Multifunctional ultrafast hydrogel for arthroscopic surgery: An enzymatically cross-linked, nanosilicate-reinforced system promotes dual cartilage and subchondral bone regeneration

Abstract

Background

Arthroscopic repair of osteochondral (OC) defects using injectable hydrogels remains highly challenging due to the high-pressure, water-filled environment of the joint during arthroscopic surgery. Conventional hydrogels exhibit slow gelation kinetics, prolonged setting times, poor adhesion to wet tissues, and insufficient mechanical strength, rendering them prone to washout throughout the procedure.

Methods

To address these limitations, we incorporated a small amount of transglutaminase (TG) and synthetic lithium silicate nanoplatelets (SN) into a gelatin–oxidized starch (GelS) precursor and evaluated the regenerative performance of the resulting hydrogel under simulated arthroscopic conditions. In vivo, the hydrogels were implanted into osteochondral defects in rats to assess their repair efficacy.

Results

The GelS-TG-SN hydrogel demonstrated ultrafast enzymatic gelation, robust underwater adhesion, and significantly enhanced mechanical strength. It was cytocompatible, displayed anti-inflammatory activity, and supported context-dependent dual-lineage differentiation of Synovial-derived stem cells (SDSCs) chondrogenesis in a cartilage-like niche and PI3K–Akt–mediated osteogenesis in a vascular-like niche. Following 8-week implantation, it enabled coordinated regeneration of cartilage and subchondral bone, recapitulating native osteochondral architecture.

Conclusion

The GelS-TG-SN nanocomposite hydrogel offers a compelling strategy for effective osteochondral regeneration in arthroscopic surgical environments.

The Translational Potential of this Article

This hydrogel platform offers an elegant and clinically accessible solution for arthroscopic osteochondral repair. Its ultrafast gelation—achieved in under one minute—combined with resilient adhesion under constant irrigation enables seamless intraoperative application without auxiliary instrumentation. By capitalizing on the body's intrinsic osteochondral gradient, a single injection orchestrates synchronized regeneration of cartilage and subchondral bone. Such integration of procedural simplicity with inherent regenerative bioactivity underscores its promise as a genuinely “one-step” therapy ready for clinical translation.

Graphical abstract

Schematic illustration of the GelS-TG-SN hydrogel for arthroscopic osteochondral repair. Gelatin and oxidized starch were crosslinked by TG to form the injectable GelS-TG hydrogel, followed by incorporation of SN to obtain GelS-TG-SN. The hydrogel showed rapid gelation, wet adhesion, and enhanced osteochondral regeneration.

1. Introduction

Osteochondral defect, involving damage to both the articular cartilage and underlying subchondral bone, represents a significant clinical challenge in orthopedics [1,2]. These defects, often resulting from trauma, osteochondritis dissecans, or degenerative conditions, possess poor intrinsic healing capacity due to the avascular nature of cartilage and the highly vascularized subchondral bone [3,4]. Left untreated, they frequently progress to debilitating pain, joint dysfunction, and ultimately osteoarthritis, imposing a substantial burden on both patients and healthcare systems [5].

Arthroscopic surgery has increasingly become the preferred minimally invasive technique for diagnosing and treating intra-articular pathologies, including osteochondral defects [6]. Injectable hydrogels are of particular interest for cartilage tissue engineering since they can be applied in minimally invasive procedures. However, a major drawback of current clinical hydrogel systems is their slow gelation kinetics in situ after injection in an aqueous environment, as well as the need to resist arthroscopic irrigation pressure (∼10 kPa) required for further surgical procedures. Conventional hydrogel application typically drains the irrigation fluid which disrupts the fluid-based visual field essential for arthroscopic surgery [[7], [8], [9]]. Some studies have attempted to expedite gelation times by developing equipment that enables crosslinking using light [10]. However, these procedures add complexity to the overall operation, making in situ gelling under arthroscopic conditions technically demanding and cost-prohibitive. Therefore, there is an urgent need to develop a hydrogel capable of rapid and spontaneous underwater gelation using existing arthroscopic equipment, while maintaining stability under typical irrigation pressure.

Enzymatic crosslinking is an attractive method to induce in situ hydrogel formation as it is chemically mild. Transglutaminase (TG) is an enzyme which catalyzes covalent crosslinking between glutamine and lysine residues within gelatin [11]. Our previous studies further demonstrated that TG promotes the formation of stable isopeptide bonds in gelatin/oxidized starch (GelS) prepolymer systems, leading to accelerated crosslinking and a significant reduction in hydrogel gelation time [12]· Through the synergistic effect of reversible Schiff base bonds and irreversible isopeptide bonds, the resulting GelS-TG double-network formed hydrogel quickly, and possesses a dense structure, which effectively restricts excessive water infiltration and excessive network swelling. Moreover, the rapid formation of irreversible isopeptide bonds not only reinforces the hydrogel's internal network but also facilitates rapid covalent integration with the surrounding collagen‐rich tissue matrix. This bonding mechanism ensures robust gelation and structural integrity in aqueous environments, making it suitable for arthroscopic surgery applications [13]. It should be noted, however, that the amino groups on gelatin participate in competitive reactions—being targeted both by aldehyde groups from oxidized starch (OS) by lysine residues during TG-mediated crosslinking. Although the addition of TG markedly accelerates the overall gelation process, the finite number of available amino groups ultimately constrains the final crosslinking density and mechanical strength of the GelS hydrogel.

Tissue regeneration is also helped by bioactive materials [14]. Synthetic lithium silicate nanoplatelets (SNs; chemical formula: Na⁺_0.7[(Si₈Mg_5.5Li_0.3) O₂₀(OH)₄] ⁻_0.7), are biodegradable nanoscale silicates with promising bioactivity. SNs have been widely utilized as mechanical reinforcements in polymer-based composites, with evidence showing that their incorporation into hydrogels significantly improves mechanical strength [15]. SN also act as bioactive agents to enhance chondrogenic and osteogenic differentiation in mesenchymal stromal cells [16]. Studies have demonstrated its strong pro-chondrogenic effects during early stages of stem cell differentiation, while subsequently promoting endochondral ossification under pro-angiogenic conditions [17,18]. The dual bioactivity of SN aligns well with the requirements for integrated osteochondral regeneration [19].

We hypothesize that incorporating TG enzyme and SN will enhance the gelation kinetics and mechanical strength and adhesive properties of the hydrogel, thereby facilitating its arthroscopic application. Furthermore, incorporating SN brings dual-lineage bioactivity to the hydrogel system, thereby facilitating osteochondral defect repair (Fig. 1).

Fig. 1.

Schematic illustration of the design and functional properties of the GelS-TG-SN hydrogel. The hydrogel was synthesized via enzymatic crosslinking of gelatin and oxidized starch mediated by transglutaminase (TG), with incorporation of silicate nanoplatelets (SN). The resulting composite hydrogel meets key requirements for arthroscopic application under aqueous pressure (∼10 kPa), exhibiting rapid gelation, strong tissue adhesion, and high mechanical strength. OS: Oxidized starch; SN: Synthetic lithium silicate nanoplatelets; TG: Transglutaminase; Gel: Gelatin; OC defect: Osteochondral defect.

2. Materials and methods

2.1. Animals

Seventy-two adult male Sprague Dawley rats (200 ± 20 g) were purchased from the Vital River Laboratory Animal Technology Company (Beijing, China). All animals were housed in pathogen free facility at Southern University of Science and Technology under controlled conditions (22 ± 2 °C, 50 ± 5% humidity, 12 h light/dark cycle) with free access to food and water. Rats were randomly assigned to experimental groups, and all analyses were performed in a blinded manner (investigators were blinded except Q.L.).

