Injectable thermosensitive hydrogel incorporating exosome-loaded chitosan microspheres for immunomodulation and cartilage regeneration

Abstract

Background

Chronic inflammation and impaired immune microenvironment are key barriers to effective cartilage regeneration. Herein, we developed an integrated, injectable thermosensitive hydrogel system termed CS@Gel-Mg. This system modularly combines exosome-loaded chitosan microspheres (CM@Exo)​ embedded within a Mg²⁺-coordinated poly(N-isopropylacrylamide)-based ionic liquid hydrogel (Gel-Mg), aiming to modulate local immune response and promote cartilage repair.

Methods

Hydroxypropyl-modified chitosan microspheres were synthesized via emulsion crosslinking and effectively loaded with bioactive exosomes. The Gel matrix was fabricated using an ionic liquid monomer (IL-C4) copolymerized with NIPAM and N-vinylimidazole, followed by Mg²⁺ complexation. Physicochemical properties, including morphology (SEM/TEM), surface chemistry (XPS/FTIR), and thermoresponsive behavior (DSC/TGA), were systematically characterized. The effects of CS@Gel-Mg on immune and cartilage regeneration were evaluated in vivo and in vitro using Western blot, qPCR, histological staining, flow cytometry, gait analysis, and asymmetric weight-bearing analysis.

Results

In vitro, the system demonstrated excellent cytocompatibility, promoted ATDC5 cell proliferation and migration, and inhibited apoptosis. The hydrogel also reprogrammed macrophage polarization from M1 to M2 phenotype, as confirmed by qPCR and flow cytometry. In vivo, using a rat full-thickness cartilage defect model, CS@Gel-Mg exhibited improved weight-bearing function, reduced inflammatory infiltration, and superior cartilage histological scores compared to control groups. Mechanistically, immunofluorescence and Western blot analyses suggested enhanced expression of SOX9, COL2A1 and COL1A1, along with suppression of inflammatory cytokines. Furthermore, SOX9 knockdown in vivo diminished the therapeutic effect, confirming its regulatory role in the immuno-chondrogenic axis.

Conclusion

This exosome-integrated thermoresponsive hydrogel provides a multifunctional platform for immune modulation and cartilage regeneration. Our findings highlight the potential of CS@Gel-Mg as a promising therapeutic strategy for osteochondral repair.

Graphical abstract

1. Introduction

Articular cartilage exhibits extremely limited intrinsic regenerative ability due to its avascularity, sparse chondrocytes, and dense extracellular matrix (ECM), which collectively hinder nutrient diffusion and inhibit repair after injury [[1], [2], [3], [4]]. Clinically, reparative strategies such as microfracture, autologous chondrocyte implantation (ACI), and osteochondral allografts have been widely employed; however, they suffer from issues including donor-site morbidity, immune rejection, fibrocartilage formation, and long-term degeneration [[5], [6], [7], [8]]. Consequently, developing injectable, biocompatible materials with in situ tissue repair potential remains a pressing and unmet need [9,10].

Thermosensitive hydrogels, notably those based on chitosan or poly(N-isopropylacrylamide) (PNIPAM), undergo sol–gel transition at body temperature for minimally invasive delivery and conform to irregular defect geometries [[11], [12], [13]]. For example, encapsulating exosomes within a hydrogel matrix has also proven to be a feasible approach to enhance cartilage repair [14,15]. However, the repair capacity of a single carrier system remains limited within the complex microenvironment of the defect. Notably, ionic crosslinking strategies, such as incorporation of divalent cations (e.g., Mg²⁺, Ca²⁺) can not only enhance the mechanical stability of the material, supporting the maintenance of chondrocyte phenotype, but also modulate the local immune microenvironment by promoting the polarization of macrophages from the pro-inflammatory M1 type toward the reparative M2 type, thereby contributing to improved regenerative outcomes [[16], [17], [18], [19]].

Exosomes, nanoscale extracellular vesicles secreted by mesenchymal stem cells (MSCs), contain a cargo of miRNAs, proteins, and growth factors capable of stimulating chondrocyte proliferation, reducing apoptosis, facilitating matrix synthesis, and modulating inflammation [[20], [21], [22]]. However, their rapid clearance and non-targeted distribution in vivo severely limit their therapeutic efficacy [23]. To improve the delivery efficacy, various carrier strategies have been developed. Among these, microsphere delivery systems enable controllable loading and release kinetics of bioactive agents. Furthermore, chitosan microspheres can provide structural reinforcement and sustained exosome release when embedded in hydrogels [24,25]. Composite systems combining exosome-loaded microspheres with thermoresponsive hydrogels have demonstrated precision delivery, tunable rheology, and multifunctionality [26].

Successful cartilage repair requires both modulation of the immune microenvironment and promotion of chondrogenesis. The balance between M1 and M2 macrophage phenotypes plays a critical role in this process. M1 macrophages exhibit a pro-inflammatory phenotype, releasing inflammatory mediators such as interleukin-1β (IL-1β), IL-6, and tumor necrosis factor-α (TNF-α), which initiate and amplify inflammatory responses. In contrast, M2 macrophages display an anti-inflammatory phenotype and secrete cytokines including IL-10, thereby promoting vascular remodeling and tissue healing [27]. Studies have shown that polarization of macrophages toward the M2 phenotype helps establish a local anti-inflammatory and pro-regenerative microenvironment in cartilage defects [28]. Therefore, an integrated strategy combining immunomodulatory and chondro-inductive functions is of considerable importance.

In this work, we introduce a multifunctional modular hydrogel composed of exosome-loaded hydroxypropyl chitosan microspheres embedded within a Mg²⁺-coordinated, thermoresponsive PNIPAM-based ionic liquid hydrogel. We investigate (Scheme 1):

• ①Physicochemical properties and thermosensitivity of the composite system;
• ②In vitro effects on chondrocyte viability, migration, uptake, apoptosis, and macrophage polarization;
• ③In vivo regenerative capacity in a rat full-thickness cartilage defect model, including functional recovery, histological evaluation, and expression analysis of SOX9, COL2A1, and COL1A1.

Scheme 1.

a) Schematic illustration of the fabrication of CS@Gel-Mg composite hydrogel. Hydroxypropyl chitosan (CSPO)-based microspheres (CM) were loaded with BMSC-derived exosomes and incorporated into a thermosensitive ionic liquid hydrogel coordinated with Mg²⁺ to form CS@Gel-Mg. b) Therapeutic mechanism of CS@Gel-Mg composite hydrogel. Upon intra-articular injection and gelation at physiological temperature, the system enabled sustained release of Mg²⁺ and exosomes. c) CS@Gel-Mg promoted SOX9 activation and upregulation of chondrogenic matrix proteins, thereby enhancing cartilage regeneration. Concurrently, the composite hydrogel promotes the polarization of macrophages from M1 to M2 phenotypes, alleviates inflammation, and creates a favorable environment for cartilage repair.

This integrated hydrogel platform offers controlled exosome delivery, localized immunomodulation, and mechanically supportive scaffolding. It presents a strong translational potential for cell-free cartilage tissue engineering, bridging material innovation with immune and molecular regenerative mechanisms.

