Repair of spinal cord injury using a time-specific four-dimensional multifunctional hydrogel with anti-inflammatory and neuronal differentiated microenvironments

Abstract

Spinal cord injury (SCI) is a severe central nervous system (CNS) condition that often leads to permanent disability. The repair of SCI presents significant challenges globally, primarily due to serious inflammatory damage in the early stage and limited neural regeneration in the long-term stage. In response to these challenges, this study developed a novel time-specific four-dimensional multifunctional SilMA hydrogel (4DMSH) that releases Houttuynia cordata extract (HCT) in the early stage of post-implantation to combat inflammation and a sustained release of neurotrophin-3 (NT-3) in the long-term stage to promote neuronal differentiation of endogenous neural stem cells (eNSCs) for neuronal regeneration. As expected, the time-specific 4DMSH significantly mitigated inflammatory responses, leading to a shift from a pro-inflammatory to a neural regenerative environment, and enhanced the differentiation of eNSCs into neurons, thereby effectively improving the recovery of motor, sensory, and autonomic functions after SCI. Therefore, this study presents a novel time-specific 4DMSH that creates anti-inflammatory and neuroactive microenvironments, contributing to efficient neuronal regeneration and SCI repair.

---

1. Introduction

Spinal cord injury (SCI) is a form of severe central nervous system (CNS) damage¹ resulting from trauma (e.g., traffic accidents or falls) or compressive lesions (e.g., myelitis or tumors),² which can lead to severe sensory, motor, and autonomic dysfunction, as well as complications such as pulmonary atelectasis, pneumonia, venous thrombosis, and depression. These issues not only drastically reduce the quality of life for patients but also impose a substantial economic burden on families and society.^3–5 However, achieving efficient neuronal regeneration and functional recovery after SCI remains a considerable challenge worldwide.

Intense inflammation in the early stage after SCI results in extensive neural cell death and significantly impairs the physiological functions of the spinal cord, leading to a state of severe injury and making recovery extremely difficult.^6,7 Briefly, microglia, the resident immune cells of the CNS, are rapidly activated⁸ after SCI and secrete various pro-inflammatory factors, including tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and interleukin-1β (IL-1β), during the early stage,^9–11 which exacerbate local tissue damage¹² and create an inhibitory microenvironment of neural regeneration. Therefore, restricting early inflammatory responses caused by activated microglia following SCI is critical for protecting undamaged neural cells and promoting SCI repair. In addition, insufficient neuronal regeneration and supplementation is another important reason for the difficult recovery of neurological function after SCI.¹³ Endogenous neural stem cells (eNSCs) differentiate into neurons to repair damaged neural functions.¹⁴ However, insufficient neuronal differentiation of eNSCs in the long-term stage after SCI presents a significant challenge for SCI repair.¹⁵ Hence, enhancing the activation of eNSCs and increasing the proportion of eNSCs that differentiate into neurons are key approaches for achieving satisfied SCI repair.¹⁶ Furthermore, traditional anti-inflammatory and eNSC neuronal differentiation methods lack time-specific and synergetic functions, which restrict SCI repair and neural functional recovery.¹⁷ Notably, although the application of exogenous neural stem cells (eNSCs) has demonstrated their potential in neuronal differentiation and SCI repair, the source of eNSCs severely limits the possibility of their clinical application. Therefore, it is urgent to develop a cell-free time-specific anti-inflammatory and neuronal differentiation strategy for serious inflammatory damage in the early stage and the limitation of neural regeneration in the long-term stage.¹⁸

In SCI repair, hydrogels serve as three-dimensional scaffolds¹⁹ that mimic the extracellular matrix,²⁰ enabling sustained drug release over extended periods through gradual degradation.^21–23 SilMA hydrogel is recognized for its excellent biocompatibility and biodegradation properties,^24–26 effectively mimicking the extracellular matrix,²⁷ and creating an optimal three-dimensional environment that enhances the survival, proliferation, and differentiation of NSCs.^28,29 However, the lack of anti-inflammatory and neural induction microenvironments limits its application in SCI repair. Houttuynia cordata extract (HCT), a natural compound derived from traditional Chinese medicine, is noted for its potent anti-inflammatory properties that can significantly reduce the release of pro-inflammatory factors and rectify the imbalanced inflammatory microenvironment.^30,31 The mechanism by which HCT has an anti-inflammatory effect has been widely demonstrated, and many experiments have shown that HCT can effectively down-regulate inflammatory mediators, including IL-6, IL-8, COX-2, and INOS,^32,33 which may have anti-inflammatory effects on microglial cell polarisation after spinal cord injury. Additionally, neurotrophin-3 (NT-3) is a critical neurotrophic factor that promotes the survival of NSCs, facilitates neuronal differentiation, and supports axonal growth, playing a vital role in the later stages of SCI repair.^34,35 It has been demonstrated that NT-3 prevents neuronal apoptosis and stimulates intrinsic regenerative capacity within neurons through activation of the PI3K/AKT and MAPK/ERK signalling pathways via the TrkC receptor.^36–39

To address the aforementioned challenges, NT-3 was incorporated into PLGA to prepare microspheres for long-term release and neuronal induction. Subsequently, we integrated PLGA microspheres and HCT into a SilMA hydrogel, thereby constructing a novel cell-free, time-specific, four-dimensional multifunctional SilMA hydrogel (4DMSH) (Fig. 1A). This strategy involves the early release of HCT to mitigate inflammation and the sustained release of NT-3 to promote neural regeneration. Specifically, the loose network of 4DMSH was designed to facilitate the rapid release of HCT post-implantation, effectively regulating the intense inflammation observed during the early stages of SCI. In contrast, NT-3 was slowly released from the PLGA microspheres to ensure the prolonged promotion of neural regeneration. As anticipated, this strategy could significantly reduce inflammation, promote neuronal regeneration, and enhance functional recovery after SCI (Fig. 1B).

Fig. 1 Schematic of 4DMSH preparation and its use of 4DMSH in the treatment of spinal cord injuries. (A) Schematic of SilMA hydrogel, PLGA microspheres containing NT-3 and 4DMSH preparation. (B) Schematic of 4DMSH-treated spinal cord injury mice.

2. Materials and methods

2.1 Experimental animals

The experimental animals used in this study were female C57BL/6J mice aged 6–8 weeks and weighing 18–25 g, which were procured from Jiangsu Jicui Yaokang Co., Ltd. Mice were maintained in a specific pathogen-free (SPF) environment at the Laboratory Animal Center of Shandong University. They had unrestricted access to food and water, with environmental temperature regulated between 20 °C–25 °C and relative humidity maintained at 40%–60%. All animal handling and experimental procedures were conducted in accordance with the “Regulations on the Administration of Laboratory Animals” issued by the National Science and Technology Commission and strictly followed the guidelines established by the Ethics Committee of Shandong University. Approval for all animal procedures was granted by the Ethics Committee of Shandong University (approval no. ecsbmssdu24019). All personnel involved in animal handling received training to ensure compliance with ethical standards and animal welfare.

2.2 Preparation of SilMA hydrogel and PLGA microspheres

The SilMA hydrogel used in this study was procured from Engineering For Life (Cat# EFL-SilMA-001). These hydrogels were processed through a series of steps including dissolution, filtration, and crosslinking via UV light, rendering them suitable for applications both in vitro and in vivo.

