The present study was conducted in compliance with the Declaration of Helsinki. The clinical specimens were obtained between July and October 2017 in Zhejiang Provincial People' Hospital, People's Hospital of Hangzhou Medical College. Informed written consent from all the participants was obtained. ADSCs were collected and isolated from adipose tissue during elective liposuction surgery of a healthy donor. The activated SFs were isolated from an obese patient diagnosed with OA in middle-aged male subjects (55–70 yrs). Adipose tissue (~1.5 g) was harvested from the subcutaneous adipose tissue and washed with PBS. The tissue was cut into strips and digested with collagenase (final concentration 1 mg/ml in 25 ml PBS) at 37°C for 45 min, after which 25 ml of DMEM (Invitrogen; Thermo Fisher Scientific, Inc.) was added to neutralize collagenase activity. The digested tissues were then filtered with a 0.22 µm filter and centrifuged at 800 × g for 6 min at 25°C and the supernatant was discarded. The resulting pellet contained ADSCs. ADSCs were seeded at 5×10⁴ cells/cm² in 60 cm² tissue culture dishes and cultured with DMEM containing 10% FBS (Invitrogen; Thermo Fisher Scientific, Inc.), 100 units/ml penicillin and 100 µg/ml streptomycin. After 1 week of culture, the cells were harvested with 0.25% trypsin-EDTA, centrifuged 800 × g at 25°C for 6 min n and washed twice with PBS. The ADSCs were then used for co-culture assays. To evaluate the multipotent potential of the ADSCs, the ADSCs were cultured under inductive conditions, where osteogenesis and chondrogenesis were promoted, according to an established protocol (13). The cell lineage and differentiation state were evaluated using immunohistochemical staining and quantitative PCR reactions.

The isolation and culture of synovial tissues was performed using a previously established method (14). In brief, synovial tissues were minced, digested for 30 min at 37°C in PBS containing 0.1% trypsin, followed by digestion with 0.1% collagenase in DMEM with 10% FBS for 1 h. The suspension was filtered and centrifuged 800 × g at 25°C for 10 min. The pellet collected contained the cells of interest. The cells were cultured for 7 days in DMEM supplemented with 10% FBS, 25 mM HEPES, 100 U/ml penicillin, 100 µg/ml streptomycin and 2.5 µg/ml amphotericin B (Gibco; Thermo Fisher Scientific, Inc.). The non-adherent cells were carefully removed. Subsequently, the cells were harvested, resuspended in DMEM/Ham's F12 medium containing 10% human allogenic serum, plated in cell culture dishes (diameter = 15 cm), and allowed to attach for about 4–6 days. Moreover, primary periosteal cells were isolated from four independent donors. In brief, the periosteal flap was rinsed with Hanks solution (Biochrom) three times, minced and digested for 3 h in Dulbecco's modified eagle medium (DMEM)/Ham's F12 medium (Biochrom) containing 10,000 U/ml collagenase II (Biochrom), 10% human allogenic serum (German Red Cross), 2.5% Hepes (Biochrom) and 1% penicillin/streptomycin solution (Biochrom). Subsequently, the cells were harvested, resuspended in DMEM/Ham'sF12 medium containing 10% human allogenic serum, plated in cell culture dishes (diameter = 15 cm), and allowed to attach for about 4–6 days. Phenotype analysis of the expression of synovial fibroblasts markers, as well as that of synovial fibroblasts features previously reported at a tissue level, was conducted by flow cytometry in synovial fibroblasts, either negatively isolated from primary culture or obtained from conventional fourth passage.

Briefly, adipose tissues were digested by collagenase using the EpiQuik Whole Cell Extraction kit (Invitrogen; Thermo Fisher Scientific, Inc.). The stromal vascular fractions (SVFs) were then resuspended into MSCs medium (Invitrogen; Thermo Fisher Scientific, Inc.). These cells were then sub-cultured to the 5th passage and used for experiments.

For exosome isolation, ADSCs between passage 4 and 8 were expanded in traditional 2D adherent cell culture and then seeded on BioNOC II micro carriers (Sigma-Aldrich, Merck KGaA). For seeding, 8×10⁶ cells resuspended in 50 ml of DMEM with 10% FBS, 1% L-glutamine and 1% P/S and layered on top of 2 g of sterilized BioNOC II in a 250 ml Erlenmeyer flask (CLS431144, Sigma-Aldrich, Merck KGaA). The cells were maintained in static incubation during the first 16 h and were then supplemented with additional 200 ml of medium. The supernatant was ultracentrifuged using a W32Ti rotor (L-80XP; Beckman Coulter, Inc.) at 110,000 × g for 70 min to pellet the exosomes. The pellet was washed in PBS and centrifuged for a second time at 110,000 × g for 70 min. The PBS was removed and the exosomes were re-suspended in 100 µl nuclease-free water. All centrifugation key steps were performed at 4°C.