2. Materials and methods

2.1. Materials

Chitosan (medium molecular weight), sodium hydroxide (NaOH), isopropanol, tetramethylammonium hydroxide (TMAOH), propylene oxide (PO), acetone, ethanol, hydrochloric acid (HCl), adipic acid (AA), N-hydroxysuccinimide (NHS), 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC·HCl), petroleum ether, Span 80, N-isopropylacrylamide (NIPAM), N-vinylimidazole (VIM), 1-bromobutane, azobisisobutyronitrile (AIBN), dimethyl sulfoxide (DMSO), magnesium sulfate (MgSO₄), and ethyl acetate were purchased from Aladdin (Shanghai, China) unless otherwise noted. All reagents were of analytical grade and used as received.

2.1.1. Synthesis of hydroxypropyl-modified chitosan (CSPO)

To improve water solubility and functionalizability, chitosan was hydrophilically modified. Briefly, 3.00 g of NaOH was dissolved in 6 mL of deionized water to prepare a 33 wt% NaOH solution. Then, 2.00 g of chitosan was added to this alkaline solution, mixed thoroughly, and refrigerated overnight. The alkaline chitosan mixture was thawed and transferred to a three-necked round-bottom flask containing 40 mL of isopropanol. After stirring at room temperature for 30 min, 1 mL of tetramethylammonium hydroxide was added dropwise, and stirring was continued for another 30 min. The mixture was gradually heated to 65 °C, followed by the slow addition of 20 mL of propylene oxide. The reaction was maintained under reflux with continuous stirring for 5 h. Afterward, the reaction solution was cooled to room temperature, and its pH was adjusted to 7.0 using hydrochloric acid. Acetone was added while stirring to precipitate the white solid, which was then filtered, washed with ethanol, and dissolved in deionized water. The solution was dialyzed (MWCO 7000 Da) against deionized water for 3 days (changing water 3–4 times per day), followed by freeze-drying to obtain CSPO.

2.1.2. Preparation of chitosan microspheres (CM)

CM were fabricated using a water-in-oil (W/O) emulsion crosslinking method. First, 0.12 g of CSPO was dissolved in 4 mL of deionized water. Cross-linking agents were added sequentially: adipic acid (15 mol%), NHS (16.5 mol%), and EDC·HCl (22.5 mol%), all calculated based on the molar amount of CSPO repeating units. The mixture was emulsified in petroleum ether containing Span 80, using a volume ratio of aqueous phase:emulsifier:oil = 1:1:20. The emulsion was stirred mechanically at 600 rpm at room temperature for 4 h. Upon completion, the reaction mixture was centrifuged and washed with absolute ethanol (3–5 times) to remove residual oil and surfactant. The final chitosan microspheres were collected by centrifugation and dried under vacuum.

2.1.3. Preparation of exosome-loaded chitosan microspheres (CM@Exo)

BMSCs were isolated from rat femurs and tibias, and cultured in exosome-free DMEM/F12 medium (GIBCO, USA) supplemented with 10% FBS (GIBCO, USA) and 1% penicillin–streptomycin (Gibco, USA) for 48 h, after which the conditioned medium was collected. The medium was then replaced, and cells were cultured for an additional 48 h before the medium was collected again. Exosomes were extracted from the above-mentioned medium using the ExoQuick kit (System Biosciences, USA). Following extraction, the exosomes were characterized by Western blot analysis for the presence of specific vesicle markers (CD63, CD81, TSG101) and the absence of common cellular contaminants (Calnexin, GM130) to confirm their identity and purity prior to loading.

For exosome incorporation, 0.12 g of CSPO was dissolved in 3 mL of deionized water, and 1 mL of exosome extract was added and mixed thoroughly. The emulsion and crosslinking procedures were identical to those described above for CM preparation. The resulting CM@Exo were centrifuged, washed, and dried.

2.1.4. Synthesis of ionic liquid IL-C4

The ionic liquid (IL-C4) was synthesized via alkylation of N-vinylimidazole. Equimolar amounts of VIM and 1-bromobutane were mixed and stirred at room temperature for 24 h. The crude product was then precipitated by dropwise addition into ethyl acetate under stirring. The precipitate was washed three times with ethyl acetate, collected by filtration, and dried using rotary evaporation for 8 h at room temperature. The resulting IL-C4 was stored in a dry desiccator until use.

2.1.5. Preparation of thermosensitive hydrogels (gel and Gel-Mg)

NIPAM (0.887 g), VIM (0.016 g), and IL-C4 (0.097 g) were weighed into an ampoule, followed by 0.2 wt% of the thermal initiator AIBN and 2 g of DMSO. The mixture was sonicated for 10 min to achieve full dissolution and dispersion, degassed under nitrogen for 10 min, sealed, and then polymerized at 70 °C for 6 h in an oil bath. The obtained polymer was washed three times with acetone to remove unreacted monomers and dried via rotary evaporation.

For hydrogel formation, 0.5 g of the obtained polymer was dissolved in 4.5 g of deionized water to form a 10 wt% gel precursor, denoted as Gel. For ionic complexation, 10 mg of MgSO₄ was dissolved in 4.5 g of deionized water and used to dissolve 0.5 g of the polymer, forming Gel-Mg after stirring for 24 h at room temperature to allow complete Mg²⁺ coordination.

2.1.6. Preparation of composite hydrogels (CS@Gel and CS@Gel-Mg)

To fabricate composite injectable hydrogels, 10 mg of CM@Exo were added to 1 g of Gel or Gel-Mg, respectively. Each mixture was stirred at room temperature for 24 h to ensure uniform dispersion of microspheres within the gel network. The resulting exosome-loaded composite hydrogels were denoted as CS@Gel and CS@Gel-Mg, respectively. A summary of the key materials and their abbreviations is provided below.

Abbreviation	Full Name/Description
CSPO	Hydroxypropyl-modified chitosan. A hydrophilically modified chitosan derivative with improved water solubility, used as the base material for microsphere fabrication.
CM	Chitosan microspheres. Spherical particles fabricated from CSPO via a water-in-oil emulsion crosslinking method.
CM@Exo	Exosome-loaded chitosan microspheres. CM that had encapsulated exosomes during their preparation.
Gel	Thermosensitive ionic liquid hydrogel. A poly(N-isopropylacrylamide)-based hydrogel synthesized with an ionic liquid (IL-C4), exhibiting temperature-sensitive sol-gel transition.
Gel-Mg	Mg2+-coordinated thermosensitive ionic liquid hydrogel. The Gel hydrogel after coordination with Mg2+ ions, forming a more stable network.
CS@Gel	Composite injectable hydrogel. Formed by uniformly dispersing exosome-loaded chitosan microspheres (CM@Exo) within the blank Gel matrix.
CS@Gel-Mg	Composite injectable hydrogel. Formed by uniformly dispersing exosome-loaded chitosan microspheres (CM@Exo) within the Mg2+-coordinated Gel (Gel-Mg) matrix.

