Curcumin-loaded dynamic crosslinked injectable hydrogel for annulus fibrosus repair and disc degeneration prevention

Abstract

Objective

Annulus fibrosus (AF) defects following discectomy often induce biomechanical instability and inflammatory microenvironment deterioration, key drivers of recurrent herniation and accelerated degeneration. Given the limited self-repair capacity of the AF, this study aimed to develop a novel composite hydrogel integrating immediate mechanical support, tissue adhesion, and sustained anti-inflammatory release to promote functional AF repair.

Methods

A composite hydrogel was prepared by incorporating curcumin-loaded liposomes (LIP@CUR) into a Schiff-base–crosslinked network formed between 4-arm polyethylene glycol (4aPEG-OPA) and gelatin. The hydrogel was comprehensively characterized in terms of microstructure, rheological behavior, mechanical properties, degradation profile, and in vitro drug release. Biocompatibility was evaluated using cytotoxicity and hemolysis assays. For in vivo assessment, male Japanese White rabbits (2.5 ± 0.2 kg) were obtained from Wangdu Tonghui Breeding Co., Ltd. (Hebei, China; License No. SCXK (Ji) 2021-006), and an AF defect model was established. Therapeutic efficacy was evaluated by maintaining disc height and by quantifying inflammatory responses (IL-6, TNF-α) and assessing extracellular matrix components (COL I/II) using radiography and histological analyses.

Results

LIP@CUR exhibited uniform particle size (159.68 ± 1.98 nm) with ~ 70% encapsulation efficiency. The composite hydrogel demonstrated rapid in situ gelation (~ 60 s) and a compressive modulus (114.06 ± 13.94 kPa) matching native AF tissue, alongside robust tissue adhesion strength (~ 50 kPa). In vitro assessments confirmed low swelling, controllable degradation, and sustained curcumin release without cytotoxicity. In vivo, the hydrogel group effectively maintained intervertebral disc height and improved MRI signal intensity. Histological and semi-quantitative analyses confirmed a significant downregulation of inflammatory cytokines (TNF-α and IL-6, p < 0.001), alongside enhanced collagen deposition, facilitating structural regeneration of the AF defect.

Conclusion

This study successfully developed a functionalized composite hydrogel featuring dual-controlled release, mechanical compatibility, and anti-inflammatory activity. By synergistically providing physical closure and modulating the immune microenvironment, this material offers a promising clinical strategy for preventing postoperative disc recurrence and enhancing AF repair.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13018-026-06909-x.

Introduction

Lumbar disc herniation (LDH) and associated intervertebral disc degeneration (IVDD) are prevalent spinal disorders globally, representing a primary etiology of low back pain and radiculopathy [1–2]. Currently, this condition constitutes a global health crisis rather than a region-specific issue. Globally, hundreds of millions of individuals suffer from LDH and IVDD, making them leading causes of years lived with disability (YLDs). Beyond the severe physical suffering of patients, these disorders impose a staggering socioeconomic burden. The chronic nature of the disease significantly diminishes global workforce productivity through increased absenteeism and early retirement. Consequently, the massive indirect costs associated with lost labor and prolonged medical care severely negatively impact the gross domestic product (GDP) and healthcare systems of nations worldwide [3].Currently, discectomy remains the gold standard for relieving nerve compression symptoms [4]. However, this procedure typically involves the removal of herniated tissue without repairing the annulus fibrosus (AF) defect, leading to a postoperative re-herniation rate ranging from 5 to 26% [5–7]. The AF is the core structure maintaining the biomechanical integrity of the intervertebral disc (IVD); its deficiency can trigger local stress imbalances, accelerate disc degeneration, and induce secondary herniation [8–10]. Unfortunately, the intrinsic self-repair capacity of the AF is extremely weak due to its avascular nature and the difficulty of regenerating its unique lamellar collagen structure. Furthermore, the pathological microenvironment, characterized by inflammation and the excessive accumulation of reactive oxygen species (ROS), further inhibits the healing process [11]. Therefore, developing a novel repair strategy that combines mechanical support with microenvironmental modulation to promote AF regeneration is an urgent clinical challenge.

The AF possesses a complex hierarchical architecture, consisting of 15–25 concentric lamellae rich in type I/II collagen and proteoglycans [12]. These collagen fibers are arranged at alternating angles to form a three-dimensional network that provides nonlinear viscoelasticity and structural stability [13]. Functionally, the AF is divided into an outer zone, responsible for resisting tensile loads, and an inner zone that buffers pressure in synergy with the nucleus pulposus [14]. AF rupture is the central pathological event in disc degeneration. It is driven by mechanical stress—such as chronic cumulative damage or acute trauma [15]—and exacerbated by physiological degeneration, including the downregulation of collagen synthesis and the upregulation of matrix metalloproteinases (MMPs) [16]. Crucially, the biochemical microenvironment plays a pivotal role in preventing repair. Pro-inflammatory cytokines (e.g., IL-1β, TNF-α) activate the NF-κB pathway to suppress matrix synthesis and induce apoptosis [17], while oxidative stress damages mitochondrial function and accelerates cellular senescence [18].

Current clinical interventions for AF defects, such as suture techniques and annular closure devices, have shown limited efficacy in restoring the physiological function of the disc. While suturing can achieve anatomical closure and potentially reduce recurrence rates [19], it fails to restore the tensile strength of the AF and may cause secondary mechanical injury [20]. Implantable devices (e.g., Barricaid) have been developed to physically block re-herniation [21], but concerns regarding device dislodgement, vertebral endplate damage, and high costs have hindered their widespread adoption [22]. Similarly, radiofrequency thermal annuloplasty provides symptomatic relief but does not address large structural defects or promote tissue regeneration [23]. Consequently, tissue engineering strategies utilizing biomaterials have emerged as a promising alternative.

Hydrogels are particularly attractive for AF repair due to their high water content, tunable mechanics, and biocompatibility [24]. Various hydrogel systems, including collagen [25], PLGA [26], and decellularized matrix [27], have been investigated for their ability to promote tissue regeneration. However, from a biomechanical perspective, an ideal hydrogel intended to address this issue must adopt a rigorous mechanical approach. The native spine is a highly dynamic environment subject to continuous mechanical loading. Thus, a successful AF repair material must not only be biocompatible but also possess a compressive modulus specifically matching that of the native AF tissue to withstand complex axial and torsional forces. Furthermore, it must exhibit robust interfacial wet-tissue adhesion strength to prevent material extrusion and migration under high intradiscal pressure—properties that traditional hydrogels rarely combine. Recently, a hydrogel system crosslinked by phthalaldehyde-terminated four-arm polyethylene glycol (4aPEG-OPA) and gelatin was reported to exhibit superior adhesive and mechanical properties [28], offering a robust candidate material for annular closure.

