greatly improves the utilization efficiency of loaded Mg, allowing the same volume of Mg particles to release greater quantities of H₂ and Mg^{2 +}, which actively participate in immunomodulation and neural repair following SCI. This results in dual synergistic effects:

1. Antioxidant and anti‐inflammatory effects of H₂: The in situ sustained release of H₂ efficiently scavenges hydroxyl radicals (·OH), inhibits lipid peroxidation, and blocks key inflammatory cascades, thereby mitigating the neuroinflammatory response. Although the primary oxidative burst typically peaks within the first 24–72 h post‐injury, persistent oxidative stress in the sub‐acute phase serves as a critical barrier to neural regeneration. Aligning with the sustained hydrogen release kinetics of our MN‐Mg system, the significant reduction of ROS observed at Day 7 confirms the material's capability to combat this long‐term oxidative damage. This sustained antioxidative efficacy is essential for preventing the transition from acute inflammation to chronic neurotoxicity, thereby creating a permissive microenvironment for subsequent tissue repair.

2. Immunomodulatory and neurorestorative functions of Mg^{2 +}: Sustained‐release Mg^{2 +} modulates immune cell activity, promotes M2 microglial polarization, and simultaneously accelerates axonal regeneration, remyelination, and angiogenesis. This spatiotemporal synergy of dual active components constitutes a dynamic “antioxidant‐anti‐inflammatory‐pro‐regenerative” regulatory network.

It is important to acknowledge the inherent challenge in experimentally decoupling the individual contributions of H₂ and Mg^{2 +} within a magnesium‐based hydrolysis system, as they are generated simultaneously [36]. While a strict “H₂‐only” or “Mg^{2 +}‐only” control group with identical release kinetics and mechanical properties is technically unfeasible, the therapeutic efficacy of the MN‐Mg system likely stems from the complementary mechanisms of these two bioactive agents. Existing evidence suggests a temporal and functional division of labor: H₂, as a highly permeable gas, rapidly diffuses to scavenge toxic free radicals in the acute phase of SCI. This “cleansing” of the oxidative microenvironment is a prerequisite for regeneration. Concurrently and subsequently, the sustained release of Mg^{2 +} acts to block NMDA receptors, reducing excitotoxicity, and promotes downstream signaling pathways associated with axonal elongation and synaptic plasticity. Therefore, rather than a simple additive effect, the MN‐Mg system functions as an integrated therapeutic platform where H₂ mitigates the hostile pathological barriers (inflammation/oxidation), thereby maximizing the regenerative potential of Mg^{2 +}. Future studies utilizing advanced gene‐editing tools to selectively block H₂ or Mg^{2 +} downstream targets may further dissect this complex interplay.

Within the complex regulatory network of neuroinflammation following SCI, this study identified AP‐1 within the MAPK pathway as a core molecular hub through which H₂ mitigates microglial oxidative stress and suppresses the inflammatory cascade. Our multifaceted experiments confirmed that H₂ significantly inhibited the activation of the JNK and p38 sub‐pathways, without exerting a significant effect on the ERK pathway. This selective regulation aligns with the central role of JNK/p38 in mediating oxidative stress responses, resulting in a marked reduction in c‐Fos and c‐Jun protein levels. As AP‐1 is a heterodimeric transcription factor composed of c‐Fos and c‐Jun, its reduced activity directly prevents the transcription of downstream pro‐inflammatory genes (e.g., Il6, Cxcl1). Crucially, the rescue experiment using an AP‐1 agonist provided mechanistic insight into the “ROS‐Inflammation Positive Feedback Loop.” We observed that artificial activation of AP‐1 not only reversed the anti‐inflammatory effects of H₂ but also significantly restored high ROS levels. This suggests that H₂ does not merely act as a passive scavenger; it disrupts a vicious self‐perpetuating cycle. Under pathological conditions, initial ROS activates AP‐1, which drives the expression of pro‐inflammatory cytokines. These cytokines, in turn, can induce secondary ROS generation—likely via the upregulation of NADPH oxidases (NOX) or the induction of mitochondrial dysfunction—thereby further fueling oxidative stress. By blocking the AP‐1 node, H₂ effectively breaks this loop. The rescue experiment confirms that reactivating AP‐1 reignites this inflammatory cascade, leading to the re‐accumulation of ROS and the repolarization of microglia toward the neurotoxic M1 phenotype. It is important to acknowledge the pharmacological complexity of the rescue experiment. We utilized 15(S)‐HPETE to reactivate the AP‐1 pathway. While 15(S)‐HPETE is a lipid metabolite with potential pleiotropic effects (including PPAR activation), our in vivo results demonstrated that its administration significantly re‐elevated ROS levels and exacerbated histological damage. This indicates that in the acute microenvironment of SCI, the pro‐inflammatory and pro‐oxidant signaling triggered by 15(S)‐HPETE (via AP‐1 and downstream inflammatory cascades) dominated over any potential cytoprotective effects. While this supports the critical role of the MAPK/AP‐1 axis identified in our RNA‐seq analysis, future studies using more specific genetic modulation (e.g., AP‐1 knock‐in or overexpression vectors) would be valuable to dissect the precise molecular hierarchy.