SilMA hydrogel preparation. First, a 0.25% (w/v) initiator stock solution is prepared by dissolving 0.05 g of LAP (lithium phenyl-2,4,6-trimethylbenzoylphosphinate) (Solarbio, Cat#YS148350) in 20 ml of PBS (Solarbio, Cat#P1022), stirring the mixture in a water bath at 40 °C–50 °C for 1 hour while protecting it from light. The stock solution was then stored at 4 °C. Next, the SilMA solution was prepared by weighing the required amount of dry SilMA using an electronic balance, placing it in a centrifuge tube, and adding the initiator stock solution. The mixture was stirred at room temperature for 1 hour to obtain an 8% SilMA solution, which was stored at 4 °C. Once the 8% SilMA solution had been fully dissolved, it was filtered using a sterile 0.22 μm needle filter in a biosafety cabinet. The solution is subsequently injected into multi-well plates (50 μL per well for 96-well plates, 100 μL per well for 48-well plates, and 300 μL per well for 24-well plates). The solution was exposed to a 405 nm light source for 30 seconds to induce gelation.^40,41 To ensure complete coverage, the culture medium was added to the wells and incubated at 37 °C for 5 minutes; thereafter, the medium was aspirated, and the wells were washed with PBS. Cell suspensions were then added for culture. For hydrogels intended for implantation, the filtered solution was crosslinked with light and trimmed to the appropriate size.

PLGA microsphere preparation. To prepare the aqueous phase, 5 mg of NT-3 (Sigma, Cat#N1905) was dissolved in 300 μL of deionized water. Next, 500 mg of polylactic acid (Solarbio, Cat#P0400) was dissolved in 4 ml of dichloromethane to form the oil phase. The aqueous and oil phases were then combined and sonicated at 100 W in an ice bath for 1 minute to achieve uniform dispersion. This mixture was slowly injected into 100 ml of a 1% polyvinyl alcohol (Solarbio, Cat#YZ-100579) solution using a syringe, followed by sonication at 200 W for 3 minutes in an ice bath to form the emulsion. The emulsion was stirred at low speed (400 rpm) using a magnetic stirrer for 3–5 hours to ensure complete dispersion of the dichloromethane. Finally, the mixture was frozen and centrifuged at high speed (12 000 rpm for 20 minutes) to separate the PLGA microspheres.⁴² The microspheres were washed with deionized water three times, freeze-dried, sterilized by irradiation at 60 °C, and stored at 4 °C.

2.3 Preparation of slowrelease SilMA hydrogel scaffolds

Following the procedure used to prepare the SilMA hydrogel solution, the solution was dissolved and filtered to eliminate impurities. PLGA microspheres containing NT-3 (10 μg ml⁻¹) and HCT (5 μg ml⁻¹) were then added. The mixture was vortexed and then injected into the well plate. Gelation was induced by exposing the preparation to a 405 nm light source for 30 seconds, after which the same post-gelation procedures as previously described were followed.

2.4 Characterization of the SilMA hydrogel and PLGA microspheres

SilMA hydrogel characterization. The SilMA hydrogel was characterized at 25 °C using a HAAKE MARS rheometer to evaluate its mechanical properties. The elastic modulus (G′) and viscous modulus (G′′) were calculated based on the test results. The mechanical properties of the hydrogels were also analyzed using a GT TCS-2000 single-column tensometer. The samples (n = 3) were cylindrical with a diameter of 10 mm and a height of 3 mm. The compression speed was set to 1 mm min⁻¹, and the test was completed at the fracture point on the stress–strain curve. The Young modulus was determined from the slope of the stress–strain curve.

Swelling tests. The SilMA hydrogel and 4DMSH were evaluated after light irradiation (365 nm) using previously reported methods with some modifications.⁴³ Briefly, the hydrogel samples were fully immersed in a PBS solution at 37 °C to perform the swelling test.

PLGA microsphere characterization. The size and morphology of the PLGA microspheres were observed using scanning electron microscopy (SEM, Philips XL-30) at an accelerating voltage of 10 kV.

Drug release detection in vitro. Each drug carrier group was weighed to 20 mg and subsequently dissolved in 20 ml of PBS solution. The tubes were then placed in a shaking incubator set at 37 °C. On days 1, 3, 7, 14, 21, 28, 35, 42, 49, and 56, the tubes were centrifuged at 2000 rpm for 10 minutes. One milliliter of the supernatant was collected and replaced with 1 ml of fresh PBS. The samples were stored, and drug concentrations were measured using ELISA kits. A drug release curve was plotted to illustrate the release profile over time.

2.5 Cell culture and isolation of mouse NSCs

Cells were cultured in a complete medium consisting of DMEM (Gibco, Cat#10566016), 10% FBS (CellMax, Cat#SA101.02), and 1% penicillin–streptomycin (Gibco, Cat#15140163) and passaged when they reached 70%–80% confluence. For cryopreservation, cells were detached, resuspended in fetal bovine serum containing DMSO (Solarbio, Cat#D8371), and stored at −80 °C. To isolate mouse NSCs, the cerebral cortex was extracted from embryos, enzymatically digested, and then filtered. Cells were cultured in a proliferation medium, with medium changes performed on day 3, and neurospheres were passaged on day 7. Subsequently, the cells were dissociated and expanded for further experiments.

2.6 CCK8 assay

First, the 96-well plate was coated with poly-L-lysine (Solarbio, Cat#YS153798) and incubated in a humidified incubator at 37 °C for 2 hours. After incubation, thoroughly wash the plate with PBS solution. Next, digest and resuspend the NSCs, adjusting the cell concentration to 5 × 10⁴ cells per ml. Add 100 μL of the cell suspension to each well of the 96-well plate. Once the cells had adhered, the medium was replaced with fresh culture medium containing varying concentrations of SilMA hydrogel solutions (0%, 1%, 2%, 4%, 8%, and 16%). The culture medium should be replaced every two days. On the third day of culture, discard the old medium and add 100 μL of culture medium containing 10% CCK8 (Beyotime, Cat#C0042) reagent to each well. Incubate the plate in a 5% CO₂, 37 °C incubator for 2 hours. After incubation, place the plate into a microplate reader, ensuring that no air bubbles are present in the wells, and measure the absorbance at 450 nm. Perform all steps under minimal light exposure to protect the reagents.

2.7 Live/dead cell staining

The 24-well plates were coated with poly L-lysine and incubated in a humidified incubator at 37 °C for 2 h. The plates were then washed with PBS and the NSCs were digested and resuspended to a concentration of 5 × 10⁴ ml⁻¹. The plates were then rinsed with PBS, digested and resuspended with NSCs, and the cell concentration was adjusted to 5 × 10⁴ ml⁻¹. 100 μL of cell suspension was added to each well. After cell attachment, transwell chambers containing SilMA hydrogel and 4DMSH were added to the SilMA and 4DMSH groups so that the hydrogel was in the upper chamber and the NSCs were in the lower chamber. The medium should be replaced every two days. After 3 days of culture, discard the medium and prepare the Calcein AM/PI (Beyotime, Cat#C2015M) working solution, adding 250 μL to each well. Incubate the plate in a humidified incubator (5% CO₂, 37 °C) for 1 hour. After incubation, observe the staining using a fluorescence microscope and take photographs. It is important to perform the entire procedure under minimal light exposure to protect the fluorescence reagents.