2.1.7. Characterization of composite hydrogels

The morphology and microstructure of CM, CM@Exo, Gel and Gel-Mg were examined using Scanning Electron Microscopy (SEM; Phenom ProX, Thermo Fisher Scientific) and Transmission Electron Microscopy (TEM; JEOL JEM-2100, Japan). The average diameter of microspheres and the internal structure of the hydrogels were analyzed using ImageJ software.

Dynamic Light Scattering (DLS) and Zeta potential measurements were performed using a Zetasizer Nano ZS90 (Malvern Instruments) to determine particle size distribution and surface charge of the microspheres. The hydrodynamic diameter and surface potential were averaged over three independent batches.

Fourier-transform infrared spectroscopy (FTIR) was conducted on a Nicolet iS10 FTIR spectrometer (Thermo Fisher Scientific) to analyze chemical bonding and confirm successful chitosan modification and hydrogel network formation.

Thermogravimetric analysis (TGA) and Differential Scanning Calorimetry (DSC) were carried out on a TGA 550 and DSC Q2000 (TA Instruments, USA) to evaluate the thermal stability and phase transition properties of hydrogels, particularly the lower critical solution temperature (LCST). The encapsulation efficiency (EE) of exosomes within the chitosan microspheres was quantified by analyzing the characteristic mass loss differences between CM and CM@Exo in their TGA profiles, and this method provides an indirect estimate based on total mass loss.

X-ray Photoelectron Spectroscopy (XPS) was performed with an ESCALAB 250Xi (Thermo Fisher Scientific) to verify the elemental composition and the coordination interaction of Mg²⁺ with the polymer matrix.

Gelation behavior was evaluated by vial inversion and rheological analysis. Rheological properties including storage modulus (G′) and loss modulus (G″) were measured using a rheometer (TA Instruments, DHR-2) at 37 °C under oscillatory mode.

Inductively coupled plasma optical emission spectroscopy (ICP-OES) (Optima 8300, PerkinElmer, USA) was used to determined Mg²⁺ content and coordination efficiency in the Gel-Mg hydrogel. The total added Mg²⁺ was measured by digesting the precursor solution. The coordinated Mg²⁺ was determined after dialyzing the hydrogel to remove unbound ions, followed by lyophilization and digestion. Both digests were analyzed against Mg standards with Yttrium as an internal standard at 285 nm. Coordination efficiency was calculated as (M_loaded/M_total) × 100%.

2.2. Cell culture and treatments

Primary murine bone marrow-derived macrophages (BMDMs) were harvested from the femurs and tibias of 6–8-week-old C57BL/6 mice. Briefly, bone marrow cells were flushed using ice-cold PBS, filtered through a 70 μm cell strainer, and cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, and 20 ng/mL recombinant mouse M-CSF (PeproTech, Cat#315-02) for 7 days to induce M0 macrophages. ATDC5 cells were obtained from ATCC and cultured in DMEM (Gibco, Cat#11965092) supplemented with 10% FBS and 1% penicillin-streptomycin under standard conditions (37 °C, 5% CO₂).

2.3. Cell viability and apoptosis assays

ATDC5 cells were seeded in 96-well plates at 5 × 10³ cells/well and incubated with serial dilutions of CS@Gel or CS@Gel-Mg for 24, 48, and 72 h. Cell viability was evaluated using the Cell Counting Kit-8 (CCK-8, Dojindo, Cat#CK04) following the manufacturer's instructions. For apoptosis detection, cells were treated with various formulations for 24–72 h and stained with Annexin V-FITC/PI (Thermo Fisher Scientific, Cat#V13242), followed by analysis on a BD FACSCanto II flow cytometer.

2.4. Confocal microscopy for cellular uptake

To assess uptake of exosome-loaded microspheres, CS@Gel or CS@Gel-Mg were labeled with FITC or Rhodamine B prior to use. BMDMs were seeded on confocal dishes (NEST) at a density of 1 × 10⁴ cells/dish and incubated with labeled hydrogels for 6 or 24 h. Cells were then fixed, counterstained with DAPI (Thermo Fisher Scientific), and imaged using a Leica TCS SP8 confocal laser scanning microscope.

2.5. Flow cytometry for macrophage polarization

To evaluate immunomodulatory effects, BMDMs were seeded in 6-well plates and treated with CS@Gel, CS@Gel-Mg, or LPS/IFN-γ to induce M1 phenotype. After 48 h, cells were harvested and stained with anti-CD86-PE (BioLegend, Cat#105012) and anti-CD206-FITC (BioLegend, Cat#141704), incubated at 4 °C for 30 min, and analyzed using a BD flow cytometer. M1/M2 ratios were calculated based on the CD86/CD206 expression profiles.

2.6. Quantitative real-time PCR

Total RNA was extracted from macrophages using TRIzol reagent (Invitrogen, Cat#15596018). Reverse transcription was performed using the PrimeScript RT reagent kit (Takara), and qPCR was carried out using SYBR Premix Ex Taq II (Takara) on a QuantStudio 3 real-time PCR system. Primer sequences for IL-6, TNF-α, IL-1β and GAPDH were designed using Primer-BLAST. The 2^–ΔΔCt method was used to quantify relative mRNA expression levels.

Gene	Forward Primer (5′–3′)	Reverse Primer (5′–3′)
GAPDH	AAGGTGAAGGTCGGAGTCAAC	GGGTGGAATCATATTGGAACA
IL-6	ACTCACCTCTTCAGAACGAATTG	CCATCTTTGGAAGGTTCAGGTTG
TNF-α	CCCTCACACTCAGATCATCTTCT	GCTACGACGTGGGCTACAG
IL-1β	GCAACTGTTCCTGAACTCAACT	ATCTTTTGGGGTCCGTCAACT

2.7. In vivo rat cartilage defect model

All animal experiments were conducted in compliance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of Shanghai Tong Ren Hospital (Approval ID: A2024-030-01). Male Sprague–Dawley rats (8 weeks old, 220–250 g) were randomly divided into six groups: Ctrl, Model, Gel, Gel-Mg, CS@Gel, and CS@Gel-Mg (n = 5 per group). All rats were anesthetized with isoflurane. The rats in Ctrl group underwent surgical exposure of the right knee joint without creating a cartilage defect, whereas those in the other groups received full-thickness cartilage defects (2 mm in diameter and 1.5 mm in depth) surgically created in the femoral trochlea of the right knee. Intra-articular injections (50 μL) were administered once weekly for three consecutive weeks. At 8 weeks post-surgery, the animals were sacrificed, and the knee joints were harvested for subsequent histological and imaging analyses.

2.8. SOX9 knockdown and overexpression model in vivo

To assess the contribution of SOX9 in vivo, a subset of rats received intra-articular injections of lentiviral vectors encoding SOX9 shRNA (GeneChem, China) or SOX9 overexpression (pLVX-SOX9, Addgene) two weeks prior to treatment initiation. The efficiency of gene modulation was confirmed by qPCR and Western blot from harvested cartilage tissues.

