Therapeutic effects of hypoxia-preconditioned cartilage progenitor cell-exosomes on osteoarthritis through autophagy activation and macrophage polarization modulation

Abstract

Osteoarthritis (OA) is a prevalent degenerative joint disease characterized by progressive cartilage loss, pain, and functional disability. While tissue engineering holds promise for cartilage repair, obtaining high-quality seed cells with optimized functional activity remains a bottleneck. Exosomes derived from cartilage progenitor cells (CPCs) have emerged as potent cell-free therapeutic candidates. Here, we demonstrate that exosomes from hypoxia-preconditioned CPCs (H-Exos) possess superior regenerative capabilities compared to their normoxic counterparts. specifically by promoting proliferation and migration while suppressing catabolism in IL-1β-treated ATDC5 cells. Mechanistically, we show that hypoxic preconditioning enriches miR-222-3p in H-Exos, which targets Rab1A to inhibit mTORC1 signaling and restore autophagy, thereby enhancing chondrocyte anabolism. Notably, we also identified a distinct immunomodulatory function: H-Exos were efficiently internalized by macrophages (RAW264.7), driving their polarization toward an anti-inflammatory phenotype via inhibition of the NF-κB signaling pathway. In a rat OA model, intra-articular delivery of a GelMA/H-Exos composite hydrogel significantly attenuated cartilage destruction and subchondral bone remodeling, concomitant with favorable modulation of synovial macrophage polarization. Collectively, this study elucidates a dual protective mechanism—involving the miR-222-3p–Rab1A–mTORC1–autophagy axis and macrophage reprogramming—and presents a promising biomaterial-based strategy for comprehensive OA therapy.

Graphical abstract

Dual mechanisms of H-Exos in osteoarthritis therapy. H-Exos enriched with miR-222-3p target chondrocytes to activate autophagy via Rab1A/mTORC1 and reprogram macrophages via NF-κB inhibition.

Highlights

• •Hypoxia-preconditioned exosomes (H-Exos) show superior bioactivity.
• •miR-222-3p in H-Exos targets Rab1A to restore chondrocyte autophagy.
• •H-Exos drive M1-to-M2 macrophage polarization by inhibiting NF-κB.
• •GelMA/H-Exos composite hydrogel attenuates OA progression in vivo.

1. Introduction

Osteoarthritis (OA) is a chronic degenerative joint disease characterized by progressive degradation of articular cartilage, along with subchondral bone remodeling, synovial inflammation, and osteophyte formation [1,2]. As the most prevalent musculoskeletal disorder worldwide, OA affects millions of individuals and represents the leading cause of global joint disability, posing substantial socioeconomic burdens [1,3]. The pathogenesis and progression of OA involve complex interactions among multiple risk factors, including aging, obesity, gender, joint trauma, and mechanical stress, which collectively drive inflammatory responses and cartilage degeneration [2,4,5]. Conventional therapies for OA mainly focus on symptom management, failing to effectively slow disease progression or promote cartilage regeneration, often necessitating joint replacement in advanced stages. [6,7]. These limitations underscore the urgent need for disease-modifying therapies to halt or reverse OA progression [8].

Stem cell-based therapies have emerged as a promising strategy for treating various degenerative conditions [9]. The search for optimal cell sources for cartilage repair has extensively investigated mesenchymal stromal cells (MSCs), including bone marrow-derived MSCs (BMSCs), synovium-derived MSCs (SMSCs), and adipose tissue-derived MSCs (ADSCs), among others [10,11]. Recently, researchers have isolated and identified a novel resident cell population in the superficial zone of articular cartilage, named cartilage progenitor cells (CPCs) [[12], [13], [14]]. These cells possess characteristic stem-like properties including sustained self-renewal capacity, typical mesenchymal stem cell marker expression, and the unique ability to undergo chondrogenic differentiation while maintaining phenotype stability [15,16]. Compared to MSCs, CPCs not only demonstrate superior chondrogenic differentiation capacity and enhanced migratory potential, but critically maintain stable phenotypes during in vitro chondrogenesis—resisting hypertrophy to persistently produce hyaline-like cartilage [15,17]. These properties establish CPCs as a promising cell source for cartilage tissue engineering.

Exosomes are currently a focal point in the stem cell research [18,19]. Exosomes are small (30-150 nm diameter), bilayer membrane vesicles that carry biologically active cargo including mRNAs, microRNAs (miRNAs), proteins, signaling cytokines, and lipids—that enable them to mediate intercellular communication and modulate recipient cell functions [[20], [21], [22]]. Compared to stem cells, exosomes exhibit greater stability, lower immunogenicity, higher cellular uptake efficiency, and more convenient routes of administration [23]. However, their low yield and compositional heterogeneity present significant challenges for therapeutic application. One of the key scientific obstacles in exosome-based therapy is to develop effective strategies to maximize their therapeutic efficacy [24].

It is well known that oxygen concentration is a key regulator of cellular behavior in the extracellular microenvironment. Hypoxia has been shown to enhance MSC proliferation, survival and differentiation [[25], [26], [27]]. Notably, hypoxia-preconditioned CPCs exhibit greater chondrogenic potential and more effectively maintain cartilage homeostasis than normoxia-cultured CPCs [28]. In addition to regulating cellular behavior, hypoxia has also been shown to modulate the characteristics and therapeutic potential of exosomes derived from MSCs. Recent studies reported that hypoxia-preconditioned BMSC-derived exosomes enhanced the biological function of articular chondrocytes and more effectively promoted tissue-engineered cartilage repair [[29], [30], [31]]. Like BMSCs, CPCs naturally reside in the relatively hypoxic environment (∼1% O₂) of the joint cavity, yet are typically cultured under normoxic conditions (∼21% O₂) in vitro [32]. It is hypothesized that hypoxic preconditioning can also enhance the biological function of CPC-derived exosomes (CPC-Exos), improving their capacity for cartilage repair and regeneration.

Beyond cartilage degradation, the pathogenesis of OA involves a complex interplay within the synovial immune microenvironment, where macrophage polarization plays a pivotal role [33]. Accordingly, multifunctional biomaterial platforms that simultaneously modulate inflammation and support cartilage regeneration have attracted increasing attention for OA therapy [34]. While the direct chondroprotective effects of stem cell exosomes are recognized, their specific impact on this inflammatory network, and whether they possess a defined immunomodulatory capacity, remains an area of active investigation.

This study was designed to elucidate the mechanisms through which H-Exos ameliorate OA. We first characterized their enhanced bioactivity on IL-1β-stimulated ATDC5 cells, identifying the key miRNA and its target pathway responsible for chondroprotection. Parallelly, we investigated their immunomodulatory capacity by examining macrophage polarization. The therapeutic potential of H-Exos was ultimately validated in a rat OA model using a GelMA hydrogel delivery system.

2. Materials and methods

2.1. CPCs isolation and identification

CPCs were isolated from the superficial zone of mouse articular cartilage using a fibronectin adhesion method, as previously described [28,[35], [36], [37]]. Briefly, the femoral head cartilage was harvested, minced, and digested with 0.25% trypsin for 30 min, followed by 0.25% type II collagenase digestion for 4 h at 37 °C to obtain single-cell suspensions. The cells were seeded at a density of 4000 cells/mL onto fibronectin-coated plates in DMEM/F12 medium. After a 30-min incubation, non-adherent cells were removed. The adherent cells were cultured in standard growth medium (DMEM/F12 [Gibco, USA] supplemented with 10% fetal bovine serum [FBS; Thermo, USA], 2.5 mM L-glutamine [Sigma-Aldrich, USA], 2 μg/mL amphotericin B [Selleckchem, USA], 62 μg/mL ascorbic acid [Macklin, China], and 100 IU/mL penicillin/streptomycin [P/S; Sigma, USA]). Passage 2 cells were used for subsequent experiments.

