A bioinspired anisotropic anti-inflammatory scaffold enhances spinal nerve regeneration and neural circuit reconstruction via FGF13/Ca²⁺/CaMK2A/CREB pathway

Abstract

Spinal cord injury (SCI) induces severe neurological impairment, exacerbated by secondary inflammation and disrupted neural circuitry. Inspired by the spinal cord's electromechanical microenvironment, we developed a biomimetic conductive nerve scaffold via directional freeze-casting of gelatin methacryloyl (GelMA) hydrogel incorporated with N-acetylcysteine-modified silver nanowires (NAC-AgNWs). The scaffold exhibits axially aligned microchannels, tunable mechanical strength, and conductivity akin to native spinal tissue. In a rat model of complete spinal cord transection (2 mm), the scaffold exhibited dual therapeutic effects: (1) early-stage anti-inflammatory modulation (mediated by the synergistic interplay between AgNWs and NAC), and (2) sustained neural reconstruction, evidenced by robust axonal bridging across the lesion, synapse reformation, and significant functional recovery. Integrated transcriptomic analyses revealed the FGF13/Ca²⁺/CaMK2A/CREB axis as the activated pathway driving neurite outgrowth and neural circuit reconstruction. This biomaterial design establishes a novel therapeutic paradigm for SCI repair, integrating structural guidance, immunomodulation, and activation of pro-regenerative signaling.

Graphical abstract

Highlights

• •NAC-AgNWs enhance conductivity in an aligned, soft, biomimetic scaffold.
• •NAC-AgNWs synergistically drive anti-inflammatory M2 microglial polarization.
• •Functional recovery through novel FGF13/Ca²⁺/CaMK2A/CREB signaling activation.
• •Repair strategy integrating immunomodulation with electroactive neural guidance.

1. Introduction

The spinal cord is a critical component of the central nervous system, essential for transmitting sensory and motor signals between the brain and peripheral nerves [1,2]. Spinal cord injury (SCI) leads to persistent neurological dysfunction due to the extremely limited regenerative capacity of the adult central nervous system. In addition to the initial mechanical disruption, secondary pathological processes—including excessive inflammation, demyelination, and failure of axonal reconnection—further compromise functional recovery [3,4]. These multifactorial barriers make SCI repair particularly challenging and highlight the need for therapeutic strategies that simultaneously address structural disruption and the hostile post-injury microenvironment. Biomaterial-based scaffolds have been widely explored as physical bridges to support axonal regrowth across lesion gaps and to modulate the injury milieu [5,6].

Scaffolds mimicking the extracellular matrix (ECM) play a vital role in SCI repair by offering structural and biochemical support [7,8]. Hydrogel scaffolds are especially attractive due to their high water content, tissue-like softness, and tunable mechanical properties, which are essential for minimizing mechanical mismatch with spinal cord tissue. However, many existing hydrogel systems primarily provide passive structural support and lack the ability to integrate additional biological functions, such as directional guidance, electrical compatibility, and immunomodulation, which are increasingly recognized as critical determinants of successful neural regeneration. The spinal cord's anisotropic architecture, validated by diffusion tensor imaging (DTI), demands aligned microstructures to guide axonal regrowth, while its bioelectrical nature requires conductive interfaces to restore neural signaling [9,10].

This study aimed to develop an anisotropic conductive hydrogel scaffold emulating the electromechanical properties of the spinal cord extracellular matrix (ECM), with biomimetic conductivity (0.02–0.60 S/m) and modulus (100–3000 Pa) [[11], [12], [13]], to mitigate early-stage inflammation and enhance spinal cord regeneration (Scheme 1). Gelatin methacryloyl (GelMA) is a collagen-derived hydrogel that retains cell-adhesive motifs and enables controllable mechanical properties through photo-crosslinking. These features allow GelMA to closely mimic key aspects of the spinal cord extracellular matrix, including softness, hydration, and biological permissiveness, making it a suitable matrix for neural tissue engineering. Nevertheless, GelMA alone is electrically insulating and therefore insufficient to recapitulate the bioelectrical characteristics of neural tissue, which play an important role in axonal guidance and signal transmission [[14], [15], [16]]. To address this limitation, conductive fillers can be incorporated into hydrogel matrices to provide electrically compatible interfaces. Silver nanowires (AgNWs) are particularly advantageous in this context due to their high-aspect-ratio morphology, which enables the formation of percolating conductive networks at relatively low loading levels. Compared with particulate conductive additives, AgNWs can achieve effective conductivity while minimizing alterations to bulk mechanical properties. Moreover, their one-dimensional geometry is well suited for integration with anisotropic scaffold architectures, allowing directional electrical conduction aligned with axonal growth pathways [[16], [17], [18], [19], [20]]. Silver nanomaterials have attracted increasing interest in neural scaffold design owing to their combined antibacterial activity, immunomodulatory potential, and ability to confer electrical conductivity to otherwise insulating matrices [21,22]. Silver nanowires, in particular, demonstrate superior biocompatibility compared to other nanostructures, attributed to their modulated ionic release kinetics and high-aspect-ratio morphology [23].

Scheme 1.

Preparation process and molecular mechanism of bioinspired conductive anti-inflammatory A-NAC-AgNW-GM scaffolds.

The practical application of silver nanowires (AgNWs) in conductive scaffolds is constrained by their inherent limitations—high contact resistance and poor dispersion stability, both of which degrade bulk conductivity [24]. To address these challenges, this study employs N-acetylcysteine (NAC) surface modification, leveraging its thiol groups to form robust sulfur-silver coordination bonds, which stabilize nanowire interfaces and minimize electron transfer barriers. Concurrently, the carboxyl and amino moieties in NAC impart zwitterionic characteristics to AgNWs, enhancing their colloidal stability within hydrogel matrices [25]. This dual-functional modification not only optimizes nanowire dispersion but also establishes continuous conductive pathways, synergistically enhancing scaffold performance while maintaining biocompatibility. Beyond its role as a material modifier, N-acetylcysteine (NAC) serves as a potent therapeutic agent, capable of scavenging free radicals and attenuating ROS-induced neural apoptosis [[26], [27], [28], [29]].

In this study, we combined GelMA with N-acetylcysteine-modified silver nanowires (NAC-AgNWs) and fabricated an anisotropic conductive hydrogel scaffold using directional freeze-casting (Scheme 1). This design leverages the complementary properties of GelMA and AgNWs to achieve a soft, cell-permissive matrix with physiologically relevant electrical conductivity and aligned microchannel architecture. By further introducing NAC modification to improve nanowire dispersion and stability, the scaffold was designed to provide structural guidance, electrical compatibility, and microenvironmental modulation in a unified platform for spinal cord repair. By coupling NAC-AgNW synergistic immune-metabolic reprogramming with topography-electroactive dual-guidance, this design paradigm provides a novel strategy for engineering biomaterials that synergistically address neuroinflammation and neural regeneration in spinal cord repair.

2. Materials and methods

2.1. Fabrication of aligned GelMA scaffolds (A-GelMA)

Aligned GelMA (EFL-GM-90, Engineering for Life, China) hydrogels were prepared by unidirectional freeze-casting as previously reported [30]. GelMA was dissolved in deionized water at a concentration of 25 mg/mL. Redox-initiated polymerization was achieved by adding ammonium persulfate (APS, 2 mg/mL) and tetramethylethylenediamine (TEMED, 3 μL/mL) (Aladdin, China). The precursor solution was injected into custom polytetrafluoroethylene molds (inner diameter: 8 mm; outer diameter: 12 mm; height: 30 mm) and vertically placed on a cooling platform maintained at −40 °C with partial liquid immersion to generate a unidirectional thermal gradient. After 20 min of directional freezing, the samples were transferred to −20 °C and maintained for 24 h to complete redox crosslinking. The resulting cryogels were demolded, rinsed thoroughly with deionized water (three cycles), and stored in phosphate-buffered saline (PBS) at 4 °C prior to use.

2.2. Preparation of NAC-functionalized silver nanowires (NAC-AgNWs)

Silver nanowires (AgNWs; diameter: 90 nm; length: 20–60 μm; Xfnano Materials Tech Co., Ltd., China) were dispersed in deionized water to obtain a suspension with a concentration of 2 mg/mL. In parallel, N-acetylcysteine (NAC; Beyotime Biotechnology, Shanghai, China) was dissolved in deionized water to prepare a 5 mM solution. The AgNW suspension and NAC solution were mixed at a volume ratio of 1:1 and homogenized by ultrasonication. The resulting mixture was incubated in a shaking water bath at 30 °C and 300 rpm for 12 h, followed by static incubation at room temperature for an additional 12 h to facilitate NAC adsorption onto the AgNW surface. Excess and unbound NAC molecules were removed by centrifugation (3500×g, 5 min), repeated twice. The purified NAC-functionalized silver nanowires (NAC-AgNWs) were collected for subsequent experiments.

2.3. Fabrication of conductive aligned scaffolds

The GelMA solution was prepared by dissolving GelMA powder in deionized water at a concentration of 25 mg/mL. Ammonium persulfate (APS, 2 mg/mL) and tetramethylethylenediamine (TEMED, 3 μL/mL) were then added to initiate redox polymerization, yielding a homogeneous precursor containing NAC-functionalized silver nanowires (NAC-AgNWs) at a final concentration of 8 mg/mL.

The pre-gel solution was poured into custom-fabricated polytetrafluoroethylene (PTFE) molds (inner diameter: 8 mm; outer diameter: 12 mm; height: 30 mm) and vertically positioned on a cooling platform partially immersed in liquid at −40 °C. This configuration established a unidirectional thermal gradient along the longitudinal axis for 20 min, guiding ice crystal alignment. After complete freezing, the samples were transferred to a −20 °C freezer and maintained for 24 h to allow redox crosslinking and stabilization of the anisotropic microstructure. The resulting cryogels were demolded, washed thoroughly with deionized water (three cycles), and stored in phosphate-buffered saline (PBS) until further characterization.

2.4. Morphological characterization of scaffolds

The hydrogels were sectioned into thin slices and examined using scanning electron microscopy (SEM; SU8600, Hitachi, Japan). Elemental distributions of silver (Ag) and sulfur (S) were analyzed by energy-dispersive X-ray spectroscopy (EDS) coupled to the SEM system. Ultraviolet–visible (UV–Vis) absorption spectra of AgNWs and NAC-AgNWs were acquired using a UV–Vis spectrophotometer. The hydrogels were sectioned into thin slices and examined using scanning electron microscopy (SEM; SU8600, Hitachi, Japan). Elemental distributions of silver (Ag) and sulfur (S) were analyzed by energy-dispersive X-ray spectroscopy (EDS) coupled to the SEM system. Ultraviolet–visible (UV–Vis) absorption spectra of AgNWs and NAC-AgNWs were acquired using a UV–Vis spectrophotometer.