Animals were anesthetized via inhalation of isoflurane (induction: 4–5%, maintenance: 1.5–2.5%) delivered in oxygen through a vaporizer and a fitted facemask. After ensuring the absence of pedal reflexes, the knee region was shaved and disinfected with povidone iodine. A medial parapatellar incision was made to expose the femoral trochlea. A full thickness osteochondral defect (1.5 mm in diameter, 3 mm in depth) was created in the trochlear groove using a high-speed drill. The defect was irrigated with sterile saline to remove debris. Subsequently, 6 μL of the pre mixed hydrogel composite containing synovial derived stem cells (SDSCs) was injected into the defect site. After waiting 5 min for gelation, the joint capsule and skin were closed in layers using absorbable sutures and staples, respectively. After surgery, animals were placed on a warming pad until fully recovered from anesthesia. For postoperative analgesia, meloxicam (0.05 mg/kg) was administered subcutaneously every 24h for the first 48 h. Animals were singly housed for the first 3 days to prevent wound interference and were monitored daily for signs of infection, discomfort, or impaired mobility. Sutures/staples were removed 10 days postoperatively. All animals were euthanized at 4 or 8 weeks after surgery by an overdose of pentobarbital sodium (150 mg/kg). All animal procedures were approved by the Institutional Animal Care and Use Committee of Southern University of Science and Technology (Approval No. SUSTech JY202401103).

2.2. Isolation and cultivation of human SDSCs

As previously reported, SDSCs were isolated from synovial tissues obtained during total knee arthroplasty (TKA) procedures which were approved by the ethics committee of Shenzhen People’s Hospital (LL-XB-2024001-01) [20]. The synovial tissues were rinsed three times with PBS containing penicillin–streptomycin and then digested using 0.075% type XI collagenase for 1 h at 37 °C with shaking. The resulting cell suspension was filtered through a 100 μm strainer and centrifuged at 800×g for 5 min. Cells were maintained at 37 °C in a 5% CO₂ atmosphere and the medium was refreshed every 2–3 days until 70–80% confluence was reached. SDSCs at passage three were used for subsequent experiments [20].

2.3. Synthesis of gelatin/oxidized starch (GelS) co-delivery hydrogels

The procedure for synthesizing oxidized starch (OS) was based on established methods [12]. A 4.8% (WT) soluble starch solution was prepared in distilled water, and its pH was lowered to 3.0 using sulfuric acid. Sodium periodate (NaIO₄) was then introduced into the solution at a 1:2 (w/w) starch-to-periodate ratio. This mixture was reacted for 5 h in an ice bath. The final product was purified through a 4-day dialysis process against distilled water, followed by lyophilization to yield the pure oxidized starch.

GelS hydrogels were synthesized through a Schiff base reaction using gelatin and oxidized starch. Solutions of oxidized starch (2.5, 5, and 10% wt) were combined with a 20% wt gelatin solution in a 1:1 wt ratio at room temperature. TG or SN was individually incorporated into GelS, yielding GelS-TG and GelS-SN, respectively. For the co-delivery system (GelS-TG-SN), SN was first blended into the oxidized starch solution and then mixed into the gelatin solution with TG prior to crosslinking (Fig. 1). A CCK-8 assay was subsequently performed to identify the most effective SN concentration in hydrogels. The mixtures were cast into preformed PDMS molds (8.0 mm diameter × 2.0 mm thickness) and permitted to crosslink in situ, with gelation time monitored. After crosslinking, all hydrogel samples were rinsed in deionized water to remove any unreacted components.

2.4. Characterization of GelS-based co-delivery hydrogel

The morphology, thickness, chemical structure, and ion release profile of the hydrogel were observed and analyzed by SEM, Fourier transform infrared spectrometer(FTIR), and Rheological test and high-performance liquid chromatography-tandem mass spectrometry (HPLC-ICP-MS) method.

• 1)SEM

The microstructural features of the hydrogels from various groups and their interfaces with cartilage tissue were examined by SEM (SU8100, Hitachi, Japan). Sample preparation involved rapid freezing in liquid nitrogen, followed by vacuum drying. A thin layer of gold was subsequently applied to the samples using an ion sputter coater (MC1000, Hitachi, Japan). Imaging of both the surface and internal porous architecture was performed at an accelerating voltage of 5 kV.

• 2)FTIR analysis

For identification of functional groups presenting in the hydrogel, the samples were lyophilized. The resulting freeze-dried material was pulverized, uniformly blended with potassium bromide (KBr), and pressed into transparent pellets. FTIR spectroscopy (Bruker Vertex 70) was subsequently used to acquire spectral data at room temperature.

• 3)Rheological test

The gelation kinetics of the hydrogels was evaluated using a Discovery HR-2 rheometer (TA Instruments, DE, UK) equipped with an 8 mm diameter steel parallel-plate geometry to monitor the storage (G′) and loss (G″) moduli. Prepolymers were thoroughly mixed and immediately loaded onto the instrument. Real-time crosslinking was assessed at 24 °C using varying concentrations of prepolymer and TG. For frequency sweep measurements, hydrogel samples were formed by pipetting the precursor solution onto a hydrophobic Teflon substrate. Tests were conducted at 24 °C under a constant strain of 1%, with angular frequencies ranging from 0.1 to 100 rad/s. Given that previous studies have indicated the temperature during arthroscopy is close to room temperature, the experiments in this study were conducted under ambient conditions to simulate the surgical environment [21,22].

• 4)Mechanical properties

The compressive mechanical properties of the hydrogel samples were evaluated on an electromagnetic mechanical test system (M−100, CARE Measurement & Control). Uniaxial compression was applied at a constant displacement rate of 0.03 mm/s until reaching a maximum displacement of 0.5 mm.

• 5)Ionic release kinetics of SN in vitro

To quantify the total ionic content across various samples, hydrogels were fabricated following the procedure above. After immersion in the solution, the hydrogel specimens were retrieved. The residual solution was subsequently centrifuged, and the resulting supernatant was collected for content analysis via high-performance liquid chromatography (HPLC).

2.5. Adhesiveness evaluation

Lap shear test was performed according to a previous method [10]. Briefly, a 3.5 × 2.5 cm fresh hog casing was attached to an 8 × 2.5 cm glass slide with cyanoacrylate glue. Then, 100 μL of gel precursor was uniformly dispersed on the surface of the hog casing. A second hog casing was attached to a glass slide and placed on the first glass slide. The obtained test samples were allowed the gel precursor to gel in situ. One side of the sample was fixed, and increasing pull strength was applied to the other side by an electromagnetic mechanical test system (M−100, CARE Measurement & Control). Then, the tissue adhesive strength of the hydrogel was calculated. Samples were tested to determine the adhesive strength of each group.

To comprehensively evaluate the adhesive properties of the hydrogel with OC defect, three ex vivo tissue models—rabbit femoral condyle osteochondral defect, mouse femoral fracture, and mouse calvarial defect—were established for simulated testing.

• 1)Rabbit femoral condyle osteochondral defect model:

Fresh rabbit femoral condyles were used to create a full-thickness osteochondral defect (3 mm diameter) in the trochlear groove using a sterile drill. A total of 25 μL hydrogel precursor was injected into the defect and crosslinked for 5 min at room temperature. To simulate the typical 10 kPa water pressure during arthroscopy, adhesive retention was evaluated while connected to an adjustable hydraulic water gun and the hydraulic pressure was recorded by a hydraulic gauge.

• 2)Mouse femoral fracture model:

A transverse mid-diaphyseal fracture was created in ex vivo mouse femurs. About 20 μL hydrogel was applied to the reduced fracture site. The stability and hydrogel integrity were evaluated to preliminarily assess bone adhesive performance.

• 3)Mouse calvarial defect model:

A circular calvarial defect (1.5 mm diameter) was prepared and covered with ∼10 μL hydrogel. After 5 min of gelation, pulsed irrigation was applied to evaluate adhesion on avascular bone surfaces.