This mechanistic discovery endows the H₂‐releasing, magnesium‐based nanosystem (MN‐Mg) with breakthrough advantages in AP‐1–targeted therapy. Compared to traditional small‐molecule AP‐1 inhibitors, this system achieves a threefold synergistic effect via H₂:

1. Precision and Selectivity: It selectively inhibits oxidative stress–activated AP‐1 subtypes, avoiding global transcriptional interference.

2. Upstream and Downstream Blockade: It concurrently scavenges ROS (the upstream initiator) and inhibits AP‐1 activity (the downstream effector), effectively severing the feedback loop at both ends.

3. Biological Safety: It replaces chemical inhibitors with an endogenous gas molecule (H₂), significantly reducing the risk of off‐target toxicity.

The origin of the increased Nestin‐positive neural stem/progenitor cells (NSPCs) observed in the MN‐Mg group is a critical aspect of the regenerative process. Based on the spatial distribution of these cells—clustering around the central canal and extending toward the lesion epicenter—we postulate that the MN‐Mg system facilitates a coupled mechanism of local proliferation followed by migration. Endogenous ependymal cells lining the central canal represent a primary reservoir of NSPCs in the spinal cord. However, in the hostile microenvironment of typical SCI (characterized by high ROS and pro‐inflammatory cytokines), these cells often fail to proliferate or differentiate into functional neurons. Our data suggests that the sustained release of H₂ from MN‐Mg effectively scavenges ROS and mitigates the inflammatory storm, thereby protecting the NSC niche and preserving the proliferative capacity of endogenous ependymal cells. Furthermore, the distinction between proliferation and migration is bridged by the physical properties of the scaffold. The HAMA microneedles likely act as a physical guidance cue, while the remodeled, less dense glial scar (discussed previously) reduces the mechanical and chemical barriers to migration. Consequently, the Nestin‐positive cells are not only triggered to proliferate at the central canal but are also enabled to migrate into the lesion site to participate in tissue repair. This suggests that MN‐Mg creates a ‘permissive highway’ that mobilizes endogenous regenerative potential.

While the glial scar has traditionally been viewed as a physical and chemical barrier to axonal regeneration, recent studies emphasize its dual role in spinal cord injury repair. In the acute phase, reactive astrocytes form a protective barrier that seals the lesion core, restricting the spread of inflammatory cells and toxic debris. However, in the subacute and chronic phases, a dense, permanent scar can physically impede axon growth. Therefore, the therapeutic goal should be to remodel the scar into a permissive state rather than to completely inhibit its formation [37]. Our results showed that MN‐Mg treatment significantly reduced the density of the dense glial scar (GFAP⁺ area) while promoting axonal outgrowth (NF200⁺). This effect is likely mediated through the modulation of astrocyte phenotypes. Reactive astrocytes are heterogeneous and can be polarized into two distinct phenotypes: the neurotoxic A1 phenotype, induced by NF‐κB and MAPK signaling, and the neuroprotective A2 phenotype, which secretes neurotrophic factors [38]. Oxidative stress and the MAPK/AP‐1 pathway are known drivers of A1 astrocyte activation [39]. By continuously releasing H₂ to scavenge ROS and inhibit the MAPK/AP‐1 axis (as confirmed by our RNA‐seq and rescue experiments), MN‐Mg may prevent the excessive polarization of astrocytes toward the A1 phenotype. Instead, it likely promotes a shift toward the A2 phenotype or a more ‘pro‐regenerative’ state, thereby attenuating the formation of a dense, inhibitory scar while preserving the beneficial functions of astrocytes. This ‘scar remodeling’ strategy creates a permissive microenvironment that facilitates the penetration of regenerating axons across the lesion site.