2.9. Gait and behavioral analysis

At weeks 0, 4, and 8 post-surgery, weight-bearing asymmetry was assessed using an incapacitance tester (IITC Life Science), and gait parameters were recorded using CatWalk XT (Noldus) to evaluate functional recovery. Stride length and limb pressure distribution were quantified.

2.10. Micro-CT imaging

Femoral joints were scanned using a micro-CT scanner (Bruker SkyScan 1176, 9 μm resolution) to evaluate trabecular bone regeneration. Bone volume to tissue volume (BV/TV), trabecular thickness (Tb.Th), number (Tb.N), and separation (Tb.Sp) were analyzed using CTAn software.

2.11. Histological and immunohistochemical analysis

Tissue samples were fixed in 4% paraformaldehyde, decalcified, and paraffin-embedded. Sections were stained with H&E, Safranin O/Fast Green, and Toluidine Blue to assess cartilage matrix and morphology. Immunohistochemical staining was conducted using anti-SOX9 (Abcam, Cat#ab185966), COL2A1 (Abcam, Cat#ab34712), COL1A1 (Abcam, Cat#ab34710) anti-TNF-α (Abcam, Cat# ab6671), anti-IL-6 (Abcam, Cat# ab9324), and anti-IL-1β (Cell Signaling Technology, Cat#12703), followed by HRP-conjugated secondary antibodies and DAB chromogen development.

2.12. Immunofluorescence

For detection of M1/M2 polarization and chondrogenic markers, paraffin sections were deparaffinized and subjected to antigen retrieval. Primary antibodies included anti-iNOS (Abcam, Cat# ab15323), anti-CD206 (Thermo Fisher, Cat#PA5-46994). Sections were incubated with Alexa Fluor-conjugated secondary antibodies and DAPI, and visualized using a Zeiss LSM 710 confocal microscope.

2.13. Western blotting

Cartilage tissues were homogenized in RIPA buffer with protease/phosphatase inhibitors. Proteins were separated by SDS-PAGE, transferred to PVDF membranes, and probed with antibodies against SOX9, COL2A1, COL1A1, MMP-2, CD31, VEGF, α-SMA, Aggrecan, GAG, CD63, CD81, TSG101, Calnexin, GM130, and GAPDH. Bands were visualized with ECL detection (Thermo, Cat#32106) and quantified using ImageJ.

Target Protein	Antibody	Brand	Catalog Number	Host	Dilution
GAPDH	Anti-GAPDH	Cell Signaling Technology	#5174	Rabbit	1:5000
SOX9	Anti-SOX9	Abcam	ab185966	Rabbit	1:1000
COL2A1	Anti-Collagen II	Abcam	ab34712	Rabbit	1:1000
COL1A1	Anti-Collagen I	Proteintech	14695-1-AP	Rabbit	1:1000
MMP-2	Anti-MMP-2	Proteintech	10373-2-AP	Rabbit	1:1000
CD31	Anti-CD31	Proteintech	11265-1-AP	Rabbit	1:2000
VEGF	Anti-VEGF	Proteintech	81323-2-RR	Rabbit	1:5000
α-SMA	Anti-α-SMA	Proteintech	14395-1-AP	Rabbit	1:2000
Aggrecan	Anti-Aggrecan	Proteintech	84819-5-RR	Rabbit	1:1000
GAG	Anti-GAG	Abcam	ab100970	Rabbit	1:1000
CD63	Anti-CD63	Proteintech	25682-1-AP	Rabbit	1:1000
CD81	Anti-CD81	Proteintech	27855-1-AP	Rabbit	1:1000
TSG101	Anti-TSG101	Proteintech	28283-1-AP	Rabbit	1:1000
Calnexin	Anti-Calnexin	Proteintech	10427-2-AP	Rabbit	1:3000
GM130	Anti-GM130	Proteintech	11308-1-AP	Rabbit	1:5000

2.14. Statistical analysis

All quantitative data are presented as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism 10.0 software (GraphPad Software Inc., USA). For comparisons involving a single independent variable across multiple groups, one-way analysis of variance (ANOVA) followed by Tukey's post hoc test was used. For comparisons involving repeated measurements over time, a repeated-measures two-way ANOVA followed by Šidák's multiple comparisons test was employed. A p-value of <0.05 was considered statistically significant and marked with p < 0.05 (∗), p < 0.01 (∗∗) and p < 0.001 (∗∗∗).

3. Results

3.1. Preparation and characterization of exosomes, chitosan microspheres, and thermosensitive hydrogels

The morphology of BMSC-derived exosomes is shown in Figure S1A. Nanoparticle tracking analysis (NTA) revealed the size distribution of exosomes (Figure S1B). The identity and purity of the extracted exosomes were confirmed by Western blot analysis, which showed positive expression of exosome-specific markers (CD63, CD81, and TSG101) and absence of contaminant markers from major cellular compartments (Figure S1C). The chemical modifications and synthetic routes were shown in Figure S2A, CSPO was obtained from native chitosan through hydrophilic modification. TEM images (Figure S2B) revealed morphological differences between chitosan and CSPO, while ¹H NMR (Figure S2C) and FTIR spectra (Figure S2D) confirmed the successful introduction of hydroxypropyl groups. Acid–base titration further indicated that the amino content of CSPO reached 4.1 mmol/g (Figure S2E). In addition, the synthesis of injectable thermosensitive gels was verified. The reaction route and molar ratios are illustrated in Figure S3A, and structural characterization was confirmed by ¹H NMR (Figure S3B), ¹³C NMR (Figure S3C), and mass spectrometry (Figure S3D).

Based on these modifications, chitosan microspheres (CM) and exosome-loaded chitosan microspheres (CM@Exo) were subsequently fabricated using an emulsion crosslinking method with adipic acid (AA), NHS, and EDC-HCl as crosslinking/activating agents. Morphological characterization confirmed successful preparation: SEM (Fig. 1A) and TEM (Fig. 1B) images revealed uniform spherical microspheres (∼4 μm in diameter). The encapsulation efficiency was estimated to be 19.88 wt% based on TGA (Fig. 1C). Elemental mapping and XPS analysis revealed signals for both amine (-NH₂) and amide (-NH-CO-) groups, with a distinct increase in the amide signal, confirming the successful encapsulation of exosomes (Fig. 1D–F). Collectively, these results indicate that the microspheres retained structural stability and effectively encapsulated exosomes.

Fig. 1.

Characterization of CM and exosome-loaded chitosan microspheres CM@Exo. (A) SEM image of chitosan microspheres loaded with exosomes. (B) TEM image of chitosan microspheres loaded with exosomes. (C) Image showing the indirectly estimated mass fraction of exosomes in microspheres, determined by TGA. (D-F) XPS elemental analysis of exosomes in microspheres. Dynamic light scattering (DLS) (G) and Zeta potential (H) of CS@Gel and CS@Gel-Mg. (I) Thermosensitive gels and thermosensitive gels complexed with Mg²⁺, sol-gel transition images. (J) SEM images of thermal gels and thermal gels complexed with Mg²⁺. (K) Images of gel degradation UV curves of thermal gels and thermal gels complexed with Mg²⁺. (L) DSC images of thermal gels and thermal gels complexed with Mg²⁺. (M) TGA thermal decomposition temperature images of thermal gels and thermal gels complexed with Mg²⁺. (N) Fourier transform infrared spectral images of blank microspheres (CM), encapsulated exosome microspheres (CM@Exo), blank thermal gels (Gel), composite Mg²⁺ thermal gels (Gel-Mg), doped microsphere thermal gels (CS@Gel), and doped microsphere and composite Mg²⁺ thermal gels (CS@Gel-Mg).