To identify CPCs, we evaluated their multilineage differentiation potential and performed flow cytometry analysis based on previous methods [16,38,39]. Briefly, CPCs were cultured in adipogenic, osteogenic, or chondrogenic differentiation media (Sigma-Aldrich, USA) following the manufacturer's protocol. Multilineage differentiation potential was assessed using Oil Red O (adipogenesis), Alizarin Red (osteogenesis), and Alcian Blue staining (chondrogenesis). Additionally, CPCs were characterized by flow cytometry (CytoFLEX, Beckman Coulter, USA) with PE-conjugated CD34, CD90, and CD105 antibodies, along with FITC-conjugated CD45 antibodies (Biolegend, USA).

2.2. Hypoxia treatment and exosome isolation

CPCs at 80% confluence were washed with PBS and then cultured in DMEM/F12 supplemented with 10% exosome-depleted FBS (VivaCell Biosciences, Shanghai, China) under normoxic (21% O₂/5% CO₂/74% N₂) or hypoxic (1% O₂/5% CO₂/94% N₂) conditions using an oxygen-controlled incubator for 48 h. The culture medium was collected for isolation of exosomes. The supernatant was sequentially centrifuged at 300×g for 10 min, 1200×g for 20 min, and 10,000×g for 30 min at 4 °C, followed by filtration through a 0.22 μm membrane filter (Merck-Millipore, Germany) to remove cells and debris. The filtrate was ultracentrifuged at 140,000×g for 90 min at 4 °C. The exosome pellet was resuspended in PBS and centrifuged again at 140,000×g for 90 min. Finally, the pelleted exosomes were resuspended in PBS for storage at − 80 °C. Transmission electron microscopy (TEM), Nanoparticle Tracking Analysis (NTA), and western blotting were used to characterize exosomes.

2.3. Cell culture

The mouse chondrocyte cell line ATDC5, which originates from teratoma fibroblasts, is characterized by rapid proliferation and serves as a well-established in vitro model for chondrocyte studies [40,41]. ATDC5 cells were procured from Jennio Biotech (GuangZhou, China) and cultured in a monolayer in DMEM/F12 medium (Gibco, USA) comprising 5% FBS (Thermo, USA) and 1% P/S (Sigma, USA). To induce ATDC5 cells into mature chondrocytes, the cells were treated with an ITS supplement containing recombinant insulin (1.0 mg/ml), transferrin (0.55 mg/ml), and sodium selenite (0.5 μg/ml) at 1 × working concentration for 14 days. Induction medium was supplemented with TGF-β1 (10 ng/mL). To establish an in vitro model of OA-like chondrocytes, ATDC5 cells were treated with interleukin-1 beta (IL-1β; Peprotech, USA) at 10 ng/mL for 24 h. This IL-1β-induced ATDC5 model effectively mimics the destructive microenvironment of progressive OA and serves as a valuable tool for mechanistic studies in isolated cells [42].

RAW 264.7 cells purchased from Shanghai Fuyu Biotechnology Co., Ltd., and cultured in RPMI 1640 medium (Gibco, USA) supplemented with 10% FBS (Thermo, USA) and 1% penicillin-streptomycin (Sigma, USA) at 37 °C in a 5% CO₂ incubator, with medium refreshed every two days.

To model inflammatory activation, macrophages were stimulated with 100 ng/mL lipopolysaccharide (LPS; Sigma-Aldrich, USA) in serum-free RPMI 1640 containing 1% penicillin-streptomycin for 24 h.

2.4. Internalization of exosomes

Equal amounts of N-Exos and H-Exos were labeled using the lipophilic dye DiO (3,3′-dioctadecyloxacarbocyanine perchlorate; Sigma-Aldrich, USA) according to the manufacturer's instructions. Excess dye from the labeled exosomes was removed by ultracentrifugation at 100,000g for 1 h at 4 °C. Exosomes were then washed three times and resuspended in PBS. Exosomes labeled with DiO were co-cultured with ATDC5 cells for 24 h, then fixed with 4% paraformaldehyde. Nuclei were stained with Hoechst 33342 (10 μg/ml, Beyotime, China). Uptake of exosomes was observed by confocal laser scanning microscopy (Zeiss LSM710, Germany).

For visualization of uptake by macrophages, equal amounts of N-Exos and H-Exos were labeled with the membrane dye PKH26 (Sigma-Aldrich, USA) following the manufacturer's protocol. Unincorporated dye was removed by ultracentrifugation at 100,000×g for 1 h at 4 °C. The PKH26-labeled exosomes were resuspended in PBS and added to the culture medium of RAW 264.7 macrophages. After 24 h of co-culture, cells were fixed with 4% paraformaldehyde, and nuclei were stained with DAPI. Internalization was confirmed using confocal laser scanning microscopy (Zeiss LSM710, Germany).

2.5. Cell viability determination

The effects of N-Exos and H-Exos on proliferation of IL-1β-induced ATDC5 cells were evaluated using the Cell Counting Kit-8 (CCK-8; Sigma, USA) assay. Cells were seeded in 96-well plates at a density of 4000 cells/well and treated with PBS, N-Exos, or H-Exos (1 × 10⁹ particles/mL). After 24, 48, or 72 h of treatment, CCK-8 reagent was added to each well and incubated for 4 h in the dark. Absorbance at 450 nm was measured using a microplate reader (Thermo Fisher Scientific, USA).

A scratch wound healing assay was used to assess migration. ATDC5 cells were seeded at a density of 2 × 10⁵ cells per well in six-well plates and cultured until they reached approximately 90% confluence. A straight scratch was then introduced in each well using a 200 μL pipette tip. After washing the wells twice with PBS to remove debris, fresh serum-free medium containing PBS, N-Exos, or H-Exos (1 × 10⁹ particles/mL) was added. Scratch closure was documented using a light microscope at 24 h, and migration rates were subsequently analyzed with ImageJ software.

2.6. Immunofluorescence staining

Cells subjected to various treatments were fixed with 4% paraformaldehyde for 20 min. After fixation, the cells were permeabilized using 0.1% Triton X-100 for 10 min and then blocked with 5% bovine serum albumin (BSA) for 1 h. Subsequently, the samples were incubated overnight at 4 °C with primary antibodies specific to the target proteins. On the following day, the slides were incubated with Alexa Fluor-488 or Alexa Fluor-594-conjugated secondary antibodies (Invitrogen, USA) for 1 h, followed by DAPI staining at room temperature for 30 min. Final images were captured using a fluorescence microscope (Zeiss Axio Vert.A1, Germany), and semi-quantitative analysis was performed with ImageJ software.

2.7. Western blotting

After different treatments, cells were washed three times with PBS, and total protein was extracted using RIPA lysis buffer (Beyotime, China) supplemented with protease and phosphatase inhibitors. Protein concentration was determined using the BCA assay kit (Beyotime, China). Equal amounts of protein (20 μg per sample) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto 0.22 μm polyvinylidene fluoride (PVDF) membranes. Membranes were blocked with 5% bovine serum albumin (BSA) for 30 min and incubated overnight at 4 °C with the following primary antibodies: Collagen II (COL II; Santa Cruz, USA), Aggrecan (Proteintech, USA), MMP13 (Proteintech, USA), SOX9 (Abcam, USA), Rab1A (Abcam, USA), p-mTOR (CST, USA), mTOR (CST, USA), p-S6K (CST, USA), S6K (CST, USA), p-4E-BP1(Abcam, USA), 4E-BP1(Abcam, USA), Beclin1 (Abcam, USA), P62 (Abcam, USA), LC3 I/II (CST, USA), p-IκBα (CST, USA), IκBα (CST, USA), p-P65 (CST, USA), P65 (CST, USA), and β-actin (Abcam, USA). After washing, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) detection reagent (Beyotime, China) and analyzed using ImageJ software.