This differentiated strategy suggests a promising avenue for the treatment of neuroinflammatory diseases driven by oxidative stress [40, 41]. Given the central role of the AP‐1 complex in propagating neurodegenerative pathologies, the ability of our system to modulate this specific axis warrants further investigation in broader contexts. For instance, in Alzheimer's disease models, literature suggests that targeted inhibition of AP‐1 may reduce the expression of enzymes involved in amyloid plaque deposition. Similarly, in multiple sclerosis, blocking AP‐1–mediated upregulation of matrix metalloproteinases could theoretically help maintain blood–brain barrier integrity. While direct evidence is pending, the mechanism demonstrated here—suppressing AP‐1–activated microglial cytotoxicity—offers a conceptual framework that might be applicable to protecting dopaminergic neurons in Parkinson's disease [42].

Despite the promising results, this study has several limitations that must be addressed before clinical translation. First, the safety and efficacy were evaluated in a rodent model over an 8‐week period. While this is a standard endpoint for assessing functional recovery in rats, it does not fully capture the long‐term biocompatibility required for human implants. Specifically, although we did not observe gas cavity formation or compression cysts in our histological analysis, the long‐term effects of chronic H₂ release and Mg^{2 +} accumulation in a larger, closed spinal canal remain to be verified. Future studies should utilize large animal models (e.g., pigs or non‐human primates) with extended follow‐up periods (>6 months) to rigorously assess systemic Mg^{2 +} levels, potential gas entrapment, and long‐term material degradation. Second, while we identified the MAPK/AP‐1 axis as a key target using RNA‐seq and rescue experiments, the precise molecular interplay—such as how H₂ differentially modulates specific AP‐1 dimer combinations (e.g., c‐Fos/c‐Jun vs. JunB/JunD)—requires further elucidation. Future directions include developing ‘smart’ responsive materials that trigger release only within the acidic microenvironment of the injury site, and constructing co‐delivery systems that integrate gene‐silencing technologies to enhance the suppression of refractory inflammation. This study serves as a proof‐of‐concept for a multi‐target neurorestorative strategy, laying a theoretical foundation for next‐generation SCI therapies.

4. Conclusion

This study confirms that H₂ inhibits AP‐1 transcriptional activity via the MAPK pathway and establishes AP‐1 as the central target linking H₂‐mediated blockade of oxidative stress to attenuation of the neuroinflammatory storm. Through the spatiotemporal synergistic delivery of H₂ and Mg^{2 +}, the MN‐Mg system constructs a three‐tiered synergistic regulatory network:

1. Upstream Level: H₂ scavenges ROS and inhibits JNK/p38 phosphorylation, blocking c‐Fos/c‐Jun activation;

2. Core Level: AP‐1 inactivation directly suppresses transcription of pro‐inflammatory mediators such as IL‐6 and CXCL1, thereby remodeling the M1/M2 microglial balance.

3. Microenvironmental Level: Mg^{2 +} synergistically enhances expression of neuroregeneration‐associated proteins, accelerating tissue repair.

This dynamic “antioxidant‐anti‐inflammatory‐pro‐regenerative” synergistic mechanism overcomes the therapeutic impasse of the inflammation‐regeneration imbalance in SCI and elevates AP‐1–targeted therapy to a new paradigm in neuro‐immunomodulation. This discovery not only provides a novel perspective on how gas molecules regulate epigenetic networks but also inspires future development of targeted therapies for neuroinflammation by modulating AP‐1 dimer specificity or chromatin accessibility. By integrating this approach with smart responsive materials or cell‐based therapies, the efficiency of neural function recovery may be further enhanced.