3.2. Thermosensitive properties and degradation behavior of gels

Dynamic light scattering (DLS) analysis further demonstrated that the average hydrodynamic diameter of CS@Gel was larger than that of CS@Gel-Mg, indicating that Mg²⁺ coordination contributed to a more compact gel network (Fig. 1G). Consistently, Zeta potential measurements showed that Mg²⁺ incorporation increased the surface charge stability of the gels, which may contribute to improved colloidal stability in aqueous environments (Fig. 1H).

Injectable thermosensitive hydrogels were prepared and compared with Mg²⁺-complexed counterparts. Both gels exhibited rapid sol–gel transition between 25 and 37 °C within ∼5 min and maintained good injectability (Fig. 1I). SEM analysis revealed porous structures (∼4 μm for blank gels), which became denser with reduced pore size (< 2 μm) upon Mg²⁺ complexation (Fig. 1J). The Mg content in Gel-Mg was quantified by ICP-OES, which showed a Mg content of 0.3 wt%, corresponding to a coordination efficiency of 82%, the calibration curve was shown in Figure S3E.

Functional analyses confirmed that Mg²⁺ complexation significantly enhanced gel stability. UV degradation curves showed slower mass loss for Mg²⁺-complexed gels, even after 336 h (Fig. 1K). DSC analysis indicated a phase transition shift from 35.5 °C to 37.2 °C (Fig. 1L), while TGA revealed improved thermal resistance (Fig. 1M). Moreover, FTIR spectra confirmed structural differences among microspheres, exosome-loaded microspheres, blank gels, Mg²⁺-complexed gels, and composite gels (Fig. 1N). Sustained Mg²⁺ release was observed over 72 h, supporting their potential as long-acting biomaterials.

3.3. The composite hydrogel promote chondrocyte proliferation and migration and inhibit apoptosis

To evaluate the effect of CS@Gel-Mg on chondrocytes, cell viability was assessed by incubating ATDC5 cells with different concentrations of CS@Gel-Mg for 24 h. Cell viability was highest at a CS@Gel-Mg concentration of 1.5 mg/mL, followed by 2 mg/mL (Fig. 2A). For subsequent experiments, a concentration of 2 mg/mL was selected, as cell viability remained above 90%, and higher concentrations were more favorable for experimental outcomes. After treating ATDC5 cells with different formulations and incubating them for 24 h, 48 h, and 72 h, the CS@Gel-Mg group exhibited significantly higher cell viability compared to other treatment groups, demonstrating its superior ability to promote cartilage regeneration (Fig. 2B). Next, the apoptosis of ATDC5 cells was assessed by flow cytometry analysis. Chondrocytes treated with CS@Gel-Mg for 24 h, 48 h, and 72 h exhibited significantly lower apoptosis rates compared to the other four treatment groups (Fig. 2C and D). Furthermore, the migration assay revealed a substantially greater number of migrated chondrocytes in the CS@Gel-Mg group compared to the other groups, demonstrating the effectiveness of CS@Gel-Mg in promoting chondrocyte migration and regeneration (Fig. 2E and F).

Fig. 2.

In vitro assessment of the ability of CS@Gel-Mg to promote chondrocyte repair. (A) Relative viabilities of ATDC5 cells incubated with CS@Gel-Mg at different concentrations for 24 h (n = 6). (B) Relative cell viability of ATDC5 cells treated with different groups for 24 h, 48 h and 72 h respectively (n = 6). (C) Flow cytometric analysis on the apoptosis levels of ATDC5 cells treated with different treatments for 24 h, 48 h and 72 h, respectively (n = 3). Q1, necrotic cells; Q2, late apoptotic cells; Q3, live cells; Q4, early apoptotic cells. (D) Quantitative analysis of apoptosis rate of ATDC5 cells treated with different groups for 24h, 48h and 72h (n = 3). (E) Representative images and quantitative analysis of the migration of injured ATDC5 cells after different groups of treatments (n = 3). Statistical analysis for A, B, D, and E was performed using one-way ANOVA followed by Tukey's test. The data are presented as mean ± SD. ∗∗P < 0.01, ∗∗∗P < 0.001.

3.4. Modulation of macrophage polarization and inflammatory response by composite hydrogel

Considering the potential role of CS@Gel-Mg in macrophage polarization, we first evaluated the cellular compatibility of CS@Gel-Mg in mouse macrophages. As shown in Fig. 3A, red fluorescence (Cy5.5 dye-labeled exosomes) colocalized with green fluorescence (Phalloidin FITC-labeled F-actin) in both M1 and M2 macrophages, indicating exosome uptake, and M1 macrophages exhibiting stronger phagocytic capacity. Next, to evaluate the immunomodulatory capacity of the microsphere-loaded hydrogel, we first assessed its impact on macrophage polarization in vitro. qPCR analysis revealed that treatment with CS@Gel-Mg significantly reduced the mRNA levels of classical pro-inflammatory cytokines TNF-α and IL-1β in unstimulated (M0) BMDMs compared to control and other groups (Fig. 3B–D), indicating an anti-inflammatory effect of this composite hydrogel.

Fig. 3.

Effects of microsphere-loaded hydrogel on inflammation and macrophage polarization. (A) Phagocytosis of exosomes by M0 and M1 macrophages (n = 3). (Scale bar: 50 μm). (B-D) mRNA expression levels of IL-6, TNF-α, and IL-1β in M0 macrophages after treatment with different groups (qPCR, n = 3); (E, F) Flow cytometry analysis of surface markers CD86 (M1) and CD206 (M2) to evaluate polarization status. (G-I) Expression levels of IL-6, TNF-α, and IL-1β in LPS/IFN-γ-induced M1 macrophages after hydrogel treatment (qPCR, n = 3); (J, K) Flow cytometry analysis of CD86 and CD206, and corresponding quantitative bar graph. Statistical analysis for B-E and G-J was performed using one-way ANOVA followed by Tukey's test. The data are presented as mean ± SD. ∗ P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P <0.0001. ns, no significance.

Flow cytometry analysis further confirmed the hydrogel's immunomodulatory function. In the M0 macrophage population, treatment with CS@Gel-Mg induced a marked shift toward the M2 phenotype, as indicated by increased expression of the M2 marker CD206 (Fig. 3E and F), suggesting enhanced polarization toward a reparative, anti-inflammatory phenotype.