Adipose-derived stem cell-derived Exos suppress inflammatory markers in activated SFs. The concentration of TNF-α and IL-10 levels in chondrocytes was detected with an ELISA kit (DTA00C and D1000B, R&D Systems) according to the manufacturer's instructions.

Patient-derived articular chondrocytes were isolated and seeded at a density of the 5,000 cells/well in 12-well plate and cultured for 5–7 days until confluency was reached. Chondrocytes were exposed to 100 µM H₂O₂ with or without exosomes (1, 5 and 10×10¹⁰ particles/ml) for 1 h at 37°C. The resultant cells were then subjected to further analysis. The effect of chondrocytes and exosomes on cell apoptosis was determined using an Annexin V-FITC Apoptosis Detection kit [cat. no. 1035-100, Hangzhou Multisciences (Lianke) Biotech Co., Ltd.] and flow cytometry. Annexin V-FITC binding and PI staining [cat. no. 1035-100, Hangzhou Multisciences (Lianke) Biotech Co., Ltd.] were detected using a flow cytometer and analyzed with FACSDiva v8.0.3 (Becton-Dickinson) software. The experiments were performed independently three times.

RNA samples were extracted and quantified using the Qubit HS RNA kit (cat. no. 2019-0023, Thermo Fisher Scientific, Inc.) with a Qubit 3.0 Fluorometer (v2.0.3 software, Thermo Fisher Scientific, Inc.), following the manufacturer's instructions. The synthesis of complementary DNA was performed using the PrimeScript™ RT reagent kit [cat. no. 2312, Hangzhou Multisciences (Lianke) Biotech Co., Ltd.]. RT-qPCR was performed using a human miRNA RT-qPCR detection kit specifically for Homo sapiens (hsa)-miR-145, hsa-miR-221, hsa-miR-194, hsa-miR-101 and RNA U6 small nuclear 6 pseudo-genes (RNU6B; cat. no. 33456, Guangzhou RiboBio Co., Ltd.). RNU6B was used for normalization of miRNA expression levels. qPCR was performed using SYBR^® and GAPDH was used for normalization of the results. The sequences of target gene were retrieved from GenBank (http://www.ncbi.nlm.nih.gov/genbank/) and miRBase (http://www.mirbase.org/). The primers were designed using Primer Designer 2.0, and the sequences are shown below. The miR-specific primers were: hsa-miR-145 AGCGAGTGCAGTGGTGAAA, hsa-miR-221GAGTGGGGGTGGGACATAAA, hsa-miR-3188 GTGCGGATACGGGGAAAA, hsa-miR-194 TGGGCTGGGTTGGGAAA and hsa-miR-101 GGCCCAGTGGGGGGAA. The PCR conditions: 95°C pre-denaturation for 15 min, followed by 40 cycles of denaturation at 94°C for 15 sec, annealing at 55°C for 30 sec and extension at 70°C for 30 sec. Relative gene expression was analyzed using the 2^−ΔΔCq method. U6 small nuclear RNA was used as a miRNA internal control. miRNA and U6 (cat. no. HmiRQP9001) primers were purchased from iGeneBio (GeneCopoeia, Inc.). All primers were synthesized by Beijing Liu He Synthetic Genomics Ltd. U6 was used as the internal control.

All western blotting analyses were performed according to established protocols (13). The following primary antibodies were obtained from Cell Signaling Technology, Inc.: Interleukin (IL)-6 (cat. no. 8904, dilution, 1:1,000), nuclear factor-κB (NF-κB) (cat. no. D14E12, dilution, 1:2,000), tumor necrosis factor-α (TNF-α) (cat. no. D2D4, dilution, 1:1,500), IL-10 (cat. no. D13A11, dilution, 1:2,000) and β-actin (cat. no. 12E5, dilution, 1:5,000). The transcription factor SRY-box transcription factor 9 (Sox9) primary antibody was obtained from Abcam (cat. no. D8G8H, dilution, 1:1,000). Collagen type II (Col II cat. no. E819H, Abcam, dilution, 1:1,000) was analyzed using 5% SDS-PAGE gels with an anti-Col II primary antibody (cat. no. E819H, Abcam, dilution, 1:1,000). Next, the membrane was washed and incubated with goat anti-rabbit secondary antibody (cat. no. 33456, Abcam, dilution 1:5,000). Finally, an enhanced chemiluminescence (ECL, Biovision) kit was used to observe protein bands in a ChemiDoc XRS Plus luminescent image analyzer (Bio-Rad Laboratories, Inc.). β-actin was used as the internal control in the experiment, and therefore, the ratio of the integrated optical density of the targeted protein bands to that of β-actin in every group was determined as the final relative protein expression intensity. All experiments were replicated 3 times.