5. Experimental Section

5.1. Fabrication of Microneedles

Preparation of HAMA Hydrogel Precursor Solution: A 0.25% (w/v) photoinitiator solution was prepared by dissolving lithium phenyl‐2,4,6‐trimethylbenzoylphosphinate (LAP, 0.025 g, EFL‐Biotech, Suzhou, China) in 10 mL of phosphate‐buffered saline (PBS) within a light‐protected brown bottle. The dissolution was facilitated by heating the mixture in a water bath at 40°C–50°C for 15 min with intermittent vortexing. Subsequently, hyaluronic acid methacrylate (HAMA, 0.5 g, 150 kDa, EFL‐Biotech) was added to the LAP solution. The mixture was stirred at room temperature for 0.5–1 h until a homogenous 5% (w/v) HAMA precursor solution was formed. The final solution was sterilized by passing it through a 0.22 µm syringe filter (Millipore, Burlington, MA, USA) and stored at 4°C in the dark before use.

Polydimethylsiloxane (PDMS) microneedle molds were utilized for fabrication. An aliquot of 600 µL of the 5% (w/v) HAMA precursor solution was cast into each mold. The molds were then placed in a vacuum desiccator at room temperature and subjected to 2–3 cycles of degassing to remove air bubbles. Following degassing, the molds were transferred to an oven and dried at 30°C–35°C for 5–6 h to evaporate the majority of the solvent. Three types of MNs were fabricated: MN‐HAMA: The hydrogel‐filled mold was directly exposed to 405 nm UV light for 30 s to initiate photocrosslinking. The resulting MN‐HAMA patch was then carefully demolded. MN‐Mg and MN‐MgCl₂: Elemental magnesium (Mg) powder (40 µm, Sigma‐Aldrich Sigma‐Aldrich) or magnesium chloride (MgCl₂, Sigma‐Aldrich) was uniformly distributed onto the surface of the partially dried hydrogel within the molds. The molds were further dried in an oven at 30°C–35°C for an additional 5–6 h to facilitate the sedimentation of particles into the needle tips. The hydrogel was subsequently crosslinked by exposure to 405 nm UV light for 30 s, followed by demolding. After demolding, all MN patches were equilibrated at ambient humidity for 1 h, and then re‐dried in an oven at 30°C–35°C for 2 h to ensure structural stability.

5.2. Material Characterization

1. Scanning Electron Microscopy (SEM): The morphology and dimensions of the dried MNs from each group were examined using a scanning electron microscope (SEM, Zeiss GeminiSEM 300, Carl Zeiss, Germany). Prior to imaging, samples were sputter‐coated with a thin layer of gold for 60 s.

2. In Vitro Hydrogen (H₂) and Magnesium Ion (Mg^{2 +}) Release Assay: Hydrogen (H₂) Release Assay: To quantify H₂ generation, 3 mg of Mg powder or 5 MN‐Mg patches were placed in separate sealed penicillin bottles (15 mL volume) containing 10 mL of distilled water (pH 7.0). At predetermined time points (0, 1, 2, 3, 4, 5, 6, and 7 days), 1 mL of gas was collected from the headspace of each bottle. The H₂ concentration was measured using gas chromatography (GC, Agilent 8890 with a thermal conductivity detector (TCD).Magnesium Ion (Mg^{2 +}) Release Assay: To measure Mg^{2 +} release, 3 mg of Mg powder or 5 MN‐Mg patches were enclosed in dialysis bags. The bags were immersed in sealed penicillin bottles containing 10 mL of distilled water (pH 7.0). At the same designated time points, the external solution was collected, and the concentration of Mg^{2 +} was determined using an inductively coupled plasma mass spectrometer (ICP‐MS, Agilent 7800).

3. Scanning Transmission Electron Microscopy with Energy Dispersive X‐ray Spectroscopy (STEM‐EDS) Analysis: The elemental composition and distribution on the surface of Mg powder and MN‐Mg patches were analyzed. Samples were soaked in deionized water for 3 and 7 days, respectively, followed by lyophilization. The dried samples were dispersed in deionized water and deposited onto carbon‐coated copper grids. Analysis was performed using a scanning transmission electron microscope (STEM, FEI Talos F200X) equipped with an energy‐dispersive X‐ray spectroscopy (EDS) detector. Elemental maps were generated for oxygen (O, red) and magnesium (Mg, green).

4. X‐ray Photoelectron Spectroscopy (XPS): XPS analysis (Thermo Scientific K‐Alpha) was conducted to investigate the chemical valence states and binding energies of Mg and O elements in the MN‐Mg and elemental Mg samples. This analysis aimed to characterize the chemical bonding environment of magnesium.