• 4)Knee arthroscopic operation in vitro

To assess the injectability, in situ gelation, and adhesive performance of the hydrogel under conditions simulating fluid-irrigated knee arthroscopy, an arthroscopic simulation was performed using a freshly isolated porcine knee joint. A customized arthroscopic setup consisting essentially of a light source and a smartphone-based imaging system was employed. Two 8 mm portals were established on the medial and lateral aspects of the joint. An arthroscope coupled to a display unit was introduced into the articular cavity under continuous saline irrigation. After the creation of a cartilage defect, the adhesive hydrogel precursor was delivered into the defect site via a custom-designed hollow cannula.

2.6. Cell viability, spreading, proliferation, and migration assays

Cell viability was assessed using a Live/Dead Viability assay kit. Fluorescence microscopy was employed to visualize and image stained cells cultured on the hydrogels. Quantification of viable (green) and non-viable (red) cells was performed using ImageJ software. Data are presented as mean ± standard deviation derived from a minimum of three independent replicates. Viability percentage was calculated as the ratio of live cells to the total number of cells. To evaluate cell spreading, samples were stained with Alexa Fluor™ 594-phalloidin for F-actin and DAPI for nuclei. After rinsing with PBS, cell morphology was captured by microscopy, and the average cell area was measured using ImageJ.

Cell migration was examined via a scratch wound healing assay. SDSCs were seeded in a 10-cm dish at 1 × 10⁵ cells/mL. Upon reaching approximately 70% confluency, a linear scratch of ∼2.0 mm in width was introduced using a pipette tip. Hydrogel discs (8 mm diameter, 2 mm height) were then placed in the dish. Wound closure was monitored at predetermined intervals using an inverted microscope (Brunel SP99F, UK) and the percentage of gap closure was quantified to evaluate migratory behavior.

2.7. Macrophage polarization

To investigate macrophage polarization, Raw 264.7 cells were first stimulated with lipopolysaccharide (LPS) to induce the M1 phenotype. Subsequently, the culture medium was replaced with hydrogel extracts to assess their immunomodulatory effects. Cells were harvested for analysis 24 h after the treatment with hydrogel extracts.

2.8. Alcian blue staining and quantification

The influence of hydrogel on mineralized extracellular matrix deposition under both 2D and 3D culture conditions was evaluated using Alcian blue staining. After 7 and 14 days of culture, cells and constructs were stained with 2% (w/v) Alcian blue S solution. For quantitative assessment, the stained samples were treated with 10% (v/v) acetic acid for 30 min. The acidified solution was then neutralized with 10% ammonium hydroxide, and the absorbance of the resulting supernatant was measured at 405 nm using a microplate reader.

2.9. Gene expression measurement

To evaluate the biomarker expression levels of cells encapsulated within the hydrogel samples, quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed to examine the expression of genes associated with chondrocyte hypertrophy and macrophage polarization, including COL X, MMP13, iNOS, and CD206. The primer sequences used are provided in Supplement Table 1. On day 14 and 28, cell-laden hydrogels were retrieved from culture and lysed with Trizol reagent (Invitrogen). Subsequent qRT-PCR analysis was carried out using a Bio-Rad MyiQ2 thermocycler with SYBR® Green Supermix (Bio-Rad).

2.10. Immunofluorescence staining

We used F4/80, CD206, iNOS, COL II, and SOX9 antibodies to label the target proteins to further evaluate protein expression levels associated with macrophage and chondrogenic differentiation in the hydrogel. Hydrogel samples were first fixed with paraformaldehyde (4%) and permeabilized with Triton X-100 solution. Then, 4% normal goat serum was used to block the nonspecific antigens, primary and secondary antibody solutions were used to incubate the hydrogel samples for 2 and 1h, at 37 °C with 5% CO². Finally, actin and cellular nuclei were marked with Alexa Fluor 594/488 phalloidin and DAPI, respectively. The staining results were observed and recorded using a confocal microscope, and the target protein-labeled area was quantified using ImageJ software (v.1.44).

2.11. In vivo osteogenic and chondrogenic assay

SD rats were randomly assigned to four experimental groups: GelS, GelS-TG, GelS-SN, and GelS-TG-SN. A 1.5 mm femoral defect model was established, into which the hydrogel was injected and allowed to crosslink for 5 min before being subjected to hydraulic flushing to assess retention (Supplementary Video 1).

• 1)Histological assessment

For paraffin embedding, tissue samples underwent decalcification in 19% EDTA (pH 7.4) for 8 weeks and were then processed using an automated tissue processor (Leica ASP 300S) for embedding in paraffin. Sectioning was performed at a thickness of 10 μm with a microtome. The sections were stained with hematoxylin and eosin (H&E) for general histological evaluation, Safranin O/Fast Green to evaluate cartilage morphology, and Alcian blue for glycosaminoglycan detection. Digital images of stained sections were acquired with an Aperio CS2 slide scanner (Leica). Cartilage repair in each group was assessed based on the International Cartilage Repair Society (ICRS) scoring criteria.

• 2)MRI

All imaging was performed on a 9.4 T preclinical MRI system (μMR 9.4 T, United Imaging Life Science Instruments, China). High-resolution proton density-weighted fast spin-echo images with fat suppression were acquired in both sagittal and coronal orientations. The imaging parameters were as follows: repetition time (TR) = 2000 ms, echo time (TE) = 15.36 ms, echo train length (ETL) = 5, field of view = 18 × 38 mm², matrix size = 176 × 372, slice thickness = 1 mm, no gap between slices, voxel size = 0.1 × 0.1 × 1 mm³, and number of averages = 2. Following image acquisition, sagittal cross-sections through the repair site were selected for further evaluation of cartilage regeneration and repair quality. Joint degeneration was graded based on the Whole-Organ Magnetic Resonance Imaging Score (WORMS) system, which assesses cartilage signal intensity, morphology, bone marrow lesions, and bone attrition. Higher WORMS scores indicate more severe degenerative changes. Three independent blinded evaluators performed all scoring.

• 3)Micro-CT evaluation

Knee joints from all groups were examined by micro-computed tomography (SkyScan1178, Bruker microCT, Kontich, Belgium). Anteroposterior radiographic images were acquired using the following parameters: voltage 100 kV, current 200 μA, resolution 10 μm, and exposure time 623 ms. Quantitative analysis of the subchondral bone architecture included BV/TV and trabecular number (Tb.N). Three-dimensional reconstructions in sagittal planes for each group were generated using DataViewer software (Bruker micro-CT).

2.12. RNA sequencing and data analysis

Cell specimens were collected from 24-well plates (GelS-TG vs. GelS-TG-SN) on day 14. Total RNA was isolated using Trizol reagent and its concentration was measured with a ND-2000 spectrophotometer (NanoDrop Technologies). Sequencing libraries were prepared with the BGISEQ-500 platform (igeneCode Biotech, Beijing), and high-quality reads were aligned to the rat reference genome. Differential gene expression analysis was carried out with the DESeq2 package in R, with significantly differentially expressed genes (DEGs) identified under the threshold of |log₂ (fold change) | ≥ 2 and adjusted p-value ≤0.05. Visualization of expression patterns was achieved through heatmaps generated with the R heatmap package. Functional annotation of DEGs was conducted via Gene Ontology (GO) and KEGG pathway analyses using the clusterProfiler R package.

2.13. Statistical analysis

Data were expressed as means ± SD and analyzed with GraphPad Prism (version 9.0; GraphPad Software, California). Intergroup comparison was subject to one-way analysis of variance followed by Tukey post-hoc tests. Statistical significance was set at P < 0.05.