2.8 Edu cell proliferation assay

NSCs were inoculated into 24-well plates and co-cultured with transwell chambers equipped with SilMA hydrogels or 4DMSH, as previously described. The EdU working solution was prepared by diluting the 2× EdU stock solution (20 μM) with NSC culture medium at a 1 : 500 ratio. An equal volume of the 2× EdU working solution was added to each well, and the cells were incubated in a humidified incubator at 37 °C with 5% CO₂ for 48 hours. Following incubation, the culture medium was removed, and 200 μL of 4% paraformaldehyde (Solarbio, Cat#P1110) was added to each well. The plate was then incubated at room temperature for 15 minutes. The wells were washed three times with PBS, with each wash lasting 5 minutes. After fixation, 200 μL of a blocking solution containing 5% bovine serum albumin (BSA) (Beyotime, Cat#ST023) and 0.5% Triton X-100 (Solarbio, Cat#T8200) were added to each well, followed by incubation at room temperature for 15 minutes. The wells were again washed thrice with PBS, with each wash lasting 5 minutes. A click reaction mixture was prepared by combining 2.15 mL of the click reaction buffer, 100 μL of CuSO₄, 5 μL of azide 488, and 250 μL of the click additive solution. A total of 200 μL of the click reaction mixture was added to each well, and the plate was gently shaken to ensure even coverage of the samples. The plate was incubated at room temperature in the dark for 30 minutes. Finally, the click reaction mixture was aspirated, and the wells were washed three times with PBS, with each wash lasting 5 minutes. Hoechst 33342 (1000×) was diluted at a ratio of 1 : 1000 with PBS. After aspirating the PBS, 200 μL of the 1× Hoechst 33342 solution was added to each well and incubated in the dark at room temperature for 10 minutes. The Hoechst solution was then aspirated, and the wells were washed three times with PBS, with each wash lasting 5 minutes. After staining, the plate was observed under a fluorescence microscope to assess the staining effect. Images were captured, and the entire process was conducted under minimal light to protect the fluorescence signal.

2.9 Analysis of mRNA expression using real-time quantitative polymerase chain reaction (RT-qPCR)

BV2 microglial cells were seeded onto poly-L-lysine-coated 6-well plates and activated with lipopolysaccharide (LPS) to induce inflammation. Simultaneously, the Transwell system was employed, placing SilMA hydrogel or 4DMSH in the upper chamber and BV2 microglial cells in the lower chamber to evaluate the effect of the hydrogels on inflammation over 24 hours. After 24 hours of co-culture with the hydrogels, RNA was extracted and purified from BV2 microglial cells using a cell/tissue RNA extraction kit (Solarbio, Cat#R1200). NSCs were seeded onto poly-L-lysine-coated 6-well plates and cultured in a neural stem cell differentiation medium. The Transwell system was employed using the same protocol as described above to treat the NSCs. After 7 days, RNA was extracted and purified from NSCs using a cell/tissue RNA extraction kit (Solarbio, Cat#R1200). RNA was reverse-transcribed into cDNA using the RevertAid RT reverse transcription kit (Thermo Scientific, Cat#K1622). RT-qPCR was performed to determine the expression levels of target genes. Relative gene expression levels were analyzed using the 2^−ΔΔCT method. The primer sequences are listed in Table 1.

Table 1 Sequences of RT-PCR primers

Gene	Forward (5′ to 3′)	Reverse (5′ to 3′)
GAPDH	GAAGGTGAAGGTCGGAGT	GAAGATGGTGATGGGATTTC
Tuj1	TGTGATGGTTGGTGGTGGAG	GAGAGTTGCCGACGAACTCG
GFAP	CGAGTTCAGTATCCGCTGGA	TCCACAACTGGCGACGAGTA
Arg-1	GGCGAGAGGGAGACAGTTTT	TCTTGTTGAGGTAGGCGGAA
iNOS	GGAAGGAGGAGACAGAGTGA	GAGGAGTGGGAGGAAGAGT

---

2.10 Establishment of the mouse SCI model

In this study, a complete transection SCI model was established by surgically severing the T10 spinal segment in aged 6–8 weeks female C57BL/6J mice (n = 6 per group). Mice were acclimated for three days before surgery, during which food and water were withheld for 12 hours before the procedure. Under isoflurane anesthesia, a 3 cm midline incision was made on the back, and the T10 spinous process and lamina were removed to expose the spinal cord. A 2 mm complete transection of the T10 segment was performed, and the incision was subsequently closed after achieving hemostasis. In the sham group, only the spinous process and lamina were removed without severing the spinal cord. For the SilMA group, the injury site was filled with SilMA hydrogel scaffolds, whereas in the 4DMSH group, the scaffolds contained 4DMSH. Postoperative care included daily monitoring of the general condition, food intake, and signs of infection. Cefuroxime sodium (1 mg) (Solarbio, Cat#C7340) was administered intraperitoneally for 5 days, and the bladder was manually expressed until spontaneous urination was restored. The presence of muscle spasms in the tail and hind limb twitching confirmed the successful establishment of the SCI model.

2.11 Behavioral testing

The Basso Mouse Scale (BMS) was used to assess the recovery of hindlimb motor function following SCI. BMS scores were recorded on days 1, 7, 14, 28, 35, 42, 49, and 56 post-SCI. To mitigate the influence of circadian rhythms, all assessments were conducted at a consistent time each day. Each mouse was placed in an unstimulated open environment and allowed to walk freely for 3 minutes while hindlimb movement was observed and documented. Following each observation, the mouse was returned to its cage and the testing area was cleaned to eliminate any excreta that could impact subsequent assessments.⁴⁴ Scoring was performed by researchers who were blinded to the group and treatment conditions, with a focus on joint movement and gait coordination of the hindlimbs. BMS scores were assigned based on the observed performance. Data were systematically collected and organized according to group assignment for subsequent statistical analysis.

2.12 Gait analysis

Eight weeks post-injury, the CatWalk XT system (Noldus Information Technology, Wageningen, The Netherlands) was used to assess gait parameters.^45,46 The system comprises a transparent glass walking platform illuminated by green LED lights and a high-speed camera capable of capturing real-time footprints. The foot placement and movement patterns were recorded in a dimly lit and quiet environment. Before testing, the animals underwent habituation to the system. On the testing day, each animal completed three successful trials without interruptions. Data were analyzed using CatWalk XT software, which generated a range of gait-related parameters.

2.13 Electrophysiological measurements

Eight weeks after surgery, mice were anesthetized with isoflurane, and sensory evoked potentials (SEP) were recorded using the BL-420 Data Acquisition and Analysis System (TECHMAN SOFT, China) to assess the recovery of motor and sensory nerve function in each group (n = 3 mice per group).⁴⁷ The detailed procedure was as follows: to assess SEP, the sciatic nerve was stimulated, and recording electrodes were placed in the contralateral cortex at the same depth. A ground wire was connected between the stimulating and recording electrodes, and the BL-420 system was used for stimulation. The intensity of the single pulse stimulation was uniformly set to 10 mA for continuous stimulation. The SEP amplitude was measured using the TM_wave bioinformatics acquisition and analysis software.

2.14 Preparation of frozen sections

Eight weeks post-SCI, mice were anesthetized with isoflurane. The chest cavity was subsequently opened, and the right atrium was incised, followed by perfusion through the left ventricle with phosphate-buffered saline (PBS) and then with 4% paraformaldehyde. Once the limbs exhibited rigidity, the spine was excised and immersed in 4% paraformaldehyde for 24 hours. The spinal cord was dissected and fixed in 4% paraformaldehyde for an additional 24 hours. Following fixation, spinal cord tissue was subjected to gradient dehydration in 20% and 30% sucrose solutions until the tissue sank. The dehydrated spinal cord was trimmed to eliminate excess dura mater and other surrounding tissues, retaining approximately 1 cm of tissue on either side of the injury site. The tissue was subsequently embedded in an optimal cutting temperature (OCT) compound (Solarbio, Cat#4583), labeled, and stored at −80 °C for 24 hours. The embedded spinal cord tissue was mounted on a support within a cryostat and sectioned to a thickness of 10 μm. The sections were placed on slides and labeled according to their group assignment. The frozen sections were stored at −20 °C for short-term use and were processed promptly for subsequent staining procedures.