5. Mechanical Testing: The elastic modulus of the microneedles from each group was measured using a universal testing machine.

6. Penetration Test: To visualize tissue penetration, Methylene Blue (0.1% w/v, Sigma‐Aldrich) was incorporated into the HAMA precursor solution. After fabrication, the Methylene Blue‐loaded MNs were gently pressed into freshly excised ex vivo rat spinal cord tissue for 5 min. Following removal of the patch, the tissue was observed under a stereomicroscope to visualize the deposition of blue dye and identify any residual microneedle fragments, confirming successful penetration and delivery.

5.3. Cell Culture

Primary rat neural stem cells (NSCs), isolated from the brains of E14 Sprague‐Dawley rat fetuses, were purchased from Hasenbio (Wuxi, China). NSCs were cultured in Rat Neuron Complete Medium (Hasenbio) and passaged every 7–10 days. The medium was half‐replaced every 3 days. The RAW 264.7 murine macrophage cell line was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). RAW 264.7 cells were cultured in high‐glucose Dulbecco's Modified Eagle Medium (DMEM, Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin‐streptomycin (Gibco). The medium was changed every 2–3 days. All cells were maintained in a humidified incubator at 37°C with 5% CO₂.

For indirect co‐culture experiments, materials (elemental Mg or MN patches) were placed in the upper chamber of a Transwell insert (0.4 µm pore size, Corning, NY, USA). RAW 264.7 cells or NSCs were seeded in the lower chamber of the well plate. The insert was then placed into the well, allowing for the exchange of soluble factors through the porous membrane.

1. In Vitro Biocompatibility: RAW 264.7 cells were seeded in the lower chamber of a 24‐well Transwell plate at a density of 1 × 10⁵ cells/well. Materials from each experimental group were placed in the upper inserts. After 3 days of co‐culture, cell viability was assessed using a LIVE/DEAD Viability/Cytotoxicity Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Cells were incubated with the staining solution for 15 min at 37°C in the dark. Fluorescence images of live (green, calcein‐AM) and dead (red, ethidium homodimer‐1) cells were captured using a fluorescence microscope (Axio Observer, Carl Zeiss, Germany).

2. In Vitro ROS Monitoring: RAW 264.7 cells were co‐cultured with materials as described above. To mimic an inflammatory environment, all groups except the Blank group were stimulated with 1 µg/mL lipopolysaccharide (LPS, from E. coli O111:B4, Sigma‐Aldrich). After 24 h of co‐culture, intracellular ROS levels were measured using a DCFH‐DA fluorescent probe (Beyotime, Shanghai, China). Cells were incubated with 10 µM DCFH‐DA for 30 min at 37°C. Fluorescence intensity, corresponding to ROS levels, was imaged using a laser scanning confocal microscope (LSCM, Leica TCS SP8, Leica, Germany).

3. Immunofluorescence (IF): Cells or tissue sections were fixed in 4% paraformaldehyde (PFA) for 20 min and then permeabilized with 0.2% Triton X‐100 (Sigma‐Aldrich, USA) for 15 min. Samples were subsequently blocked with 5% BSA (Beyotime, China) overnight. Primary antibodies (including anti‐CD86, anti‐CD206 (ab64693, Abcam, USA), anti‐iNOS (ab178945, Abcam, USA), anti‐Iba‐1 (ab6640, Abcam, USA), anti‐GAP43 (ab75810, Abcam, USA), anti‐MAP2 (ab5392, Abcam, USA), anti‐Tau, and anti‐Integrin β) were added and incubated at 4°C overnight. Secondary antibodies and DAPI were incubated with the samples at 37°C for 1 h. Before each step, samples were washed three times with PBS for 5 min each. Images were captured using a Leica confocal microscope.