3. Results

3.1. Hydrogel fabrication

A major challenge in hydrogel-based arthroscopic therapy is achieving rapid gelation and high mechanical stability within the dynamic, fluid-filled joint space to prevent material washout and enable precise defect filling [10]. To address the challenges of long gelation time and complex preparation conditions which are typical for underwater hydrogels, we employed enzymatic crosslinking to significantly shorten the gelation time under convenient conditions without the need for additional equipment. Based on our previous studies, gelatin can interact with OS to form a hydrogel via reversible Schiff base reaction between amino groups of gelatin and aldehyde groups of OS [12]. Simultaneously, TG acts as a highly efficient biocatalyst, rapidly forming inreversible covalent isopeptide bonds between gelatin molecules by catalyzing cross-links specifically between glutamine and lysine residues [12]. These isopeptide linkages serve as "anchors" which tightly crosslinking gelatin molecules and significantly accelerating gelation, thereby effectively preventing hydrogel dispersal in aqueous environment. As schematically illustrated, the fabrication process of the hydrogel is presented along with a representative cylindrical cross-linked sample (8 mm in diameter and 2 mm in height) and its simple chemical composition (Fig. 1, Fig. 2B). The successful oxidation of starch was confirmed by Fourier transform infrared spectrometer(FTIR)spectroscopy, as evidenced by the appearance of a characteristic peak at 1738 cm⁻¹ which is attributed to the C=O stretching vibration (Fig. S1).

Fig. 2.

Constructing injectable, fast-gelling, high-strength hydrogels. (A) Representative SEM images of hydrogels from different groups. Scale bars: 400 μm. (B) Hydrogel of GelS-TG-SN. (C) FTIR spectra of the hydrogels. (D-E) EDS elemental analysis of GelS and GelS-TG-SN hydrogels. (F) Injection capability of the hydrogel through a syringe needle. (G) Gelation time of different hydrogel groups. (H) Rheological time sweep showing the gelation process of GelS-TG-SN. (I) Inversion test demonstrating gelation of GelS and GelS-TG-SN. (J) Final elastic modulus of various hydrogels. (K) Stress–strain curves of the hydrogels. (L) Rheological viscoelastic properties assessed through oscillatory measurements. (M) Deformation behavior of hydrogels under mechanical loads. (N) EDS quantitative analysis of trace element content within the GelS-TG-SN hydrogel. (O-Q) In vitro degradation profiles and corresponding ion release kinetics of magnesium and silicon of GelS-TG-SN hydrogel. ∗∗p < 0.01, ∗∗∗p < 0.001.

3.2. Hydrogel characterization

To confirm the successful incorporation of TG and SN into the hydrogel, a series of characterization techniques were employed. SEM imaging revealed that the GelS-TG-SN hydrogels exhibited a porous microstructure with uniformly dispersed granular material (Fig. 2A). FTIR analysis displayed a characteristic peak at 1651 cm⁻¹, which corresponds to Schiff base bonds in the base GelS hydrogel (Fig. 2C). Additional peaks observed at 1541 cm⁻¹ and 1551 cm⁻¹ were assigned to primary amide bonds (in GelS and GelS-SN) and secondary amide bonds (in GelS-TG and GelS-TG-SN), respectively. Moreover, a distinct peak at 646 cm⁻¹ was identified as the Mg–O vibration originating from SN (Fig. 2C), which is consistent with previous literature reports [23]. Elemental analysis further confirmed the incorporation, showing elevated levels of Mg and Si in both GelS-TG-SN and GelS-SN hydrogels compared with GelS-TG and GelS hydrogels (Fig. 2D–E, S2A-D). The spectral features closely resembled those of pure SN (Fig. 2D–E, S2A-D), providing additional evidence of successful integration. Collectively, these results demonstrated the effective incorporation of both TG and SN into the GelS hydrogel system.

• 1)Gelation time

The gelation time of GelS varied with the presence and concentration of TG. A notably short gelation time (<5 min, Fig. S2G) was observed for the 5% GelS group. The crosslinking rate correlated positively with the concentrations of both OS and TG. Increasing the OS concentration from 1.25% to 5.00% reduced the gelation time by 58% (Fig. S2G). The incorporation of TG (GelS-TG) shortened the crosslinking time to 118 ± 23 s, and the further addition of SN (GelS-TG-SN) accelerated gel formation to 60 ± 6.5 s, enabling in situ gelation within osteochondral defects (Fig. 2F–G, Fig. S2G–I). Rheological measurements showed a crossover between G′ and G″ within 1 min, consistent with the approximately 1-min gelation time determined by vial inversion tests (Fig. 2H and I). Both TG and SN contributed to the reduction in gelation time (Fig. 2G), with TG having a more pronounced effect. This is attributed to the role of TG as a highly efficient biological catalyst that rapidly forms robust covalent isopeptide bonds between gelatin molecules, dramatically increasing crosslinking density and accelerating network formation.

• 2)Injectability

Injectability is a crucial property of adhesive hydrogels for minimally invasive procedures such as arthroscopy. Following the mixture of equal volumes of gelatin and an oxidized starch precursor solution (containing 2.5% TG and 3% SN), the resulting solution was smoothly extruded from a syringe. Upon deposition into a culture dish or glass bottle, it formed a stable gel that remained intact without dissolving (Fig. 2F–S2F). Moreover, when injected into cartilage defects in SD rats, the GelS-TG-SN hydrogel precursor underwent successful in situ gelation, conformed to the defect shape, integrated well with the surrounding tissue, and exhibited strong wet adhesion even after repeated rinsing or immersion in aqueous medium (Fig. S2F and Supplement Video 1).

• 3)Mechanical reinforcement

Given the persistent intra-articular pressure of approximately 10 kPa, hydrogel strength must be sufficient to withstand such hydraulic stress required for further surgical procedures after gelling in situ following injection in a low aqueous environment [24]. Although both TG and SN contributed to enhancing the final gel strength, SN played a more dominant role in mechanical reinforcement (Fig. 2J and K). Mechanistically, SN functions as a rigid nanoplatelet filler dispersed within the polymer network, providing substantial mechanical enhancement through physical interactions. This resulted in a marked improvement in compressive modulus and overall strength. Specifically, the incorporation of SN (GelS-SN) increased the compressive modulus to 9666 ± 1527 Pa, while the further addition of TG in GelS-TG-SN elevated the modulus to 11033 ± 1001 Pa in approximately 5 min (Fig. 2J–S2E and Supplementary Video 2-3). Stress–strain curves indicated that both GelS-TG-SN and GelS-SN exhibited superior elastic behavior compared to the GelS and GelS-TG groups. Notably, the GelS-TG-SN hydrogel maintained structural integrity without failure even under a stress of 10 kPa (Fig. 2K–M). The greater reinforcing effect of SN over TG is attributed to the limited availability of amino groups in gelatin. Although TG promotes gelation via the formation of secondary amide bonds, it does not significantly increase crosslinking density.

• 4)Self-healing capability and ion release

The self-healing capability of the hydrogel composite was further evaluated. A crescent-shaped adhesive hydrogel sample was bisected into two segments, one of which was stained red. The two pieces were brought into contact and incubated at 24 °C for 20 min, after which they successfully fused into a continuous, unified structure (Fig. S2J). The healed hydrogel retained structural integrity and resisted external mechanical forces (Fig. S2K). These results indicate that the GelS-TG-SN hydrogel exhibits rapid and efficient self-healing properties, attributable to the presence of dynamic covalent Schiff base bonds and reversible hydrogen bonding [25]. Compared to traditionally crosslinked hydrogels, self-healing hydrogels can endure prolonged service life in high-pressure aqueous environments such as during arthroscopy, while also providing an adaptive network structure conducive to cell proliferation and controlled gel degradation.

Furthermore, GelS-TG-SN exhibited an extended degradation period of 42 days, accompanied by sustained release of magnesium and silicon ions. In contrast, the other groups degraded more rapidly, within 7 days (GelS), 21 days (GelS-SN), and 35 days (GelS-TG), respectively (Fig. 2N–Q).