2.15 Immunohistochemistry staining

The frozen sections were washed in TPBS for 10 minutes to eliminate excess OCT compound. The staining area was delineated using a histochemistry pen. The sections were then placed in a humidified chamber, and 4% paraformaldehyde was added to the histochemistry circle to fix the tissue at room temperature for 15 minutes. The sections were then washed twice with TPBS, with each wash lasting 10 minutes. A blocking solution containing 5% BSA and 0.5% Triton X-100 was subsequently applied to the sections and incubated for 1 hour at room temperature to inhibit non-specific binding and expose intracellular antigens. After blocking, the sections were washed washes with TBST, each for 10 minutes. The primary antibodies used were anti-Iba1 (ab283319, mouse, Abcam), anti-GFAP (3670, mouse, CST), and anti-β3-tubulin (5568, rabbit, CST), all diluted to 1 : 200. The primary antibodies were applied to the sections, which were then covered with a thin membrane and incubated overnight at 4 °C. The following day, the sections were washed three times with TBST, with each wash lasting 10 minutes. After washing, fluorescent secondary antibodies were applied: goat anti-mouse IgG H&L (Alexa Fluor®555, ab150117, Abcam) and goat anti-rabbit IgG H&L (Alexa Fluor®488, ab150080, Abcam), both diluted to 1 : 400. The sections were incubated for 1 hour at room temperature, followed by three washes with TBST, each for 10 minutes. DAPI-containing anti-fade mounting medium was then added to the spinal cord tissue, and coverslips were carefully placed on top, ensuring that no bubbles formed. The edges of the coverslips were sealed with nail polish. The stained sections were examined using an Olympus VS120 microscope at 20× magnification, and the immunofluorescence staining effects were scanned and photographed.

2.16 Hematoxylin and eosin (H&E) staining

The frozen sections were washed three times with TBST for 5 minutes each to eliminate excess OCT compound. The sections were then immersed in hematoxylin staining solution (PHYGENE, China) for approximately 1 minute, followed by rinsing under running water until the tissue exhibited a blue color. Care was taken to avoid direct water flow into the tissue. Subsequently, the sections were placed in a 1% hydrochloric acid–ethanol differentiation solution for 30 seconds and rinsed under running water for 1 minute. The sections were then immersed in eosin staining solution (PHYGENE, China) for 2 minutes, followed by another rinse under running water for 1 minute. The sections were sequentially immersed in 80%, 95%, and absolute ethanol, with each step lasting 30 seconds. The sections were then immersed in absolute ethanol, followed by xylene I and xylene II, each for 1 minute. Finally, the sections were mounted with neutral resin and observed under an Olympus VS120 microscope (Olympus, Japan).

2.17 Statistical methods

All experimental results are presented as the mean ± standard deviation (SD) from at least three independent experiments. Comparisons between groups were performed using a one-way analysis of variance (ANOVA). Data analysis and graph plotting were performed using GraphPad Prism 8.0. A p-value of less than 0.05 was considered statistically significant. “ns” indicates no significant difference, while * denotes p < 0.05, ** denotes p < 0.01, and *** denotes p < 0.001.

3. Results and discussion

3.1 SEM of SilMA, SilMA@HCT, and 4DMSH

This study used SEM to investigate the microstructure of the SilMA hydrogel and evaluate its potential for filling lesion areas and delivering drugs and cells for SCI repair (Fig. 2A). The SEM analysis indicated that the SilMA hydrogel possesses a highly porous network structure with pore sizes exceeding 10 μm, which is ideal for accommodating therapeutic agents and microspheres, thereby facilitating substance exchange in vivo and promoting tissue regeneration. Furthermore, the loose network structure of the SilMA hydrogel may enable the rapid release of HCT and modulate inflammation during the early stages of SCI. Concurrently, PLGA microspheres that encapsulate NT-3 can prolong the release of NT-3 at the SCI site, ensuring its sustained presence and high concentration throughout the long-term repair process. SEM analysis also confirmed that SilMA@HCT and 4DMSH maintained their original porosity and pore sizes, indicating that the incorporation of HCT and NT-3 did not compromise the structural integrity of the SilMA hydrogel. These findings affirm the successful fabrication of a structurally sound SilMA hydrogel capable of effectively delivering HCT and NT-3 for SCI repair.

Fig. 2 Characterization of the four-dimensional multifunctional SilMA hydrogel (4DMSH). (A) Scanning electron microscopy (SEM) images of the SilMA hydrogel showing various components. The size and morphology of PLGA microspheres, indicated by arrows, can be observed in 4DMSH (scale bar = 100 μm). (B) Absorbance spectra of the SilMA hydrogel with various components. (C) Viscosity measurements of the SilMA hydrogel with various components. (D) Swelling ratios of the SilMA hydrogel with various components. (E) Modulus values of SilMA hydrogel with various components. (F) Intensity measurements of the SilMA hydrogel with various components. (G) Energy profiles of the SilMA hydrogel containing various components. (H) EDS layered images of the SilMA hydrogel with various components, scale bar = 100 μm.

3.2 Physicochemical and mechanical characterization of SilMA, SilMA@HCT, and 4DMSH

Fig. 2B shows the FTIR spectra of SilMA, SilMA@HCT, and 4DMSH, highlighting the characteristic peaks associated with the SilMA structure. The additional peaks observed for SilMA@HCT and 4DMSH indicate the successful incorporation of HCT and NT-3. Fig. 2C illustrates the shear-thinning behavior of the hydrogels, revealing no significant differences in viscosity; this confirms that the addition of HCT and NT-3 does not adversely affect the rheological properties, thereby ensuring ease of injection and maintaining structural integrity in vivo. Fig. 2D displays comparable swelling profiles for all hydrogels, suggesting that the inclusion of HCT and NT-3 does not compromise water absorption, which is essential for promoting cell infiltration and tissue repair. Fig. 2E shows an analysis of the mechanical properties (storage modulus G′ and loss modulus G′′), indicating similar strength across all formulations. XRD analysis (Fig. 2F) demonstrates that the incorporation of HCT and NT-3 does not alter the crystallinity of the hydrogel. EDS analysis (Fig. 2G) confirms a consistent elemental composition (C, O, N) across the groups, while EDS layered images (Fig. 2H) illustrate a uniform distribution of these elements. Collectively, these results affirm that the SilMA hydrogel, even when loaded with HCT and NT-3, preserves its structural, mechanical, and chemical integrity, rendering it a robust platform for drug delivery and tissue engineering in SCI repair. In summary, HCT and sustained-release microspheres of NT-3 were incorporated into the SilMA hydrogel to form 4DMSH for time-specific SCI repair. The crosslinked loose network of the hydrogel not only mitigates damage to the eNSCs⁴⁸ but also facilitates the rapid release of HCT post-implantation, thereby addressing the intense inflammation characteristic of the early stages of SCI, which is recognized as a critical period for nerve damage. Meanwhile, NT-3 was slowly released from the PLGA microspheres to achieve sustained promotion of neural regeneration. Therefore, we successfully constructed a time-specific 4DMSH by releasing HCT early in response to anti-inflammation and releasing long-term NT-3 to enhance neural regeneration.