4. Flow Cytometry: RAW 264.7 cells from the Sham, Control, MN‐HAMA, MN‐MgCl₂, MN‐Mg, and MN‐Mg + AP1 agonist groups were collected. Samples were incubated with fluorescently labeled antibodies: CD11b (Biolegend, 201805, 1:200), CD86 (Biolegend, 200307, 1:200), and CD206 (Santa Cruz, sc‐58986, 1:200) at 4°C for 30 min, followed by fixation with 1% PFA. The proportions of CD11b⁺/CD86⁺ and CD11b⁺/CD206⁺ RAW 264.7 cells were analyzed using a BD FACSAria III flow cytometer. Data were analyzed with FlowJo 7.6 software by first gating on CD11b‐positive cells and then gating on CD86‐positive and CD206‐positive subpopulations.

5. Quantitative Real‐Time PCR (qRT‐PCR): RAW 264.7 cells were co‐cultured with materials and stimulated with LPS (1 µg/mL). At 1, 3, 5, and 7 days, total RNA was extracted using the RNeasy Mini Kit (Qiagen, Hilden, Germany). cDNA was synthesized using the iScript cDNA Synthesis Kit (Bio‐Rad, Hercules, CA, USA). qRT‐PCR was performed on a QuantStudio 7 Flex Real‐Time PCR System (Applied Biosystems, Foster City, CA, USA) using SYBR Green Master Mix (Applied Biosystems). The expression levels of pro‐inflammatory (TNF‐α) and anti‐inflammatory (IL‐10) genes were analyzed. GAPDH was used as the housekeeping gene for normalization. The 2⁻ΔΔCt method was used for relative quantification. All primer sequences are listed in Table 1.

TABLE 1.

Primer sequences of each gene was showed below.

Target	Forward	Reverse
IL‐10	ACTGCTATGTTGCCTGCTCTTAC	TGTGGGTCTGGCTGACTGG
TNF‐α	CCTCTCTCTAATCAGCCCTCTG	GAGGACCTGGGAGTAGATGAG
actin	CACCCGCGAGTACAACCTTC	CCCATACCCACCATCACACC

5.4. Animal Experiments

1. Rat SCI Model Establishment and Material Implantation: Female Sprague‐Dawley (SD) rats (approx. 220 g) were purchased from Xuzhou Medical University and randomly divided into 5 groups (n = 6 per group). Animal care, all surgical procedures, and perioperative treatments were performed according to protocols approved by the Ethics Committee of Xuzhou Medical University. An SCI model was established using a vascular clip compression method. Rats (6–8 weeks old) were anesthetized by intraperitoneal injection of 2% pentobarbital (50 mg/kg). A 2 cm longitudinal incision was made over the T9 vertebra. Paraspinal muscles were bluntly dissected, and a laminectomy was performed at T9 using small rongeurs to expose the spinal cord. A vascular clip was inserted from both sides to compress the spinal cord for 8 s. The area was rinsed with saline, and any blood or fluid was dried with a cotton ball. MN‐HAMA, MN‐MgCl₂, or MN‐Mg patches were implanted. The fascia and skin were sutured in layers. Penicillin (2 × 10⁵ units) was administered intramuscularly for seven consecutive days. Bladders were manually emptied every 12 h until normal urination was restored. The sham group underwent only laminectomy to expose the spinal cord. The control group received SCI but no material implantation.

2. In Vivo Biocompatibility: To assess local tissue reaction, MN patches from each group were implanted subcutaneously on the backs of healthy SD rats. After 14 days, the skin tissue surrounding the implant was excised, fixed in 4% PFA, embedded in paraffin, and sectioned. Sections were stained with Hematoxylin and Eosin (H&E) for histopathological analysis. The density of inflammatory cells around the implant site was semi‐quantified using ImageJ software (NIH, Bethesda, MD, USA).

3. Motor Function Assessment: Hindlimb motor function recovery was evaluated using the Basso, Beattie, and Bresnahan (BBB) motor function score and footprint analysis. The BBB score ranges from 0 (no hindlimb movement) to 21 (completely normal motor function). Assessments were performed weekly post‐surgery by two blinded investigators. The Inclined Plane Test (IPT) score was also used to increase the specificity and sensitivity of the experiment. For footprint analysis, the forelimbs and hindlimbs of the rats were dipped in red and black ink, respectively. The rats were then allowed to walk along a narrow strip of white paper. Footprints were evaluated by measuring parameters such as hindlimb rotation angle, relative position of forelimbs and hindlimbs, stride length, and base of support (distance between left and right hindlimbs).