Osteogenic differentiation was assessed using alcian blue staining. Briefly, tissue was embedded in paraffin and sections (4.0 µm thick) were mounted onto glass slides, cleared using xylene, and gradually rehydrated. Antigen retrieval was performed on slides using sodium citrate buffer (10 mM sodium citrate and pH 6) at 100°C for 20 min. Slides were cooled and rinsed in distilled water to remove citrate solution. Slides were blocked using Super Block (Thermo Fisher Scientific, Inc.) for 1 h at room temperature. For slides stained with alcian blue staining, sections were stained with alcian blue solution (Sigma-Aldrich, Merck KGaA) and Weigert's Iron Hematoxylin (Sigma-Aldrich, Merck KGaA) according to the manufacturer's protocol. Alcian blue Stain kit (Connective Tissue Stain; Abcam) was also used to stain sections according to the manufacturer. For slides stained with hematoxylin and eosin (H&E), sections were cleared and rehydrated followed by staining with H&E, according to standard protocols. Sections were then dehydrated gradually and mounted for imaging. All images were acquired using a Nikon Digital Imaging System.

After removing the culture medium, the cells were rinsed twice with PBS and fixed with 95% ethyl alcohol and 5% isopropyl alcohol for 2 h at 4°C. Subsequently, the cells were rinsed twice with deionized water. The cells were treated with 2.5% silver nitrate solution in deionized water at room temperature for 24 h, rinsed twice with deionized water, and 1 ml freshly prepared 0.5% hydroquinone solution in deionized water was applied to the cell layers on microscope slides and incubated for 2 min in the dark at room temperature. The hydroquinone solution was removed and 1 ml freshly prepared 5% sodium thiosulfate solution in deionized water was added for 2 min at room temperature. The cells were washed again with deionized water. The staining was semi-quantitatively evaluated under a light microscope using the color intensity of the Von Kossa staining. Under a microscope, measurements were made in 10 consecutive visual areas at 400× magnification.

The effect of exosome stimulation on chondrocytes was measured using the 5-ethynyl-2′-deoxyuridine (EdU)-488 Cell Proliferation kit (Guangzhou RiboBio Co. Ltd.) and flow cytometry, according to the manufacturer's instructions. Briefly, normal chondrocytes were seeded into 48-well plates and cultured with exosomes (400 µg/ml) for 48 h. EdU working solution, consisting of 150 µl complete culture medium for chondrocytes supplemented with 0.15 µl EdU, was added per well and incubated at 37°C for 3 h. The cells were then trypsinized, washed with PBS, fixed in 4% paraformaldehyde (PFA) for 15 min at room temperature, neutralized with 2 mg/ml glycine for 2 h and washed again with PBS. Permeabilization was performed using 0.4% Triton X-100 for 5 min at room temperature and the cells were washed twice with PBS. The labelled chondrocytes were resuspended in the staining solution from the kit, incubated for 10 min at room temperature in the dark, washed twice with 0.4% Triton X-100 and then resuspended in PBS for flow cytometry analysis and analyzed with FACSDiva v8.0.3 software (Becton-Dickinson).

A chondrocyte migration assay was performed using Transwell inserts. In total, 5×10⁴ chondrocytes were seeded into the upper chambers of the Transwell (8 µm pores) in 10% FBS chondrocyte culture DMEM media, while the lower chamber was filled with 600 µl chondrocyte culture medium containing exosomes (400 µg/ml). The chondrocytes were allowed to migrate at 37°C for 12 h. The non-migrated chondrocytes remaining on the upper surface of the insert were carefully removed. The migrated chondrocytes on the bottom surface of the insert were fixed with 4% PFA for 15 min at room temperature and stained with 0.5% crystal violet for 10 min at room temperature, before washing with PBS. In total, five fields were randomly selected and photographed (magnification, ×100) using a Leica light microscope (Leica Microsystems, Inc.).

Statistical analysis was performed with SPSS v11.0 (SPSS, Inc.). Data represents assays performed a minimum of three times in triplicates. Data are presented as the mean ± SD. Comparisons of macroscopic and histological scores were performed using the Mann-Whitney U test. Comparison made among multiple groups were performed using ANOVA and Tukey's post-hoc test. P<0.05 was considered to indicate a statistically significant difference.

The multipotency of ADSCs obtained from an obese patient with OA was assessed. ADSCs were cultured under different inductive conditions. For example, under osteogenic conditions, ADSCs differentiated along the osteogenic lineage, as shown by the increased expression of runt-related transcription factor 2 (Runx2) mRNA (~1,000-fold higher than the control) and the von Kossa staining intensity of deposited calcium phosphate. (Fig. 1A and B). The increased Col II mRNA (~250-fold higher than the control) and Alcian Blue staining was a method to detect the glycosaminoglycans in cells that were indicators of chondrogenic differentiation (Fig. 1A and B). These data provided ex vivo evidence that the primary ADSCs obtained in the present study were multipotent.