2.5. Physicochemical characterization of scaffolds

The rheological properties of the hydrogels were characterized using a HAAKE MARS rheometer (Thermo Fisher Scientific). Strain amplitude sweep tests were performed over a strain range of 0.01%–100% at a fixed angular frequency of 10 rad/s to determine the storage modulus (G′) and loss modulus (G″). Frequency sweep measurements were subsequently conducted within a frequency range of 0.1–10 Hz. Time-dependent viscoelastic behavior was evaluated by time sweep tests at a frequency of 1 Hz, a strain of 1%, and a temperature of 25 °C for 10 min.

The swelling behavior of the hydrogels was assessed by recording the equilibrium wet weight (Weq) after immersion in phosphate-buffered saline (PBS) for 24 h and the initial dry weight (Wd). The swelling ratio (SR) was calculated according to the following equation $\left(1\right)$:

$$\begin{array}{c}SR=\left(Weq-Wd\right)/Weq\times 100\%\end{array}$$ (1)

An in vitro degradation assay was conducted to evaluate the degradation behavior of the hydrogels. The hydrogels were incubated in collagenase II solution (5.0 U mL⁻¹ in PBS, pH 7.4; Solarbio, China) at 37 °C in a shaker incubator (80 rpm). At predetermined time points (0, 1, 2, 3, 6, and 12 h), samples were collected, rinsed with PBS, lyophilized, and weighed. The mass retention was calculated according to the following equation $\left(2\right)$:

$$\begin{array}{c}Massretention=Wt/Wo\times 100\%\end{array}$$ (2)

Where $Wo$ and $Wt$ denote the initial dry weight and the remaining dry weight of the hydrogel at time $t$, respectively.

Mechanical properties were evaluated using a universal testing machine (INSTRON, MA, USA). Cylindrical hydrogel specimens (5 mm in diameter and 5 mm in height) were compressed at a displacement rate of 1 mm/min using a 1000 N load cell. Young's modulus was calculated from the linear region of the stress–strain curves between 5% and 10% strain. All data were analyzed using GraphPad Prism software (version 8.4.0).

2.6. Electrochemical characterization of scaffolds

To enable a reliable comparison of conductive performance among different hydrogel formulations, electrical and electrochemical properties were evaluated using complementary techniques. The four-point probe measurement was used to quantify the bulk electrical conductivity of the hydrogels under steady-state conditions, whereas cyclic voltammetry and electrochemical impedance spectroscopy were used to compare the potential-dependent current response and the frequency-dependent charge-transport behavior, respectively.

The four-point probe system (RTS-9, Guangzhou Four-Point Probe Technology Co., China) was used to measure the electrical conductivity of hydrogels. Electrochemical analyses were conducted using a CHI660E workstation (Beijing Chinese Science Days Technology Co., China) configured with a three-electrode setup: hydrogels (10 × 10 × 1 mm³) served as the working electrode, a platinum wire functioned as the counter electrode, and an Ag/AgCl electrode was utilized as the reference electrode. Electrochemical impedance spectroscopy (EIS) measurements were performed at open-circuit potential over a frequency range of 0.01 Hz to 10 kHz. Cyclic voltammetry (CV) was performed in a 0.1 M phosphate-buffered saline (PBS) solution at pH 7.4, with a potential range of −0.1 to 1.0 V and a scan rate of 10 mV/s.

2.7. Assessment of NAC sustained release from scaffolds

Conductive aligned hydrogels were prepared from 1 mL of a homogeneous NAC-AgNWs precursor as described above and immersed in 1 mL of culture medium. At predetermined time points (0, 24, 48, …, 192 h), the entire medium was collected and replaced with an equal volume of fresh medium. The collected samples containing released N-acetylcysteine (NAC), characterized by free sulfhydryl (–SH) groups, were analyzed.

The cumulative NAC concentration was quantified using a 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB) assay kit (Beyotime Biotechnology, Shanghai, China) according to the manufacturer's instructions. The cumulative amount of released NAC at each time point was calculated using the following equation $\left(3\right)$:

$$\begin{array}{c}Cx=Ct+Ax\end{array}$$ (3)

where Cx: total NAC concentration at timepoint X, Ct: cumulative concentration before X, and Ax: incremental concentration at X.

2.8. Cellular cultivation

PC12 and BV2 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin. Cells were maintained under standard culture conditions at 37 °C in a humidified atmosphere containing 5% CO₂. The culture medium was refreshed every 24 h to ensure adequate nutrient supply.

2.9. Cell vitality and cell compatibility

Hydrogels were pre-equilibrated in sterile phosphate-buffered saline (PBS) and sterilized by ultraviolet (UV) irradiation for 30 min. PC12 cells were seeded onto the hydrogels when reaching approximately 70% confluence. After an initial attachment period of 30 min, fresh culture medium was gently added to facilitate stable cell adhesion.

Cell viability was evaluated using a LIVE/DEAD Viability/Cytotoxicity Kit (Beyotime Biotechnology, Shanghai, China), containing calcein-AM (10 μg/mL) for labeling live cells (green fluorescence) and ethidium homodimer-1 (EthD-1, 5 μg/mL) for labeling dead cells (red fluorescence). Fluorescence images were acquired using a laser scanning confocal microscope (Eclipse Ti-S, Nikon, Japan). Cell morphology and viability were quantitatively analyzed using ImageJ software. At days 1 and 3, cells were incubated with Cell Counting Kit-8 (CCK-8, Beyotime Biotechnology, Shanghai, China) working solution for 1 h, after which the absorbance at 450 nm was measured to quantify cellular metabolic activity. For cytoskeletal visualization, we used FITC-phalloidin (1 h) and imaged via confocal microscopy. To assess migration, we put a sterile polydimethylsiloxane (PDMS) strip (0.4 cm wide) in the middle. Cells grew on one side for 12 h and the strip was removed. After 48 h of serum-free incubation, migrated cells were stained with FITC-phalloidin and quantified using ImageJ.

2.10. Assessment of the antioxidant capacity of scaffolds

Prior to the experiments, hydrogel samples were sterilized by ultraviolet (UV) irradiation for 30 min. PC12 cells were seeded in 24-well plates at a density of 1 × 10⁵ cells per well and cultured overnight. Transwell inserts containing hydrogel samples (Φ 4 mm × 2 mm) were then placed into the wells and co-cultured with PC12 cells for 24 h.

To establish an oxidative stress microenvironment, the culture medium was replaced with fresh medium supplemented with 0.1 mM tert-butyl hydroperoxide (TBHP), a concentration that reduced PC12 cell viability to near the IC50 level in preliminary dose–response experiments (Fig. S 4A-C). After 6 h of TBHP exposure, intracellular reactive oxygen species (ROS) levels were assessed using 2′,7′-dichlorofluorescin diacetate (DCFH-DA, 10 μM; Beyotime Biotechnology, Shanghai, China). Cells were incubated with the probe at 37 °C for 30 min, followed by fluorescence imaging using a fluorescence microscope. Fluorescence intensity was analyzed to evaluate the antioxidant capacity of the scaffolds.

Fig. 4.

In Vitro Study on the Immunomodulatory Effects of Different Scaffolds on BV2 Cells. (A) RT-qPCR results of gene expression levels of pro-inflammatory factors IL-6, iNOS, and CD86 in BV2 cells. (B) RT-qPCR assay of the expression levels of anti-inflammatory cytokines IL-10, Arg-1, and CD206 in BV2 cells. (C-D) Representative immunofluorescence images of iNOS(red), Arg-1(red), and Iba-1(green) of BV2 cells in each group. Scale bar: 50 μm. (E-F) Quantitative analysis of iNOS and Arg-1 fluorescence intensity in each group. (G) Western blot analysis of iNOS and Arg-1 protein expression in BV2 cells after corresponding treatments. (H-I) Statistical analysis of iNOS and Arg-1 protein intensity. ∗ represents P < 0.05. ∗∗ represents P < 0.01. ∗∗∗ represents P < 0.001. ∗∗∗∗ represents P < 0.0001. “ns” stands for “no statistical significance”. Data are represented as mean ± SD (n = 3).

2.11. In vitro evaluation of macrophage polarization

BV2 microglial cells were seeded in 24-well plates at a density of 1 × 10⁵ cells per well and cultured overnight. Transwell inserts containing hydrogel samples (Φ 4 mm × 2 mm) were then placed into the wells and co-cultured with BV2 cells for 24 h. To induce an inflammatory phenotype, lipopolysaccharide (LPS, 1 μg/mL) was added to the culture medium, and cells were further incubated for 24 h.

Following stimulation, macrophage polarization was evaluated by analyzing the expression of inflammation-related genes and proteins. Quantitative reverse transcription polymerase chain reaction (qRT-PCR), immunofluorescence staining, and Western blotting were performed to assess the relative expression levels of pro- and anti-inflammatory markers.

2.12. Quantitative reverse transcription polymerase chain reaction (qRT-PCR)

Total RNA was isolated from cells with the EZ-press RNA Purification Kit (EZBioscience). Subsequently, complementary DNA (cDNA) was generated utilizing a reverse transcription kit from Bio-Rad. Quantitative real-time PCR was conducted with PowerUp SYBR Green Master Mix (Applied Biosystems, Thermo Fisher Scientific). The primer sequences utilized in these studies are enumerated in Table S 1of the supplementary information.

2.13. Spinal cord transection model and treatment

Adult female SD rats (220-230 g body weight) came from Wenzhou Medical University's Experimental Animal Center. The Wenzhou Medical University Experimental Animal Ethics Committee approved all procedures (wydw2025-0412).

We used sixty rats divided equally into four 15-rat groups: spinal cord injury group, A-GelMA group, A-AgNW-GM group, and A-NAC-AgNW-GM group. Rats got 2% pentobarbital sodium through belly injections for sleep. Surgeons removed back bone sections at T9-10 to show the spinal cord, then cut out a 2 mm piece. After stopping blood loss, they put hydrogel in the damaged area and sewed up muscle layers and skin separately. For three days after surgery, rats received penicillin shots under skin and had their bladders pressed twice daily until normal peeing returned.

2.14. Assessment of functional behavior

Hindlimb locomotor recovery was evaluated weekly for 6 weeks post-injury using the Basso–Beattie–Bresnahan (BBB) scoring system. Rats were allowed to walk freely along a straight runway (5 cm × 50 cm), and hind paw movements were recorded using a high-speed camera. Kinematic analysis was performed with DeepLabCut (v2.3.5) by tracking key joint landmarks (iliac crest, hip, knee, ankle, and toe) to generate stick diagrams of hindlimb motion.

Motor evoked potentials (MEPs) were recorded to assess electrophysiological recovery. Recording electrodes were inserted into bilateral gastrocnemius muscles, with a stimulating electrode placed rostral to the lesion and a reference electrode positioned subcutaneously in the abdomen. Electrical stimulation was applied at 10 mA (0.1 ms pulse width, 1 Hz), and MEP amplitudes were recorded at 15 s intervals for quantitative analysis.