4. Immunohistochemistry (IHC) and Pathology: Ten rats per group were randomly assigned to 4‐ and 8‐week time points. At each time point, rats were euthanized by injection of 2% pentobarbital and then intracardially perfused with saline followed by 4% PFA. Spinal cord and organ tissues were harvested, embedded, and cryosectioned. Coronal sections (5 µm) of the spinal cord were stained with H&E and observed under a bright‐field microscope. The spinal cord cavity area was measured using ImageJ software, with three random fields of view selected for statistical analysis (n = 3). For IHC, frozen sections were permeabilized and blocked. They were then incubated with primary antibodies overnight at 4°C, including anti‐Iba‐1, anti‐iNOS, anti‐Arg‐1, anti‐FosB, anti‐c‐Jun, anti‐GFAP, anti‐Tuj‐1, anti‐MBP, anti‐NFH, anti‐GAP‐43, anti‐Nestin, and anti‐CD31. After incubation with secondary antibodies at 37°C for 1 h and DAPI staining for 15 min, images were captured by microscopy.

5. Transcriptome Analysis: 3 days post‐injury, rats (n = 3 per group) from the MN‐MgCl₂ and MN‐Mg groups were euthanized. A 5 mm segment of spinal cord tissue at the T9 injury site was harvested, immediately flash‐frozen in liquid nitrogen, and stored at −80°C. Total RNA was extracted, and library preparation and sequencing were performed by BGI (Wuhan, China). Differentially expressed genes (DEGs) were identified using the DESeq2 package in R. Genes with a fold change > 2 and a p‐value < 0.05 were considered significant. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed on the DEGs.

6. Western Blotting: Total protein was extracted from cells using RIPA lysis buffer (Beyotime, China) containing protease inhibitors (Beyotime, China) and phosphatase inhibitors (Beyotime, China). Protein concentration was determined using a BCA Protein Assay Kit (Solarbio, China). Equal amounts of protein (10 µg) were separated by 10% SDS‐PAGE (Beyotime, China) and transferred to polyvinylidene fluoride (PVDF) membranes. After blocking with 5% non‐fat milk for 1 h, membranes were incubated with primary antibodies overnight at 4°C, including anti‐CD206 (ab64693, Abcam, USA), anti‐iNOS (ab178945, Abcam, USA), anti‐IL‐1β (32165, SAB, USA), anti‐Tuj‐1 (A17913, ABclonal, China), anti‐GFAP (ab7260, Abcam, USA), anti‐NFH (ab207176, Abcam, USA), anti‐GAP43 (ab75810, Abcam, USA), and anti‐MAP2 (ab5392, Abcam, USA). The next day, membranes were incubated with secondary antibodies for 1 h. Protein bands were visualized using an Enhanced Chemiluminescence (ECL) kit (Beyotime, China) and analyzed with ImageJ software.

5.5. Statistical Analysis

The normality of the data distribution was assessed using the Shapiro‐Wilk test. Data are presented as the mean ± standard deviation (SD). The specific sample size (n) for each experimental group is indicated in the corresponding figure legends. Statistical comparisons between multiple groups were performed using one‐way analysis of variance (ANOVA) followed by Tukey's post hoc test for multiple comparisons. All statistical tests were two‐sided, and a p‐value < 0.05 was considered statistically significant. Statistical analyses and graph generation were performed using Prism 9.0 or Origin 2021 software.

Funding

National Natural Science Foundation of China (82120108017, 82272501, 82572769, 82372484, 82302683); Project 333 of Jiangsu Province (2069999); the Natural Science Foundation of Jiangsu Province (BK20250123, BK20230215); China Postdoctoral Science Foundation (2025M782215); Suzhou Basic Research Pilot Program: Outstanding Young Scholars Lead Project (SSD2025025); Suzhou Gusu Health Talent Program (GSWS2023093); Soochow University Medical +X project (ML12202923); Key Laboratory of Spine and Spinal Cord Injury Repair and Regeneration (Tongji University), Ministry of Education (No. BZ2024003);A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.

All animal experiments were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) and approved by the Laboratory Animal Management and Ethics Committee (Approval No. IACUC‐2022095048).

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File: smll72828‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

All data associated with this study are present in the paper or the Supplementary Materials. Any additional information required to reanalyze the data reported here is available from the corresponding authors upon reasonable request.