Beyond physical repair, effective treatment requires modulating the hostile inflammatory and oxidative microenvironment. Curcumin (Cur), a bioactive compound derived from turmeric [29], exhibits potent anti-inflammatory and antioxidant activities [30]. Studies have shown that Curcumin can inhibit inflammatory cytokines and matrix-degrading enzymes in IVD cells [31], promote wound healing via ROS scavenging [32], and attenuate tissue remodeling in other organs [33]. However, the clinical application of Curcumin is severely limited by its hydrophobicity and low bioavailability. Liposomes, as amphiphilic nanocarriers, offer an effective solution by encapsulating Curcumin to enhance its stability, cellular uptake, and sustained release [34].

Based on these considerations, this study aims to develop a multifunctional therapeutic strategy that integrates mechanical sealing with biological microenvironment modulation. We engineered a high-strength, low-swelling composite hydrogel based on the 4aPEG-OPA/Gelatin system, loaded with Curcumin-encapsulated liposomes (LIP@CUR). This system is designed to provide immediate physical closure of the AF defect while sustaining the release of Curcumin to scavenge ROS and suppress inflammation. We hypothesize that this “mechanical-biological” synergistic approach will effectively prevent postoperative recurrence and promote functional regeneration of the annulus fibrosus.

2. Materials and methods

2.1 Preparation of curcumin-loaded liposomes (LIP@CUR)

The liposomes were prepared using the film hydration-ultrasonication method. Soy lecithin (250 mg) and cholesterol (20 mg) were dissolved in 25 mL of anhydrous ethanol, and 4 mg of curcumin was added. The mixture was stirred at 40 °C under dark conditions until dissolved. The organic solvent was removed using a rotary evaporator at 40 °C and − 0.1 MPa, forming a uniform lipid film on the bottle wall. Phosphate-buffered saline (PBS) was added, and the solution was hydrated at 40 °C for 30 min. The resulting suspension was treated with ice-cold sonication (power 250 W, 3 s on, 3 s off) to reduce the particle size. After standing at 4 °C for 2 h, the suspension was centrifuged at 3000 rpm for 10 min, and the supernatant was collected as LIP@CUR, which was stored in the dark at 4 °C.

2.2 Preparation of 4aPEG-OPA

4aPEG-OPA (4-arm polyethylene glycol with o-phthalaldehyde (OPA) terminal groups) was synthesized according to a previously reported method. Briefly, 1,3-dimethoxy-1,3-dihydroisobenzofuran-5-carboxylic acid was dissolved in dichloromethane and activated with 4-dimethylaminopyridine (DMAP) for 20 min. Afterward, end-hydroxylated 4-arm polyethylene glycol was added, and the reaction was allowed to proceed for 24 h. The product was deprotected using a trifluoroacetic acid (1:1 v/v) solution and purified by dialysis against water using a dialysis membrane (MWCO 3500 Da), followed by freeze-drying to yield 4aPEG-OPA as a white solid. The chemical structure was confirmed by proton nuclear magnetic resonance (¹H-NMR).

2.3 Preparation of 4aPEG-OPA/Gelatin/LIP@CUR Hydrogel

Gelatin (120 mg/mL) was prepared by dissolving freeze-dried gelatin in PBS (37 °C) to form a 12% (w/v) solution, after dialyzing in deionized water for 2–3 days. The freeze-dried 4aPEG-OPA powder was dissolved in the LIP@CUR suspension to obtain a 12% (w/v) 4aPEG-OPA/LIP@CUR precursor solution. Equal volumes of the two solutions were quickly mixed using a dual-syringe or pipette, and the Schiff-base reaction between the amine groups of gelatin and the aldehyde groups of OPA crosslinked the mixture into a gel, forming the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel (Fig. 1B). For the comparison groups, the free curcumin-loaded hydrogel (4aPEG-OPA/Gelatin/CUR) was prepared by replacing the LIP@CUR suspension with free curcumin dissolved in PBS containing 2% Tween-80. Additionally, the blank control hydrogel (4aPEG-OPA/Gelatin) was prepared by simply replacing the LIP@CUR suspension with pure PBS.

Fig. 1.

(A) Molecular structure of 4aPEG-OPA and the corresponding ¹H-NMR spectrum of the synthesized product. (B) Schematic illustration of the chemical mechanisms: the synthesis of the 4aPEG-OPA crosslinker (via esterification) and the subsequent dynamic Schiff-base crosslinking reaction between the dual aldehyde groups of 4aPEG-OPA and the primary amine groups of gelatin. The schematic diagram in B was drawn using the Figdraw platform (Authorization ID: AlOIUbe8a8). (C) Particle size distribution of LIP@CUR nanoparticles. (D, E) Cryogenic scanning electron microscopy (cryo-SEM) images showing the microstructure of the hydrogels.(F) Transmission electron microscopy (TEM) image of LIP@CUR nanoparticles

2.4 Characterization of materials

Liposome characterization

Dynamic light scattering (DLS) was used to measure the hydrodynamic particle size, polydispersity index (PDI), and Zeta potential of the liposomes. Prior to measurement, the LIP@CUR suspension was diluted in PBS (pH 7.4) for particle size and PDI analysis to simulate physiological conditions. For Zeta potential measurements, to avoid the severe charge shielding effect caused by the high ionic strength of PBS, the samples were highly diluted in deionized water immediately prior to the test. This rapid measurement post-dilution effectively minimizes osmotic stress-induced liposome rupture while allowing accurate assessment of the intrinsic surface charge.Transmission electron microscopy (TEM) was used to observe the microstructure. The encapsulation efficiency of curcumin was determined using UV-Vis spectroscopy according to the method outlined in Supporting Information 1.1.