3.3. In situ injection of hydrogel under arthroscopy

Although the hydrogel possesses the properties of rapid gelation and high mechanical strength, the presence of hydraulic pressure during arthroscopy makes its ability to adhere rapidly in situ equally critical [10]. Furthermore, the hydrogel must withstand irrigation water pressure of 10 kPa after gelation without being washed away [10,26]. Therefore, we first investigated its underwater adhesive properties and further simulated the arthroscopic environment to validate the feasibility of in situ gelation and resistance to hydraulic pressure during underwater injection.

• 1)
  Underwater adhesion and stability of GelS-TG-SN hydrogel

  • i)
    Underwater adhesive properties of hydrogels to cartilage and bone.
    To evaluate the adhesive properties of hydrogels to cartilage, an ex vivo rabbit chondral defect model was employed. The GelS-TG-SN hydrogel exhibited excellent injectability which rapidly gelled within 1 min and maintained strong adhesion even when the inverted defect model was being shaking underwater (Fig. 3A and Supplementary Videos 4). In contrast, the unmodified GelS hydrogel showed insufficient defect-filling capability (Fig. 3A). Then, the chondral defect sealed with GelS-TG-SN hydrogel was subjected to irrigation — using a connection to an adjustable hydraulic water gun and the hydraulic pressure was recorded by a hydraulic gauge (Fig. 3B and Supplementary Videos 5). In comparison, the unmodified GelS hydrogel was easily dislodged under hydraulic flow (Fig. 3B,yellow arrow). The enhanced performance of GelS-TG-SN is attributed to the synergistic effects of TG and SN (Fig. 3C): TG facilitates rapid formation of isopeptide bonds with collagen fibers in cartilage via enzymatic crosslinking as well as Schiff base linkages, while SN contributes through electrostatic interactions and hydrogen bonding with the collagen network [27]. Furthermore, SEM observation revealed a tight interface between the hydrogel and the cartilage (Fig. 3C and Fig. S3A–B). Together, these mechanisms ensure rapid fixation, mechanical stability, and strong tissue adhesion, meeting the critical demands of arthroscopic applications.
    To further assess bone adhesion, ex vivo murine femoral fracture and calvarial defect models were employed. The hydrogel also exhibited reliable underwater adhesion in both models (Fig. 3D and Supplementary Video 6). The adhesion mechanism is likely similar to that observed in cartilage, involving multifunctional physical and chemical interactions.
  • ii)
    Quantitative evaluation of hydrogel adhesive
    For quantifying the underwater tissue adhesion capability of the hydrogel, lap-shear testing was performed using hydrogel samples adhered to porcine intestinal membrane which was mounted on glass slides (Fig. 3E and F). The adhesive strength of the GelS-TG-SN group (22.54 ± 2.65 kPa) was significantly higher than those of the other groups (GelS: 3.13 ± 1.15 kPa; GelS-TG: 13.61 ± 1.56 kPa; GelS-SN: 8.33 ± 1.53 kPa), confirming its superior underwater adhesive performance compared to conventional hydrogels [26].

• 2)Application of GelS-TG-SN Hydrogels in a Simulated Arthroscopic Setting

Fig. 3.

Evaluation of hydrogel adhesion of bone and cartilage under wet conditions. (A–B) Injection and adhesion assessment of the hydrogel in an ex vivo rabbit osteochondral defect model under water pressure. (C) SEM image showing the interface between the hydrogel and cartilage tissue. Scale bar: 40 μm. (D) Adhesion performance of the hydrogel in ex vivo murine femoral fracture and calvarial defect models under aqueous conditions. (E–F) Lap-shear tests evaluating the underwater adhesive strength of different hydrogels. (G–H) Simulation of arthroscopic surgery in an ex vivo porcine knee joint: (G) injection and adhesion of the hydrogel for osteochondral repair under water; (H) validation of the hydrogel's resistance to water flow after arthrotomy. ∗p < 0.05, ∗∗∗p < 0.001.

To better simulate arthroscopic conditions, an ex vivo porcine knee joint model was employed to mimic actual arthroscopic surgery. Specifically, an endoscope was used to simulate the arthroscope, and a water bag was suspended 1 m above the operative plane to generate a hydrostatic pressure equivalent to 10 kPa (Fig. 3G and H). The hydrogel retained its injectability through a syringe, along with structural stability and adhesive performance under aqueous conditions (Fig. 3G and H and Supplementary Video 7 and 8). Furthermore, to closely mimic postoperative irrigation, the hydrogel was subjected to high-pressure flushing (the hydraulic pressure was recorded by a hydraulic gauge of an adjustable hydraulic water gun). During the experiment it remained firmly adhered without dislodgement or structural failure (Fig. 3G and H).

3.4. Biocompatibility and anti-inflammatory evaluation of the hydrogel

• 1)Cell viability and proliferation assays

SDSCs were selected for this study due to their distinct practical and biological advantages. Based on our previous research, SDSCs exhibit superior chondrogenic potential compared to other mesenchymal stem cell sources—such as those from adipose tissue or umbilical cord—making them particularly suitable for promoting cartilage regeneration in osteochondral defects [20]. Moreover, SDSCs can be readily harvested during routine arthroscopic procedures, ensuring minimal invasiveness and enhancing clinical feasibility [28,29].

To determine the optimal concentration of SN, CCK-8 assays revealed that cell viability of SDSCs at a density of 100 × 10⁶ cells/mL was compromised when SN concentrations exceeded 3% (Fig. 4A). However, at suitable concentrations (2.5% TG and 3% SN), the incorporation of SN did not increase the cytotoxicity of the hydrogel. After 7 days of co-culture with SDSCs, cells encapsulated within GelS-based hydrogels maintained high viability. Consistent with live/dead staining results, no significant difference in cells viability was observed among the different hydrogel groups (Fig. 4B and D). Cytoskeletal staining further showed that SDSCs within the hydrogel matrices displayed well-extended morphology (Fig. 4C and E). Additionally, scratch wound assays demonstrated that extracts of the hydrogels significantly enhanced cells migration rates, suggesting that released bioactive ions contribute to promoting stem cell motility (Fig. 4S).

• 2)SN exerts anti-inflammatory effects

Fig. 4.

Biocompatibility and macrophage polarization evaluation of the hydrogel. (A) CCK-8 assay was used to determine the optimal concentration of SN. (B) Live/dead staining of stem cells. Scale bar: 200 μm. (C) Cytoskeletal staining of stem cells. Scale bar: 20 μm. (D) Quantification of cell viability. (E) Quantification of cell spreading and morphology. (F) Immunofluorescence staining of F4/80, iNOS, and CD206 in RAW264.7 cells co-cultured with different hydrogel groups following LPS stimulation. Scale bar: 40 μm. (G-H) qPCR analysis of mRNA expression levels of CD206 and iNOS. ∗∗p < 0.05.

Patients with osteochondral defects often exhibit significant inflammatory responses within the joint cavity, which substantially impedes the repair process [30]. Therefore, mitigating inflammation is crucial to establish a conducive microenvironment for stem cell differentiation and successful osteochondral regeneration.

After establishing an LPS-stimulated RAW264.7 inflammatory model, co-culture experiments with the hydrogels revealed that both GelS-TG-SN and GelS-SN significantly increased the expression of the M2 marker CD206 while reducing the proportion of iNOS⁺ M1 macrophages. Immunofluorescence staining clearly demonstrated this phenotypic shift (Fig. 4F). Quantitative PCR further confirmed the downregulation of the pro-inflammatory gene iNOS in the GelS-TG-SN (0.67 ± 0.15) and GelS-SN (0.60 ± 0.20) groups compared to the GelS (3.17 ± 0.84) and GelS-TG (3.73 ± 0.71) groups. Concurrently, the anti-inflammatory gene CD206 upregulated in the GelS-TG-SN (5.17 ± 0.35) and GelS-SN (5.97 ± 0.42) groups relative to the GelS (0.43 ± 0.15) and GelS-TG (0.63 ± 0.15) groups (Fig. 4G and H). These results indicate that SN plays a key role in macrophage phenotypic reprogramming, underpinning its anti-inflammatory functionality.