2.8. Quantitative RT-PCR analysis (qRT-PCR)

RNA was extracted from cells or exosomes using TRIzol reagent (Invitrogen, USA) according to the manufacturer's instructions. RNA was reverse-transcribed into cDNA using oligo-dT primers (Takara, Japan) for mRNA or the TaqMan® MicroRNA Reverse Transcription Kit (Applied Biosystems, USA) for miRNA. Gene expression was quantitatively analyzed using SYBR Green Premix Ex Taq real-time PCR reagents (Takara, Japan). The relative expression levels of mRNA or miRNA were normalized to GAPDH and U6, respectively. Fold changes were calculated using the 2−ΔΔCt method. Sequences are listed in Supplementary Table 1.

2.9. Transmission electron microscopy

Treated cells were fixed with 2.5% glutaraldehyde (Solarbio) at 4 °C. Subsequent processing and imaging were performed by specialized technicians using a transmission electron microscope (Carl Zeiss, Germany). Autophagic vesicles were qualitatively assessed in captured images.

2.10. miRNA sequencing

Exosomal RNA was extracted from N-Exos and H-Exos using a total exosome RNA isolation kit (Thermo Fisher, USA), following the manufacturer's instructions. Small RNAs, including miRNAs, were enriched by PAGE, and sequencing adaptors were ligated to their 3′ and 5′ ends. After reverse transcription and PCR amplification, miRNA libraries were constructed and sequenced on a high-throughput sequencing platform. The resulting data were used to compare the miRNA expression profiles between N-Exos and H-Exos.

2.11. Dual-luciferase reporter assay

Cells were transfected with luciferase reporter vectors containing the full-length wild-type or mutant 3ʹ-UTR of Rab1A mRNA, together with miR-222-3p mimics or mimic negative control (mimic-NC), using Lipofectamine 2000 (Invitrogen, CA, USA). Luciferase activity was assessed 72 h after transfection using a dual-luciferase reporter assay system (Promega, USA).

2.12. RNA-sequencing

Total RNA was extracted from cells using TRIzol reagent (Invitrogen, USA). The quality and integrity of the RNA samples were verified, and qualified samples were used to construct sequencing libraries. The libraries were subjected to high-throughput sequencing on an Illumina NovaSeq 6000 platform (Illumina, USA). Gene expression levels were quantified based on Reads Per Kilobase per Million mapped reads (RPKM).

2.13. Synthesis of GelMA/H-Exos hydrogel

Gelatin (10 g) was dissolved in 100 mL of Dulbecco's phosphate-buffered saline (D-PBS) at 55 °C, followed by dropwise addition of 8 mL methacrylic anhydride under continuous stirring. After 2 h of reaction, 400 mL of D-PBS was added to terminate the reaction. The mixture was then transferred into dialysis tubing (MWCO 7000) and dialyzed against deionized water for 5 days. The final product was lyophilized and stored at −20 °C.

To fabricate GelMA/H-Exos hydrogel, GelMA (100 mg) was dissolved in 1 mL of PBS containing 1 mg of lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP; Sigma-Aldrich, USA) photoinitiator at 37 °C. H-Exos solution (500 μL, 1 × 10¹⁰ particles/mL) was added and gently mixed. The mixture was cast into Teflon molds (10 mm diameter, 3 mm depth) and crosslinked under 365 nm light (10 mW/cm²) for 1 min. For biological applications, GelMA was sterilized via UV irradiation and filtered through a 0.2 μm membrane.

2.14. Characterization of GelMA/H-Exos hydrogels

The degree of methacrylation (DoM) of GelMA was determined by ¹H NMR spectroscopy (Bruker, Germany; 400 MHz). Lyophilized GelMA was dissolved in DMSO-d₆, and ¹H NMR spectra were recorded.

The surface morphology of GelMA and GelMA/H-Exos hydrogels was observed under a scanning electron microscope (SEM, Sirion 2000, FEI, United States). The storage and loss modulus of the hydrogels were determined at 37 °C and 10% strain for 220 s with a TA Rheometer (Discovery Hybrid Rheometer-2).

The biodegradation of the hydrogels was quantitatively assessed through gravimetric analysis. For swelling measurements, hydrogels were incubated in PBS at 37 °C. The swelling ratio was calculated as:

$$\text{Swelling}\text{ratio}(\%)=\left[\left({\mathrm{W}}_{\mathrm{t}}-{\mathrm{W}}_0\right)/{\mathrm{W}}_0\right]\times 100\%$$

where:

W₀ = initial dry weight of hydrogel.

W_t = swollen weight at time point.

For degradation measurements, hydrogels were immersed in PBS containing 2 U/mL collagenase (24-well plate, 37 °C). The degradation rate was determined as:

$$\text{Degradation}\text{ratio}(\%)=\left[\left({\mathrm{W}}_0‐{\mathrm{W}}_1\right)/{\mathrm{W}}_0\right]\times 100\%$$

where:

W₀ = initial weight.

W₁ = remaining weight after degradation at time point.

Exosome release kinetics from GelMA hydrogels were quantified using a BCA protein assay (Beyotime, China). GelMA/H-Exos hydrogels were incubated in PBS in 24-well plates at 37 °C. At specific time points, the supernatant was collected and replenished with fresh PBS. Cumulative exosome release was expressed as the percentage of total protein content relative to the initial load, with absorbance measured at 562 nm.

The biocompatibility of GelMA and GelMA/H-Exos hydrogels was assessed using ATDC5 cells in vitro and major organ histology in vivo. Cell viability was evaluated through live/dead staining (Calcein-AM/PI) after 3 days of culture, while proliferation rates were quantified using CCK-8 assays (Sigma, USA) at 24, 48, and 72-h timepoints. For in vivo evaluation, heart, liver, spleen, lung, and kidney tissues were collected after 4 weeks of treatment, fixed in 4% paraformaldehyde, and stained with H&E for histocompatibility evaluation.

The in vitro immunological safety of the hydrogels was evaluated using RAW264.7 macrophages. RAW264.7 cells were cultured on GelMA and GelMA/H-Exos hydrogels for 24 h, followed by Live/Dead staining for viability and phalloidin staining for adhesion morphology. After 24 h, culture supernatants were collected for ELISA quantification of TNF-α, IL-6, and IL-10 (R&D Systems, USA), and RT-qPCR was performed to assess the expression of CD163 and CD86.

2.15. Animal experiments

All animal experimental protocols were approved by the Animal Ethics Committee of Southwest Medical University (20240220-012). Male 10-week-old Sprague-Dawley rats underwent destabilization of the medial meniscus (DMM) and anterior cruciate ligament transection (ACLT) to induce OA, as previously described [43]. Briefly, under general anesthesia, the right knee joint was exposed through a medial parapatellar approach. For DMM surgery, the medial meniscotibial ligament was carefully transected to destabilize the medial meniscus, while ACLT was performed by complete transection of the ACL at its tibial insertion. Joint stability was confirmed after the procedure, and the wound was closed in layers. Rats in the sham group underwent skin and muscle incision without joint manipulation.