Rheology and gelation time

The macroscopic gelation time was determined using the test tube inversion method. A rotational rheometer (37°C, 1% strain, 1 Hz frequency) was used to monitor the storage modulus (G’) and loss modulus (G”) as a function of time. The gelation time was determined from the intersection point of the G’ and G” curves, with the time corresponding to the peak platform of the G’ curve indicating the time required for the sample to reach its maximum strength.

Microstructure and mechanical properties

The hydrogel was fixed on a conductive adhesive-coated sample stage and rapidly immersed in ultra-low-temperature liquid nitrogen for 30 s, followed by sublimation at 90 °C for 10 min and rewarming to − 140 °C. The sample was then gold-coated for observation using a cryo-scanning electron microscope (SEM) (operating at 5 kV). For the compressive strength test, custom-made PTFE circular molds with an inner diameter of 8 mm, thickness of 5 mm, and side openings of 2 mm were used. A 250 µL hydrogel sample was injected into the molds and placed in a 37 °C water bath for 2 h. The stress (F) was recorded using a universal testing machine (22–25 °C, trigger force 0.004 N, testing speed 2 mm/min, 50 N sensor). The compressive stress, strain, and compressive modulus were calculated according to the standard equations detailed in Supporting Information 1.2. The compression modulus was determined from the slope at the 10–20% strain region on the stress-strain curve.The adhesion strength was calculated using the universal testing machine under conditions specified in Supporting Information 1.2.

In vitro release, swelling, and degradation

Curcumin solutions with concentrations of 1.25, 2.5, 5, 10, and 20 µg/mL were prepared using PBS containing 2% Tween-80 as the solvent, and an absorbance standard curve was constructed. The release rate of curcumin from 4aPEG-OPA/Gelatin/CUR and 4aPEG-OPA/Gelatin/LIP@CUR was measured according to the method in Supporting Information 1.3. The equilibrium swelling ratio and in vitro degradation at 37 °C (PBS, pH 7.4) were determined following the detailed protocols in Supporting Information 1.4 and 1.5, respectively.

Antioxidant activity

The antioxidant activity of the hydrogel extract was evaluated using the DPPH radical scavenging assay as outlined in Supporting Information 1.6.

2.5 Biocompatibility evaluation

Rabbit annulus fibrosus cell isolation and culture

This study was conducted in accordance with the “Guidelines for Ethical Review of Animal Welfare in China” (GB/T 35892 − 2018) and the “Regulations for the Administration of Laboratory Animals in China.” The protocol was approved by the Animal Ethics Committee of the School of Basic Medical Sciences, Jilin University (Approval No. SYXK 2017-003). All animal procedures were performed following relevant institutional and national regulations. Male Japanese white rabbits were subjected to inhalation anesthesia using isoflurane. Upon confirmation of deep anesthesia and the loss of the pedal withdrawal reflex, the animals were humanely euthanized via air embolism while fully unconscious. This procedure was strictly monitored to ensure the amelioration of animal suffering. Following euthanasia, rabbit thoracolumbar intervertebral discs were collected, and the annulus fibrosus tissue was separated and cut into 1 mm³ pieces. These pieces were incubated with 3 mL of 0.25% trypsin at 37 °C in a shaking incubator for 15 min. After centrifugation at 1500 rpm for 5 min, the precipitate was collected. Then, 5 mL of 0.1% type II collagenase was added, and the mixture was incubated at 37 °C for 3–4 h. The resulting cloudy solution was collected into a centrifuge tube and centrifuged at 1200 rpm for 5 min. The supernatant was discarded, and the pellet was collected and resuspended in a T25 culture flask. The solution was gently pipetted and filtered through a 200-mesh cell strainer. The filtrate was centrifuged at 1000 rpm for 5 min, the supernatant was discarded, and the pellet was resuspended in culture medium. This centrifugation and resuspension step was repeated twice to obtain the annulus fibrosus cells. The cell suspension was plated at a defined density in culture flasks containing 10% fetal bovine serum (FBS), 1% penicillin-streptomycin solution, and DMEM/F12 medium. The cells were incubated in a cell culture incubator for 3 days, after which the medium was replaced. Once the cells adhered, the medium was changed every 2 days, and microscopic observation was performed. The non-adherent cells were removed by differential adhesion, resulting in a high-purity annulus fibrosus cell population.

Polymer CCK-8 assay

Rabbit annulus fibrosus cells were seeded in cell culture medium (90% DMEM/F12, 10% FBS, 1% penicillin-streptomycin solution) and incubated at a density of 7000 cells per well in a 96-well plate for 24 h until the cells adhered. The medium was then replaced with a mixture of PBS solution containing 10% polymer and 90% fresh culture medium, and cells were cultured for an additional 24 h. A control group without polymer was also included. After 24 h, the culture medium containing polymer was removed, and fresh medium containing CCK-8 reagent was added. The plate was incubated in a cell incubator for 2 h. The culture medium was transferred to a new 96-well plate, and absorbance at 450 nm was measured. Cell viability was calculated according to the formula in Supporting Information 1.7. For direct cell seeding on pre-formed hydrogel surfaces, cells were co-cultured for 1, 3, and 5 days, and proliferation and survival were assessed using the CCK-8 and live/dead cell fluorescence staining assays.

Hydrogel CCK-8 assay

The hydrogel solution was first filtered through a 0.2 μm filter to sterilize. Hydrogel was mixed in a 48-well plate and allowed to gel for 30 min. Rabbit annulus fibrosus cells were seeded at 2000 cells per well in the hydrogel and incubated in a cell culture incubator. The medium was changed on days 1, 3, and 5, followed by the addition of fresh medium containing CCK-8 reagent for 2 h. Absorbance at 450 nm was measured, and cell viability was calculated according to the formula in Supporting Information 1.7.

Live/dead cell staining assay

Hydrogel was prepared in 24-well plates, and rabbit annulusfibrosus cells were seeded at 20,000 cells per well in 1 mL ofculture medium. After culturing for 5 days, the medium wasremoved, and the cells were washed twice with PBS. Then, 250 μLof live/dead cell staining reagent was added to each well, and theAccepted manuscriptACCEPTED MANUSCRIPTcells were incubated in the dark for 30 min. Fluorescentimages were captured using a fluorescence microscope.