2.15. Cortical spinal tract (CST) tracing

We tracked corticospinal tract (CST) axons using 10% biotin dextran amine (BDA). Two weeks before euthanizing the rats, we injected BDA (0.5 μL per spot) into six locations in each brain hemisphere under anesthesia. Injection sites [31] were: 1.2 mm deep from the cerebral surface, 0.5 mm, 1.0 mm, and 2.0 mm posterior to the bregma, and 1.0 mm and 2.0 mm lateral to the midline (sagittal suture). After injecting, we left the needle in place for 5 min to let BDA spread, then slowly pulled it out. Fourteen days later, we deeply anesthetized the rats, flushed their hearts with saline, and fixed tissues with 4% formaldehyde. We counted BDA-labeled axons in equally magnified images to measure nerve regrowth density.

2.16. RNA sequencing (RNA-seq) analysis

Spinal cord samples were collected from the SCI group and A-NAC-AgNW-GM group at 6 weeks after surgery. Using the lesion center as the reference point, spinal cord segments including 5 mm rostral and 5 mm caudal to the injury site were carefully harvested. The collected tissues were immediately frozen in liquid nitrogen and sent to OE Biotech (Shanghai, China) for RNA sequencing. Data analysis was performed using the OE Cloud platform provided by OE Biotech.

2.17. Histopathological examination

At 6 weeks post-injury, rats were deeply anesthetized and transcardially perfused with saline followed by 4% paraformaldehyde. Spinal cord segments encompassing the lesion site (±5 mm) were harvested, cryoprotected, and sectioned into 20 μm-thick sagittal sections using a freezing microtome. Sections were stained with hematoxylin and eosin (H&E), Nissl, and Luxol Fast Blue (LFB) according to the manufacturers’ protocols (Solarbio, China).

Bladder tissues were collected and subjected to Masson's trichrome staining (Solarbio, China) to evaluate tissue remodeling. Major organs, including heart, liver, spleen, lung, and kidney, were harvested and processed for H&E staining to assess the in vivo biosafety of the implanted scaffolds.

2.18. Western blot analysis

We collected proteins using RIPA buffer with protease and phosphatase inhibitors. We measured protein amounts with a BCA kit. Next, we separated proteins by size using SDS-PAGE. We transferred them to PVDF membranes (Millipore, Germany).

After blocking membranes with 5% skim milk, we incubated them overnight at 4 °C with these primary antibodies: iNOS (1:1000, Affinity,AF0199), Arg-1 (1:1000, Affinity,DF6657), CD68 (1:1000, Santa Cruz, sc-20060), FGF13 (1:1000, Affinity, DF4699), CaMK2A (1:1000, abcam,ab134041), p-CaMK2A (1:1000, Affinity,AF3493), CREB (1:1000, Affinity,AF6188), p-CREB (1:1000, Affinity,AF3189), β-actin (1:5000, Zenbio,P60709). We washed membranes three times with Tris-buffered saline Tween 20 (TBST), then treated them with HRP-linked secondary antibodies for 2 h at room temperature. Finally, we scanned bands using Cytiva's Amersham ImageQuant 800 and analyzed data with ImageJ.

2.19. Immunofluorescence examination of cellular and tissue samples

We fixed treated BV2 cells on glass slides. We permeabilized the membranes and blocked them with goat serum. The cells stayed with primary antibodies overnight, then secondary antibodies at room temperature. We stained nuclei with DAPI and sealed the slides.

For spinal cords, we flushed them with saline and 4% paraformaldehyde through the heart. These tissues soaked in 30% sucrose before freezing. A CryoStar NX70 machine (Thermo Fisher) made thin frozen slices. We stained these slices like the cells. We took pictures with a confocal microscope. Key antibodies used: iNOS (1:500, Affinity, AF0199), Arg-1 (1:500, Affinity, DF6657), CD68 (1:500, Santa Cruz, sc-20060), Iba-1 (1:1000, Abcam, ab178846), GFAP (1:1000, Abcam, ab4674), MAP-2 (1:500, Affinity, AF4081), MBP (1:500, Affinity, AF4085), 5-HT (1:500, Sigma, S5545), TH (1:1000, Abcam, ab112), Neun (1:1000, Abcam, ab104224), PSD95 (1:1000, Abcam, ab238135), Syn (1:500, Affinity, BF0348), CaMK2A (1:500, Cell Signaling, 50049), Cav1.2(1:500,proteintech, 21774-1-AP). Fluorescent secondary antibodies matched each animal type (Abcam).

2.20. ICP-MS analysis of tissue silver content

Animals were euthanized at days 1, 7, and 42 post-injury, and major organs were collected. Approximately 0.1 g of tissue was digested in nitric acid using microwave-assisted digestion. Silver content was quantified by ICP-MS (Aglient 7800).

2.21. Primary cortical neuron isolation and culture

Primary cortical neurons were isolated from embryonic Sprague–Dawley rats (E16–E18). Cerebral cortices were dissected, mechanically dissociated, and plated on poly-L-lysine–coated dishes or scaffolds. After initial attachment, the medium was replaced with Neurobasal medium supplemented with B27 and GlutaMAX (Thermo Fisher). Neurons were maintained at 37 °C with 5% CO₂ and used for subsequent experiments at the indicated time points.

2.22. Lentiviral transduction of primary cortical neurons

Lentiviral transduction was employed to overexpress FGF13 in primary cortical neurons using lentiviral vectors encoding rat FGF13 or the corresponding negative control (GeneChem, Shanghai, China). For transduction efficiency validation, neurons were first plated on standard culture dishes at an appropriate density and exposed to lentiviral particles at the indicated multiplicity of infection in the presence of polybrene. After 12 h of incubation, the viral medium was replaced with fresh Neurobasal medium, and neurons were further cultured for 72 h to allow stable transgene expression. FGF13 overexpression was further confirmed by Western blot analysis.

For scaffold-based experiments, primary cortical neurons were seeded directly onto the scaffolds and allowed to attach for 12 h prior to lentiviral transduction. Lentiviral particles encoding FGF13 were then added directly to the culture medium, and transduction was performed in situ on the scaffolds for 12 h. The infection medium was subsequently replaced with fresh Neurobasal medium, and neurons were maintained for an additional 72 h before subsequent experiments.

2.23. Statistical analysis

All quantitative data are presented as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism 8.4.0. Comparisons among multiple groups were conducted using one-way or two-way analysis of variance (ANOVA), followed by Bonferroni post hoc tests where appropriate. A p value < 0.05 was considered statistically significant.

3. Results

3.1. The physicochemical characteristics and manufacturing of biomimetic conductive scaffolds

The NAC modification strategy for silver nanowires (AgNWs) primarily involves the formation of Ag–S bonds through coordination between NAC's thiol groups (-SH) and silver atoms. Simultaneously, the carboxyl and amino groups of NAC enhance hydrophilicity and biocompatibility. NAC-modified AgNWs exhibited stable dispersion in aqueous solution (Fig. S 1A). Both AgNWs and NAC-AgNWs showed distinct localized surface plasmon resonance (LSPR) absorption in the 350–400 nm range. Following NAC modification, the absorption peak exhibited a slight red shift with reduced intensity, confirming successful surface functionalization (Fig. S 1B). EDS elemental mapping demonstrated complete spatial overlap between sulfur (yellow) and silver (green) signals (Fig. 1A and B), providing direct evidence for the formation of Ag–S coordination bonds.

Fig. 1.

Preparation and characterization of A-NAC-AgNW-GM. (A) SEM and EDS of NAC-AgNWs. Scale bar: 10 μm. (B) SEM and EDS of NAC-AgNW. Scale bar: 1 μm. (C) Longitudinal SEM images of A-GelMA and R-GelMA. Scale bar: 400 μm. (D) Transverse SEM images of A-GelMA and R-GelMA. Scale bar: 400 μm. (E) SEM and EDS of A-AgNW-GM; scale bar: 500 μm (F) SEM and EDS of A-NAC-AgNW-GM. Scale bar: 500 μm. (G) Pore size distribution chart of frozen oriented scaffolds. (H) The corresponding EDS spectrum of A-AgNW-GM. (I) The corresponding EDS spectrum of A-NAC-AgNW-GM.

We fabricated aligned GelMA scaffolds (A-GelMA) using unidirectional freezing. GelMA precursor solution was injected into custom Teflon tubes (8 mm inner/12 mm outer diameter, 10 mm height). The tubes were pressed against −40 °C coolant to create axial temperature gradients, forming aligned ice crystals. Redox polymerization using solid APS achieved covalent crosslinking during freezing. To match spinal cord stiffness (100-3000 Pa) [11], we tested 2.5%, 3%, 4%, 5% GelMA concentrations. The 2.5% group showed 1.9 kPa modulus (Fig. S 1C-D), selected for spinal applications. Randomly porous GelMA scaffolds (R-GelMA) made in −40 °C freezer. SEM revealed A-GelMA's parallel grooves versus R-GelMA's disorder (Fig. 1C and D), with uniform 150-200 μm pores (Fig. 1G). Similarly, by injecting the NAC-AgNWs precursor solution into the same custom-made Teflon tube, we obtained an axially aligned GelMA scaffold doped with NAC-AgNWs (A-NAC-AgNW-GM). The conductivity of the spinal cord ranges from 0.02 to 0.60 S/m [11,13]. Consequently, we determined that the conductivities of NAC-AgNWs solutions at concentrations of 4 mg/ml, 8 mg/ml, 12 mg/ml, and 16 mg/ml were 0.126 S/m, 0.247 S/m, 0.267 S/m, and 0.185 S/m respectively (Fig. S 2A). It was observed that conductivity tended to decrease with increasing NAC-AgNWs concentration, potentially due to agglomeration within the scaffold disrupting the conductive network. Considering that higher conductive filler loading may influence cellular responses, cytocompatibility was further examined using Live/Dead staining(Fig. S 2B-C). Hydrogels containing 4 and 8 mg/ml NAC-AgNWs exhibited a high proportion of live cells with only sporadic dead cells observed. When the NAC-AgNWs concentration was increased to 12 mg/ml, a visible increase in dead cells was detected. This trend became more evident at 16 mg/ml, indicating a progressive reduction in cell viability at higher concentrations. Based on these observations, 8 mg/ml NAC-AgNWs was selected for subsequent experiments, as this concentration maintained good cell viability while providing enhanced electrical conductivity. To confirm successful incorporation of NAC-AgNWs into hydrogels, we performed energy-dispersive X-ray spectroscopy (EDS) elemental mapping. Comparative analysis between AgNW-doped aligned GelMA scaffold (A-AgNW-GM) (Fig. 1E and H) and A-NAC-AgNW-GM (Fig. 1F and I) revealed distinct Ag signal patterns - while both exhibited green Ag signatures, A-NAC-AgNW-GM displayed continuous silver distribution along the scaffold axis. The AgNW-GM showed minimal yellow S signals (likely residual APS), whereas A-NAC-AgNW-GM demonstrated intense, uniformly distributed S signals overlapping extensively with Ag patterns. This yellow-green spatial co-localization verifies homogeneous NAC-AgNW integration within the aligned matrix (Fig. 1E–F, H-I).