3.3 Degradation and controlled release profiles of SilMA, SilMA@HCT, and 4DMSH

To evaluate the biodegradation and drug release capabilities of the SilMA hydrogel, we assessed the degradation profiles and release kinetics. Fig. 3A shows the degradation profiles of the SilMA hydrogel (SilMA, SilMA@HCT, 4DMSH) over a 20-day period. Gradual mass loss across all groups indicates similar biodegradation rates, confirming that the inclusion of NT-3 within PLGA microspheres and HCT does not significantly impact the hydrogel's controlled degradation. This consistency provides prolonged structural support during tissue repair. Fig. 3B illustrates the cumulative release profiles of HCT and NT-3 from 4DMSH. HCT is released more rapidly, reaching 80% by day 14, which offers early-stage anti-inflammatory effects. In contrast, NT-3 exhibited a slower, sustained release over 56 days, providing long-term neurotrophic support for neural regeneration. This time-specific release optimizes the therapeutic timing of SCI recovery. The gradual decrease in the release rates of HCT and NT-3 over time may be attributed to the reduced osmotic driving force as the concentration gradient between the hydrogel and its surroundings diminishes during drug release. Fig. 3 illustrates the capability of the SilMA hydrogel for controlled biodegradation and sustained therapeutic delivery, facilitating immediate and long-term regenerative outcomes. These results underscore the effective design of the SilMA hydrogel for SCI applications. The rapid release of HCT in the initial stages is crucial for mitigating inflammation, but it can be detrimental to neuronal survival and regeneration.⁴⁹ The sustained release of NT-3 ensures continuous neurotrophic support throughout the recovery process, thereby enhancing the likelihood of successful neural regeneration.^50,51 This dual-phase release mechanism not only meets the immediate needs of the injury site but also facilitates long-term recovery strategies. Future studies should explore the optimization of drug ratios within 4DMSH to further enhance its therapeutic efficacy.

Fig. 3 Degradability and drug release profiles of 4DMSH. (A) Remaining mass ratio of the SilMA hydrogel with various components over the first 20 days. (B) Cumulative drug release rates of Houttuynia cordata extract (HCT) and neurotrophin-3 (NT-3) over the first 56 days.

3.4 Biocompatibility and cytotoxicity of 4DMSH

Biocompatibility is essential for biomaterials used in regenerative medicine. To evaluate the effects of SilMA hydrogel and its drug-loaded formulations on the viability of NSCs, a CCK8 assay was conducted. NSCs were cultured with various concentrations of SilMA hydrogel, and their viability was assessed after 3 days. Cell viability, measured using the CCK8 reagent, revealed no significant difference between cells cultured with up to 8% SilMA and the control group (Fig. 4B), indicating that all tested concentrations exhibited excellent biocompatibility. Consequently, the 8% SilMA hydrogel was selected for subsequent in vitro and in vivo experiments to confirm its suitability and safety. In addition to cell viability, it is imperative to evaluate whether the hydrogel induces cytotoxicity or cell death, as cell survival is crucial for successful transplantation and tissue regeneration.⁵² Live/dead staining using Calcein AM and propidium iodide (PI) was performed to assess cell membrane integrity and esterase activity. Calcein AM stained live cells green, whereas PI marked dead cells red. NSCs were cultured with the three SilMA hydrogel formulations for 3 and 7 days. No significant cell death was observed in any group, indicating that all SilMA formulations demonstrated excellent biocompatibility and did not exhibit cytotoxicity (Fig. 4C), thereby supporting their suitability for application in vivo. To confirm that the SilMA hydrogel does not interfere with the hydrogel's cellular functions, an EdU assay was performed. The results indicated that the drug-loaded SilMA hydrogels promoted NSC proliferation compared with the unloaded hydrogels, suggesting that the inclusion of therapeutic agents, such as HCT and NT-3, not only supports cell survival but also enhances their proliferative capacity (Fig. 4D). This finding is significant because effective neural regeneration depends on the survival and proliferation of NSCs at the injury site. In summary, these findings indicate that unloaded and drug-loaded SilMA hydrogels are biocompatible, non-cytotoxic, and support NSC viability, survival, and proliferation. Furthermore, the drug-loaded formulations provide an added advantage by promoting cell proliferation, making them promising candidates for in vivo SCI repair applications.

Fig. 4 Biocompatibility and cytotoxicity assessment of 4DMSH. (A) Schematic of cell culture using Transwell system. (B) CCK8 assay results of neural stem cells. (NSCs) viability with SilMA hydrogel at various concentrations; ns: no significance. (C) Live/dead staining results of NSCs co-cultured with SilMA hydrogel of different components, with live cells stained green and dead cells stained red, scale bar = 10 μm. (D) EdU staining results of NSCs co-cultured with SilMA hydrogel of different components, scale bar = 10 μm.

3.5 4DMSH promotes NSC differentiation into neurons and induces microglial polarization towards the M2 phenotype in vitro

RT-qPCR analysis revealed that 4DMSH regulates NSC differentiation and microglial polarization. In terms of NSC differentiation, the expression of Tuj1 (a neuronal marker) was significantly upregulated in the 4DMSH group compared with the control and SilMA groups (Fig. 5A). In contrast, the expression of GFAP (an astrocyte marker) was significantly downregulated (Fig. 5B), indicating a reduction in astrocytic differentiation. These findings indicate that 4DMSH promotes NSC differentiation into neurons while inhibiting astrocytic differentiation. Regarding microglial polarization, the expression of Arg-1 (an M2 marker) was significantly increased (Fig. 5C), whereas that of iNOS (an M1 marker) was significantly decreased in the 4DMSH group (Fig. 5D). These findings suggest that 4DMSH modulates the inflammatory response by promoting the transition of microglia to the reparative M2 phenotype, thereby enhancing the regenerative microenvironment. This dual action of 4DMSH—promoting neurogenesis while simultaneously regulating the inflammatory response—highlights its potential as a therapeutic strategy for spinal cord injury repair. Further studies are required to explore the long-term effects of 4DMSH on functional recovery and the underlying mechanisms in vivo.

Fig. 5 4DMSH promotes differentiation of NSCs into neurons and microglia to M2. (A) Tuj1 expression was higher in the 4DMSH group than in the TCP and SilMA groups. (B) GFAP expression in the 4DMSH group was lower than that in the TCP and SilMA groups. (C) Arg-1 expression in the 4DMSH group was higher than that in the TCP and SilMA groups. (D) 4DMSH group had lower iNOS expression than the TCP and SilMA groups.

3.6 Modulation of microglial immune response by 4DMSH in vivo

4DMSH significantly modulates the microglial immune response in vivo 1-week post-SCI, as demonstrated by Iba-1 immunofluorescence staining and semi-quantitative analysis. In the SCI and SilMA groups, Iba-1 staining revealed a high density of activated microglia surrounding the injury site. Conversely, the 4DMSH group exhibited a marked reduction in Iba-1⁺ cells, indicating diminished microglial activation (Fig. 6A). The semi-quantitative analysis further revealed a significant decrease in Iba-1 signalling in the SilMA group compared with the SCI group (Fig. 6B). The temporal release of HCT from 4DMSH effectively suppressed inflammation and fostered a reparative environment, thereby facilitating neural tissue regeneration and functional recovery. Immunofluorescence images and quantitative data corroborate these findings, demonstrating 4DMSH's capacity to modulate microglial activity and shift the immune response toward regeneration.

Fig. 6 Evaluation of the anti-inflammatory and neuronal inductive functions of 4DMSH in vivo. (A) Immunofluorescence images showing Iba-1 (green) expression in cells from various treatment groups, scale bar = 200 μm. (B) Quantification of the relative fluorescence intensity of Iba-1, with * indicating p < 0.05 and *** indicating p < 0.001. (C) Immunofluorescence images showing Tuj1 (red) and GFAP (green) expression in cells from various treatment groups, scale bar = 200 μm. (D) Quantification of the relative fluorescence intensity of Tuj1. (E) Quantification of the relative fluorescence intensity of GFAP, with * indicating p < 0.05, ** indicating p < 0.01, and *** indicating p < 0.001.