To simulate an inflamed microenvironment, BMDMs were pre-stimulated with LPS and IFN-γ to induce M1 polarization. Following hydrogel treatment, we observed significant downregulation of IL-6, TNF-α, and IL-1β in the CS@Gel-Mg groups (Fig. 3G–I). Corresponding flow cytometry results demonstrated a clear reversal of the M1 phenotype, with reduced CD86 expression and enhanced CD206 expression (Fig. 3J and K), indicating that the hydrogel system effectively reprograms macrophages from a pro-inflammatory to a pro-regenerative state even under inflammatory conditions.

Collectively, these results suggest that the composite hydrogel carrying exosomes and Mg²⁺ can modulate macrophage behavior and suppress inflammatory responses, thereby establishing a favorable immunological environment for subsequent cartilage regeneration.

3.5. Composite hydrogel promotes cartilage repair and bone regeneration in rat models

To assess the therapeutic efficacy of the microsphere-loaded hydrogel in vivo, we established a rat full-thickness cartilage defect model and administered intra-articular injections according to the timeline shown in Fig. 4A. Animals were sacrificed at 8 weeks post-treatment for behavioral, radiological, and histological evaluations.

Fig. 4.

In vivo evaluation of the therapeutic effect of microsphere-loaded hydrogel in a rat cartilage defect model. (A) Schematic illustration of the experimental timeline: rats underwent cartilage defect modeling at week −2, followed by intra-articular treatment starting at week 0. Animals were sacrificed at week 8 for evaluation. (B) Weight-bearing distribution on the injured limb. (C) Weight-bearing on the healthy limb. (D) Stride length analysis during gait. (E) OARSI scores evaluating cartilage degradation; (F) Bone volume to tissue volume ratio (BV/TV); (G) Trabecular thickness (Tb.Th); (H) Trabecular number (Tb.N); (I) Trabecular separation (Tb.Sp). (J) Representative 3D reconstructed micro-CT images showing joint morphology in each group (Ctrl, Gel, Gel-Mg, CS@Gel, CS@Gel-Mg) (Scale bar: 200 μm). (K) Hematoxylin & eosin (H&E) staining for tissue morphology; Safranin O/Fast Green staining for proteoglycan distribution; Toluidine Blue staining for sulfated glycosaminoglycans. (scale bar: 1000 μm). Data in B, C, and E–I were analyzed by one-way ANOVA followed by Tukey's test; data in D were analyzed by repeated measures two-way ANOVA followed by Šidák's post hoc test for between-group comparisons at 8 weeks. The data are presented as mean ± SD. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. ns, no significance. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Functional recovery was first evaluated by weight-bearing asymmetry analysis. Rats in the CS@Gel-Mg group exhibited a significant improvement in weight-bearing on the injured limb (Fig. 4B), along with normalized force distribution on the contralateral healthy limb (Fig. 4C), indicating reduced pain and enhanced joint function. Stride length and gait score further showed a significantly restored in the CS@Gel-Mg group (Fig. 4D). Histological scoring using OARSI criteria demonstrated that CS@Gel-Mg treatment markedly reduced cartilage degradation (Fig. 4E). These findings suggest that the composite hydrogel effectively preserves cartilage integrity during the repair process.

Micro-CT analysis revealed significant subchondral bone regeneration in the CS@Gel-Mg group. Specifically, quantitative morphometric parameters showed a higher bone volume to tissue volume ratio (BV/TV) (Fig. 4F), increased trabecular thickness (Tb.Th) and trabecular number (Tb.N) (Fig. 4G and H), and reduced trabecular separation (Tb.Sp) (Fig. 4I), compared to other groups. These improvements were corroborated by representative 3D reconstructed micro-CT images, which illustrated a more continuous and structured subchondral bone architecture in the CS@Gel-Mg-treated group (Fig. 4J).

Histological staining provided further insight into tissue regeneration. H&E staining revealed better tissue organization and fewer inflammatory infiltrates in the CS@Gel-Mg group. Safranin O/Fast Green staining indicated a more robust deposition of proteoglycans, while Toluidine Blue staining showed enhanced accumulation of sulfated glycosaminoglycans, both indicative of hyaline-like cartilage regeneration (Fig. 4K).

Collectively, these findings confirm that the exosome- and Mg²⁺-loaded hydrogel significantly enhances both cartilage and subchondral bone regeneration, promoting structural and functional restoration of the injured joint in vivo.

3.6. Composite hydrogel inhibited inflammatory factors and regulated the polarization of macrophages

To further elucidate the therapeutic mechanisms of the microsphere-loaded hydrogel in vivo, we examined local inflammatory cytokine expression. IHC staining of cartilage sections revealed that the expression of pro-inflammatory cytokines TNF-α, IL-6, and IL-1β was markedly suppressed in the CS@Gel-Mg group compared to all other groups (Fig. 5A and B). Quantitative analysis via ELISA confirmed these findings, showing significantly reduced levels of these inflammatory mediators in joint tissue homogenates (Fig. 5C). These results indicate a strong anti-inflammatory effect exerted by the composite hydrogel.

Fig. 5.

Effects of composite hydrogel on local inflammation and macrophage polarization. (A, B) IHC of cartilage tissue for inflammatory markers TNF-α, IL-6, and IL-1β (scale bar: 1000 μm). (C) Quantification of TNF-α, IL-6, and IL-1β levels in joint tissues (ELISA). (D) IF staining of cartilage tissues showing the expression of macrophage polarization markers iNOS (M1 phenotype) and CD206 (M2 phenotype). (scale bar: 200 μm). Statistical analysis for B and C was performed using one-way ANOVA followed by Tukey's test. The data are presented as mean ± SD. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. ns, no significance.

Next, we further verified the effect of this composite microsphere-loaded hydrogel on macrophage polarization in vivo. IF staining for macrophage phenotypic markers iNOS (M1) and CD206 (M2) revealed a distinct shift in macrophage polarization. In the CS@Gel-Mg group, there was a marked decrease in iNOS-positive M1 macrophages and a concurrent increase in CD206-positive M2 macrophages (Fig. 5D).

These results suggest that the composite hydrogel effectively modulated the local immune microenvironment, favoring a reparative M2 phenotype over a pro-inflammatory M1 state. This immune shift is consistent with the observed reduction in pro-inflammatory cytokines.

3.7. The composite hydrogel regulated angiogenic activity and SOX9-mediated cartilage regeneration in vivo

To assess angiogenic activity and its possible role in subchondral repair, western blot was performed for CD31, VEGF, α-SMA, and MMP-2. Densitometric analysis revealed that these markers were moderately expressed in Gel and Gel-Mg groups and remained at a relatively controlled level in the CS@Gel-Mg group (Fig. 6A). Cartilage biomechanical testing demonstrated that the compressive modulus of cartilage tissue in the CS@Gel-Mg group was significantly elevated (Fig. 6C). The enhanced mechanical integrity suggests improved extracellular matrix deposition and structural organization within the regenerated cartilage.

Fig. 6.