Fig. 2.

Properties characterization of different scaffolds. (A) Oscillating time-sweep measurements of scaffolds. (B) Frequency-sweep measurements of scaffolds. (C) Strain-sweep measurements of scaffolds (D) Swelling ratio of scaffolds. (E) Stress-strain curves of scaffolds (F) Young's modulus of scaffolds. (G) The conductivity characterization of scaffolds. (H) Cyclic voltammograms of scaffolds. (I) Nyquist curves of scaffolds. ∗ represents P < 0.05. ∗∗ represents P < 0.01. ∗∗∗ represents P < 0.001. Data are represented as mean ± SD (n = 3).

We evaluated hydrogel mechanical properties through storage modulus (G′) measurements and swelling capacity analysis. Three formulations - A-GelMA, A-AgNW-GM (8 mg/ml AgNW), and A-NAC-AgNW-GM (8 mg/ml NAC-AgNW) - underwent comparative rheological testing. All groups maintained stable G' > G″ over time, confirming elastic gel stability. A-NAC-AgNW-GM exhibited lower G′ than A-AgNW-GM, likely due to NAC-induced viscoelastic buffering and anti-aggregation effects. During frequency sweep testing (0.1–10 Hz), A-GelMA exhibited a decrease in G′ at high frequencies, suggesting internal network disruption preventing reorganisation. A-AgNW-GM showed an increase in G′ at high frequencies, indicating rearrangement of internal AgNW aggregates under high-frequency shear. Conversely, A-NAC-AgNW-GM demonstrated a steady increase in G'. During strain-scanning rheometry, A-NAC-AgNW-GM exhibited more stable modulus changes compared to other groups (Fig. 2A–C). Similarly, in stress-strain curves, A-NAC-AgNW-GM withstood 80% deformation strain, demonstrating superior deformation resistance. The Young's modulus of A-NAC-AgNW-GM was approximately 3 kPa, closely approximating that of spinal cord tissue (Fig. 2E and F). After equilibration in PBS buffer for 24 h, all three scaffolds exhibited high swelling rates exceeding 90%, demonstrating excellent water absorption capacity. This enables gentle expansion to fill damaged cavities. Although AgNWs are inherently hydrophobic, their swelling behaviour was not significantly impaired. Furthermore, the surface carboxyl and amino groups formed upon NAC modification of AgNWs enhanced hydration (Fig. 2D). The degradation behavior of the hydrogels was evaluated under enzymatic conditions. As shown in Fig. S2E, A-GelMA underwent rapid mass loss and was almost completely degraded within 12 h. In contrast, A-AgNW-GM and A-NAC-AgNW-GM exhibited markedly higher mass retention throughout the observation period, indicating a more suitable and controllable degradability for spinal cord repair in vivo. This behavior is most likely associated with changes in the internal network structure of the hydrogel. The incorporation of AgNWs, with or without NAC modification, leads to the formation of a denser and mechanically reinforced polymer network, which can restrict enzyme diffusion into the hydrogel matrix. As a result, the accessibility of cleavage sites within the GelMA backbone is reduced, thereby slowing enzyme-mediated degradation.

Electrical signals are essential to the nervous system and serve as the basis for both electrophysiology and neuroscience [32]. Therefore, in this work, we used electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), and a four-point probe to assess the hydrogels' conductive qualities. The conductivities of A-GelMA, A-AgNW-GM, A-NAC-AgNW-GM exhibited conductivities of 0.03 S/m, 0.14 S/m, and 0.24 S/m, respectively. Following NAC modification of AgNWs, NAC-AgNWs reduced the agglomeration of AgNWs, enabling a continuous internal conductive network and enhancing conductivity. CV revealed anodic peaks for A-GelMA, A-AgNW-GM, A-NAC-AgNW-GM exhibited anodic peaks at 0.39 V, 0.31 V, and 0.29 V, respectively, and reduction peaks at −0.8 V, −0.84 V, and −0.84 V, respectively. The A-NAC-AgNW-GM hysteresis loop exhibited a larger area, attributable to its superior internal conductive network. Similarly, electrochemical impedance spectroscopy (EIS) revealed that A-NAC-AgNW-GM exhibited the smallest arc diameter and the lowest charge transfer resistance, indicating superior conductive capability. Furthermore, the bulb experiment visually demonstrated that A-NAC-AgNW-GM produced the brightest illumination (Fig. 2G–I; Fig. S 2F). To further investigate whether NAC undergoes physical release within the scaffold to exert a sustained-release effect, we analyzed typical thiol groups post-NAC release. Employing the DTNB method to measure A-NAC-AgNW-GM solution concentrations in PBS at various time points, we found that 1 ml of A-NAC-AgNW-GM sustained-release NAC, accumulating to 1.8 mM by day 8. This provides favorable support for early spinal cord injury repair (Fig. S 2G).

3.2. In vitro biocompatibility study and antioxidant capacity of biomimetic conductive oriented scaffolds

To investigate the scaffold's cellular compatibility, PC12 cells were cultured on A-GelMA, A-AgNW-GM, and A-NAC-AgNW-GM scaffolds. Following 1 and 3 days of culture, LIVE/DEAD staining revealed predominantly viable cells on all scaffolds, with extensive calcein-AM–positive (green) signals and only sparse PI-positive (red) cells. No apparent cytotoxicity was observed upon incorporation of AgNWs or NAC-AgNWs. Notably, PC12 cells cultured on A-AgNW-GM and A-NAC-AgNW-GM exhibited enhanced cell density by day 3 compared with A-GelMA, indicating improved proliferative support (Fig. 3A and E; Fig. S 3A). Live/dead staining further confirmed excellent long-term cytocompatibility, with >90% cell viability maintained on all scaffolds at days 7 and 14 (Fig. S 3B–D). Interestingly, analysis of PC12 cell migration capacity via phalloidin staining revealed that PC12 cells cultured on A-NAC-AgNW-GM migrated significantly greater distances than those in other groups. Quantitative results revealed significantly higher cell coverage on A-NAC-AgNW-GM compared to other groups (Fig. 3B and F), indicating its superiority in supporting and promoting long-distance cell migration. PC12 cells exhibited irregular alignment on R-GelMA. Cells on the directed scaffold, on the other hand, showed properly aligned orientation along the directional axis of the scaffold. Furthermore, Fig. 3C revealed the orientation of PC12 cell long axes. Along the A-GelMA, A-AgNW-GM, and A-NAC-AgNW-GM scaffolds within ±10° were 20%, 21%, and 27%, respectively. PC12 cells exhibited optimal alignment along the A-NAC-AgNW-GM scaffold direction, potentially due to its more uniform internal conductive network structure, which guided PC12 cell growth and migration more effectively. Excessive reactive oxygen species (ROS) production at the SCI site induces oxidative stress and cytotoxic inflammation, leading to widespread neuronal cell death at the injury site [33]. To validate the A-NAC-AgNW-GM scaffold's capacity to suppress intracellular ROS excess within the SCI microenvironment, the 2,7-dichlorodihydrofluorescein diacetate (DCFH-DA) probe assessed intracellular antioxidant properties in PC12 cells co-cultured with the scaffold (as described in Methods) (Fig. 3D and G). The TBHP control and A-GelMA groups showed strong green fluorescence signals, indicating high ROS content. However, the A-AgNW-GM group exhibited reduced fluorescence intensity, attributable to the redox reactions between trace silver and reactive oxygen species (ROS) such as superoxide anion (O₂⁻•) and hydroxyl radical (·OH) [34]. More significantly, the A-NAC-AgNW-GM group exhibited the lowest fluorescence intensity, primarily attributable to the dual antioxidant effects of NAC and AgNWs, demonstrating marked ROS suppression.

Fig. 3.

In vitro biocompatibility study and antioxidant capacity of different scaffolds. (A) Calcein-AM (green) and PI (red) immunofluorescence staining of PC12 cells in different scaffold groups. Scale bar: 100 μm. (B) Fluorescence images showing the migration of PC12 cells on different scaffolds along the longitudinal direction. Scale bar: 200 μm. (C) The cell morphology of PC12 cells on the different scaffold groups and distribution charts of PC12 cells on the different scaffold groups. Scale bar: 200 μm. (D) DCFH-DA staining of PC12 cells treated with different scaffolds. Scale bar: 100 μm. (E) Quantitative analysis of the live/dead cell ratio of different groups for PC12 cells (F) Percentage of cell-covered area at different regions of the migration zone. (G) Quantitative analysis of DCFH-DA relative intensity in different scaffold groups. ∗ represents P < 0.05. ∗∗ represents P < 0.01. ∗∗∗ represents P < 0.001. ∗∗∗∗ represents P < 0.0001. “ns” stands for “no statistical significance”. Data are represented as mean ± SD (n = 3).

3.3. A-NAC-AgNW-GM promotes M2 microglia polarization

Early on in spinal cord injury (SCI), the inflammatory response is a major factor aggravating damage to neural tissue [3]. As the main immune cells that enter SCI lesions after an injury, microglia are quickly activated and release inflammatory cytokines, which further enlarges the lesion size and slows down the healing process. Microglia have been categorized into pro-inflammatory M1 and anti-inflammatory M2 subtypes [35]. Inducible nitric oxide synthase (iNOS), interleukin-6 (IL-6), TNF-α,IL-1β,and CD86 are the main markers for M1 microglia, whereas interleukin-10 (IL-10), arginase-1 (Arg-1), and CD206 are the markers for M2 microglia. A-NAC-AgNW-GM was co-cultured with microglia and examined using Western blot (WB) tests, immunofluorescence labeling, and quantitative real-time polymerase chain reaction (qRT-PCR) to assess its impact on microglial polarization in the LPS-induced inflammatory milieu. qRT-PCR results indicated that, in comparison to the control group (LPS alone), A-NAC-AgNW-GM markedly downregulated the expression of proinflammatory genes (iNOS, IL-6, CD86, TNF-α, and IL-1β)) while enhancing the expression of anti-inflammatory genes (IL-10, Arg-1, and CD206) (Fig. 4A and B; Fig. S 3E), thereby substantiating the anti-inflammatory effect of A-NAC-AgNW-GM. As demonstrated by immunofluorescence staining (Fig. 4C–F), the presence of A-NAC-AgNW-GM significantly reduced iNOS fluorescence intensity while increasing Arg-1 fluorescence intensity compared to the control group (LPS alone), indicating that A-NAC-AgNW-GM effectively induces polarization of M1 to M2 microglia. As shown in Fig. 4G–I, WB assays revealed results consistent with immunofluorescence. Thus, A-NAC-AgNW-GM suppresses M1 phenotype differentiation and promotes M2 phenotype differentiation of microglia within the inflammatory microenvironment, thereby creating a favorable anti-inflammatory microenvironment conducive to spinal cord regeneration.