Hemolysis test

According to the ISO 10993-4 (15) medical device standard, the hydrogel was soaked in saline at a 1:5 (v/v) ratio for 24 h to prepare different concentrations of extracts. Rabbit whole blood was centrifuged and washed with saline until the supernatant became colorless and transparent. The red blood cell pellet was diluted to a 1:49 (v: v) ratio in saline to obtain a red blood cell suspension. The red blood cell suspension was mixed with equal volumes of hydrogel extract at different concentrations and incubated for 2 h at 4 °C. The mixture was centrifuged at 1500 rpm for 15 min, and images were recorded. The supernatant absorbance at 540 nm was measured for the hydrogel group, negative control (0.9% saline), and positive control (cell lysis solution). Hemolysis rate was calculated according to the formula in Supporting Information 1.8.

Animal experiments

Animal model and grouping

Thirty-two healthy adult male New Zealand White rabbits were randomly divided into four groups (n = 8): the sham surgery group (only disc exposure), the defect group (created AF defect), the 4aPEG-OPA/Gelatin/CUR group, and the 4aPEG-OPA/Gelatin/LIP@CUR group.

Surgical model and material implantation

All rabbits were fasted for 12 h before surgery and anesthetized with isoflurane gas. Once fully anesthetized, the lateral abdomen was shaved and prepared. A 10 cm incision was made between the 12th rib and the iliac crest. The thoracolumbar fascia was incised to expose the psoas muscle, which was longitudinally separated to visualize the vertebral body and intervertebral disc. The L3/4 disc space was precisely located, and an 8G needle was inserted into the disc space to a depth of 5 mm. The needle was rotated 360° and held for 30 s before being carefully withdrawn. The puncture site showed visible leakage of the gel-like material, confirming successful model creation. To simulate the immediate post-discectomy scenario and evaluate acute repair efficacy (Day 0 Defect), the other two groups received the corresponding hydrogel injection for defect filling immediately after defect creation. The surgical site was closed, and the rabbits were treated with penicillin injections for infection control during recovery.

Radiological evaluation

Radiographs and MRI were performed 4 and 8 weeks post-surgery. The disc height index (DHI) was calculated to assess collapse, and degeneration was evaluated using the modified Thompson grading scale based on T2-weighted images.

Histological analysis

Rabbits were euthanized at the end of the study period, and intervertebral disc tissue was collected for H&E staining to observe structural repair. Immunohistochemistry (IHC) was performed to detect the expression of matrix proteins (COL I, COL II) and inflammatory factors (IL-6, TNF-α). For the evaluation of inflammation, semi-quantitative analysis was subsequently performed using ImageJ software to calculate the relative optical density of the positive staining areas specifically for IL-6 and TNF-α.

Statistical analysis

Data are presented as mean ± standard deviation. GraphPad Prism was used for statistical analysis via one-way ANOVA followed by Tukey’s post-hoc test. Differences were considered statistically significant when *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Results

Material synthesis and structural characterization

The 4aPEG-OPA crosslinker was synthesized via an esterification reaction between the terminal hydroxyl groups of 4-arm PEG and the carboxyl groups of the OPA precursor. After deprotection, characteristic aldehyde peaks at 10.52 and 10.61 ppm in the ¹H-NMR spectrum confirmed its successful synthesis (Fig. 1A). Subsequently, as illustrated in Fig. 1B, the hydrogel network forms through a dynamic Schiff-base reaction. The rapid nucleophilic attack by the primary amines of gelatin onto the dual aldehydes of 4aPEG-OPA enables both rapid in situ gelation and robust tissue adhesion.The prepared LIP@CUR exhibited a typical spherical vesicle structure (TEM), with an average hydrodynamic particle size of 159.68 ± 1.98 nm, a PDI of 0.247 (in PBS), and a Zeta potential of − 39.65 ± 1.46 mV (in deionized water), indicating that the liposome system is uniform and stable (Fig. 1C). The encapsulation efficiency of curcumin was 69.9 ± 2.4%. Given the highly hydrophobic nature of curcumin and the classic thin-film hydration method used, the drug is physically loaded (rather than chemically conjugated) into the liposomes. Specifically, curcumin molecules partition into and are entrapped within the hydrophobic tail region of the lipid bilayer via hydrophobic interactions and van der Waals forces. This successful physical encapsulation is strongly supported by the drastic macroscopic transformation of curcumin from being virtually completely water-insoluble to forming a uniform, stable aqueous suspension (LIP@CUR), as well as the standard UV-Vis quantification of the entrapped drug. Cryo-SEM images revealed that the composite hydrogel had a uniform, interconnected porous structure (Fig. 1D and E). Quantitative analysis showed a mean pore size of 0.51 ± 0.28 μm. This sub-micron porous network provides ample internal space for the efficient and homogeneous dispersion of the nanoscale liposomes (LIP@CUR, ~ 160 nm) before gelation. Concurrently, the relatively dense hydrogel walls subsequently restrict the physical migration of the embedded liposomes, enabling stable physical encapsulation and providing a robust structural basis for sustained drug release. Additionally, the Feret radius ratio (min/max Feret diameter) of the hydrogel pores was calculated as 0.84 ± 0.06, quantitatively confirming the high regularity and structural integrity of the hydrogel network. Furthermore, the micro-ring structure of LIP@CUR was observed under transmission electron microscopy (TEM), as shown in Fig. 1F. LIP@CUR exhibited a regular spherical structure, uniformly dispersed in the observed field. To quantify this morphology, the average aspect ratio (major/minor axis) of the liposomes was measured to be 1.09 ± 0.10, confirming their highly spherical nature.