Effects of the composite hydrogel on markers of angiogenesis and SOX-dependent chondrogenesis in vivo. (A) Expression levels of angiogenesis-related proteins (CD31, MMP-2, VEGF, and α-SMA) and their densitometric analysis (Western blot). (B) Expression levels of cartilage extracellular matrix proteins (Aggrecan, COL2A1, GAG) and densitometric quantification (Western blot). (C) Cartilage biomechanics test showing compressive modulus (MPa) for each group. (D, E) IHC staining of cartilage tissues demonstrating the expression of the key chondrogenic transcription factor SOX9, as well as cartilage matrix proteins COL1A1 and COL2A1. (scale bar: 1000 μm). Statistical analysis for A-D was performed using one-way ANOVA followed by Tukey's test. The data are presented as mean ± SD. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. ns, no significance.

Next, we explored the downstream molecules that may be regulated by CS@Gel-Mg. We found that the expression of chondrogenic matrix markers was significantly enhanced in the CS@Gel-Mg group. Aggrecan, collagen II, and glycosaminoglycans (GAGs) showed elevated protein levels, as evidenced by Western blot and corresponding densitometric quantification (Fig. 6B). Since SOX9 is a core transcription factor involved in chondrocyte differentiation and matrix synthesis, we further examined its expression [29]. IHC staining showed that the expression of SOX9 was significantly upregulated in the CS@Gel-Mg group (Fig. 6D and E). In parallel, both COL2A1 and COL1A1 were increased, with a more pronounced upregulation of COL2A1, suggesting a shift in matrix remodeling toward a hyaline cartilage-like phenotype. These results suggest that CS@Gel-Mg promoted a chondrogenic outcome, an effect that may involve the observed upregulation of the key transcription factor SOX9.

3.8. SOX9 involvement in the chondro-regenerative effect of the composite

To further examine the potential role of SOX9 in the therapeutic process, we employed SOX9 knockdown (KD) and overexpression (OE) rat models and evaluated the outcomes following intra-articular treatment with CS@Gel-Mg.

At the functional level, CS@Gel-Mg treatment led to significant improvement in limb loading (Fig. 7A and B) and gait parameters (Fig. 7C and D) at 8 weeks. Notably, SOX9 knockdown attenuated these functional improvements, while SOX9 overexpression appeared to produce additional benefit, suggesting that the functional recovery promoted by the hydrogel is closely associated with SOX9 activity.

Fig. 7.

Functional and histological validation of SOX9 involvement in hydrogel-mediated cartilage repair. (A, B) Weight-bearing distribution on the injured and contralateral healthy limbs at 2 and 8 weeks. (C) Gait score evaluation at 2 and 8 weeks. (D) Stride length analysis over 2–8 weeks. (E) Representative histological staining (HE, Safranin O/Fast Green, Toluidine Blue) showing cartilage morphology and proteoglycan deposition. (scale bar: 1000 μm). (F) IHC staining of inflammatory cytokines (TNF-α, IL-6, IL-1β) in cartilage tissues, demonstrating differential effects of CS@Gel-Mg, SOX9 knockdown, and SOX9 overexpression. (scale bar: 1000 μm). Statistical analysis for A-C, F was performed using one-way ANOVA followed by Tukey's test; data in D were analyzed by repeated measures two-way ANOVA followed by Šidák's post hoc test for between-group comparisons at 8 weeks. The data are presented as mean ± SD. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. ns, no significance. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Histological evaluation further supported these findings. HE, Safranin O/Fast Green, and Toluidine Blue staining (Fig. 7E) revealed that CS@Gel-Mg-treated defects were filled with well-organized cartilage tissue rich in proteoglycans, while KD + CS@Gel-Mg animals exhibited incomplete repair with reduced proteoglycan deposition. Conversely, OE + CS@Gel-Mg rats displayed the most extensive hyaline-like cartilage regeneration. Consistent with these observations, immunohistochemical staining for inflammatory mediators (TNF-α, IL-6, IL-1β) demonstrated pronounced suppression in the CS@Gel-Mg group, which was partially reversed by SOX9 knockdown but further potentiated by SOX9 overexpression (Fig. 7F). These results suggest that SOX9 is a critical downstream effector for hydrogel-mediated anti-inflammatory and reparative activity.

At the molecular level, immunohistochemistry and Western blot confirmed that CS@Gel-Mg treatment upregulated SOX9 and its downstream chondrogenic markers, including COL2A1, Aggrecan, and COL1A1 expression (Fig. 8A–C). SOX9 knockdown significantly blunted these effects, resulting in diminished COL2A1, aggrecan and COL1A1 expression, whereas SOX9 overexpression promoted chondrogenic profile (Fig. 8A–C). Moreover, analysis of angiogenesis-related proteins (CD31, VEGF, MMP-2, α-SMA) revealed that CS@Gel-Mg maintained a balanced vascularization response, while SOX9 modulation further fine-tuned this effect (Fig. 8D).

Fig. 8.

Molecular validation of SOX9-mediated regulation in chondrogenesis and angiogenesis. (A) IHC staining of SOX9, COL1A1, and COL2A1 in cartilage tissues. (scale bar: 1000 μm). (B, C) Western blot and densitometric quantification of SOX9, COL1A1, COL2A1, Aggrecan, and GAG expression. (D) Western blot analysis and quantification of angiogenesis-related proteins (CD31, MMP-2, VEGF, α-SMA). Data showed that SOX9 modulation is associated with ECM deposition and vascular remodeling in the defect site. Statistical analysis for A-D was performed using one-way ANOVA followed by Tukey's test. The data are presented as mean ± SD. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. ns, no significance.

Collectively, these results establish SOX9 as a key regulatory hub mediating the dual effects of CS@Gel-Mg on inflammation inhibition and chondrogenic matrix regeneration, thereby mechanistically validating the therapeutic potential of this composite hydrogel for cartilage repair.

4. Discussion

The current study presents a multifunctional, injectable thermoresponsive hydrogel (CS@Gel-Mg), incorporating exosome-loaded hydroxypropyl chitosan microspheres and Mg²⁺-coordinated IL-PNIPAM polymer. This composite hydrogel exhibited favorable physicochemical characteristics, which are essential features for intra-articular delivery. Briefly, CS@Gel-Mg promoted chondrocyte viability and proliferation, significantly influenced macrophage polarization, shifting the balance toward an M2 anti-inflammatory phenotype. In the rat cartilage defect model, CS@Gel-Mg resulted in significant improvements in both behavioral, histological parameters, and subchondral bone remodeling. Mechanistic investigation revealed that CS@Gel-Mg promotes cartilage repair by upregulating SOX9, which in turn enhances the expression of aggrecan, GAGs, COL1A1, and COL2A1. These findings underscore the therapeutic potential of integrating immunomodulatory and osteochondral regenerative strategies in a single hydrogel platform.