3.4. A-NAC-AgNW-GM markedly enhances hindlimb motor function and neural pathologies in rats

Fig. 5A depicts the in vivo experimental configuration. Fig. 5B illustrates the experimental framework for monitoring motion trajectories. Point-line diagrams are produced during the gait cycle based on five critical points: the crest, hip, knee, ankle, and toe. The figures, shown at various time intervals, illustrate the trajectory of the rat's lower limb movements, with the red segment denoting the moment the rat's lower leg is elevated from the ground (Fig. S 7A). The movement angles of the knee, hip, and ankle joints were documented during the complete movement cycle. The movement cycle was recorded along the point-line diagram as joint angles fluctuated. The A-NAC-AgNW-GM and A-AgNW-GM groups exhibited significantly superior locomotor function compared to the SCI group, including greater angular swing at the hip, knee, and ankle joints, confirming hindlimb movement recovery. In contrast, rats in the SCI and GelMA groups demonstrated noticeable dragging behavior (Fig. 5C–D, H; Fig. S 7D-G). In the footprint test (Fig. S 5B), SCI and GeMA groups exhibited pronounced hindlimb dragging, whereas A-AgNW-GM and A-NAC-AgNW-GM groups showed reduced hindlimb dragging and distinct off-ground movement. To thoroughly assess the impact of A-NAC-AgNW-GM on the incorporation of injured neurons at the injury site, we performed MEP studies to observe signal transduction efficacy. The A-NAC-AgNW-GM group produced the highest hindlimb MEP amplitude and the shortest latency, indicating superior restoration of signal transmission compared with A-GelMA group and A-AgNW-GM group (Fig. 5E; Fig. S 7B-C). Functional recovery was evaluated utilizing BBB scores at 1, 2, 3, 4, 5, and 6 weeks following the injury. One-way ANOVA followed by planned pairwise comparisons at the final time point, with the SCI group as the reference control. According to experimental results (Fig. 5G), the A-NAC-AgNW-GM group demonstrated markedly higher BBB scores than the other groups. The results indicate that the improved motor function in the A-NAC-AgNW-GM group likely arises from the formation of new circuit connections.

Fig. 5.

Treatment with A-NAC-AgNW-GM results in notable improvement of functional recovery and histological architecture in the injured rat spinal cord. (A) Schematic illustration of the in vivo therapeutic experiment. (B) Schematic illustration of hindlimb locomotor joint. (C) Color-coded stick views illustrating hindlimb movements in representative rats from each group. (D) Footstep analysis of motor hind limb movement in rats from each group. (E) Representative electrophysiological (MEP) traces of each group at week 6 post-injury. Each small division on the horizontal axis represents 5 ms, and each small division on the vertical axis represents 5 mV. Auxiliary guide lines indicate the definitions of latency (time from stimulus onset to the first detectable response) and peak-to-peak amplitude. (F) Representative H&E staining images of longitudinal spinal cord sections in each group. Scale bar: 1 mm. (G) BBB scale in each group. (H) Radar graph quantification of seven behavioral features of rats in each group. (I) Quantitative analysis of relative cavity area in longitudinal section of spinal cord. ∗ represents P < 0.05. ∗∗∗ represents P < 0.001. ∗∗∗∗ represents P < 0.0001. “ns” stands for “no statistical significance”. Data are represented as mean ± SD (n = 3).

Fig. 7.

A-NAC-AgNW-GM promotes neural regeneration. (A-D)Representative images of GFAP(green), MAP-2(red), MBP(red), 5-HT(red), TH(red)and DAPI (blue) immunofluorescence staining in each group 6 week after surgery. Scale bar: 1 mm. Scale bar (enlarged image): 50 μm. (E-I)Quantitative analysis of GFAP, MAP-2, MBP, 5-HT and TH fluorescence intensity in each group.(J) Western blot analysis of GFAP, MAP-2, MBP, 5-HT and TH protein expression in each group. (K-O) Statistical analysis of GFAP, MAP-2, MBP, 5-HT and TH protein intensity. ∗ represents P < 0.05. ∗∗ represents P < 0.01. ∗∗∗ represents P < 0.001. ∗∗∗∗ represents P < 0.0001. “ns” stands for “no statistical significance”. Data are represented as mean ± SD (n = 3).

Spinal cord function is closely linked to its highly differentiated tissue and structure; disruption of this architecture leads to severe neurological deficits [36]. Macroscopic observations (Fig. S 5A) revealed that the A-AgNW-GM and A-NAC-AgNW-GM groups exhibited a superior degree of connectivity and integration between the injured spinal cord region and surrounding intact tissue. In contrast, the SCI and A-GelMA groups remained to exhibit tissue damage, including cavities and/or tissue abnormalities. H&E staining revealed the volume of cavities in longitudinal spinal cord sections surrounding the injury site. As demonstrated by H&E staining, distinct irregular cavities were observed in the injury area of SCI rats. Compared to the SCI group, A-NAC-AgNW-GM significantly reduced the area of spinal cord injury cavities (Fig. 5F and I). Demyelination is a hallmark indicator of neural injury [37]. Therefore, LFB staining was used to label spinal cord samples, revealing markedly reduced demyelination in the A-NAC-AgNW-GM group compared to other experimental groups (Fig. S 5C-D). Nissl bodies serve as markers for neuronal loss and injury [38]. Local Nissl body content in spinal cord injury specimens was assessed via Nissl staining. In comparison to SCI and A-GelMA, both A-NAC-AgNW-G and A-AgNW-G demonstrated a considerable increase in Nissl bodies, with A-NAC-AgNW-GM displaying the highest quantity of Nissl body-positive neurons. (Fig. S 5E-F). These findings indicate that A-NAC-AgNW-GM implantation at the SCI site promotes spinal cord tissue regeneration, thereby effectively facilitating motor recovery. Fig. S 6A illustrates that no inflammatory or toxic responses were detected in the histological sections of major organ tissues across all groups, signifying that hydrogel transplantation did not inflict damage on key organs. ICP-MS analysis was performed to quantify silver accumulation in major organs (heart, liver, spleen, lungs, and kidneys) at 1, 7, and 42 days post-implantation. In both A-AgNW-GM and A-NAC-AgNW-GM groups, silver levels in all examined organs remained below 1 μg/g tissue at all time points and exhibited a clear time-dependent decline. By day 42, silver content in all organs approached background levels, with no evidence reported of persistent accumulation in the liver or kidneys (Fig. S 6B-C). Neurogenic bladder is a critical consequence in spinal cord injury patients, usually presenting as urinary incontinence, discomfort from retention, and psychological distress, significantly diminishing quality of life [39]. Therefore, we analyzed bladder functional reserve via Masson staining at 6 weeks post-surgery. Compared to SCI alone, the A-NAC-AgNW-GM group exhibited significantly thickened bladder walls, indicating that A-NAC-AgNW-GM effectively mitigates pathological damage to the bladder muscle (Fig. S 5G-H).

Fig. 6.

A-NAC-AgNW-GM alleviated inflammation of lesion site in the early stages of SCI. (A) Representative images of CD68(red) immunofluorescence staining of longitudinal spinal cord sections in each group 1 week after surgery. Scale bar: 1 mm. Scale bar (enlarged image): 100 μm. (B) Representative images of spinal cord tissue CD68(red) immunofluorescence staining in A-NAC-AgNW-GM group at 1 and 6 weeks, respectively. Scale bar: 1 mm. (C) Quantitative evaluation of CD68 fluorescence intensity. (D) Quantification of CD68 fluorescence intensity in A-NAC-AgNW-GM group at different time points after SCI. (E-F) Representative immunofluorescence images of iNOS(red) and Arg-1(purple) of longitudinal spinal cord sections in each group 1 week after surgery. Scale bar: 1 mm. Scale bar (enlarged image): 100 μm. (G-H) Quantitative analysis of iNOS and Arg-1 fluorescence intensity in each group. (I) Western blot analysis of iNOS, Arg-1 and CD68 protein expression in each group. (J-L) Statistical analysis of iNOS, Arg-1 and CD68 protein intensity. ∗ represents P < 0.05. ∗∗ represents P < 0.01. ∗∗∗ represents P < 0.001. ∗∗∗∗ represents P < 0.0001. “ns” stands for “no statistical significance”. Data are represented as mean ± SD (n = 3).

3.5. In vivo anti-inflammatory effects of A-NAC-AgNW-GM

Inflammatory responses typically trigger a cascade of secondary injuries following spinal cord injury [3]. Therefore, suppressing inflammation can create a favorable environment for spinal cord repair. In vitro experiments confirmed that A-NAC-AgNW-GM promotes M2 polarization of microglia. Subsequently, spinal cords were collected from rats in each treatment group to assess injury-induced inflammation and evaluate the implanted materials. The SCI group and A-GelMA group exhibited high CD68-positive cell density at the injury site, whereas the A-AgNW-GM group and A-NAC-AgNW-GM group showed significantly lower CD68-positive cell density compared to the SCI group, with the lowest CD68-positive cell density observed in the A-NAC-AgNW-GM group, indicating statistically significant differences. This suggests that A-NAC-AgNW-GM effectively suppresses severe inflammatory responses occurring in the early phase after spinal cord injury (Fig. 6A and C). Quantitative analysis of CD68 expression at the SCI site in the A-NAC-AgNW-GM group at 1 and 6 weeks post-surgery showed that the fluorescence intensity of CD68 in the A-NAC-AgNW-GM group at 6 weeks was slightly lower than at 1 week, but no statistically significant difference was observed (Fig. 6B and D). Immunofluorescence staining of the lesion area further evaluated the therapeutic effect of A-NAC-AgNW-GM on spinal cord injury (Fig. 6E–H). Immunofluorescence analysis revealed increased iNOS-positive and reduced Arg-1–positive cell densities in the SCI and A-GelMA groups. In contrast, A-AgNW-GM and A-NAC-AgNW-GM scaffolds markedly shifted macrophage polarization toward an Arg-1–dominant phenotype, with A-NAC-AgNW-GM exhibiting the highest Arg-1–positive cell density. Consistently, Western blot analysis demonstrated reduced expression of pro-inflammatory markers CD86 and iNOS, alongside enhanced Arg-1 protein levels in the A-NAC-AgNW-GM group (Fig. 6I–L). The findings indicate that A-NAC-AgNW-GM controls microglial polarization through the persistent release of NAC and the actions of AgNW, thereby decreasing inflammation in the early phases of damage and fostering a long-term conducive milieu for neuroregeneration.