2.16. Micro-CT analysis

At designated time points, samples were collected and fixed in 4% paraformaldehyde for 24 h. Subsequently, the knee joint specimens were placed on a microcomputed tomography (micro-CT) system (SkyScan 1176, Bruker, Germany). Scanning was performed with an X-ray voltage of 80 kV and an X-ray tube current of 200 μA, using a voxel size of 18 μm and a 1.0 mm aluminum filter.

2.17. Histology, immunohistochemistry and immunofluorescence staining

Specimens were decalcified in 10% EDTA (pH 7.4) for 4 weeks at room temperature with regular solution changes. Tissues were dehydrated through graded ethanol series, embedded in paraffin, and sectioned at 5 μm thickness. Sections were stained with hematoxylin and eosin (H&E) for general morphology, Safranin O/Fast Green for cartilage matrix evaluation, and Masson's trichrome for collagen distribution.

For immunofluorescence staining, deparaffinized sections underwent antigen retrieval in citrate buffer (pH 6.0), followed by blocking with 5% bovine serum albumin for 30 min at room temperature. The sections were then incubated overnight at 4 °C with primary antibodies against COL II, p-mTOR, p-S6K, Beclin1 and LC3 II. After washing, appropriate fluorescent secondary antibodies were applied. Nuclei were counterstained with DAPI, and images were captured using a fluorescence microscope (Zeiss Axio Vert.A1, Germany).

For immunohistochemistry, deparaffinized sections underwent antigen retrieval in citrate buffer (pH 6.0), followed by blocking with 5% bovine serum albumin for 30 min at room temperature. The sections were incubated overnight at 4 °C with primary antibodies against iNOS and CD206. After washing, HRP-conjugated secondary antibodies were applied and developed with DAB.

2.18. Statistical analysis

All data are presented as the mean ± standard deviation. Statistical significance was determined using one-way analysis of variance (ANOVA) or Student's t-test, with a P-value of less than 0.05 considered significant. GraphPad Prism 7 software was used for statistical analysis and graphing.

3. Results

3.1. Characterization of CPCs and CPC-Exos

Cells isolated from the superficial zone of mouse femoral head cartilage selectively adhered to fibronectin-coated surfaces (Fig. 1A). Visible colonies appeared after 12 days of culture, indicating the clonogenic potential of these cells.

Fig. 1.

CPC Characterization, CPC-Exos Isolation, and Therapeutic Effects on IL-1β-induced ATDC5 Cells. (A) Schematic illustration of CPC isolation. (B) Multipotent differentiation of CPCs demonstrated by adipogenic (scale bar = 50 μm), osteogenic (scale bar = 20 μm), and chondrogenic (scale bar = 10 μm) differentiation. (C) The negative expression of CD34/CD45 and positive expression of CD90/CD105 of CPCs analyzed by flow cytometry. (D) Schematic illustration of exosome isolation. (E) TEM images of N-Exos and H-Exos (scale bar = 600 nm). (F) Representative NTA profiles of N-Exos and H-Exos. (G) Western blot analysis of exosomal markers CD9 and CD63, with calnexin as negative control. (H) Cellular internalization of N-Exos and H-Exos by ATDC5 cells (scale bar = 3 μm). (I) Experimental design comparing N-Exos and H-Exos treatments in IL-1β-induced ATDC5 cells. (J) CCK-8 assay to observe the proliferation of ATDC5 cells. (K) Representative images of scratch wound healing assay to assess cell migration (scale bar = 60 μm). (L) Quantitative analysis of the migration rates. (M − O) Representative immunofluorescence staining of Aggrecan, COL II, and MMP13 (scale bar = 25 μm). (P) Quantification of immunofluorescence. (Q) mRNA expression levels of Aggrecan, COL II, MMP13, IL-1β, TNF-α, and IL-6 were detected by qRT-PCR. (R-S) Relative protein expression levels of Aggrecan, COL II, MMP13, and SOX9, along with statistical analysis. All data are presented as mean ± standard deviation (SD) from at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001.

The isolated cells demonstrated clear multilineage differentiation capacity. Oil Red O, Alizarin Red, and Alcian Blue staining confirmed successful adipogenic, osteogenic, and chondrogenic induction, respectively (Fig. 1B).

Flow cytometry analysis showed that the cells were positive for CD90 (99.8%) and CD105 (93.7%), and negative for CD34 (0.5%) and CD45 (0.26%), consistent with the surface marker profile of CPCs (Fig. 1C).

Exosomes were isolated from the culture supernatant of normoxia- or hypoxia-preconditioned CPCs via ultracentrifugation (Fig. 1D). TEM revealed that both types of exosomes exhibited the typical cup-shaped morphology with intact lipid bilayers (Fig. 1E). NTA analysis showed that the mean particle diameter was 122.7 ± 7.5 nm for N-Exos and 136.2 ± 9.4 nm for H-Exos (n = 3), with no statistically significant difference (p > 0.05) (Fig. 1F).

Western blot analysis confirmed the presence of exosomal markers CD9 and CD63 in CPC-Exos, with calnexin serving as a negative control (Fig. 1G).

To assess the uptake of CPC-Exos by recipient ATDC5 cells, fluorescently labeled N-Exos and H-Exos were co-cultured with the cells for 12 h. Confocal microscopy revealed distinct punctate fluorescent signals predominantly localized in the perinuclear cytoplasm, indicating efficient internalization and intracellular trafficking of both exosome types (Fig. 1H).

3.2. Chondroprotective effects of H-Exos on IL-1β-induced ATDC5 cells

To evaluate the protective and functional effects of H-Exos compared with N-Exos on IL-1β-induced ATDC5 cells, we assessed cell proliferation, migration, extracellular matrix (ECM) anabolism, catabolism and the inflammatory response (Fig. 1I). CCK-8 assay showed that cell proliferation was suppressed by IL-1β. The H-Exos group exhibited a higher OD value than the N-Exos group (Fig. 1J).

Scratch wound healing assays indicated enhanced cell migration in both exosome-treated groups, with a larger migrated area observed in the H-Exos group (Fig. 1K and L).

Immunofluorescence (Fig. 1M-P) and qRT-PCR (Fig. 1Q) analyses consistently showed that H-Exos increased Aggrecan and COL II expression while decreasing MMP13 levels compared to N-Exos. Furthermore, qRT-PCR analysis revealed that H-Exos more potently suppressed the IL-1β-induced expression of inflammatory cytokines (IL-1β, TNF-α and IL-6). Western blotting further confirmed upregulation of COL II, Aggrecan, and SOX9 proteins, along with downregulation of MMP13 (Fig. 1R and S). These findings suggest that H-Exos more effectively promote proliferation, migration, and ECM anabolism, while inhibiting catabolism and inflammatory response in IL-1β-induced ATDC5 cells.

3.3. MicroRNA-sequencing analysis

Exosomes, enriched in miRNAs, serve as crucial mediators intercellular communication and play essential roles in tissue regeneration. To investigate the underlying mechanisms by which H-Exos exert superior therapeutic effects of H-Exos over N-Exos in cartilage repair, we conducted high-throughput miRNA sequencing. Differential expression analysis revealed 9 significantly upregulated and 8 downregulated miRNAs in H-Exos (Fig. 2A and B). qRT-PCR validation in H-Exos identified miR-222-3p as the most abundantly expressed differential miRNA (Fig. 2C). Subsequent KEGG pathway enrichment analysis indicated that these miRNAs were primarily associated with autophagy and mTOR signaling pathways (Fig. 2D), supporting its selection for further validation experiments.