3.7 Neural regeneration using 4DMSH in vivo

4DMSH significantly promotes the differentiation of eNSCs eight weeks post-SCI, thereby enhancing SCI repair. This was evidenced by immunofluorescence staining and semi-quantitative analysis of TUJ1 (neuronal marker) and GFAP (astrocyte marker). In the SCI group, GFAP⁺ astrocytes were abundant around the injury site, indicating increased glial scar formation, whereas TUJ1⁺ neurons were notably scarce. In contrast, the SilMA-treated group exhibited a marked increase in TUJ1⁺ neurons and a reduction in GFAP⁺ astrocytes, suggesting enhanced neuronal differentiation and diminished glial scar formation. The 4DMSH group exhibited the most significant improvement, characterized by a pronounced increase in TUJ1⁺ neurons and a greater suppression of GFAP⁺ astrocytes compared with the SilMA group (Fig. 6C). The semi-quantitative analysis confirmed a significantly higher intensity of TUJ1 in the 4DMSH group (Fig. 6D), whereas GFAP intensity was reduced, indicating a favorable environment for neuronal regeneration (Fig. 6E). These findings, supported by immunofluorescence images and intensity data, illustrate that 4DMSH effectively promotes eNSC differentiation into neurons, facilitating spinal cord repair and functional recovery. The sham group, which served as the baseline, exhibited normal levels of TUJ1 and GFAP expression, further underscoring the therapeutic effect of 4DMSH in SCI.

3.8 Promotion of motor function recovery in mice by 4DMSH

The 4DMSH significantly improved motor function recovery in mice following SCI, as evidenced by the BMS scores and CatWalk gait analysis. Eight weeks post-SCI, the BMS scores indicated a significant improvement in the 4DMSH group compared with the SCI and SilMA-only groups. Mice treated with 4DMSH exhibited more coordinated hindlimb movements and enhanced motor recovery (Fig. 7A). CatWalk gait analysis demonstrated a significant increase in average running speed in the 4DMSH group relative to the SCI group, with values approaching those observed in the sham group. Furthermore, the step duration was notably reduced in the 4DMSH group, indicating improved gait efficiency. These findings, supported by statistical analysis of running speed and step duration (Fig. 7B), indicate that 4DMSH significantly enhances motor coordination and gait performance in mice following SCI. This analysis underscores the effectiveness of 4DMSH in improving motor function, suggesting its potential as a therapeutic strategy for SCI repair.

Fig. 7 Recovery of motor and sensory function in the SCI model. (A) Basso Mouse Scale (BMS) scores of four groups of mice at 8 weeks postoperatively. (B) Average running speed and running duration of the four groups of mice. Data are presented as mean ± S.E.M. *p < 0.05, **p < 0.01, and ***p < 0.001. (C) Sensory evoked potentials (SEPs) across the four groups of mice. (D) Amplitude of SEPs in the four groups, ** indicates p < 0.01 and *** indicates p < 0.001. (E) Duration of hind limb twitching in response to 55 °C stimuli for the four groups, *** indicates p < 0.001. (F) Duration of hind limb twitching in response to 4 °C stimuli for the four groups, * indicates p < 0.05 and *** indicates p < 0.001.

3.9 Promotion of sensory function recovery in mice by 4DMSH

The 4DMSH significantly promotes sensory recovery in mice following SCI, as demonstrated by SEP recordings and thermal sensitivity tests conducted at 4 °C and 55 °C. Electrophysiological assessments using SEP recordings obtained eight weeks post-SCI, revealed significantly higher amplitudes in the 4DMSH group compared with the SCI-only and SilMA-only groups (Fig. 7C). In contrast, the untreated SCI group exhibited markedly reduced SEP amplitudes, indicating impaired sensory signal transmission. However, the 4DMSH group showed substantial amplitude restoration, suggesting effective recovery of sensory pathways. Statistical analysis of amplitude confirmed this recovery, with the 4DMSH group showing a significant increase relative to the other groups, approaching the values observed in the sham group (Fig. 7D). Thermal sensitivity tests, which involved exposing mice to cold (4 °C) and hot (55 °C) stimuli while recording their response times, indicated that the SCI group had prolonged response times, reflecting sensory loss. The SilMA group exhibited moderate improvement, whereas the 4DMSH group exhibited the most significant recovery, with response times being considerably shorter than those of the SCI group. Notably, the response time of the 4DMSH group was closest to that of the sham group, indicating maximal repair of thermal sensitivity (Fig. 7E and F). Collectively, the results from the SEP and thermal sensitivity tests indicate that the 4DMSH enhances sensory nerve function at the electrophysiological level and restores behavioral responses to thermal stimuli, underscoring its therapeutic potential for sensory recovery following SCI.

3.10 Promotion of autonomic function recovery in mice by 4DMSH

4DMSH significantly enhances autonomic function recovery in mice following SCI, as demonstrated by H&E staining and bladder muscle thickness analysis. SCI frequently results in severe autonomic dysfunction, including loss of bladder control, due to disruption of neural circuits, which leads to smooth muscle atrophy and fibrosis in the bladder wall.^53–55 H&E images of the bladder wall revealed substantial disorganization and thinning of the smooth muscle layer in the SCI group, accompanied by pronounced fibrosis. In contrast, the 4DMSH-treated group exhibited a well-organized smooth muscle layer with minimal fibrosis, closely resembling the sham group (Fig. 8A). The quantitative analysis confirmed that the 4DMSH group had a significantly thicker bladder muscle layer than the SCI group (Fig. 8B). These findings indicate that 4DMSH facilitates bladder tissue regeneration and promotes the recovery of autonomic bladder function following SCI.

Fig. 8 Recovery of autonomic function and organ toxicity (A) hematoxylin and eosin (H&E) staining of the bladder and its enlargement across four groups of mice, scale bar = 1 mm and 250 μm. (B) Bladder detrusor thickness in the four groups, *** indicates p < 0.001. (C) HE staining of the heart, liver, spleen, lungs, and kidneys in four groups of mice, scale bar = 250 μm.

3.11 Organs toxicity of 4DMSH hydrogel

The biocompatibility of 4DMSH was assessed via H&E staining of major organs, including the heart, liver, spleen, lungs, and kidneys, in vivo (Fig. 8C). Histological analysis of mice treated with 4DMSH revealed no signs of inflammation, necrosis, or adverse effects in any of the major organs, indicating favorable biocompatibility. All organ tissues exhibited a normal structure, comparable to those in the sham and untreated SCI groups, thereby confirming the absence of systemic toxicity or organ damage. These findings suggest that 4DMSH effectively promotes tissue regeneration and functional recovery while remaining safe for in vivo applications with no harmful effects on major organs.

4. Discussions

In this study, SilMA hydrogel, which is known for its favorable biocompatibility, injectability, and responsiveness, was employed as a multifunctional drug delivery platform. By integrating NT-3 encapsulated in PLGA microspheres and directly loading HCT, the hydrogel achieved a spatiotemporally controlled release profile. The early release of HCT exerted rapid anti-inflammatory effects, whereas the sustained release of NT-3 supported long-term neural regeneration. This dual-release strategy enabled the four-dimensional multifunctional SilkMA hydrogel (4DMSH) to play a stage-specific role in spinal cord injury (SCI) repair. As a comprehensive therapeutic system, 4DMSH effectively improves the post-injury microenvironment, protects neural tissue by reducing early inflammation, and promotes axonal regeneration at later stages, ultimately facilitating functional recovery.