3.6. A-NAC-AgNW-GM promotes neural regeneration and neural circuit formation

Immunofluorescence staining of the regenerative region was observed 6 weeks after spinal cord injury (Fig. 7A–I). In the SCI and A-GelMA groups, prominent glial fibrillary acidic protein (GFAP) labeling was visible around the injury site, a marker of central nervous system injury, indicating the formation of dense glial scar tissue surrounding the lesion [40]. In contrast, GFAP-positive staining was reduced in the A-AgNW-GM and A-NAC-AgNW-GM groups. Mature neurons stained for microtubule-associated protein-2 (MAP-2) were observed in the A-GelMA, A-AgNW-GM, and A-NAC-AgNW-GM groups, with particularly prominent findings in the A-NAC-AgNW-GM group (Fig. 7A and F). A-NAC-AgNW-GM inhibits the development of GFAP-positive astrocytic scarring, therefore creating a conducive environment for cellular proliferation. Furthermore, numerous neurons undergoing myelination were observed in the regenerating spinal cord treated with A-NAC-AgNW-GM. Myelin formation was indicated by positive staining for the myelination marker myelin basic protein (MBP), suggesting active myelination (Fig. 7B and G). Immunoreactive 5-HT (red) signals were extremely faint in the SCI control group, indicating injury-induced disruption of the descending serotonergic pathway originating from the brainstem raphe nuclei. Sparse 5-HT-positive fibers were visible extending into the injury zone in the A-GelMA and A-AgNW-GM groups. Notably, the A-NAC-AgNW-GM group exhibited the densest and most extensive network of 5-HT-positive fibers, showing substantial colocalization with GFAP-positive areas (yellow). This indicates regenerated 5-HT-positive axons successfully traversed the injury microenvironment composed of reactive astrocytes (Fig. 7C and H). Similarly, for the dopaminergic pathway marker TH, its expression pattern resembled that of 5-HT. The SCI control group exhibited only background-level signals. All material-implanted groups showed varying degrees of TH-positive regenerating axons, with the A-NAC-AgNW-GM group demonstrating the most pronounced regenerative effect, featuring extensive invasion of TH-positive fibers into the core lesion area (Fig. 7D and I). Western blot results were in good agreement with the immunofluorescence findings, showing reduced GFAP expression and increased levels of MAP-2, MBP, 5-HT, and TH in the A-NAC-AgNW-GM group compared with the SCI group (Fig. 7J–O). These results indicate that our A-NAC-AgNW-GM composite scaffold not only supports general neuronal (MAP-2) and myelin (MBP) regeneration but also specifically promotes axonal regeneration in serotonergic and dopaminergic descending pathways critical for motor function recovery [41].

To assess whether regenerating axons could traverse the lesion zone and establish potential synaptic connections with downstream neurons, we employed anterograde tracing using biotinylated dextran amine (BDA) injected into the primary motor cortex to specifically label and track regenerating corticospinal tract (CST) axons (Fig. 8A) [42]. As shown in Fig. 8B, in the SCI control group, BDA-labeled CST axons were blocked at the injury margin by a dense GFAP-positive glial scar and failed to penetrate the core lesion area. The A-GelMA group showed limited axonal penetration, while the A-AgNW-GM group demonstrated superior axonal regeneration capacity. Most notably, the A-NAC-AgNW-GM group exhibited abundant BDA-positive regenerating axons densely traversing the entire lesion area with excellent integration into host tissue. Quantitative analysis of axonal growth distance (Fig. 8C and D) statistically validated this finding. The average axonal extension distance in the A-NAC-AgNW-GM group was considerably greater than in all other groups, suggesting that this scaffold offers the most conducive milieu for axonal regeneration. To investigate whether these regenerated axons could form functional neural circuits, we performed high-resolution imaging and colocalization analysis of synaptic markers distal to the injury site. In the A-NAC-AgNW-GM group, we observed BDA-positive axon terminals (purple) forming tight contacts with NeuN-positive neuronal cell bodies (red). Crucially, these regenerated axon terminals exhibited clear colocalization with the postsynaptic marker PSD95 (green) (Fig. 8E). Finally, 3D confocal reconstruction (Fig. 8F) provided compelling evidence: regenerated BDA-positive axon terminals (purple) were enriched with the presynaptic vesicle protein Synapsin (red) and formed typical presynaptic-postsynaptic paired structures with PSD95 (green) signals [43]. This result indicates that CST axons regenerated from the cerebral cortex not only successfully extended beyond the injury site but also formed structurally intact, functionally potent synaptic connections with local target neurons. This strongly demonstrates that the A-NAC-AgNW-GM composite scaffold represents a highly promising strategy for promoting neural regeneration and functional circuit reconstruction following spinal cord injury.

Fig. 8.

A-NAC-AgNW-GM promotes endogenous neural regeneration, neural circuit formation, and neural circuit formation. (A) (Left) Schematic illustration showing the site of BDA injection. (Right) Immunofluorescence staining of coronal section of the brain—stained for NeuN (green), BDA (red), and DAPI (blue). Scale bar: 500 μm. Scale bar (enlarged image): 20 μm. (B) Representative images of GFAP (green), BDA (red), and DAPI (blue) immunofluorescence staining in each group; vertical white dashed lines indicate the proximal border (PB), the distal border (DB), and the central part of the lesion (LC). Scale bar: 1 mm. (C) Schematic lesion sites and vertical lines were used to count the number of axons crossing at each location indicated. (D) Plot of the number of crossing axons. ∗∗ represents P < 0.01. ∗∗∗ represents P < 0.001. ∗∗∗∗ represents P < 0.0001. Repeated measures of two-way ANOVA with Bonferroni (∗ denominated as comparison between the SCI group and the A-NAC-AgNW-GM group). Data are represented as mean ± SD (n = 3). (E) Representative images of PSD95 (green), Neun (red), BDA (purple), and DAPI (blue) immunofluorescence staining in the A-NAC-AgNW-GM group. Scale bar: 1 mm. Scale bar (enlarged image): 50 μm and 15 μm. (F) Representative images of PSD95 (green), syn (red), BDA (purple), and DAPI (blue) immunofluorescence staining in the A-NAC-AgNW-GM group. Scale bar: 50 μm. Scale bar (enlarged image): 15 μm.

3.7. A-NAC-AgNW-GM promotes neuroregeneration and neural circuit formation via the FGF13/Ca²⁺/CaMK2A/CREB pathway

To elucidate the molecular mechanisms by which A-NAC-AgNW-GM facilitates spinal cord repair, we conducted RNA sequencing (RNA-seq) analysis on tissues from the injured area to compare genome-wide expression profiles between the A-NAC-AgNW-GM treatment group and the SCI control group. Whole-genome differential expression analysis revealed that A-NAC-AgNW-GM treatment induced extensive gene expression reprogramming compared to the SCI group (Fig. 9A). Gene Ontology (GO) enrichment analysis (Fig. 9B and C) indicated that these differentially expressed genes were significantly enriched in biological processes closely related to neural repair, such as synapse organization, neurotransmitter secretion, and chemical synaptic transmission, as well as molecular functions like calcium ion binding and calmodulin binding. We particularly note that the expression level of fibroblast growth factor 13 (FGF13) exhibited extremely significant upregulation. FGF13 is a non-classical member of the FGF family. Recent studies indicate it is not a typical growth factor but is primarily expressed within the nervous system. By regulating the subcellular localization and activity of voltage-gated sodium channels, it is essential for axon growth, neuronal excitability, and repair [44].

Fig. 9.

Bioinformatics analysis and validation of the impact of A-NAC-AgNW-GM on neuroregeneration and neural circuit formation. (A) Volcano plot comparing gene expression in the A-NAC-AgNW-GM group to that in the SCl group, where the x-axis represents log2 (fold change) and the y-axis represents -log10 (P-value). (B) Partial GO analysis of the differentially expressed genes (DEGs). The x- and y-axes represent the number of genes and GO terms. (C) The circular plot displays the enrichment of differentially expressed genes in neuron-related functional categories and their expression changes (log2FC). (D) Bubble charts of the KEGG enrichment analysis for DEGs. The x and y axes represent the enrichment ratio and KEGG terms, respectively. (E) GSEA analyses were conducted on the calcium signaling pathway. (F) Heatmap showing DEGs within the calcium signaling pathway. (G) Western blot analysis of FGF13, CaMK2A, p-CaMK2A, CREB, and p-CREB protein expression in each group. (H-J) Statistical analysis of FGF13, p-CaMK2A/CaMK2A, and p-CREB/CREB protein intensity.

GO and KEGG pathway enrichment analyses (Fig. 9B–D) collectively pointed to a process closely related to neuronal excitability and plasticity—the calcium signaling pathway. This pathway was significantly enriched, and GSEA analysis (Fig. 9E) further confirmed its overall activation in the A-NAC-AgNW-GM group. Fig. 9F displays significantly differentially expressed genes within the calcium signaling pathway. This suggests a potential functional link between FGF13 upregulation and calcium signaling pathway activation. Previous studies provide a plausible mechanism: FGF13 directly influences neuronal action potential firing frequency and pattern by stabilizing microtubules and regulating voltage-gated sodium channels (Nav) [44,45]. Increased neuronal electrical activity results in the activation of voltage-gated calcium channels (VGCCs), leading to heightened intracellular calcium ion (Ca²⁺) levels and the subsequent activation of downstream calcium signaling pathways [46]. Furthermore, prior studies indicate that FGF13 can influence the function and localization of L-type calcium channels (CaV1.2) [47]. Based on these bioinformatics findings and literature support, we propose a scientific hypothesis: The A-NAC-AgNW-GM scaffold specifically upregulates FGF13, elevating intracellular calcium levels and thereby activating calcium signaling pathways centered on CaMK2A. To validate this hypothesis, protein-level analyses were performed using spinal cord tissues harvested at 6 weeks post-injury (Fig. 9G–J). Western blot analysis showed that FGF13 expression was markedly reduced after spinal cord injury compared with the sham group, indicating injury-induced suppression of this pathway. Among the treatment groups, the highest FGF13 protein level was observed in the A-NAC-AgNW-GM group, which may be attributed to the aligned scaffold architecture combined with NAC-modified AgNWs. Specifically, NAC modification enhances the electrical conductivity of the AgNW network while simultaneously alleviating oxidative stress, thereby creating a more favorable microenvironment for FGF13 upregulation. Consistent trends were observed for downstream signaling components: phosphorylation of CaMK2A (p-CaMK2A) and CREB (p-CREB) was substantially decreased in the SCI group compared with sham controls, whereas A-NAC-AgNW-GM treatment prominently enhanced the phosphorylation levels of both proteins. These results indicate that the A-NAC-AgNW-GM scaffold effectively reactivates the FGF13/CaMK2A/CREB signaling cascade that is impaired following spinal cord injury. Notably, phosphorylation of CREB is a recognized indicator of its transcriptional activation, which is closely associated with gene programs governing neuronal survival, synaptic plasticity, and axonal regeneratio [48,49].

Immunofluorescence combined with BDA anterograde tracing revealed that FGF13, CaMK2A, and Cav1.2 were enriched along BDA-labeled axons crossing the lesion region. Confocal imaging showed clear spatial co-localization of FGF13 and CaMK2A with regenerating axonal trajectories, which was further supported by high-magnification views. Line-scan analyses demonstrated overlapping fluorescence intensity profiles of FGF13, CaMK2A, and Cav1.2 with BDA signals along the same axons, indicating coordinated enrichment of Ca²⁺-related signaling components within regenerating axons(Fig. S 8A-D).