Fig. 2.

Effects of H-Exos via exosomal miR-222-3p. (A) Volcano plot of differentially expressed miRNAs in H-Exos versus N-Exos. (B) Heatmap of differentially expressed miRNAs in H-Exos compared to N-Exos. (C) Verification of the top five upregulated miRNAs in H-Exos by qRT-PCR. (D) Differentially expressed miRNAs related pathway enrichment map based on KEGG enrichment analysis. (E) The miR-222-3p expression levels in ATDC5 cells were assessed by qRT-PCR. (F) CCK-8 assay to observe the proliferation of ATDC5 cells. (G) Representative images of scratch wound healing assay to assess cell migration (scale bar = 60 μm). (H) Quantitative analysis of the migration rates. (I–K) Representative immunofluorescence staining of Aggrecan, COL II, and MMP13 (scale bar = 25 μm). (L) Quantification of immunofluorescence. (M) mRNA expression levels of Aggrecan, COL II, MMP13, IL-1β, TNF-α, and IL-6 were detected by qRT-PCR. (N) The number of autophagic vacuoles was observed by TEM in the above groups (Scale bar = 6000 nm and 1000 nm). (E) Representative immunofluorescence staining of LC3 (scale bar = 25 μm). All data are presented as mean ± standard deviation (SD) from at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001.

3.4. Effect of H-Exos depends on miR-222-3p

To determine whether the biological effects of hypoxia-preconditioned CPC-derived exosomes (H-Exos) depend on exosomal miR-222-3p, CPCs were transfected with miR-222-3p inhibitors and cultured under hypoxic conditions. Exosomes and cells were subsequently harvested. qRT-PCR confirmed successful suppression of miR-222-3p in both CPCs and derived exosomes after inhibitor transfection (Supplementary Fig. S1). Building upon the established chondroprotective properties of H-Exos described previously, we investigated their functional dependence on miR-222-3p by treating IL-1β-induced ATDC5 cells with both H-Exos^NC-KD and H-Exos^{miR−222−3pKD}. Knockdown of miR-222-3p significantly attenuated the therapeutic effects of H-Exos. Specifically, CCK-8 (Supplementary Fig. S2) and scratch wound healing assays (Supplementary Fig. S3) demonstrated that the proliferative and migratory capacities enhanced by H-Exos were substantially reduced by H-Exos^{miR−222−3pKD}. Moreover, immunofluorescence staining and qPCR analysis revealed that the upregulated expression of Aggrecan and COL II induced by H-Exos was markedly decreased by H-ExosmiR-222–3pKD, while the suppression of MMP13 and inflammatory cytokines (IL-1β, TNF-α, and IL-6) was partially reversed (Supplementary Fig. 4A–B). TEM revealed autophagic vacuoles accumulation in H-Exos^NC-KD (Supplementary Fig. S5), with corresponding LC3 fluorescence intensity elevation (Supplementary Fig. S6), collectively indicating enhanced autophagy activation.

To elucidate the functional role of miR-222-3p in chondroprotection, we conducted gain and loss-of-function experiments by transfecting IL-1β-induced ATDC5 cells with miR-222-3p mimics, inhibitors, or their respective negative controls. qRT-PCR analysis confirmed successful modulation of miR-222-3p expression levels across all experimental groups (Fig. 2E). miR-222-3p overexpression significantly enhanced cell proliferation (Fig. 2F) and migration capacity (Fig. 2G and H). Furthermore, immunofluorescence staining (Fig. 2I–L) revealed that miR-222-3p overexpression markedly upregulated COL II and Aggrecan while downregulating MMP13. Meanwhile, qRT-PCR results (Fig. 2M) confirmed these matrix-related changes and further demonstrated a significant suppression of inflammatory cytokines (IL-1β, TNF-α, and IL-6), collectively mimicking the protective effects of H-Exos. Conversely, inhibition of miR-222-3p substantially attenuated these beneficial effects. TEM analysis demonstrated increased autophagic vacuoles formation in miR-222-3p mimic-treated cells (Fig. 2N), which correlated with elevated LC3 fluorescence intensity (Fig. 2O). In contrast, miR-222-3p inhibition reduced autophagic vesicles and LC3 signal, confirming the essential role of miR-222-3p in autophagy activation during chondroprotection.

3.5. Effect of miR-222–3p on ATDC5 cells via targeting Rab1A

To identify potential targets of miR-222-3p, we performed bioinformatic analysis using miRDB, TargetScan, and DIANA-microT-CDS, which predicted 163 candidate genes (Fig. 3A). Among these, Rab1A was selected for further validation. A luciferase reporter assay confirmed the direct binding of miR-222-3p to the 3′-UTR of Rab1A. Co-transfection of miR-222-3p with the wild-type (WT) Rab1A 3′-UTR construct significantly reduced luciferase activity, whereas this effect was abolished when a mutated 3′-UTR was used (Fig. 3B and C).

Fig. 3.

Effect of miR-222-3p on IL-1β-induced ATDC5 cells by targeting Rab1A and regulating autophagy via the mTORC1 signaling pathway. (A) Venn diagram of miR-222-3p target genes predicted by miRDB, TargetScan, and DIANA-microT-CDS databases (B–C) Luciferase reporter assay to verify the direct binding of miR-222-3p to the 3′-UTR of Rab1A. (D) The relative protein level of Rab1A in ATDC5 cells after different treatments. (E) Quantification of the relative gray level of the Rab1A. (F–G) miR-222-3p or Rab1A expression in ATDC5 cells transfected with miR-222-3p mimics or/and Rab1A-overexpression plasmids. (H) CCK-8 assay to observe the proliferation of ATDC5 cells. (I) Representative images of scratch wound healing assay (scale bar = 60 μm). (J) Quantitative analysis of migration rates. (K–M) Representative immunofluorescence staining of Aggrecan, COL II, and MMP13 (scale bar = 25 μm). (N) Quantification of immunofluorescence. (O) mRNA expression levels of Aggrecan, COL II, MMP13, IL-1β, TNF-α, and IL-6 were detected by qRT-PCR. (P) The number of autophagic vacuoles was observed by TEM in the above groups (Scale bar = 6000 nm and 1000 nm). (Q) Representative immunofluorescence staining of LC3. (R–S) Representative Western blot images of p-mTOR, mTOR, p-S6K, S6K, p-4E-BP1, 4E-BP1, Beclin1, P62 and LC3 I/II with corresponding quantification. (T–V) Representative Western blot images of Rab1A, p-mTOR, mTOR, p-S6K, S6K, p-4E-BP1, 4E-BP1, Beclin1, P62 and LC3 I/II with corresponding quantification. All data are presented as mean ± standard deviation (SD) from at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001.

Furthermore, Western blotting revealed that miR-222-3p mimics downregulated Rab1A protein levels in ATDC5 cells, while its inhibitor had the opposite effect (Fig. 3D and E).

To investigate miR-222-3p′s role in cartilage repair through Rab1A regulation, we performed functional studies using OE-NC and Rab1A^OE plasmids. qRT-PCR analysis confirmed successful modulation of miR-222-3p levels (Fig. 3F) and Rab1A mRNA (Fig. 3G) expression across all experimental groups. The enhanced cell proliferation (Fig. 3H) and migration capacity (Fig. 3I and J) induced by miR-222-3p mimics were significantly suppressed when Rab1A was overexpressed. Immunofluorescence assays (Fig. 3K–N) showed that miR-222-3p increased the protein levels of COL II and Aggrecan while reducing MMP13, effects reversed by Rab1AOE. Consistently, qRT-PCR analysis (Fig. 3O) confirmed these matrix-related changes at the mRNA level, as well as a significant reduction in inflammatory cytokines (IL-1β, TNF-α, and IL-6). TEM imaging showed miR-222-3p mimics markedly increased autophagic vacuoles, an effect suppressed by Rab1AOE (Fig. 3P). Consistently, LC3 immunofluorescence assays confirmed miR-222-3p-dependent autophagy activation, which was attenuated upon Rab1A overexpression (Fig. 3Q). These findings demonstrate that miR-222-3p exerts its chondroprotective effects by targeting Rab1A.