Macrophages are central to tissue regeneration through their pivotal role in coordinating immune responses [30]. Numerous studies have demonstrated the role of Mg²⁺-coordinated hydrogel systems in regulating macrophage polarization. For example, magnesium-procyanidin coordinated metal polyphenol nanoparticles were able to promote the polarization of the M2 phenotype, which had a positive effect on tendon-bone interface injury [9]. Furthermore, studies have shown that a strategy involving the pre-incorporation of Mg²⁺ and the hypoxia-mimicking agent dimethyloxalylglycine into an adaptive hydrogel can enhance the functionality of MSCs and promote macrophage polarization toward an anti-inflammatory phenotype, offering a promising approach for OA treatment [31]. Mg²⁺-incorporated multifunctional gelatin scaffold modulates the inflammatory microenvironment of burns and accelerates wound healing, which is also associated with macrophage transition from M1 to M2 [32]. Hydrogel microspheres crosslinked by Mg²⁺ reduce the secretion of pro-inflammatory cytokines induced by macrophage polarization and attenuate inflammation, which plays a key role in spinal cord repair [33]. Consistent with these studies, we found that after treatment with Mg²⁺-coordinated exosome-loaded hydroxypropyl chitosan microspheres, macrophages achieved polarization toward M2 and attenuated the inflammatory response both in vivo and in vitro. This immunoregulatory effect is consistent with previous findings that local immune modulation is critical for creating a pro-regenerative microenvironment conducive to cartilage repair [34,35].

The transcription factor SOX9 is indispensable for chondrogenesis, acting as a master regulator of chondrocyte lineage commitment and driving gene expression of collagen II and aggrecan, while concurrently inhibiting hypertrophic differentiation [36,37]. However, few studies have integrated SOX9-driven differentiation with immunomodulatory scaffold strategies for cartilage repair. We found that the composite hydrogels promoted the expression of cartilage-specific ECM components including Aggrecan, collagen II, and GAG, suggesting that the cellular phenotype underwent a shift to hyaline cartilage, consistent with SOX9's well-established role as a master regulator of cartilage development and ECM homeostasis [38]. This was further confirmed by our in vivo knockdown and overexpression models. SOX9 deletion significantly impaired the cartilage repair effects of CS@Gel-Mg, while SOX9 overexpression further enhanced these effects. These results suggest that the hydrogel's capacity to facilitate cartilage regeneration is at least partially dependent on endogenous SOX9 signaling. While both exosomes and Mg²⁺ have been individually reported to influence SOX9 expression through different pathways, such as circRNA/miRNA or HIF-1α pathway [39,40], their combined presentation within the CS@Gel-Mg composite provided a more promising and effective strategy for activating chondrogenic programs.

The experimental design employed murine cell lines for in vitro studies and a rat osteochondral defect model for in vivo validation, a common approach in proof-of-concept cartilage repair research [41]. This cross-species approach has inherent limitations, as there are significant anatomical and physiological differences between mice and rats, and between rodents and humans. Rats possess larger joints and thicker cartilage than mice, making them more suitable for surgical modeling [42]. However, rodent cartilage is notably thinner than human cartilage, and both species exhibit a greater intrinsic healing capacity, which may overestimate therapeutic outcomes relative to human translation. Furthermore, rodent gait and biomechanical loading differ substantially from humans, limiting direct extrapolation of repair durability [43]. Despite these translational considerations, the core pathways investigated, including macrophage polarization and SOX9-driven chondrogenesis, are conserved across mammalian species [44,45]. The pro-regenerative effects observed in vitro translated into significant functional improvement in the rat model, supporting the biological relevance of the CS@Gel-Mg composite. Future studies using species-matched systems or large animal models with human-like cartilage mechanics will be essential to validate these findings and assess clinical potential.

Overall, our findings demonstrate that the injectable CS@Gel-Mg system offers a comprehensive therapeutic platform by integrating bioactive cues for immune modulation, pro-chondrogenic signaling, and structural support. This strategy overcomes several limitations of current hydrogels, such as poor mechanical integrity, insufficient cellular interaction, or short-term bioactivity. However, there are some limitations to our study. While the composite hydrogel incorporating both Mg²⁺ coordination and exosome loading demonstrated optimal efficacy for cartilage repair and immunomodulation, we did not investigate whether a direct molecular-level synergy exists between Mg²⁺ and the exosomes, nor did we explore its potential association with the activation of specific signaling pathways. This represents an important direction for future research. In addition, although we preliminarily demonstrated the regenerative effect of this system and identified SOX9 as a central regulator in the process, the specific mechanism by which it regulates SOX9 remains unknown. Future studies should aim to delineate the precise molecular pathways underlying SOX9 activation, and assess long-term performance in large animal models to facilitate clinical translation.

5. Conclusion

In summary, we developed a novel injectable thermoresponsive hydrogel system (CS@Gel-Mg) integrating exosome-loaded hydroxypropyl chitosan microspheres and Mg²⁺-coordinated IL-based polymer, which demonstrated excellent physicochemical stability, biocompatibility, and controlled release behavior. This composite hydrogel not only promoted chondrocyte proliferation and reduced apoptosis in vitro, but also effectively modulated macrophage polarization toward an anti-inflammatory M2 phenotype. In a rat cartilage defect model, CS@Gel-Mg significantly enhanced functional recovery, cartilage matrix regeneration, and subchondral bone remodeling. Mechanistic investigations revealed that these therapeutic effects were closely associated with reduced local inflammation and upregulated expression of SOX9 and cartilage-specific ECM proteins. Furthermore, gain- and loss-of-function studies in vivo confirmed the pivotal role of SOX9 signaling in mediating the regenerative efficacy of the hydrogel.

These findings underscore the potential of combining immunomodulatory and pro-chondrogenic strategies within a bioresponsive material platform to enhance osteochondral repair. The translational relevance of this system lies in its minimally invasive delivery, scalable fabrication, and the use of clinically adaptable components, offering a promising approach for the treatment of cartilage injuries and degenerative joint diseases.

Ethics statement

All animal experiments were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of Shanghai Tong Ren Hospital. The experimental protocol was reviewed and approved by the IACUC of Shanghai Tong Ren Hospital (Approval ID: A2024-030-01). All procedures complied with the ARRIVE guidelines and relevant regulations on the care and use of laboratory animals.

Funding

This study was supported by the Program of Changning District Health Committee (20254Y004), the Laboratory Open Fund of Key Technology and Materials in Minimally Invasive Spine Surgery (No. 2024JZWC-YBB06), the Fundamental Research Funds for the Central Universities (No. YG2024QNA62), the Medical-Engineering Interdisciplinary Collaborative Project between Shanghai Tongren Hospital and Donghua University (Grants No. 2023DHYGJC-YBB04) , and the Shanghai Magnolia Talent Program Pujiang Project (25PJD100D).

CRediT authorship contribution statement

Li Xiong: Conceptualization, Data curation, Formal analysis, Writing – original draft. Bin Chai: Data curation, Formal analysis, Investigation, Project administration, Writing – original draft. Gaixia Kuang: Formal analysis, Investigation, Resources, Software, Supervision, Writing – review & editing. Yosuke Kaneko: Investigation, Methodology, Software, Supervision, Validation, Writing – review & editing. Gonghao Zhang: Conceptualization, Investigation, Methodology, Project administration, Software, Writing – review & editing. Yunhan Ji: Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – review & editing.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is the supplementary data to this article.

Data availability

Data will be made available on request.