To determine whether FGF13 directly influences scaffold-induced axon growth, primary cortical neurons were transduced with a lentiviral vector encoding FGF13 and cultured on different scaffolds. Western blot analysis confirmed effective upregulation of FGF13 following lentiviral transduction (Fig. S 9A-B). Immunofluorescence staining for Tuj-1 and GAP43 showed that FGF13 overexpression promoted neurite elongation (Fig. S 9C). Sholl analysis revealed a higher number of neurite intersections at extended distances from the soma in FGF13-overexpressing neurons, indicating enhanced neurite complexity and extension capacity (Fig. S 9D). Quantitative analysis further demonstrated elevated GAP43 expression in the FGF13 overexpression groups, a marker associated with axonal growth (Fig. S 9E). While aligned GelMA scaffolds supported baseline neurite growth, the effects of FGF13 overexpression were more pronounced on A-AgNW-GM and A-NAC-AgNW-GM, suggesting that FGF13 enhances the axon growth–promoting properties of the scaffold.

4. Discussion

Repairing spinal cord injury (SCI) remains an extraordinarily complex challenge, as it requires coordinated regulation of structural guidance, electrical signaling, and the hostile post-injury microenvironment characterized by inflammation and oxidative stress. Accordingly, conductive scaffolds have been widely explored for SCI repair, including systems based on carbon nanotubes (CNTs), graphene derivatives, and conducting polymers such as poly(3,4-ethylenedioxythiophene) (PEDOT), owing to their good electrical conductivity and ability to promote neural regeneration [[50], [51], [52]]. CNT-, graphene-, and PEDOT-based conductive scaffolds provide effective electrical cues for neural regeneration, but limitations in dispersibility, mechanical compliance, and biological functionality remain. These limitations underscore the necessity for conductive scaffolds that combine electrical functionality with microenvironmental modulation, which motivated the design of the NAC-AgNW–based, aligned scaffold in this work. Compared with conventional conductive materials, the NAC-modified silver nanowire (NAC-AgNW) system provides high conductivity at low loading while preserving hydrogel compliance, and its one-dimensional geometry is well suited for forming continuous conductive networks within aligned scaffolds. Importantly, NAC modification improves nanowire dispersion and enhances the electrical network structure, while simultaneously providing local antioxidant regulation, which is particularly relevant in the oxidative microenvironment following SCI. The scaffold exhibits physicochemical properties that closely match the mechanical and electrical characteristics of the spinal cord injury microenvironment, supporting early inflammatory modulation and subsequent neural circuit reconstruction. A hallmark feature of spinal cord white matter is the highly oriented arrangement of nerve fiber bundles. SEM images of the GelMA scaffold (GM), prepared via the ice-molding method, reveal its highly ordered, anisotropic, layered, porous structure. This biomimetic-oriented topology is not randomly formed but deliberately designed to provide explicit contact guidance for cell migration and axonal growth. Cell experiments intuitively demonstrate this: cells growing on A-NAC-AgNW-GM exhibit highly ordered alignment and extension of their cell bodies and synapses along the scaffold's groove direction, starkly contrasting with the random, disordered growth patterns observed on conventional hydrogels. This physical guidance is crucial for directing regenerating axons through injury sites to achieve long-distance precise regeneration, directly corresponding to in vivo experiments where neurons grow along the scaffold in a directed manner and extensively traverse injury zones [53]. The mechanical properties of biomaterials (e.g., elastic modulus, conductivity) must match those of host tissues to avoid compression from overly rigid materials or insufficient support from excessively soft ones [11,54]. Mechanical property testing indicates that while the introduction of NAC-AgNWs enhances the composite material's mechanical properties, its elastic modulus remains within the range suitable for natural spinal cord tissue. This mechanical matchingensures the scaffold effectively fills the lesion cavity, providing stable three-dimensional physical support for neural regeneration while minimizing foreign body reactions and secondary damage caused by mechanical mismatch. It thus offers a soft yet sturdy platform for nerve regeneration. Electrical matching of the electrophysiological microenvironment is critical, as the spinal cord functions as a continuous pathway for electrical signal transmission. [55]. Following injury, disrupted myelin sheaths and cellular necrosis lead to disorder in the local electrophysiological microenvironment. A core objective of introducing NAC-AgNW in this study is to reconstruct this microenvironment. Conductivity tests confirm that A-NAC-AgNW-GM exhibits significantly higher conductivity than the insulating pure gel group, with NAC-modified AgNWs showing superior conductivity to unmodified AgNW-attributed to its enhanced internal conductive network. This electrically matched property mimics the conductivity of a natural neural matrix, potentially providing a beneficial electrical microenvironment for neurons. It facilitates intercellular electrical signaling communication, guides the directed differentiation of neural stem cells, and enhances synaptic formation and function [27,56]. This is crucial for the functional reconstruction of neural circuits rather than merely structural repair.

The core challenge in repairing spinal cord injury (SCI) lies in its adverse microenvironment, where persistent pro-inflammatory (M1) polarization of macrophages drives inflammatory responses that constitute a key barrier to regeneration. Therefore, reversing the immune microenvironment from the pro-inflammatory M1 state to the anti-inflammatory, reparative M2 state is a prerequisite for successful neural regeneration. The multifunctional scaffold A-NAC-AgNW-GM developed in this study demonstrates one of its most critical biological effects: through the synergistic action of NAC and AgNWs, it efficiently modulates immune cell polarization, thereby clearing obstacles for subsequent neural regeneration. N-acetylcysteine (NAC), a potent antioxidant, induces M2 polarization primarily by fundamentally alleviating oxidative stress at injury sites [57,58]. Extensive research indicates that reactive oxygen species (ROS) surges following spinal cord injury are central drivers of persistent inflammation and M1 polarization. Previous studies have confirmed that ROS scavenging can break this vicious cycle, thereby suppressing pro-inflammatory factor release and creating a microenvironment conducive to M2 polarization [59,60]. To investigate the direct cellular effects of the material under controlled conditions, we first conducted in vitro experiments. The findings of this study align strongly with this mechanism. As demonstrated by PCR, fluorescence, and Western blot analyses, the A-NAC-AgNW-GM group exhibited significantly reduced expression of the pro-inflammatory mediator iNOS (an M1 marker) while simultaneously elevating levels of the M2 marker Arg-1. This strongly indicates that the sustained release of NAC from the scaffold effectively neutralizes ROS at the injury site, thereby shifting the immune balance toward the reparative M2 phenotype. This aligns with literature reports that NAC can alleviate neuroinflammation and promote functional recovery. More notably, this study reveals that silver nanowires (AgNWs) may exhibit positive immunomodulatory functions at specific concentrations and morphologies. Studies indicate that nanosilver can modulate the activation state of microglia, suppressing their pro-inflammatory responses while simultaneously promoting their transition toward a reparative phenotype [22,61]. Our findings provide compelling corroboration: Immunofluorescence staining revealed that under inflammatory stimulation, the A-AgNW-GM and A-NAC-AgNW-GM groups exhibited a characteristic pattern of decreased M1 phenotype (iNOS⁺) cells and a significant increase in M2 phenotype (Arg-1⁺) cells. This indicates that AgNWs incorporation does not merely suppress immune responses but actively and selectively directs immune cells toward the beneficial M2 direction. We hypothesize that NAC and AgNWs may synergistically act through distinct yet complementary mechanisms to jointly influence the M2 polarization process, thereby generating a potent synergistic effect. The distinct phenotypic regulation trends observed in vitro were further validated in vivo through more complex and robust animal model studies. In comparison to SCI controls and other material control groups, the A-NAC-AgNW-GM treatment group demonstrated the lowest total CD68 count in the injured area, a marked decrease in the percentage of cells expressing iNOS (an M1 marker), and a considerable increase in the percentage of cells expressing Arg-1 (an M2 marker). Western blot examination demonstrated a considerable reduction in iNOS protein expression in the damaged spinal cord tissue of the A-NAC-AgNW-GM group, but Arg-1 protein expression was markedly elevated. This shift in cellular phenotype from pro-inflammatory to anti-inflammatory repair directly replicates in vitro findings, conclusively demonstrating that this composite material can systematically transform the immune microenvironment in the injury site from an M1-dominated destructive state to an M2-dominated reparative state. Both in vitro and in vivo experiments demonstrated that the group containing both NAC and AgNWs exhibited the optimal effect, significantly outperforming single-component groups. This synergistic effect is key to the composite material's exceptional anti-inflammatory efficacy.

Early A-NAC-AgNW-GM regulation of a beneficial immune microenvironment directly correlates with the preservation of key neural pathways. Analysis of MAP-2 staining, a marker for neuronal dendrites and cell bodies, revealed that the intensity and extent of MAP-2 positive signals in the perilesional areas of the A-NAC-AgNW-GM treatment group were significantly superior to those in the control group. This indicates that more neuronal cell bodies and their intricate dendritic structures survived, preserving the fundamental architecture of neural networks and providing the structural foundation for signal reception and integration. Markers of descending excitatory pathways originating from the brainstem, such as serotonergic (5-HT) and noradrenergic (TH) fibers, showed significantly better preservation distal to the injury. These fibers are critical for activating spinal central pattern generators (CPGs) and maintaining motor neuron excitability. Their preservation indicates that the communication cables linking the brain to the spinal cord remain largely intact, preserving a valuable structural foundation for functional recovery [41]. More critically, these axons did not grow haphazardly but highly conformed to the oriented topology, exhibiting distinct directional extension and successfully traversing the typically impassable central lesion zone. Building upon the preservation of existing structures, promoting regenerative axon growth represents the second step in circuit reconstruction. One of the most critical pieces of evidence in this study stems from the anterograde tracing technique using BDA (biotinylated dextran amido). BDA is taken up by neuronal cell bodies and transported axonally in an anterograde direction, enabling specific, high-resolution labeling of entire axonal projection pathways [42]. Experimental results revealed that only in the A-NAC-AgNW-GM treatment group could numerous BDA-labeled fibers be clearly observed traversing the lesion center and extending into distal spinal cord tissue. This finding provides irrefutable direct evidence that the scaffold effectively supports long-distance regeneration of host corticospinal tract (CST) neuronal axons across the lesion zone. Functional regeneration of these axons depends on effective myelination. Myelin basic protein (MBP) staining revealed significantly enhanced MBP-positive signals within the lesion area of the A-NAC-AgNW-GM group. Myelin formation not only protects axons but also substantially enhances the speed and fidelity of electrical signal transmission, representing a critical step in transitioning regenerated axons from structural presence to functional utility. The final step in axonal regeneration is establishing synaptic connections with distal target neurons. Detailed high-magnification confocal imaging analysis of the distal spinal gray matter revealed strong colocalization between BDA-positive fibers and presynaptic vesicle protein markers (Syn) and postsynaptic density markers (PSD-95), providing strong morphological evidence that regenerated axons have established structural synaptic connections with host neurons [43,62]. This strongly suggests the significant potential for functional synapse formation and neural circuit reconstruction. Motor function assessment (BBB score and joint range-of-motion analysis) demonstrated that rats in the A-NAC-AgNW-GM treatment group exhibited significantly superior hindlimb motor recovery compared to all other groups. This recovery necessarily relies on regenerated, myelinated axons effectively transmitting brain commands through newly formed synapses to target motor neurons.