3.6. Effect of exosomal miR-222-3p on autophagy via the mTORC1 signaling pathway

Previous studies have demonstrated that Rab1A regulates autophagy through two distinct mechanisms: suppression of autophagosome formation via reduction of phosphatidylinositol 3-phosphate (PI3P) and activation of mTORC1 signaling. Additionally, the results from the KEGG analysis of bioinformatics revealed significant enrichment of both the mTOR signaling pathway and autophagy pathways. These results may indicate that mTORC1 might serve as the crucial regulatory node through which exosomal miR-222-3p/Rab1A exerts its chondroprotective effects. Western blot analysis demonstrated that miR-222-3p significantly reduced phosphorylation levels of mTOR, S6K and 4E-BP1, increased the LC3 II/LC3 I ratio and Beclin1 expression, and decreased P62 protein levels compared to control groups. However, these effects were significantly attenuated by the inhibition of miR-222-3p(Fig. 3R and S). Moreover, Rab1A overexpression completely abolished all miR-222-3p-induced alterations in these autophagy-related proteins (Fig. 3T–V). Therefore, we conclude that the H-Exos/miR-222-3p/Rab1A axis promotes chondroprotection through mTORC1 suppression and autophagy activation.

3.7. H-Exos modulate macrophage polarization via the NF-κB pathway

To investigate whether H-Exos influence the immune microenvironment, we first confirmed that macrophages effectively internalized both N-Exos and H-Exos (Fig. 4A). RNA sequencing of macrophages treated with H-Exos versus N-Exos revealed significant differences in their transcriptional profiles (Fig. 4B–D). KEGG and GSEA enrichment analyses identified the NF-κB signaling pathway as one of the most significantly altered processes (Fig. 4E and F). We therefore hypothesized that H-Exos regulate macrophage polarization. Immunofluorescence staining and qRT-PCR analysis demonstrated that H-Exos treatment significantly downregulated the M1 marker CD86 and the pro-inflammatory cytokine TNF-α, while upregulating the M2 markers CD163 and ARG1 (Fig. 4G–J). Furthermore, western blot analysis confirmed that H-Exos inhibited the activation of the NF-κB pathway, as evidenced by decreased phosphorylation of IκBα and P65 (Fig. 4K and L). Collectively, these results indicate that H-Exos are internalized by macrophages and promote a shift from pro-inflammatory M1 toward anti-inflammatory M2 polarization, likely through suppression of the NF-κB signaling pathway.

Fig. 4.

H-Exos regulate macrophage polarization via the NF-κB signaling pathway. (A) Cellular internalization of N-Exos and H-Exos by macrophages. (scale bar = 20 μm). (B) Column charts. (C) Volcano plot of differentially expressed genes. (D) Heatmap of differentially expressed genes. (E) KEGG pathway enrichment analysis. (F) GSEA analysis. (G-H) Representative immunofluorescence staining of CD86 and CD163 (scale bar = 20 μm). (I) Quantification of immunofluorescence intensity for CD86 and CD163. (J) mRNA expression levels of CD86, TNF-α, CD163, and ARG1 detected by qRT-PCR. (K) Western blot analysis of NF-κB signaling pathway proteins (p-IκBα, IκBα, p-P65, and P65). (L) Quantitative analysis of Western blot results. All data are presented as mean ± standard deviation (SD) from at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001.

3.8. Characterization of GelMA/H-Exos hydrogel

¹H NMR analysis confirmed the successful methacrylation of gelatin, with a degree of methacrylation of ∼75% (Fig. S8).Through UV crosslinking of GelMA with incorporated exosomes, we successfully fabricated stable GelMA/H-Exos composite hydrogels. SEM analysis revealed that the hydrogels exhibited uniform porous structures, enabling high-density exosome loading (Fig. 5A). The vial inversion test confirmed that the hydrogels possess excellent mechanical strength and self-supporting capability (Fig. 5B). Both GelMA and GelMA/H-Exos hydrogels showed similar swelling kinetics and reached equilibrium swelling within comparable timeframes (Fig. 5C). The degradation profiles of both hydrogels also showed no significant differences throughout the evaluation period (Fig. 5D). Rheological characterization revealed that the storage modulus (G′) was consistently higher than the loss modulus (G″) for both GelMA and GelMA/H-Exos hydrogels, indicating their dominant elastic characteristics. These results collectively demonstrate that the incorporation of exosomes preserves the intrinsic physicochemical characteristics of the GelMA hydrogel network (Fig. 5E). The GelMA/H-Exos composite exhibited sustained release of exosomes over an extended period, with a cumulative release rate gradually reaching 83-86% by the end of the study (Fig. 5F). This prolonged release profile ensures consistent delivery of therapeutic exosomes, providing effective and long-lasting treatment for OA during cartilage repair. In vitro evaluation using live/dead staining and CCK-8 assays demonstrated that GelMA/H-Exos hydrogels exhibited no detectable cytotoxicity while significantly enhancing cellular proliferation during the 3-day culture period (Fig. 5G and H). In vivo histological examination of major organs (heart, liver, spleen, lung, and kidney) after 4-week implantation revealed well-preserved tissue architecture without observable inflammatory responses or pathological alterations (Fig. 5I).

Fig. 5.

Characterization of GelMA/H-Exos hydrogels. (A) SEM images of exosomes adhered to GelMA/H-Exos hydrogels. (B) Vial inversion test performed to evaluate gel stability and self-supporting capability. (C) Swelling behavior of GelMA/H-Exos hydrogels measured over time. (D) Degradation profiles of GelMA/H-Exos hydrogels in PBS. (E) Rheological properties of GelMA/H-Exos hydrogels characterized by storage modulus (G′) and loss modulus (G″). (F) Cumulative release profile of exosomes from GelMA/H-Exos hydrogels. (G) Live/dead staining of ATDC5 cells cultured on GelMA and GelMA/H-Exos hydrogels for 3 days (Scale bar = 500 μm). (H) Quantitative analysis of cell proliferation on GelMA and GelMA/H-Exos hydrogels using CCK-8 assay. (I) H&E staining of major organs (heart, liver, spleen, lung, kidney) across treatment groups (Scale bar = 20 μm). Data are presented as mean ± SD of at least three replicates. *p < 0.05, **p < 0.01, ***p < 0.001.

To further evaluate immunological safety, RAW264.7 cells–hydrogel interactions were assessed after 24 h phalloidin and Live/Dead staining indicated comparable adhesion morphology and good macrophage viability on GelMA and GelMA/H-Exos (Fig. S9–S10). ELISA of culture supernatants showed no significant differences between the Control and GelMA groups, whereas GelMA/H-Exos significantly reduced cytokine levels (TNF-α, IL-6, and IL-10) compared with GelMA (Fig. S11A–C). Consistently, RT-qPCR revealed increased CD163 and decreased CD86 expression in the GelMA/H-Exos group, demonstrating an attenuated macrophage activation with a shift toward an M2-like phenotype (Fig. S11D–E).