Secreted exosomes have been reported to be involved in communication among cells and shown to participate in a wide spectrum of cellular processes, including the pathological development and progression of arthritis (5,6). In the present study, secreted exosomes were isolated from ADSCs and incubated with activated SFs obtained from an obese patient with OA. The results demonstrated that exosome-treated SFs exhibited significantly reduced levels of inflammatory biomarkers, namely IL-6, TNF-α and NF-κB, while there were increased levels of IL-10 (Fig. 2A and B). The microenvironment was also assessed by examining the synovial macrophages. It was found that ADSC-Exos had immune-suppressive effects on the microphages; the M1 type macrophage marker TNF-α was significantly downregulated by treatment with ~10¹⁰ exosomes particles/ml, while IL-10 was upregulated (Fig. 2C and D).

After establishing that exosomes from ADSCs reduced inflammation in SFs, their actions on articular chondrocytes were investigated. Primary articular chondrocytes were cultured and subjected to H₂O₂ insults in order to mimic the oxidative stress induced by M1 macrophages in the arthritic synovium (9,10). H₂O₂ treatment induced apoptosis in the articular chondrocytes. When a sufficient number of exosomes were added to the articular chondrocytes treated with H₂O₂, significantly less apoptotic chondrocytes were detected (Fig. 3A). For example, ~37% of chondrocytes treated with exosomes (Exo 10 group; 10×10¹⁰ particles/ml) survived H₂O₂ treatment (Fig. 3A). In addition, the effects of ADSC-derived exosomes on the migration of MSCs was assessed. The migratory ability of MSCs appeared to increase in a dose-dependent manner with increasing exosome treatment (Fig. 3B). The morphogenic effects of ADSC-Exos on MSCs were then investigated. It was found that MSCs incubated with exosomes had a higher differentiation potential, including osteogenesis, adipogenesis and chondrogenesis. This was reflected by the increased mRNA expression of peroxisome proliferator-activated receptor-γ (PPARγ; a transcription factor involved in adipocyte cells differentiation), Col II (a marker of chondrocyte differentiation) and Runx2 (an osteogenesis marker). Runx2 was the most elevated in the MSCs (~1,000-fold higher) following incubation with exosomes (Fig. 3C).

As ADSC-Exos induced Runx2 mRNA expression in MSCs, the chondrogenesis-promoting effect of exosomes on periosteal cells (precursor cells) was investigated. ADSC-Exo incubation led to increased Alcian Blue staining in a time-dependent manner (Fig. 4A). Periosteal cells exhibited chondrocyte morphology, with a substantial Alcian Blue staining intensity, after 10 days of incubation (Fig. 4A). The increased Alcian Blue staining was accompanied by the increased mRNA expression of chondrogenesis markers, namely Sox9, Col II and β-catenin (Fig. 4b). The increased β-catenin mRNA in the exosome-treated periosteal cells was consistent with previous studies that indicated that the Wnt/β-catenin/T cell factor-mediated transcription process served an important role in chondrocytic differentiation and proliferation (11,12). To test this hypothesis, ICG-001, an antagonist of Wnt/β-catenin signaling, was introduced along with the exosome treatment. The presence of ICG-001 appeared to completely prevent the periosteal cells from differentiating into chondrocytes, as reflected by the absence of Alcian Blue stain (Fig. 4C). Consistently, the RT-qPCR analysis of the ICG-001+exosome-treated periosteal cells revealed that the expression of markers of chondrogenesis (Sox9, Col II and β-catenin) was effectively reversed (Fig. 4D).

After observing that the ADSC-derived exosomes induced the expression of pro-chondrogenic markers, the underlying mechanism of action was investigated. Several established miRNAs responsible for promoting chondrocyte differentiation and proliferation were screened. Among these miRNAs, miR-145 and miR-221 were increased following ADSC-Exo treatment, while no significant changes were identified in the levels of miR-101 and miR-194 (Fig. 5A). The effects of miR-221 and miR-145 on periosteal cells were examined using transient transfections of their respective mimics and inhibitors into periosteal cells (Fig. S1). The mimic-transfected periosteal cells had increased viability, while the opposite was found in the cells transfected with the miR-221 inhibitor (Fig. 5B). In addition, periosteal cells transfected with the miR-145 mimic had a marked increase in alcian Blue staining compared with almost no staining in the inhibitor-transfected cells (Fig. 5C); this result was supported by western blotting, which showed that the expression of Sox9 was reduced in the inhibitor-transfected periosteal cells compared with the mimic-transfected cells (Fig. 5D).