Rheological and mechanical properties

As shown in Figs. 2A and 4aPEG-OPA/Gelatin, 4aPEG-OPA/Gelatin/CUR, and 4aPEG-OPA/Gelatin/LIP@CUR all withstood approximately 70% strain. For further analysis, the linear portion of the compression curve (the first 20% strain) was fitted, and the slope of the linear equation was used to calculate the Young’s modulus of the hydrogel. The calculated Young’s modulus of the 4aPEG-OPA/Gelatin hydrogel was 132.74 ± 10.88 kPa, that of the 4aPEG-OPA/Gelatin/CUR hydrogel was 139.12 ± 6.28 kPa, and that of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel was 114.06 ± 13.94 kPa (Fig. 2C). The addition of liposomes slightly affected the stability of some crosslinked structures. This slight decrease in compressive modulus is primarily attributed to the steric hindrance caused by the large volume of the ~ 150 nm lipid vesicles, which partially interferes with the Schiff-base crosslinking density between 4aPEG-OPA and gelatin. Additionally, the liposomes act as soft structural inclusions within the polymer network. However, the excellent compressive performance of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel (~ 114 kPa) still matches the mechanical properties of native annulus fibrosus tissue, suggesting that it provides sufficient mechanical support for defect sealing.Furthermore, the tissue adhesion properties were evaluated using lap-shear tests (Fig. 2B). As quantitatively shown in Fig. 2D, the adhesion strength of the 4aPEG-OPA/Gelatin hydrogel was 51.07 ± 3.67 kPa, that of the 4aPEG-OPA/Gelatin/CUR hydrogel was 50.29 ± 6.56 kPa, and that of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel was 50.35 ± 6.11 kPa.There was no significant difference in the adhesion strength among these three groups. Therefore, the introduction of LIP@CUR or CUR does not affect the adhesion performance of the hydrogel. Moreover, the adhesion strength of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel was significantly higher than that of commercially available fibrin glue (16.26 ± 0.71 kPa), indicating a significant advantage in adhesion performance of the hydrogel developed in this study.

Fig. 2.

(A) Compressive stress–strain curves of the hydrogels. (B) Lap-shear Strength curves of the hydrogels. (C) Compressive modulus of the hydrogels. (D) Lap-shear strength of the hydrogels. (E) Gelation time of the hydrogels. (F) Rheological characterization of the hydrogels.*p < 0.05, **p < 0.01, ***p < 0.001 ****p < 0.0001

In addition to the static mechanical and adhesive properties, the dynamic gelation process and rheological behavior of the hydrogels were evaluated to confirm their clinical applicability. As shown in Fig. 2E, the macroscopic gelation times determined by the tube inversion method at 37 °C were 61.2 ± 3.37 s, 64.6 ± 2.57 s, and 65.0 ± 1.57 s for the 4aPEG-OPA/Gelatin, 4aPEG-OPA/Gelatin/CUR, and 4aPEG-OPA/Gelatin/LIP@CUR hydrogels, respectively. There was no significant difference among the three groups, indicating that the incorporation of CUR or LIP@CUR did not substantively affect the crosslinking kinetics. Importantly, this rapid in situ gelation (~ 1 min) perfectly aligns with surgical requirements, providing a sufficient operational window for injection while avoiding material leakage or inadequate defect sealing caused by prolonged gelation.

The dynamic sol-gel transition was further monitored using time-sweep rheological analysis at 37 °C (Fig. 2F). To clearly illustrate the initial phase, the first 5 min of the test are presented. The crossover point of the storage modulus (G′) and loss modulus (G″) was observed at approximately 1.5 min. This result is highly consistent with the macroscopic tube inversion test (~ 60 s), with the slight delay primarily attributed to the time consumed during sample loading and rotor adjustment. Furthermore, prolonged rheological monitoring (Supporting Information 2) revealed that G′ continued to increase and ultimately reached a stable plateau of approximately 11,700 Pa at 40 min. This confirms that the fully crosslinked 4aPEG-OPA/Gelatin/LIP@CUR hydrogel possesses excellent and structurally stable viscoelastic properties.

Degradation, swelling, and drug release

As shown in the corrected Fig. 3A, the in vitro degradation profile of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel (monitored over time in days) exhibited a characteristic biphasic curve. During the initial phase (days 0 to 15), the hydrogel mass gradually increased, reaching a peak of approximately 135% of its initial weight. This transient mass gain is a typical phenomenon for bulk-degrading hydrogels, where the continuous physiological fluid uptake of the crosslinked network temporarily outpaces the mass loss caused by hydrolytic degradation. Following this peak hydration phase at day 15, the bulk degradation of the gelatin and dynamic Schiff-base network became the dominant process, leading to a rapid decline in mass and complete degradation by approximately day 29. Crucially, the peak hydration state observed during this early degradation process perfectly corresponds to the swelling capacity evaluated independently. Additionally, as shown in Fig. 3B, the maximum swelling ratio of 4aPEG-OPA/Gelatin/LIP@CUR was strictly controlled at 34.21 ± 2.67%. It is important to note that this value represents the additional swelling of the as-prepared hydrogel, rather than the swelling from a dry state, to accurately simulate the post-implantation behavior in vivo. This remarkably low extra-swelling behavior is primarily attributed to the highly dense Schiff-base crosslinking network and the incorporation of hydrophobic LIP@CUR nanoparticles. Consequently, this low extra-swelling property offers significant advantages for annulus fibrosus repair, as it effectively prevents severe post-implantation volume expansion that could otherwise compress the spinal cord or exiting nerve roots.As shown in Fig. 3C, the absorption maximum exhibited a noticeable blue shift during the scavenging process. This shift originates from the consumption of DPPH radicals (characteristic peak at 517 nm) and the increasing dominance of the intrinsic curcumin signal (approximately 425 nm) as the reaction progresses. Consequently, the DPPH Clearance Rate of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel exceeded 80% (Fig. 3D).The use of ethanol as a solvent facilitates the penetration of DPPH free radicals into the hydrogel network to effectively react with the loaded curcumin (detailed experimental procedures and solvent configurations are provided in Supporting Information 1.6). This demonstrates that the drug-loaded hydrogel retains excellent antioxidant properties.The drug release capacity of LIP@CUR was evaluated (Fig. 3E), showing nearly 100% release in a short time. The release behavior of CUR from 4aPEG-OPA/Gelatin/LIP@CUR and 4aPEG-OPA/Gelatin/CUR hydrogels was further examined (Fig. 3F). The 4aPEG-OPA/Gelatin/LIP@CUR hydrogel exhibited an initial burst release followed by sustained release, while the 4aPEG-OPA/Gelatin/CUR hydrogel showed higher release efficiency with a faster rate and shorter duration. Thus, the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel provided better sustained CUR release, potentially more effective in reducing postoperative inflammation.

Fig. 3.

(A) In vitro degradation profile of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogels. (B) Swelling ratio of the 4aPEG-OPA/Gelatin/LIP@CUR hydrogels. (C) UV–vis absorption spectra of DPPH solution after co-incubation with CUR (0.1 mg/mL), 4aPEG-OPA/Gelatin, and 4aPEG-OPA/Gelatin/LIP@CUR hydrogels. (D) DPPH radical-scavenging activity of CUR (0.1 mg/mL), 4aPEG-OPA/Gelatin, and 4aPEG-OPA/Gelatin/LIP@CUR hydrogels. (E, F) Cumulative release profiles of CUR from LIP@CUR and 4aPEG-OPA/Gelatin/LIP@CUR hydrogels.*p < 0.05, **p < 0.01, ***p < 0.001 ****p < 0.0001

Biocompatibility

The CCK-8 assay indicated that, as shown in Fig. 4A, the hydrogel precursor solution did not significantly affect the viability of rabbit annulus fibrosus cells within the concentration range of 0.0625 mg/mL to 1.0 mg/mL (p > 0.05), with cell viability remaining above 80%. Similarly, material extracts did not significantly affect cell viability (Fig. 4B, p > 0.05, viability > 80%). As shown in Fig. 4C, cell viability continued to increase over time, suggesting that the hydrogel does not adversely affect cell proliferation (p > 0.05). Live/dead cell staining on the fifth day is shown in Fig. 4D, where red and green labeling indicate dead and live cells, respectively. Image analysis revealed that cells proliferated well with robust attachment and a very low proportion of dead cells. Based on these quantitative proliferation assays and high cell viability, the 4aPEG-OPA/Gelatin/LIP@CUR group demonstrated excellent biocompatibility.Furthermore, the hemolysis rate was below 5%, meeting international safety standards and demonstrating excellent blood compatibility. As shown in Fig. 5A, the supernatants of all experimental groups remained colorless and transparent, consistent with the negative control (physiological saline), while the positive control exhibited distinct red color due to red blood cell rupture. Quantitative analysis (Fig. 5B) confirmed that the hemolysis rate for all groups was below 5%,(p > 0.05) which meets international safety standards for biomaterials and demonstrates that the hydrogel possesses excellent hemolytic stability.

Fig. 4.

(A) Viability of rabbit annulus fibrosus cells after co-culture with polymer solutions at different concentrations. (B) Cell viability after 24 h co-culture with hydrogel extracts. (C) Cell viability after 1, 3, and 5 days of co-culture with the hydrogels. (D) Live/dead staining of cells after 5 days of co-culture with the hydrogels. Data are presented as mean ± SD. Statistical analysis (one-way ANOVA) showed no significant differences (p > 0.05) between experimental and control groups, confirming the high biocompatibility of the tested materials

Fig. 5.

(A) Representative microscopic images of rabbit red blood cells after 2 h incubation with hydrogel extracts. (B) Hemolysis ratio of rabbit red blood cells after 2 h incubation with hydrogel extracts

Radiological evaluation

Radiological evaluations were conducted to assess the structural integrity of the intervertebral discs. X-ray imaging (Fig. 6A) visually presented the displacement of the intervertebral disc space, which, when combined with the disc height index (DHI), allowed for a precise quantitative assessment of disc height changes at different postoperative time points. As shown in Fig. 6B, at 4 weeks post-surgery, the untreated defect group exhibited significant disc collapse, yielding the lowest DHI. In contrast, the 4aPEG-OPA/Gelatin/LIP@CUR group maintained a disc height only slightly lower than that of the sham surgery group, demonstrating substantially less collapse compared to both the 4aPEG-OPA/Gelatin/CUR and defect groups. At 8 weeks post-surgery, the defect group showed continued and progressive disc height collapse, whereas the 4aPEG-OPA/Gelatin/LIP@CUR and 4aPEG-OPA/Gelatin/CUR groups exhibited significantly less disc height reduction than the defect group.Furthermore, magnetic resonance imaging (MRI) was utilized to evaluate the hydration status of the discs. As shown in Fig. 7A, at 4 weeks post-surgery, the defect group showed a blackened intervertebral disc signal, indicating severe disc degeneration and dehydration. Conversely, the 4aPEG-OPA/Gelatin/LIP@CUR group retained a nucleus pulposus signal intensity close to that of a normal disc. At 8 weeks post-surgery (Fig. 7B), the 4aPEG-OPA/Gelatin/LIP@CUR group exhibited a significantly better modified Thompson grade than the defect group.Consistent with this morphological grading, the quantitative analysis of the intervertebral disc signal intensity (Fig. 7C) revealed that the 4aPEG-OPA/Gelatin/LIP@CUR group maintained significantly higher signal intensity values compared to the puncture group (p < 0.0001). This quantitative result objectively confirms the superior preservation of water content and structural integrity within the treated discs.It is noteworthy that while the 4-week outcomes were highly promising, mild signs of disc degeneration were observed at 8 weeks compared to the 4-week mark. This progression is primarily attributed to the gradual in vivo degradation of the hydrogel matrix and the eventual depletion of the encapsulated curcumin over the extended 8-week period, which gradually diminishes the localized anti-inflammatory microenvironment. Furthermore, the inherent avascular nature of the annulus fibrosus severely limits its intrinsic regenerative capacity under continuous biomechanical loading. However, the comprehensive radiological data clearly emphasize that despite these expected long-term challenges, the structural integrity and hydration in the 4aPEG-OPA/Gelatin/LIP@CUR group remained significantly superior to those of the untreated defect group. This indicates that the composite hydrogel exerts a robust effect in delaying the progression of disc degeneration and promoting annulus fibrosus repair.

Fig. 6.

(A) X-ray images obtained at 4 and 8 weeks postoperatively. (B) Intervertebral disc height index at 4 and 8 weeks postoperatively.*p < 0.05, **p < 0.01, ***p < 0.001

Fig. 7.

(A) MRI images obtained at 4 and 8 weeks postoperatively. (B) Thompson grading scores at 8 weeks postoperatively. (C) Quantitative analysis of intervertebral disc signal intensity at 8 weeks postoperatively.*p < 0.05, **p < 0.01, ***p < 0.001 ****p < 0.0001

Histological and Immunohistochemical Results

As shown in Fig. 8A, compared to the sham surgery group, the defect group showed a significant gap in the defect site at 4 weeks post-surgery. At 8 weeks, the defect site showed significant proliferative tissue and disordered tissue structure. At 4 weeks post-surgery, the 4aPEG-OPA/Gelatin/LIP@CUR group showed a significantly reduced annulus fibrosus defect gap, and at 8 weeks post-surgery, the defect had healed. The tissue structure was more complete than in the defect group, and the proliferative tissue was significantly less than in the defect and 4aPEG-OPA/Gelatin/CUR groups. H&E staining results indicated that the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel could effectively promote the healing of the annulus fibrosus defect.

Fig. 8.

(A) H&E staining of the annulus fibrosus tissues at 4 and 8 weeks postoperatively. (B) Immunohistochemical staining of the annulus fibrosus tissues at 8 weeks postoperatively

Immunohistochemical analysis, as shown in Fig. 8B, demonstrated that the expression of TNF-α and IL-6 in the 4aPEG-OPA/Gelatin/LIP@CUR group was visibly lower than in the defect group and the 4aPEG-OPA/Gelatin/CUR group. This observation was further validated by semi-quantitative analysis (Fig. 9A, B), which showed that the relative optical densities of TNF-α and IL-6 in the 4aPEG-OPA/Gelatin/LIP@CUR group were significantly decreased compared to the untreated defect group (p < 0.001). Together, these findings robustly demonstrate that the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel possesses excellent anti-inflammatory properties. Additionally, qualitative observation of the immunohistochemical staining (Fig. 8B) revealed that the 4aPEG-OPA/Gelatin/LIP@CUR group exhibited visually denser and more continuous positive staining for both COL I and COL II compared to the untreated defect and 4aPEG-OPA/Gelatin/CUR groups. This visually more intense extracellular matrix deposition suggests that the composite hydrogel creates a favorable microenvironment for promoting collagen synthesis, which is highly beneficial for the structural regeneration of the annulus fibrosus.

Fig. 9.

(A) Semi-quantitative analysis of IL-6 expression. (B) Semi-quantitative analysis of TNF-α expression.*p < 0.05, **p < 0.01, ***p < 0.001

Discussion

Annulus fibrosus (AF) repair after discectomy is crucial to prevent recurrence. The 4aPEG-OPA/Gelatin/LIP@CUR composite hydrogel system developed in this study integrates both “physical sealing” and “biological modification” aspects in its design.

First, mechanical compatibility and tissue adhesion are the foundation of AF sealing materials. The dynamic network formed by Schiff-base bonds between 4aPEG-OPA and gelatin imparts excellent tissue adhesion (~ 50 kPa) to the material. This strength is sufficient to withstand the fluid pressure within the intervertebral disc, preventing early displacement or extrusion of the hydrogel.To rigorously evaluate the clinical translation potential of this physical sealing strategy, we employed an acute AF defect model simulating the immediate post-discectomy scenario. The initial needle puncture disrupts the concentric lamellae of the AF, causing acute depressurization of the intervertebral disc and continuous extrusion of the nucleus pulposus (NP). By injecting the hydrogel immediately after defect creation (Day 0), we subjected the material to the most demanding biomechanical conditions. This approach specifically evaluates the hydrogel’s capacity for rapid in situ gelation and instant wet-tissue adhesion within a challenging, NP-residue-rich environment, enabling the immediate physical sealing necessary to halt further leakage and prevent the onset of an irreversible degenerative cascade. Furthermore, this strategy emphasizes the critical window of acute functional repair, rather than attempting to reverse end-stage chronic degeneration. Furthermore, its compressive modulus (~ 114 kPa) is similar to that of natural AF tissue, providing moderate mechanical support to buffer axial loads and delay the loss of disc height. This was confirmed by X-ray results in the animal experiments.To ensure clinical relevance, we compared these properties with physiological loads. Although the compressive modulus (~ 114 kPa) is below that of intact AF tissue, the ~ 50 kPa adhesion strength—coupled with the confined geometry of the defect—is sufficient to resist extrusion under typical intradiscal pressures (0.1–0.5 MPa in neutral positions). This provides the necessary mechanical stability during the early stage of AF repair.

Second, to address the local inflammatory microenvironment at the defect site, we introduced curcumin-loaded liposomes (LIP@CUR). Conventional curcumin administration suffers from issues of “rapid release and rapid metabolism.” This study demonstrated that through the dual blocking effect of “liposomal encapsulation + hydrogel network,” the release period of curcumin was significantly prolonged. In vivo experiments showed that the Gelatin/LIP@CUR group exhibited lower levels of inflammatory cytokines and higher matrix collagen expression compared to the Gelatin/CUR group, strongly indicating the importance of sustained drug delivery in reversing persistent inflammation and promoting tissue regeneration.Furthermore, AF injury recovery is a highly dynamic process characterized by fluctuations in multiple biological signals, particularly a shift toward an acidic microenvironment driven by local inflammation and anaerobic glycolysis. Such environmental changes can alter the isoelectric point of the tissue matrix and affect the surface charge of implanted biomaterials, potentially leading to liposome aggregation or premature drug leakage. However, the colloidal stability of our system is safeguarded by two distinct mechanisms. First, the formulated LIP@CUR possesses a robust initial negative Zeta potential (− 39.65 ± 1.46 mV). This strong surface charge provides a broad safety margin against charge neutralization, ensuring persistent electrostatic repulsion between nanoparticles even if the local pH decreases. Second, the encapsulation of these liposomes within the dense 4aPEG-OPA/Gelatin hydrogel network provides physical steric shielding, mitigating the impact of external biochemical fluctuations. Together, these properties ensure robust colloidal stability under a dynamically changing microenvironment, effectively preventing nanoparticle aggregation and maintaining a consistent, sustained drug release profile throughout the repair process.

Moreover, the low swelling property of this hydrogel has significant clinical safety implications, as it prevents excessive swelling post-implantation, which could otherwise compress the spinal cord or nerve roots. Regarding the degradation profile, our in vitro baseline assessment in physiological simulating conditions (PBS, pH 7.4, 37 °C) demonstrated complete mass loss over approximately 30 days. Although exogenous enzymes were not included in this static hydrolytic assay to establish a baseline, the incorporation of gelatin inherently endows the hydrogel network with robust susceptibility to endogenous matrix metalloproteinases (MMPs) in vivo. Crucially, this ~ 4-week degradation timeline—particularly the notable mass drop of approximately 25% observed during the third week—is strategically engineered to precisely align with the early phase of AF soft tissue repair. This ensures that the hydrogel provides a temporary, load-bearing mechanical bridge during the initial critical period (the first two weeks), and subsequently degrades at an appropriate rate to yield sufficient spatial clearance for endogenous cell infiltration and new extracellular matrix deposition. Furthermore, our 4-week in vivo histological and immunohistochemical results (showing significant defect closure and robust Col I/II expression) robustly confirm that this specific degree of mass loss in the third week is indeed sufficient to promote adequate structural repair.Concomitant with structural degradation, the mechanical properties of the hydrogel undergo dynamic evolution. Hydrolysis and MMP-mediated cleavage of the Schiff-base network over the 4-week period lead to a reduction in crosslinking density and a subsequent decline in the bulk compressive modulus. Biomechanically, this mechanical attenuation is beneficial as it enables ‘stress transfer’ (load sharing). While the initial robust modulus (~ 114 kPa) ensures an immediate physical seal against acute intradiscal pressure, the subsequent mechanical waning allows for the progressive transfer of loads to infiltrating cells and nascent collagen networks. This mechanical stimulation is a key driver of cellular mechanotransduction, which further promotes matrix maturation and functional restoration of the AF.

For successful clinical translation, the sterilization of implantable biomaterials is a critical consideration. Traditional terminal sterilization techniques, such as autoclaving or gamma irradiation, are unsuitable for this composite hydrogel; high temperatures would irreversibly denature the gelatin and disrupt the LIP@CUR lipid bilayers, while irradiation could induce uncontrolled polymer chain scission, fundamentally altering the material’s mechanical properties and degradation profile. Therefore, the ideal clinical sterilization strategy for this system relies on aseptic processing. Specifically, the liquid precursor solutions (4aPEG-OPA and gelatin) and the LIP@CUR suspension can be individually sterilized using standard 0.22 μm sterile filtration—a protocol successfully employed prior to our in vitro cellular assays. By injecting and mixing these pre-sterilized precursors in situ under standard aseptic surgical conditions, the final hydrogel network forms without being exposed to harsh sterilizing environments. This strategy ensures that the delicate Schiff-base crosslinking kinetics, optimal compressive modulus, and sustained drug-release characteristics of the hydrogel are fully preserved for robust AF repair.Beyond sterilization, the practical method of application in a clinical setting is a fundamental determinant of the material’s translatability. To ensure surgical feasibility, the 4aPEG-OPA/Gelatin/LIP@CUR hydrogel is designed to be delivered via a dual-chamber pre-filled syringe system integrated with a static mixing tip. This configuration enables the two precursor solutions to be mixed homogeneously during the injection process, initiating the ~ 60 s in situ gelation immediately upon reaching the AF defect site. In a real-world surgical scenario, this system could be applied either during open discectomy to seal the residual annular hole or as a minimally invasive percutaneous intervention under real-time fluoroscopic or CT guidance. Such a controlled delivery approach facilitates precise anatomical filling of the irregular defect while minimizing surgical trauma. By integrating this streamlined application method with the aforementioned aseptic processing, the proposed hydrogel system transitions from a laboratory-scale material to a viable clinical tool for postoperative disc management.

Nevertheless, certain limitations must be acknowledged. First, tracking real-time in vivo biomechanical changes and mapping the exact timeline of this stress-transfer process remains technically challenging; thus, dynamic mechanical monitoring during the degradation phase was not performed. Second, regarding the in vivo evaluation, the rabbit puncture model was strategically selected because it provides a highly reproducible baseline for acute AF injury and shares fundamental biochemical and structural similarities with human discs. This makes it an exceptionally useful proof-of-concept platform for assessing the hydrogel’s biocompatibility, tissue adhesion, and acute microenvironment modulation. However, extrapolating these data directly to a real clinical scenario requires caution. Rabbits are quadrupedal, meaning their spinal mechanical loading environment fundamentally differs from the dominant axial compressive forces experienced by bipedal humans, and their disc dimensions are substantially smaller. Therefore, our current data primarily validate the early-stage biological and initial mechanical sealing efficacy of the material. Finally, the 8-week observation period is insufficient to assess ultimate outcomes after complete material resorption. Future research utilizing large animal models (e.g., ovine or porcine) that better approximate human spinal biomechanics and volume, combined with non-invasive monitoring techniques, is a mandatory prerequisite to validate the long-term durability and clinical translatability of this synergistic repair strategy.

Conclusion

In summary, this study successfully developed an injectable, and tissue-adhesive composite hydrogel loaded with curcumin-loaded liposomes. The material demonstrates good biocompatibility, suitable mechanical strength, and excellent drug release properties. Both in vitro and in vivo experiments consistently confirmed that the composite hydrogel effectively closed the AF defect, provided mechanical support, and exhibited long-lasting anti-inflammatory and antioxidant effects through a synergistic mechanism. This resulted in the effective suppression of inflammation, promotion of extracellular matrix synthesis, and structural repair of the AF defect, thereby delaying intervertebral disc degeneration. This strategy offers a promising biomaterial solution for addressing the challenges of AF repair after discectomy.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Q.Z.: Writing – original draft, Conceptualization, Methodology.M.K.: Data curation, Software.R.D.: Project administration, Investigation.X.T.: Data curation, Formal analysis, Validation.Y.Q.: Writing – review & editing, Supervision, Funding acquisition.

Funding

This work was supported by the Jilin Province Department of Finance, China (Grant No. 2020SCZT033).

Data availability

The original contributions presented in the study are included in the article and Supplementary files of the present research article.

Declarations

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.

Ethics and consent to participate

This study was conducted in accordance with the “Guidelines for Ethical Review of Animal Welfare in China” (GB/T 35892 − 2018) and the “Regulations for the Administration of Laboratory Animals in China.” Male Japanese White rabbits (2.5 ± 0.2 kg) were obtained from Wangdu Tonghui Breeding Co., Ltd. (Hebei, China; License No. SCXK (Ji) 2021-006).The protocol was approved by the Animal Ethics Committee of the School of Basic Medical Sciences, Jilin University (Approval No. SYXK 2017-003). All animal procedures were performed following relevant institutional/national regulations and the principles of the Declaration of Helsinki.

Declaration of AI

During the preparation of this work, the authors used ChatGPT solely to polish the English language and check for grammatical errors. After using this tool, the authors rigorously reviewed and edited the manuscript, and assume full responsibility for the final content of the publication. The original drafting, data curation, figure generation, and data analysis were conducted entirely manually by the authors without the assistance of AI technologies.