3.9. Therapeutic efficacy of GelMA/H-Exos in vivo

At 8 weeks post-surgery, the GelMA/H-Exos + AntagomiR-NC group showed well-preserved articular surfaces, while the therapeutic effect was significantly attenuated in the GelMA/H-Exos + AntagomiR-222-3p group, demonstrating that miR-222-3p inhibition partially abolished the cartilage-protective effects of H-Exos (Fig. 6B).

Fig. 6.

AntagomiR-222-3p reverses the therapeutic effect of H-Exos on OA in rats. (A) Experimental workflow and therapeutic evaluation of H-Exos in rat OA model. (B) Representative macroscopic images of knee joint articular surfaces. (C-E) Representative histological images of cartilage repair evaluated by Safranin O/Fast Green, Masson's trichrome, and H&E staining (Scale bar = 100 μm and 50um). (F) Quantitative ICRS scores for each group. (G) OARSI histological scores for each group. (H) Relative content of cartilage matrix for each group. (I) Representative immunofluorescence images of COL II (scale bar = 40 μm). (J) Quantification of immunofluorescence intensity for COL II. Data are presented as mean ± SD (n = 6 per group). *p < 0.05, **p < 0.01, ***p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The results of Safranin O/Fast Green, Masson and H&E histological staining revealed distinct patterns of cartilage degeneration among groups (Fig. 6C–E). The OA + AntagomiR-NC group exhibited severe articular cartilage degeneration, characterized by extensive tissue loss, marked depletion of ECM extending to the calcified zone, and evident subchondral bone erosion. The GelMA + AntagomiR-NC group demonstrated partial ECM preservation but showed irregular matrix distribution and focal surface defects. In the GelMA/H-Exos + AntagomiR-222-3p group, ECM loss was primarily observed in superficial regions. By contrast, the GelMA/H-Exos + AntagomiR-NC group maintained well-organized cartilage structure with homogeneous ECM distribution and uniform chondrocyte arrangement. Consistent with these findings, semi-quantitative analysis showed the GelMA/H-Exos + AntagomiR-NC group achieved the highest ICRS scores (Fig. 6F), the lowest OARSI scores (Fig. 6G), and the greatest relative cartilage matrix content (Fig. 6H). Moreover, immunofluorescence confirmed that the strong preservation of COL II in this group was significantly diminished by miR-222-3p inhibition (Fig. 6I and J), collectively demonstrating that the cartilage-protective effects of H-Exos are partially dependent on miR-222-3p.

Micro-CT images revealed distinct subchondral bone alterations across treatment groups (Fig. 7A). The GelMA/H-Exos + AntagomiR-NC group maintained near-normal articular architecture with smooth surfaces and minimal osteophytes. In contrast, the OA + AntagomiR-NC group exhibited severe surface erosion and prominent osteophytosis. The GelMA/H-Exos + AntagomiR-222-3p group showed intermediate degeneration, characterized by mild surface irregularities and scattered osteophytes.

Fig. 7.

Therapeutic effects of H-Exos through the miR-222-3p/Rab1A axis and macrophage regulation in rats with osteoarthritis. (A) Representative 3D and coronal micro-CT images of treatment groups. Red arrows indicate osteophytes. (B) Representative immunofluorescence images of p-mTOR, p-S6K, Beclin1 and LC3-II expression in articular cartilage. Scale bar = 20 μm. (C–F) Quantification of immunofluorescence intensity for p-mTOR, p-S6K, Beclin1 and LC3-II. (G) Heat map depicting the relative expression levels of p-mTOR, p-S6K, Beclin1 and LC3-II. (D) Representative immunohistochemistry of iNOS and CD206 (scale bar = 20 μm). (E–F) Quantification of immunohistochemistry. Data are presented as mean ± SD (n = 6 per group). *p < 0.05, **p < 0.01, ***p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Immunofluorescence analysis of autophagy signaling revealed that GelMA/H-Exos + AntagomiR-NC treatment significantly decreased the expression of p-mTOR and p-S6K while increasing the levels of Beclin1 and LC3-II (Fig. 7B). Quantitative analysis and a corresponding heatmap confirmed these trends (Fig. 7C–G), demonstrating the activation of autophagy. These effects were attenuated by AntagomiR-222-3p treatment, confirming miR-222-3p′s role in coordinating mTOR inhibition and autophagy activation during cartilage repair.

Furthermore, immunohistochemical evaluation of synovial macrophages showed that H-Exos treatment significantly reduced the M1 marker iNOS while promoting the M2 marker CD206 (Fig. 7H–J), indicating that H-Exos modulated synovial macrophage polarization toward an anti-inflammatory phenotype in vivo.

4. Discussion

OA is a multifactorial degenerative joint disorder characterized by progressive cartilage degradation, subchondral bone remodeling, and synovial inflammation, resulting in joint dysfunction and pain. The limited self-repair capability of cartilage poses significant challenges to the development of effective disease-modifying therapies for OA. Our study demonstrates that H-Exos exhibited greater biological activity than N-Exos. Specifically, H-Exos significantly enhanced proliferation, migration, and ECM anabolism, while inhibiting catabolism and inflammation response of ATDC5 cells. Mechanistically, we explored the role of miR-222-3p and its target Rab1A in regulating the mTORC1 signaling pathway and activating autophagy. Moreover, H-Exos modulated the immune microenvironment by promoting macrophage polarization toward an anti-inflammatory phenotype via the NF-κB pathway. In vivo, the cartilage-protective and bone-stabilizing effects of GelMA/H-Exos were mediated by miR-222-3p, as evidenced by their significant attenuation upon AntagomiR-222-3p administration. Additionally, H-Exos treatment additionally promoted anti-inflammatory macrophage polarizatio, contributing to a comprehensive therapeutic outcome. To our knowledge, this is the first study to investigate the therapeutic efficacy and underlying mechanisms of hypoxia-preconditioned CPC-Exos for OA treatment, demonstrating their dual role in promoting chondrocyte function and modulating immune responses, thus providing a promising strategy for cartilage regeneration and disease modification.

Stem cells derived from diverse sources, including bone marrow, embryonic tissue, and adipose tissue, have emerged as valuable candidates in tissue engineering due to their potent self-renewal capacity and multilineage differentiation potential. CPCs, a specialized type of stem cell, demonstrate superior chondrogenic differentiation capacity, enhanced migratory potential, and, importantly, maintain a more stable chondrogenic phenotype compared to BMSCs and other MSCs [[15], [16], [17]]. Therefore, CPC‐based cell therapy has great potential for the treatment of cartilage degeneration. In recent years, cell-free exosome therapy has gained considerable attention in regenerative medicine, as exosomes offer advantages such as low immunogenicity, high stability, and absence of tumorigenicity compared to stem cell transplantation [19,44]. Despite this, research on CPC-Exos for OA treatment remains limited. Thus, this study focuses on investigating the therapeutic potential of CPC-Exos, particularly those derived under hypoxic preconditioning, aiming to provide new insights into cell-free therapies for cartilage regeneration.