3.5. SN-enhanced chondrogenic differentiation of SDSCs

Furthermore, beyond its anti-inflammatory effects, SN has been reported to synergistically enhance the chondrogenic differentiation of stem cells during early stages [17,31]. To evaluate this, SDSCs were incorporated into the hydrogel. Under three-dimensional (3D) culture conditions, Alcian blue staining revealed markedly higher glycosaminoglycan (GAG) deposition in the SN-containing group compared to the control on days 7 and 14, with uniform staining distribution observed throughout the gel construct (Fig. 5A and D). A consistent trend was observed in two-dimensional (2D) culture, where the supernatant from the hydrogel was co-cultured with SDSCs. The intensity of extracellular matrix blue staining was notably stronger in the SN group at both 7 and 14 days (Fig. 5B and E). Immunofluorescence staining further elucidated the underlying mechanism: after 14 days of co-culture, the SN group exhibited higher expression of SOX9 and stronger COL II signals compared to groups without SN (Fig. 5C–F, and G).

Fig. 5.

SN promotes chondrogenic differentiation of SDSC. (A) Three-dimensional Alcian blue staining of stem cell-laden hydrogels containing lithium silicate nanoplatelets (SN) after 7 and 14 days of culture. (B) Two-dimensional Alcian blue staining of stem cells co-cultured with hydrogels for 7 and 14 days. (C) Immunofluorescence staining of collagen type II (Col II) and SOX9 expression in stem cell-laden hydrogels with SN after 14 days. Scale bar: 40 μm. (D) Quantitative analysis of stem cell-laden hydrogels containing lithium silicate nanoplatelets (SN) after 7 and 14 days of culture. (E) Quantification of stem cells co-cultured with hydrogels for 7 and 14 days. (F, G) Quantitative analysis of collagen type II (Col II) and SOX9 expression in stem cell-laden hydrogels with SN after 14 days. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

3.6. SN-enhanced endochondral ossification of SDSCs in a vascularized microenvironment

In rat osteochondral defect model, immunofluorescence staining revealed spatially distinct microenvironments between the subchondral bone and cartilage regions. Specifically, high expression of EMCN and VEGFA was detected in the subchondral bone (Fig. 6A–C). Growing evidence suggests that vascularization plays a pivotal role in driving endochondral ossification [17,18]. To examine the effect of SN on stem cell differentiation under vascularized conditions resembling the subchondral bone niche, a pro-angiogenic environment was simulated by supplementing VEGFA (50 ng/mL) during a 14-day co-culture period with SDSC-laden hydrogels (GelS-TG-SN and GelS-TG) (Fig. 6D).

Fig. 6.

SN promotes chondrogenic differentiation of SDSC in vascularized environment. (A-C) Fluorescence staining validating the pro-angiogenic microenvironment in the subchondral region following osteochondral defect in rat model. C: Cartilage; SCB: Subchondral bone; OC defect: Osteochondral defect. Scale bars: 200 μm. (D) Schematic diagram of the experimental design. (E) Relative mRNA expression of COL X and MMP13 in Gels/TG and Gels/TG-SN groups after 14 days of culture in VEGF-containing medium. (F–G) RNA sequencing analysis of Gels/TG-SN under VEGF-supplemented and VEGF-free conditions: (F) heatmap of differentially expressed genes, (G) volcano plot. (H) GO enrichment analysis of differentially expressed genes. ∗∗∗p < 0.001.

The results showed that VEGFA stimulation markedly enhanced the expression of COL X, a hypertrophic chondrocyte marker, and significantly increased MMP13 levels in the GelS-TG-SN group compared with GelS-TG-SN without VEGFA (Fig. 6E). In contrast, although VEGFA also existed, the GelS-TG group without SN exhibited no significant upregulation of COL X or MMP13 compared to the SN-containing hydrogel. These findings indicate that both VEGFA and SN are required for the robust induction of COL X and MMP13 expression (Fig. 6E). Elevated expression of COL X and MMP13 signifies progression toward endochondral ossification, further supporting that GelS-TG-SN promotes hypertrophic differentiation in a vascularized subchondral bone-like microenvironment. Consistent with our above findings, the presence of SN in hydrogels without VEGFA led to upregulated expression of COL II and SOX9. In contrast, the addition of VEGFA markedly downregulated the expression of both markers (Fig. 5 and Fig. S5).

To further explore the molecular mechanisms, RNA sequencing was performed to compare the transcriptional profiles of SDSCs within GelS-TG-SN under pro-angiogenic versus control conditions. Heatmap and Volcano plot analyses revealed substantial differences in gene expression (Fig. 6F and G). Specifically, VEGFA exposure under chondrogenic conditions did not promote stable cartilage differentiation; rather, it induced a shift toward a hypertrophic cartilage program. This transition was marked by the upregulation of hypertrophy-related genes such as COL X, coupled with the downregulation of the chondrogenic transcription factor SOX9 (Fig. 6F). Concurrently, the volcano plot demonstrated a clear shift toward upregulated osteogenic and hypertrophic genes under VEGFA stimulation, prominently including RUNX2, a master regulator of osteogenesis (Fig. 6G). GO enrichment analysis indicated that SN selectively enhanced the PI3K–Akt signaling pathway under pro-angiogenic conditions (Fig. 6H).

In summary, within a vascularized subchondral bone–like microenvironment, SDSCs encapsulated in GelS-TG-SN undergo hypertrophic differentiation potentially mediated through activation of the PI3K–Akt signaling pathway.

3.7. GelS-TG-SN enhances OC defect repair in rats

The rat articular cavity features distinct microenvironments: avascular cartilage layer and a highly vascularized subchondral bone region [32,33]. While SN are known for their bidirectional differentiation potential, it is unknown if this is retained across such different microenvironments within a defect. To assess the OC repair capacity of our hydrogel, we created a critical-sized defect (1.5 mm diameter) in the rat femoral condyle⁴⁰ and performed an in situ injection of the hydrogel loaded with SDSCs at 100 × 10⁶ cells/mL [34].

• 1)GelS-TG-SN promotes articular cartilage regeneration in a rat OC defect

Cartilage repair was first evaluated 8 weeks after hydrogel injection (Fig. 7A, red rectangle). By 8 weeks post-operation, defects in the GelS-TG-SN group were filled with neocartilage-like tissue, whereas other groups primarily exhibited fibrous, whitish tissue (Fig. 7B). The GelS-TG-SN group showed the best cartilage repair outcomes at 8 weeks, as indicated by higher ICRS and Modified O'Driscoll scores compared to other SDSC-loaded hydrogels (Fig. 7C–E). Safranin O/Fast Green and Alcian blue staining further revealed that GelS-TG-SN increased the regenerated tissue area by approximately 4.5-fold compared to GelS and 3.8-fold compared to GelS-TG. The GelS-SN hydrogel also promoted regeneration, demonstrating a 2.2-fold increase over GelS and a 1.7-fold increase over GelS-TG (Fig. 7F and G). The original defect boundary, marked by yellow triangle, remained visible in some samples (Fig. 7D–F and Fig. S8). Transverse and sagittal MRI views confirmed superior cartilage repair in the GelS-TG-SN group, supported by the lowest WORMS scores (Fig. 7H and I).

Fig. 7.

GelS-TG-SN promotes articular cartilage regeneration. (A) Schematic illustration of hydrogel injection for OC defect repair in a rat model. C = cartilage; S = subchondral bone. Red rectangle: cartilage defect region in OC defect. (B) Gross appearance of the repaired defects at 8 weeks post-operation. Scale bars in first row: 2 mm; Scale bars in second row: 500 μm (C) ICRS scoring of cartilage repair. (D–G) Histological staining with Safranin O/Fast Green and Alcian Blue and quantification of MODS scoring and Repair area. Yellow triangle: Original defect boundary of OC defect. Scale bars in first row: 500 μm; Scale bars in second row: 200 μm. (H–I) Coronal and sagittal MRI views of femoral condyles and corresponding WORMS cartilage scores. Scale bars: 1.5 mm ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Importantly, no statistically significant differences were observed between the GelS-TG-SN and Sham groups across these evaluation parameters, suggesting that GelS-TG-SN treatment effectively restored cartilage repair to a level comparable with native tissue.