This persuasive phenotypic evidence demonstrates that A-NAC-AgNW-GM successfully intervenes in the degenerative progression of spinal cord damage. To surpass traditional phenotypic descriptions and enhance our comprehension of the fundamental global molecular networks and core mechanisms, we utilized transcriptome sequencing (RNA-seq) technology for an unbiased, genome-wide analysis of gene expression profiles in injured spinal cord tissue. This research strategy enabled us to systematically decipher the intrinsic molecular programs by which A-NAC-AgNW-GM promotes repair, starting from the source of transcriptional regulation. RNA-seq data demonstrated substantial and significant alterations in the gene expression profile of spinal cord tissue in the A-NAC-AgNW-GM therapy group relative to the control group. Gene Ontology (GO) enrichment analysis jointly demonstrated that the therapy significantly increased gene groups related to axonal regeneration, synapse assembly, neural projection guidance, and myelination. This finding strongly correlates with our morphologically observed axonal regeneration and functional recovery outcomes, providing transcriptional-level validation of the treatment's efficacy. Fibroblast growth factor 13 (FGF13) was identified as one of the most markedly increased differentially expressed genes. Establishing it as the core molecular target for mechanistic exploration in this study was a logical choice due to its unique functional properties aligning closely with material characteristics. FGF13 is a non-classical member of the fibroblast growth factor family (not secreted extracellularly). Its core function in the nervous system is regulating microtubule stability and voltage-gated sodium channels (Nav) [44,45,63]. It directly guides axonal directional extension by stabilizing the cytoskeleton within the axon growth cone. Simultaneously, it maintains neuronal excitability and the fidelity of electrical signal transmission by regulating the aggregation and function of sodium channels in the axon initial segment. This functional characteristic exhibit remarkable complementarity and synergy with the biomimetic design of our material. This scaffold employs cryo-oriented fabrication to create a highly ordered, directional topological structure, providing physical contact guidance for axonal growth. Meanwhile, FGF13 upregulation delivers chemical propulsion from within the cell: by stabilizing microtubules, it enhances the growth cone's responsiveness to external physical guidance cues, enabling neurons to extend more efficiently along the scaffold's grooves. In other words, the material's physical guidance provides the road for axonal regeneration, while FGF13 upregulation enhances the neuron's navigational ability. This synergy significantly improves the efficiency of directed regeneration. Concurrently, one core objective of introducing silver nanowires (AgNWs) in this study was to impart electrical conductivity to the scaffold, thereby reconstructing the disordered electrophysiological microenvironment in the injured region. FGF13 directly enhances the neurons' intrinsic electrical excitability and signal conduction capacity by regulating sodium channel function. This implies that the material provides an externally supportive matrix for electrical signal transmission, while FGF13 internally enhances the neurons' intrinsic ability to generate and conduct electrical signals. This synergistic internal-external enhancement is crucial for reconstructing functional neural circuits. FGF13 uniquely integrates two major functions: cytoskeletal dynamics and electrophysiological regulation. This precisely corresponds to the two core biomimetic properties of this material: directional guidance and electrical conduction. Our transcriptomic data links this to biomaterial therapy, revealing a novel material-mediated mechanism that promotes precise axonal regeneration by upregulating FGF13. The result contributes new knowledge to the field rather than merely validating known pathways. To further investigate FGF13's upstream regulation and downstream effects, we performed KEGG pathway enrichment analysis and GSEA on differentially expressed genes. Results consistently indicate that the calcium signaling pathway is one of the most significantly enriched. This enrichment strongly suggests that Ca²⁺ signaling mediates multiple beneficial effects following A-NAC-AgNW-GM treatment. FGF13 directly influences neuronal electrical excitability and action potential generation by regulating voltage-gated sodium channels (Nav). Action potentials serve as the primary trigger for opening voltage-gated calcium channels (VGCC), thereby inducing Ca²⁺ influx [44,45,63]. Crucially, studies have demonstrated that FGF13 directly binds to the intracellular domain of L-type voltage-gated calcium channels (L-VGCC, primarily the CaV1.2 subtype) [47]. By stabilizing the pore-forming α1 subunit and promoting its membrane localization, FGF13 significantly enhances channel function and increases Ca²⁺ influx. Elevated intracellular Ca²⁺ concentration is key to activating downstream signaling. Ca²⁺ binds to calmodulin, specifically activating calmodulin-dependent kinase IIα (CaMK2A). CaMK2A is widely distributed throughout the cytoplasm, but its core functional sites are located at the synapse. CaMK2A is a pivotal regulatory molecule in the process of neurorepair. It promotes neuronal survival, axonal regeneration, and the functional reconstruction of neural networks through multiple mechanisms: regulating cytoskeletal dynamics, enhancing the expression of neurotrophic factors, maintaining synaptic plasticity, and integrating calcium signaling with gene transcription [[64], [65], [66]]. Phosphorylated CaMK2A (p-CaMK2A) subsequently phosphorylates and activates cAMP response element-binding protein (CREB) [48]. p-CREB, a vital transcription factor, translocates to the nucleus to commence transcription of many genes essential for neuronal survival and axonal regeneration [29,67].

The present results provide convergent molecular and spatial evidence that the A-NAC-AgNW-GM scaffold engages a neuron-intrinsic Ca²⁺-dependent growth program centered on the FGF13–CaMK2A–CREB axis. Western blot analysis demonstrated that scaffold implantation significantly elevated FGF13 expression and enhanced CaMK2A and CREB phosphorylation in vivo, indicating activation of downstream transcriptional signaling associated with axonal growth. Importantly, immunofluorescence combined with BDA anterograde tracing revealed that FGF13, CaMK2A, and the L-type calcium channel Cav1.2 were co-enriched along regenerating axons traversing the lesion site, rather than being diffusely distributed in surrounding tissue. This spatial coupling supports a direct role of calcium signaling within regenerating axons. Complementary gain-of-function experiments in primary cortical neurons further established the functional relevance of FGF13, as lentiviral overexpression markedly enhanced neurite extension and GAP43 expression, particularly in A-NAC-AgNW-GM scaffold. Together, these findings suggest that electrical matching provided by the anisotropic NAC-AgNWs network promotes local calcium signaling, while FGF13 acts as a molecular amplifier linking scaffold-derived biophysical cues to intrinsic axon growth machinery. This integrated mechanism provides molecular insight into how anisotropic conductive hydrogels interface with neuronal signaling pathways to facilitate spinal cord regeneration.

Recent studies in bioelectrically active biomaterials demonstrate that anisotropic conductive scaffolds integrating structural guidance with directional electrical signal propagation can promote axonal alignment and functional recovery, highlighting spatial organization of conductive elements as a critical design parameter [[68], [69], [70], [71]]. Beyond passive conductivity, emerging bioelectrical and piezoelectric strategies further indicate that electrical cues can actively modulate the injury microenvironment and engage intracellular signaling relevant to neural repair [67]. Recent studies indicate that regulation of intracellular Ca²⁺ dynamics can activate CaMK2A/CREB-dependent transcriptional programs, providing a mechanistic framework for how electrical stimulation translates into gene expression changes associated with neural growth and plasticity [72,73]. In this context, our study identifies FGF13 as a molecular mediator linking scaffold-derived electrical cues to neuron-intrinsic axon growth programs. By combining anisotropic architecture with a percolated conductive network, the scaffold not only provides directional guidance but also engages FGF13 associated CaMK2A/CREB signaling, offering mechanistic insight into electrically guided spinal cord regeneration.

Future studies will be required to evaluate the long-term durability of functional recovery and scaffold integration beyond the current 6-week observation window. Expanding validation to clinically relevant contusion or compression models may further strengthen the translational relevance of this platform. In addition, direct in vivo assessment of electrical signal propagation within the injured spinal cord could provide mechanistic insight into how electrical matching contributes to neural circuit reconstruction. Finally, although FGF13 is implicated as a mediator of scaffold-induced regeneration, its in vivo necessity warrants further investigation.

5. Conclusion

This study successfully designed and constructed a multifunctional bionic neural scaffold, A-NAC-AgNW-GM, integrating directional topology, immunomodulation, and electrical conductivity. This material effectively induces glial cells to polarize toward the anti-inflammatory M2 phenotype, significantly reducing inflammation levels in the injured area and creating a crucial favorable microenvironment for neural regeneration. Simultaneously, the material's directed structure provides physical guidance for axonal growth. Combined with its active immunomodulatory and electrically conductive properties, it significantly promotes long-distance regeneration of host axons across the injury site. These axons ultimately form synaptic connections with distal neurons, reconstructing functional neural circuits and driving remarkable recovery of motor function in experimental animals. Transcending phenotypic observation, this study employs transcriptome sequencing and protein validation to propose and validate for the first time that FGF13/Ca²⁺/CaMK2A/CREB constitutes a core signaling axis mediating the therapeutic efficacy of this material. By upregulating FGF13, the material modulates Ca²⁺ influx, thereby activating the CaMK2A/CREB signaling pathway—closely linked to synaptic plasticity and cell survival—achieving precise molecular-level regulation of intrinsic neuronal regeneration programs.

In summary, A-NAC-AgNW-GM, through its unique biomimetic design and multifunctional synergy, successfully achieves active regulation of the early inflammatory microenvironment and effective reconstruction of neural circuits in the later stages of spinal cord injury. This offers a highly promising, comprehensive therapy approach for spinal cord damage restoration.

Ethics approval and consent to participate

All protocols and animal experiments were conducted in strict accordance with the Animal Care and Use Committee of Wenzhou Medical University (No. WYDW-2025-0412).

CRediT authorship contribution statement

Minghao Jiang: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft. Wenjie Lu: Conceptualization, Formal analysis, Investigation, Methodology, Validation, Writing – original draft. Junyu Zhuang: Formal analysis, Investigation, Methodology, Validation. Jiahui Song: Formal analysis, Investigation, Methodology, Validation. Yanfang Zhao: Formal analysis, Investigation, Validation. Cheng Zhou: Formal analysis, Investigation, Validation. Yangbo Zhou: Investigation, Validation. Weizhi Shu: Investigation, Validation. Zhongwei Zhu: Investigation, Validation. Lelin Jiang: Investigation, Validation. Ping Wu: Writing – review & editing. Aimin Wu: Writing – review & editing. Sunren Sheng: Project administration, Supervision, Writing – review & editing. Sipin Zhu: Conceptualization, Funding acquisition. Zhouguang Wang: Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the supplementary data to this article:

Data availability

Data will be made available on request.