Exosomes serve as critical carriers and conveyors of genetic material, proteins, and lipids, mediating intercellular communication and modulating recipient cell behavior [45]. Consequently, regulating exosome biogenesis, release, cargo sorting, and uptake has emerged as a promising strategy to intervene in disease progression. Hypoxia has been reported to not only increase the quantity of exosome release but also alter their cargo composition [45]. Recent studies have demonstrated that exosomes derived from hypoxia-preconditioned stem cells significantly enhance chondrocyte function, reduce inflammation, and promote cartilage repair [29,31,46]. In the present study, we observed no significant differences in morphology, size distribution, or surface markers between H-Exos and N-Exos, indicating that hypoxic culture did not alter their basic physical characteristics. However, H-Exos demonstrated markedly enhanced biological activity both in vitro and in vivo, evidenced by their superior capacity to promote ATDC5 cell proliferation, migration, and ECM anabolism, as well as their more potent suppression of catabolism under inflammatory conditions. These findings underscore that hypoxia enhances the therapeutic potential of CPC-Exos by modulating their bioactive cargo.

Among exosomal components, microRNAs (miRNAs) represent a major functional cargo that regulates recipient cell behavior primarily through post-transcriptional gene regulation. Emerging evidence indicates that hypoxic preconditioning can alter the miRNA profile of exosomes, thereby enhancing their therapeutic potential [29,47,48]. In this study, high-throughput sequencing revealed a distinct miRNA expression signature in exosomes derived from hypoxia-treated CPCs, with miR-222-3p significantly upregulated. Several reports have implicated a protective role of miR-222-3p in OA. For instance, miR-222-3p has been shown to suppress IL-1β-induced chondrocyte injury [49]. A clinical study reported that circulating miR-222-3p levels were inversely correlated with radiographic severity in patients with hand OA, suggesting a potential protective role in disease progression [50]. Notably, Wang et al. reported that miR-222-3p was upregulated in extracellular vesicles derived from CPCs, and suggested that miR-222-3p may play a critical role in promoting chondrocyte proliferation and migration [51]. Consistent with these findings, our study found that H-Exos mediate their therapeutic effects primarily via miR-222-3p, which promotes proliferation, migration, and ECM synthesis while inhibiting catabolic processes in ATDC5 cells.

miRNAs exert their regulatory functions primarily by binding to the 3′-untranslated region (3′-UTR) of target mRNAs, thereby modulating gene expression at the post-transcriptional level [10,52]. Through integrated bioinformatic analysis, Rab1A emerged as a potential target of interest. Notably, recent studies have demonstrated that Rab1A, a member of the Rab GTPase family predominantly localized in the endoplasmic reticulum and Golgi apparatus, accelerates OA progression by inhibiting autophagy via activation of the mTORC1-S6K signaling pathway [53,54]. Rab1A facilitates vesicle transport between the endoplasmic reticulum and Golgi and has been shown to activate mTORC1 signaling at the Golgi in response to amino acid stimulation, while concurrently impeding autophagic vacuolesformation by reducing phosphatidylinositol 3-phosphate (PI3P) biogenesis [53,55]. Consistent with these findings, our KEGG pathway enrichment analysis revealed that the predicted target genes of the differentially expressed miRNAs were predominantly enriched in autophagy-related and mTOR signaling pathways, among other pathways. Based on these data, Rab1A was selected as a candidate target gene. Further validation by a dual-luciferase reporter assay confirmed that Rab1A is a direct target of miR-222-3p. Our results demonstrate that miR-222-3p exerts chondroprotective effects by downregulating Rab1A, thereby promoting ATDC5 cell proliferation, migration, and ECM anabolism, while suppressing catabolism.

Autophagy is a conserved cellular degradation process that maintains cellular homeostasis by eliminating damaged organelles and misfolded proteins [56]. It is implicated in various degenerative diseases and has been shown to regulate epigenetic modifications, metabolic processes, stem cell differentiation, and apoptosis [56,57]. Increasing evidence suggests that autophagy plays a critical role in protecting chondrocytes from degradation, and its dysregulation is closely associated with the pathogenesis of OA [[58], [59], [60]]. The mechanistic target of mTOR kinase plays a central role in the regulation of autophagy. It forms two distinct multi-protein complexes, mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2), with mTORC1 being the key regulator of autophagy. Accumulating evidence indicates that mTORC1 acts as a major negative regulator of autophagy and integrates various upstream signals to modulate autophagic activity [61]. Consequently, inhibition of mTORC1 to activate autophagy has emerged as a promising therapeutic strategy for OA [62]. Our results demonstrated that hypoxia-derived exosomal miR-222-3p enhances autophagy in ATDC5 cells by directly targeting Rab1A, leading to the inhibition of mTORC1 signaling.

Beyond their chondroprotective effects, H-Exos demonstrated remarkable capacity to reprogram the synovial immune microenvironment—a critical advance given the established role of macrophage polarization in OA progression [33]. While current literature emphasizes that synovial macrophages exist along a spectrum of functional states and their imbalance drives chronic inflammation, our findings provide a tangible therapeutic strategy to correct this imbalance. The observed shift toward an anti-inflammatory phenotype, characterized by suppressed M1 markers and enhanced M2 marker, coupled with NF-κB pathway inhibition, suggests H-Exos effectively modulate the core signaling axis governing macrophage polarization. This immunomodulatory function, achieved through a cell-free approach, is particularly significant considering recent challenges in macrophage-targeted therapies [63]. Unlike strategies involving macrophage depletion, H-Exos offer a more nuanced immunomodulation, potentially avoiding such pitfalls. Furthermore, the use of GelMA hydrogel for localized H-Exos delivery aligns with emerging biomaterial-based approaches for precise immune microenvironment regulation. This multifaceted strategy, simultaneously addressing cartilage degradation and synovial inflammation, represents a promising direction for next-generation OA therapies.

This study has several limitations. First, among the nine upregulated miRNAs identified, only miR-222-3p was selected for functional validation, leaving the potential contributions of the other miRNAs unexamined. Second, specific agonists or inhibitors of the mTOR pathway were not employed to more precisely delineate the mechanistic link between autophagy and mTOR signaling. Third, we did not include a free exosome injection group as a control in our in vivo experiments. The focus of this study is to evaluate GelMA-exosome composite delivery systems, though it may limit our ability to assess the effect of free exosomes on cartilage repair. Free exosomes are rapidly cleared from the joint cavity after injection, significantly reducing their therapeutic efficacy compared to exosome-loaded hydrogels.

5. Conclusion

This study systematically elucidates the dual mechanisms of H-Exos in osteoarthritis treatment. In chondrocytes, H-Exos deliver miR-222-3p to target Rab1A, inhibiting mTORC1 signaling and activating autophagy, thereby enhancing proliferation and extracellular matrix anabolism while suppressing catabolism. Simultaneously, H-Exos modulate the synovial immune microenvironment by promoting macrophage polarization toward an anti-inflammatory phenotype via NF-κB pathway inhibition. Delivered via an injectable GelMA hydrogel, this integrated strategy significantly attenuates OA progression by concurrently addressing cartilage degradation and synovial inflammation. These findings establish H-Exos as a promising cell-free therapeutic with dual regenerative and immunomodulatory functions.

Ethics approval

All animal experimental protocols were approved by the Animal Ethics Committee of Southwest Medical University (20240220-012).

CRediT authorship contribution statement

Peng Zhou: Data curation, Formal analysis, Investigation, Methodology, Software, Writing – original draft. Xu Liu: Data curation, Investigation, Resources. Lingzhi Li: Project administration, Software, Validation. Yimin Du: Validation, Visualization. Juncai Liu: Resources, Software, Validation. Xiangtian Deng: Investigation, Methodology, Validation. Zan Chen: Resources, Software, Validation. Shiyi Chen: Resources, Validation. Zhong Li: Supervision, Validation, Visualization, Writing – review & editing. Zheng Li: Data curation, Methodology, Writing – review & editing. Yanwei He: Funding acquisition, Project administration, Supervision, Writing – review & editing.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.