• 2)GelS-TG-SN promotes endochondral ossification in subchondral bone in a rat OC defect

Furthermore, we conducted histological staining and micro-CT analysis at 2, 4, and 8 weeks after hydrogel implantation into OC defects to evaluate whether SN can promote endochondral ossification of the SDSC-laden hydrogel within the vascularized microenvironment of the subchondral bone (Fig. 8A, red rectangle). Micro-CT analysis showed a significant increase in BV/TV in the GelS-TG-SN group (4 weeks: 19.25 ± 1.48%; 8 weeks: 35.00 ± 6.56%)—values significantly higher than those in the GelS and GelS-TG groups (Fig. 8B–G). A similar increasing trend was observed in Tb.N. Notably, no statistically significant differences were detected between the GelS-TG-SN and Sham groups at 8 weeks in either BV/TV or Tb.N, suggesting that GelS-TG-SN treatment effectively restored subchondral bone architecture to a level comparable to that of native bone.

Fig. 8.

GelS-TG-SN promotes endochondral ossification and subchondral bone regeneration in osteochondral defects. (A) Schematic diagram of hydrogel implantation in a rat osteochondral defect model. (B–G) Micro-CT analysis and quantitative results of BV/TV and Tb.N in the four different groups at 4 and 8 weeks. Scale bars: 2 mm. (H–I) Safranin O/Fast Green and Alcian blue staining of the subchondral region (highlighted by red rectangles) at 2-, 4-, and 8-weeks post-implantation. Yellow triangle: Hypertrophic cartilage. Purple triangle: Vascular invasion. Scale bar in middle column: 500 μm. Scale bars in right column: 250 μm ∗p < 0.05.

Histological evaluation (Safranin O/Fast Green and Alcian blue staining) revealed that GelS-TG-SN hydrogels accelerated chondrogenic differentiation of stem cells by 2 weeks, which was accompanied by the appearance of hypertrophic chondrocytes (Fig. 8H and I, yellow triangle). By 4 weeks, chondrogenic cells within the subchondral defect underwent progressive ossification accompanied by synchronized vascular invasion (Fig. 8H and I, purple triangle, Fig. S6–9). At 8 weeks, complete ossification of the subchondral bone was achieved, while the overlying cartilage layer maintained a stable chondrogenic phenotype and extracellular matrix composition. This spatiotemporally heterogeneous regeneration pattern corresponds to the distinct microenvironments at the cartilage–bone interface. In contrast, the GelS and GelS-TG groups exhibited incomplete subchondral bone repair across all time points (Fig. 8H–I, Fig. S7-9).

3.8. In vivo cell fate tracking, mechanistic validation, biosafety, and functional assessment

To directly verify the contribution of implanted stem cells to tissue regeneration, immunofluorescence staining for human nuclear antigen (HNA) was performed to trace the fate of transplanted human SDSCs in vivo. At 8 weeks post-implantation, distinct HNA-positive signals were detected within the defect region in the GelS-TG-SN group, confirming the long-term survival, engraftment, and active participation of implanted SDSCs in the repair process (Fig. S10). In contrast, minimal or absent signals were observed in other groups.

Furthermore, comparative analysis between SDSC-loaded and cell-free hydrogels revealed markedly inferior cartilage repair in the absence of exogenous cells, as demonstrated by gross observation, histological staining, MRI evaluation, and ICRS scoring (Fig. S10). These findings provide compelling evidence that SDSCs represent the primary functional stem cell population driving osteochondral regeneration within the hydrogel system.

To further validate the proposed molecular mechanism underlying dual-lineage differentiation in vivo, immunohistochemical analysis was performed on subchondral bone sections at 2 weeks post-implantation. The GelS-TG-SN group exhibited significantly elevated expression of phosphorylated PI3K (p-PI3K) and phosphorylated Akt (p-Akt) compared with other groups. Notably, the upregulation of these signaling markers was accompanied by increased expression of the hypertrophic cartilage marker COL X (Fig. S11). The coordinated upregulation of p-PI3K, p-Akt, and COL X provides in vivo evidence that the PI3K–Akt signaling pathway is functionally activated during the osteogenic phase of repair, promoting endochondral ossification within the vascularized subchondral microenvironment. These findings are consistent with the RNA sequencing results and in vitro mechanistic analyses.

Systemic biosafety was assessed by harvesting major organs—including heart, liver, spleen, lung, and kidney—at 8 weeks post-implantation, followed by H&E staining. No observable histopathological abnormalities, inflammatory infiltration, or tissue damage were detected in any experimental group (Fig. S12), indicating favorable medium-term biocompatibility of the GelS-based hydrogels. While these results demonstrate satisfactory safety within the 8-week observation period, long-term biosafety evaluation remains necessary for future translational studies.

Beyond morphological and structural restoration, functional recovery was evaluated through gait analysis at 8 weeks post-operation. Rats treated with the GelS-TG-SN hydrogel exhibited gait parameters comparable to those of the Sham group, whereas other treatment groups displayed varying degrees of functional impairment (Fig. S13). These results further demonstrate that the nanocomposite hydrogel not only promotes structural regeneration of osteochondral tissue but also restores joint function to a level approaching that of native tissue.

4. Discussion

OC defects, commonly resulting from traumatic injury or osteoarthritis [35]. Arthroscopic repair of OC defects remains particularly difficult, as the continuous fluid irrigation required for clear visualization during surgery creates a high-pressure aqueous environment that complicates the use of conventional injectable hydrogels. Conventional hydrogels are often limited by slow gelation kinetics, which impedes in situ formation and strong tissue adhesion, leading to premature displacement under irrigation [30]. Thus, there is an urgent need for advanced biomaterials that enable true osteochondral regeneration under minimally invasive, aqueous surgical conditions. In this study, we developed an enzymatically crosslinked nanocomposite hydrogel, GelS-TG-SN, specifically designed to address the critical limitations of existing injectable hydrogels for arthroscopic application. Our results demonstrate that this injectable hydrogel combines rapid, underwater gelation and robust adhesion with potent dual-lineage bioactivity guided by the natural osteochondral microenvironment, effectively promoting integrated osteochondral regeneration in a rat model.

A key hurdle for hydrogel-based arthroscopic therapy is achieving rapid gelation within the dynamic, fluid-filled joint space to prevent washout and ensure precise defect filling [10]. In this study, the hydrogel formed within 1 min—significantly faster than enzyme-free crosslinked hydrogels and previously reported self-crosslinking hydrogel systems [36]. Although this gelation time is slightly longer than that of some photo-crosslinked hydrogels, the mild enzymatic crosslinking conditions used in our study eliminate the need for additional equipment, making it more suitable for current arthroscopic surgical setups. Moreover, under simulated arthroscopic high-pressure irrigation conditions, the hydrogel exhibited favorable in situ gelation and resistance to washout. This may be attributed to the dual role of the enzyme, which not only promotes rapid collagen crosslinking within the hydrogel but also enhances binding between the hydrogel and native collagen in bone and cartilage, facilitating rapid gelation and strong adhesion to osteochondral defects. Additionally, Schiff base linkages, in combination with enzymatically formed anchoring structures, further increased the crosslinking density. Given the necessary interval between hydrogel pre-mixing and in situ injection, the 1-min gelation time also provides a practical window for surgical manipulation.