This study introduces a novel 4D hydrogel system that simultaneously addresses multiple challenges in SCI treatment by combining bioactive molecule delivery with tunable release kinetics. The spatiotemporal coordination of anti-inflammatory and neuroregenerative factors within a single scaffold represents a significant advancement in the field of spinal cord tissue engineering. Furthermore, the use of SilMA as a base material contributes to excellent structural adaptability and in situ gelation, which is critical for minimally invasive applications in clinical settings.

In terms of axonal regeneration, the use of TEM could provide direct ultrastructural evidence of nerve repair at the injury site. Therefore, introducing TEM technology in follow-up studies is crucial for visualizing fine axonal structures and validating the regenerative effects of 4DMSH.

Regarding its anti-inflammatory effects, although in vitro PCR results showed that 4DMSH promoted a phenotypic shift of microglia from pro-inflammatory (M1) to anti-inflammatory (M2), in vivo validation remains lacking. Future studies should employ qPCR or western blotting to analyze spinal cord tissue samples and further elucidate the anti-inflammatory mechanisms of 4DMSH in vivo.

5. Conclusions

In this study, we developed a novel cell-free 4DMSH by early release of HCT for anti-inflammatory and long-term release of NT-3 to enhance neural regeneration. Briefly, the loose network of the 4DMSH could quickly release HCT after implantation to regulate the intense inflammation of SCI in the early stage, while the NT-3 was slowly released from the PLGA microspheres to achieve sustained promotion of neural regeneration. The current 4DMSH was biocompatible and provided a favorable environment for the survival and proliferation of eNSCs. Moreover, 4DMSH modulated the immune response, promoted eNSC differentiation into neurons, and significantly improved motor, sensory, and autonomic function recovery in mice with SCI, highlighting its potential for SCI repair. Histological analysis revealed no adverse effects on major organs, confirming the safety of 4DMSH. Therefore, this study developed a novel time-specific four-dimensional multifunctional SCI repair strategy for efficient neural regeneration and functional recovery.

Author contributions

Ruizhi Zhang: data curation, investigation, formal analysis, and writing of the original draft. Chenbo Zou: data curation, formal analysis, and investigation. Linlin Jiang: data curation and investigation. Baoshuai Bai: data curation, conceptualization, and writing review. Chunlin Li: conceptualization and writing review. Chi Zhang: data curation. Hua Zhao: writing review. Shaohui Zong: writing review. Hao Li: writing review and supervision. Kai Jiang: writing review and supervision. Hengxing Zhou: supervision, data curation and funding acquisition. Shiqing Feng: writing review, conceptualization, supervision, funding acquisition and data curation.

Conflicts of interest

There are no conflicts to declare.

Data availability

All data generated or analyzed during this study are included in this published article and are available from the corresponding author.

Acknowledgements

We thank the Translational Medicine Core Facility of Shandong University for consultation and instrument availability that supported this work. This study was supported by grants from the National Natural Science Foundation of China (82372413), the Major Basic Research Project of the Shandong Provincial Natural Science Foundation (ZR2024ZD13), the Taishan Scholars Program of Shandong Province-Young Taishan Scholars (tsqn201909197), the Taishan Scholars Program of Shandong Province-Pandeng Taishan Scholars (tspd20210320) and the Scientific Research Cooperation Seed Fund of Shandong University (26010287395002).

References

1. J. S. Lin and I. Fox, Plast. Aesthetic Res., 2024, 24, 2957–2981.
2. B. Fan, Z. Wei and S. Feng, Bone Res., 2022, 10, 35.
3. B. Q. Shi, S. Lu, H. R. Yang, S. Mahmood, C. H. Sun, N. Malek, W. H. A. Kamaruddin, S. Saidin and C. C. Zhang, BMEMat, 2024, 3 DOI:10.1002/bmm2.12111.
4. B. C. Chan, S. M. Cadarette, W. P. Wodchis, M. D. Krahn and N. Mittmann, J. Spinal Cord Med., 2019, 42, 184–193.
5. M. D. Stillman, J. Barber, S. Burns, S. Williams and J. M. Hoffman, Arch. Phys. Med. Rehabil., 2017, 98, 1800–1805.
6. T. M. O'Shea, J. E. Burda and M. V. Sofroniew, J. Clin. Invest., 2017, 127, 3259–3270.
7. M. V. Sofroniew, Nature, 2018, 557, 343–350.
8. X. Zha, G. Zheng, T. Skutella, K. Kiening, A. Unterberg and A. Younsi, Neural Regener. Res., 2025, 20, 454–463.
9. W. K. Kim, Y. Kan, D. Ganea, R. P. Hart, I. Gozes and G. M. Jonakait, J. Neurosci., 2000, 20, 3622–3630.
10. M. Mukaino, M. Nakamura, O. Yamada, S. Okada, S. Morikawa, F. Renault-Mihara, A. Iwanami, T. Ikegami, Y. Ohsugi, O. Tsuji, H. Katoh, Y. Matsuzaki, Y. Toyama, M. Liu and H. Okano, Exp. Neurol., 2010, 224, 403–414.
11. L. Yang, P. C. Blumbergs, N. R. Jones, J. Manavis, G. T. Sarvestani and M. N. Ghabriel, Spine (Phila Pa 1976), 2004, 29, 966–971.
12. N. Philips, M. Portillo-Esnaola, P. Samuel, M. Gallego-Rentero, T. Keller and J. Franco, Plast. Aesthetic Res., 2022, 9, 4.
13. M. A. Anderson, T. M. O'Shea, J. E. Burda, Y. Ao, S. L. Barlatey, A. M. Bernstein, J. H. Kim, N. D. James, A. Rogers, B. Kato, A. L. Wollenberg, R. Kawaguchi, G. Coppola, C. Wang, T. J. Deming, Z. He, G. Courtine and M. V. Sofroniew, Nature, 2018, 561, 396–400.
14. G. Wang, Q. Li, S. Liu, M. Li, B. Liu, T. Zhao, B. Liu and Z. Chen, Mater. Today Bio, 2024, 28, 101250.
15. J. Zhang, J. Shang, H. Ding, W. Li, Z. Li, Z. Yuan, H. Zheng, Y. Lou, Z. Wei, H. Zhou, S. Feng, X. Kong and N. Ran, Stem Cell Rev. Rep., 2024, 20, 1854–1868.
16. C. Xing, S. Liu, L. Wang, H. Ma, M. Zhou, H. Zhong, S. Zhu, Q. Wu and G. Ning, J. Transl. Med., 2024, 22, 723.
17. G. Chen, K. Tong, S. Li, Z. Huang, S. Liu, H. Zhu, Y. Zhong, Z. Zhou, G. Jiao, F. Wei and N. Chen, Bioact. Mater., 2024, 35, 135–149.
18. M. Liu, W. Zhang, S. Han, D. Zhang, X. Zhou, X. Guo, H. Chen, H. Wang, L. Jin, S. Feng and Z. Wei, Adv. Mater., 2024, 36, e2313672.
19. C. Ma, P. Zhang and Y. Shen, J. Neurorestoratol., 2019, 7, 196–206.
20. T. C. Ho, C. C. Chang, H. P. Chan, T. W. Chung, C. W. Shu, K. P. Chuang, T. H. Duh, M. H. Yang and Y. C. Tyan, Molecules, 2022, 27, 2902.
21. O. A. Carballo-Molina and I. Velasco, Front. Cell. Neurosci., 2015, 9, 13.
22. L. H. Nguyen, M. Gao, J. Lin, W. Wu, J. Wang and S. Y. Chew, Sci. Rep., 2017, 7, 42212.
23. M. Y. Chai, W. W. Zhong, S. T. Yan, T. Ye, R. Zheng, Z. L. Yang and X. T. Shi, BMEMat, 2024, 2, e12108.
24. L. Q. Zhou, Z. Y. Wang, D. H. Chen, J. M. Lin, W. W. Li, S. Y. Guo, R. C. Wu, X. Zhao, T. T. Lin, G. Chen and W. E. Liu, Mater. Des., 2022, 217, 110670.
25. J. Amirian, J. K. Wychowaniec, E. A. Zendehdel, G. Sharma, A. Brangule and D. Bandere, Biomacromolecules, 2023, 24, 2957–2981.
26. X. Y. Shi, X. Y. Wang, W. Shen and W. F. Yue, Biochem. Biophys. Res. Commun., 2023, 650, 62–72.
27. S. Maji and H. Lee, Int. J. Mol. Sci., 2022, 23, 2662.
28. N. Zhu, Y. P. Zhuang, W. J. Sun, J. Wang, F. Wang, X. Y. Han, Z. Y. Han, M. Ni, W. G. Cui and Y. Qiu, Mater. Today Adv., 2024, 21, 100465.
29. V. K. Kuna, A. Lundgren, L. O. Anerillas, P. Kelk, M. Brohlin, M. Wiberg, P. J. Kingham, L. N. Novikova, G. Andersson and L. N. Novikov, Int. J. Mol. Sci., 2022, 23, 8746.
30. K. Moorthy, K. C. Chang, H. C. Huang, W. J. Wu and C. K. Chiang, Antioxidants, 2024, 13, 32.
31. J. Liu, J. B. Zou, J. Y. Wang, R. Wang, S. C. Zhai, X. Chang, X. F. Zhang, J. Sun, F. Luan and Y. J. Shi, Int. J. Biol. Macromol., 2024, 279, 135230.
32. N. Charachit, A. Sukhamwang, P. Dejkriengkraikul and S. Yodkeeree, Antioxidants, 2022, 11, 221.
33. W. Li, P. Zhou, Y. Zhang and L. He, J. Ethnopharmacol., 2011, 133, 922–927.
34. H. M. Tuinstra, M. O. Aviles, S. Shin, S. J. Holland, M. L. Zelivyanskaya, A. G. Fast, S. Y. Ko, D. J. Margul, A. K. Bartels, R. M. Boehler, B. J. Cummings, A. J. Anderson and L. D. Shea, Biomaterials, 2012, 33, 1618–1626.
35. J. C. Stanwick, M. D. Baumann and M. S. Shoichet, J. Controlled Release, 2012, 160, 666–675.
36. Z. W. Zhu, G. Li, G. G. Wu, Y. J. Zhang, Y. R. Bai, B. Q. Lai, Y. Ding, X. Zeng, Y. H. Ma, S. Liu, R. Wang, J. H. Liang, Y. B. Xu, B. He and Y. S. Zeng, Biomaterials, 2025, 315, 122949.
37. S. Maher, K. Wynne, V. Zhernovkov and M. Halasz, Sci. Data, 2024, 11, 1111.
38. M. Ulrichsen, N. P. Goncalves, S. Mohseni, S. Hjaeresen, T. L. Lisle, S. Molgaard, N. K. Madsen, O. M. Andersen, A. F. Svenningsen, S. Glerup, A. Nykjaer and C. B. Vaegter, Front. Cell. Neurosci., 2022, 16, 856734.
39. A. K. Barui, P. Jhelum, S. K. Nethi, T. Das, D. Bhattacharya, V. B, S. Karri, S. Chakravarty and C. R. Patra, Bioconjugate Chem., 2020, 31, 895–906.
40. J. Yao, Y. Chen, X. Zhang, J. Chen, C. Zhou, J. Jiang, H. Zhang and K. Wu, Mater. Today Bio, 2024, 25, 100947.
41. H. Liu, Y. Jiao, T. Forouzanfar, G. Wu, R. Guo and H. Lin, Biomater. Adv., 2024, 160, 213856.
42. X. Wu, Y. Huo, Z. Ci, Y. Wang, W. Xu, B. Bai, J. Hao, G. Hu, M. Yu, W. Ren, Y. Zhang, Y. Hua and G. Zhou, Appl. Mater. Today, 2022, 27, 101478.
43. B. Bai, Y. Liu, J. Huang, S. Wang, H. Chen, Y. Huo, H. Zhou, Y. Liu, S. Feng, G. Zhou and Y. Hua, Adv. Sci., 2024, 11, e2305580.
44. M. B. Graeber, W. Li and M. L. Rodriguez, FEBS Lett., 2011, 585, 3798–3805.
45. J. C. Heinzel, V. Oberhauser, C. Keibl, N. Swiadek, G. Langle, H. Frick, J. Kolbenschlag, C. Prahm, J. Grillari and D. Hercher, Front. Neurosci., 2020, 14, 593545.
46. X. Fan, L. Shi, Z. Yang, Y. Li, C. Zhang, B. Bai, L. Chen, E. E. Yilihamu, Z. Qi, W. Li, P. Xiao, M. Liu, J. Qiu, F. Yang, N. Ran, Y. Shang, J. Liu, T. Zhang, X. Kong, H. Liu, H. Zhou and S. Feng, Adv. Sci., 2024, 11, e2309305.
47. C. Zhao, T. Zhou, M. Li, J. Liu, X. Zhao, Y. Pang, X. Liu, J. Zhang, L. Ma, W. Li, X. Yao and S. Feng, Neural Regener. Res., 2024, 19, 434–439.
48. X. He, R. Wang, F. Zhou and H. Liu, Int. J. Biol. Macromol., 2024, 256, 128031.
49. X. Liu, Y. Zheng, Q. Wang, L. Zhao, Z. Zhang, H. Wang, Y. Yang, N. Song, J. Xiang, Y. Shen and S. Fan, J. Controlled Release, 2023, 364, 601–617.
50. D. M. McTigue, P. J. Horner, B. T. Stokes and F. H. Gage, J. Neurosci., 1998, 18, 5354–5365.
51. B. S. Bregman, M. McAtee, H. N. Dai and P. L. Kuhn, Exp. Neurol., 1997, 148, 475–494.
52. Z. Tan, L. Xiao, J. Ma, K. Shi, J. Liu, F. Feng, P. Xie, Y. Dai, Q. Yuan, W. Wu, L. Rong and L. He, Sci. Adv., 2024, 10, eado9120.
53. H. Zhu, J. D. Guest, S. Dunlop, J. X. Xie, S. Gao, Z. Luo, J. E. Springer, W. Wu, W. Young, W. S. Poon, S. Liu, H. Gao, T. Yu, D. Wang, L. Zhou, S. Wu, L. Zhong, F. Niu, X. Wang, Y. Liu, K. F. So and X. M. Xu, Neural Regener. Res., 2024, 19, 2773–2784.
54. V. Stenimahitis, M. Gharios, A. Fletcher-Sandersjoo, V. G. El-Hajj, A. Singh, A. Buwaider, M. Andersson, P. Gerdhem, C. Hultling, A. Elmi-Terander and E. Edstrom, Sci. Rep., 2024, 14, 20945.
55. K. Bai, Y. Hou, Z. Zhang, F. Yuan, X. Huang, P. Liu, X. Zou and J. Sun, Biomolecules, 2024, 14, 1141.

---

Footnote

† These authors contributed equally to this work.

---