Beyond its superior physical properties, the GelS-TG-SN hydrogel demonstrates compelling bioactive functionality essential for osteochondral repair, may largely mediated by the sustained release of bioactive ions (Mg²⁺, Si⁴⁺) from SN. Mg²⁺ could play a central role in modulating cellular behavior, exhibiting significant immunomodulatory effects by promoting macrophage polarization toward the pro-regenerative M2 phenotype—evidenced by elevated CD206 and reduced iNOS expression—thereby establishing an anti-inflammatory microenvironment conducive to healing [37]. This observation is consistent with previous studies establishing the critical role of anti-inflammatory processes in successful osteochondral repair [38].

Furthermore, within this anti-inflammatory milieu, SN synergistically enhances the chondrogenic differentiation of stem cells. Consistent with previous studies, SN promotes chondrogenesis in the early phase and facilitates endochondral ossification at later stages—a process critically dependent on vascular invasion [17,31,39,40]. Accordingly, SN orchestrates a microenvironment-dependent regenerative response: in the hypoxic environment of the articular cartilage layer, it promotes and maintains stable chondrogenesis. Articular cartilage effectively prevents vascular ingrowth under hypoxic conditions (with oxygen tension as low as ≈5–7%, compared to 13–17% in blood vessels) through HIF-1α-mediated signaling, secretion of anti-angiogenic factors, and physical barrier mechanisms—consistent with our experimental observations showing no vascular invasion into the cartilage layer [32,33,41]. In contrast, within the pro-angiogenic environment of the subchondral region, SN stimulates chondrocyte hypertrophy, which ultimately drives the osteogenic differentiation of stem cells [39,42]. This sequential process was corroborated by immunohistochemical staining, which revealed that over time, the subchondral region in the GelS-TG-SN group underwent progressive remodeling: initial cartilaginous tissue was gradually invaded by blood vessels, followed by chondrocyte hypertrophy and eventual endochondral ossification.

This compartment-specific regeneration is attributed to the ability of SN to selectively enhance the PI3K-Akt signaling pathway under VEGFA stimulation, leading to upregulation of hypertrophic markers (COL X, MMP13) and ultimately facilitating the endochondral ossification process necessary for bone formation [43]. Although the GelS-SN group exhibited osteogenic effects comparable to those of GelS-TG-SN, its chondrogenic performance was suboptimal. This can be explained by the absence of TG, which resulted in inferior adhesive strength of the hydrogel. Consequently, the hydrogel failed to fully seal the defect area and maintain the hypoxic microenvironment essential for chondrogenesis, thereby compromising cartilage-layer formation.

Moreover, the SDSCs for this study was based on their distinct practical and biological advantages. SDSCs can be readily harvested during routine arthroscopic procedures, ensuring minimal invasiveness and clinical feasibility [28,44]. More importantly, according to our pervious study, they exhibit superior chondrogenic potential compared to other mesenchymal stem cell sources, such as those derived from adipose tissue or umbilical cord, making them particularly suitable for promoting cartilage regeneration in osteochondral defects [20].

The successful outcomes observed in this study—integrated repair with hyaline-like cartilage and well-vascularized subchondral bone—can thus be attributed to a strategic combination of factors: 1) Rapid, enzyme-mediated gelation kinetics suitable for aqueous arthroscopy; 2) Enhanced mechanical properties and unprecedented underwater adhesion provided by the nanocomposite structure; 3) The dual-lineage bioactivity of SN ions, directing stem cell fate towards chondrogenesis in the cartilage zone and osteogenesis via endochondral ossification in the bone zone; and 4) The creation of an immunomodulatory microenvironment that supports regeneration.

Nevertheless, this study has certain limitations. First, although we observed favorable repair outcomes and short-term biocompatibility within the 8-week observation period, the long-term biosafety of the material remains to be fully established. In particular, for the degradable nanosilicate components, their ultimate metabolic fate over extended periods in vivo, potential accumulation in major organs, and possible delayed immune responses still require systematic evaluation through animal studies lasting several months or longer. Such investigations should include monitoring of animal survival status, hematological indices, and histopathological analysis of vital organs such as the heart, liver, spleen, lungs, and kidneys. Future work will therefore focus on addressing these aspects to provide a comprehensive safety profile for the clinical translation of this material.

While the GelS-TG-SN hydrogel demonstrates promising attributes for arthroscopic osteochondral repair, several important challenges must be addressed to advance its clinical translation. First, the long-term stability of the regenerated tissue under full physiological loading remains to be established. Although our 8-week rat model showed excellent integration and matrix deposition, human joints endure complex, cyclical mechanical stresses over decades. Future studies employing large animal models subjected to controlled weight-bearing protocols over 6–12 months will be essential to evaluate functional durability and structural resilience. Second, the biocompatibility and clearance pathways of the degradation products warrant further investigation. While the SN and gelatin/starch components have shown favorable cytocompatibility in vitro and in vivo, the systemic fate and potential accumulation of silicate ions or polymer fragments over extended periods require detailed pharmacokinetic and toxicological assessment in relevant preclinical models. Third, scalability and manufacturing consistency present a critical translational hurdle. Reproducible synthesis of the nanocomposite—particularly ensuring uniform dispersion of silicate nanoplatelets and consistent enzymatic crosslinking activity—must be achieved at a clinically relevant scale. Developing robust quality control measures for sterility, gelation kinetics, and mechanical properties will be vital for regulatory approval. Finally, the practical delivery and handling of the hydrogel in a real surgical setting deserve attention. While the material displayed excellent injectability and rapid setting in our simulated arthroscopy setup, integration with standard arthroscopic instruments and optimization of delivery systems will enhance its usability and adoption by surgeons.

In summary, we have engineered a multifunctional nanocomposite hydrogel, GelS-TG-SN, that addresses the challenges associated with arthroscopic repair of osteochondral defects in terms of underwater gelation and osteochondral induction. Its rapid gelation, high underwater adhesive strength, and ability to synchronously promote cartilage and bone regeneration through ion-mediated dual bioactivity make it a promising candidate for minimally invasive integrated osteochondral repair. This study provides a biomaterial strategy that bridges the gap between demanding surgical environments and complex biological regeneration requirements (Fig. 9).

Fig. 9.

The hydrogel is engineered for arthroscopy, demonstrating rapid underwater gelation while also providing critical biological functions. It supports dual-lineage differentiation—promoting both chondrogenesis and osteogenesis for integrated osteochondral repair—by effectively leveraging the inherent vascular gradient within the defect environment.

Ethical statement

All animal procedures were approved by the Institutional Animal Care and Use Committee of Southern University of Science and Technology (No. SUSTech-JY202401103).

As previously reported, SDSCs were isolated from synovial tissues obtained during total knee arthroplasty (TKA) procedures which were approved by the ethics committee of Shenzhen People's Hospital (LL-XB-2024001-01) [20].

Authorship contribution

Q.L. and G.L. designed the research study and wrote the manuscript; Q.L., J.W. and Y.L. conducted the experiments, acquired and analyzed data; D.D. and Y.Y. executed histological staining; Y.F. analyzed mechanical test data; S.F., H.Z., J.L., C.Y., and Y.W., analyzed RNA-seq data analysis; D. W., C.L. supervised the project; G.L. coordinated the research framework and finalized the manuscript. All authors reviewed and approved the final version.

Data and materials availability

Any data associated with this study are available from the corresponding author upon reasonable request.

Declaration of generative AI in scientific writing

The artificial intelligence tool of ChatGPT, Gemini and DeepSeek were used only to improve readability and language.

Funding statement

This work was supported by grants from the National Natural Science Foundation of China (82172463), Shenzhen Medical Research Fund (A2402016), Shenzhen Medical Research Fund (C2501025), Shenzhen Science and Technology Innovation Commission (KQTD20200820113012029) and Guangdong Provincial Key Laboratory of Advanced Biomaterials (2022B1212010003), Science and Technology Projects in Guangzhou (2025A04J4792）.

Declaration of competing interest statement

The author(s) have no conflicts of interest relevant to this article.

Appendix A. Supplementary data

The following are the Supplementary data to this article: