Chondrogenic primed extracellular vesicles activate miR-455/SOX11/FOXO axis for cartilage regeneration and osteoarthritis treatment

Abstract

Osteoarthritis (OA) is the leading cause of disability worldwide. Considerable progress has been made using stem-cell-derived therapy. Increasing evidence has demonstrated that the therapeutic effects of BMSCs in chondrogenesis could be attributed to the secreted small extracellular vesicles (sEVs). Herein, we investigated the feasibility of applying engineered EVs with chondrogenic priming as a biomimetic tool in chondrogenesis. We demonstrated that EVs derived from TGFβ3-preconditioned BMSCs presented enriched specific miRNAs that could be transferred to native BMSCs to promote chondrogenesis. In addition, We found that EVs derived from TGFβ3-preconditioned BMSCs rich in miR-455 promoted OA alleviation and cartilage regeneration by activating the SOX11/FOXO signaling pathway. Moreover, the designed T3-EV hydrogel showed great potential in cartilage defect treatment. Our findings provide new means to apply biosafe engineered EVs from chondrogenic primed-BMSCs for cartilage repair and OA treatment, expanding the understanding of chondrogenesis and OA development modulated by EV-miRNAs in vivo.

Fig. 1. Identification of T3-EV and un-EV.

A Illustration of study design. B Representative images showing the morphology of T3-EV and un-EV visualized under transmission electron microscopy (TEM). C Particle size distribution of h T3-EVs measured using nanoparticle tracking Analysis (NTA). D Quantification of surface markers of EVs evaluated by western blotting. BMSCs served as the control in the western blot analysis of surface markers of EVs. E Representative fluorescence micrograph of PKH26 (red)-labeled EVs internalized by primary BMSCs. The labeled EVs were co-incubated with BMSCs for 24 h. F GAG staining with Safranin-O and Toluidine blue staining of BMSCs treated with EVs for 21 days. G Chondrogenesis was defined with immunostaining of SOX9, ACAN, and COL2A1(red). Counterstaining with F-actin (green) and DAPI (blue) was applied. H Three different experiments with the same BMSCs and the same EVs were performed. Chondrogenic gene expression (n = 3 for each) was assayed with qRT-PCR for SOX9, ACAN, and COL2A1. I–K Quantification of deposited GAGs and collagens (n = 3 for each) was also performed with the same BMSCs and the same EVs to demonstrate the chondrogenic lineage committed by the BMSCs. Treatment with saline served as control. *P < 0.05, **P < 0.01, ***P < 0.001, NS not significant.

Fig. 2. Discovery of T3-EV-associated miRNAs by microarray.

A Heatmap of clustering dysregulated miRNA expression profiles with microarray in T3-EVs compared to untreated un-EV control. B Volcano plot of miRNA expression profiles and miR-455 (red dot) was most significantly upregulated in T3-EV. C–E All differentially expressed miRNAs(DEG with fold change >2 or <0.5, p value <0.01) were subjected to gene ontology (GO) analysis for C biological processes, D molecular function, and E cellular component. F Significantly enriched pathways for target genes of miRNAs enriched within T3-EV in Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. G miRNAs expression elevation (n = 3 for each) with the same EVs was validated with qRT-PCR. *P < 0.05, **P < 0.01, ***P < 0.001, NS not significant.

Fig. 3. Activation of FOXO signaling by T3-EV.

A Heatmap of clustering dysregulated mRNA expression profiles with microarray in T3-EV-treated BMSCs compared to un-EV treated control. B Volcano plot of mRNA expression profiles in T3-EV-treated BMSC recipient. C Dysregulated typical chondrogenic markers derived from the microarray results with T3-EV treatment. D All differentially expressed genes were subjected to gene ontology (GO) analysis (DEG with fold change >2 or <0.5, p value <0.01). BP biological processes, MF molecular function, CC cellular component. E Significantly enriched pathways for dysregulated genes enriched with T3-EV treatment in KEGG pathways. F, G FOXO1 expression(red) assayed with western blotting and fluorescent immunostaining (red for FOXO1; green for cytoskeleton) in T3-EV treated BMSC recipient. Cells were counterstained with DAPI for the nucleus (blue). Treatment with saline served as control. H Quantification of gene expression with the same BMSCs and the same EVs (n = 3 for each) with qRT-PCR for Chondrogenic genes (SOX9, ACAN, COL2A1, and MMP13) and FOXO signaling-related genes (FOXO1, Gadd45a, p27, and Cathepsin L). *P < 0.05, **P < 0.01, ***P < 0.001, NS not significant.

Fig. 4. SOX11 is a direct target of miR-455 to modulate chondrogenesis.

A All predicted genes were compiled for Venn analysis to search for the potential targets of miR-455. B miRNA–mRNA network using the Cytoscape software was constructed for miRNA-455. C Sequence of wild-type (WT) and mutant (Mut) SOX11 binding sites for miRNA-455 (left) and conservation level of miR-455 sequence among species (right). D, E SOX11 expression(red) assayed with fluorescent immunostaining (red for SOX11; green for cytoskeleton) and qRT-PCR in T3-EV treated BMSC recipient (n = 3 for each with the same BMSCs and the same EVs). F Fluorescence micrograph of Cy3 (red)-labeled miR-455 mimic internalized by BMSCs. G Luciferase reporter assay analysis results (n = 3 for each) to confirm the direct interaction between miR-455 and SOX11 binding sites. Relative luciferase reporter activity was assessed for co-transfected SOX11 WT (or Mut) with miR-455 mimics or inhibitors in cultured primary BMSC cells. miRNA-455 mimics control and inhibitor control served as negative controls. H–J miR-455 and SOX11 expression(red) with H, J) qRT-PCR and I fluorescent immunostaining in BMSCs transfected with miR-455 mimics or inhibitor (n = 3 for each). K, L Fluorescent immunostaining was also conducted on anabolic and catabolic markers ACAN (green) and MMP13 (green) in BMSCs transfected with miR-455 mimics or inhibitors. Cells were counterstained with DAPI for the nucleus (blue). *P < 0.05, **P < 0.01, ***P < 0.001, NS not significant.

Fig. 5. MiR-455 regulates chondrogenesis and OA development by modulating the SOX11/FOXO signaling pathway.

A Schematic representation of how the miR-455/SOX11/FOXO signaling pathway might mediate chondrogenesis and the therapeutic effects of T3-EV in OA treatment and cartilage regeneration. B Cultured primary human BMSCs were transfected with miR-455 mimics, miR-455 inhibitor, their negative controls, control siRNA or SOX11 siRNA for 72 h, respectively, and the expression levels of chondrogenic markers SOX9, ACAN, COL2A1, MMP13, and SOX11/FOXO signaling pathway markers SOX11, FOXO1, Gadd45a, p27, and Cathepsin L were assessed with western blot. C Rescue experiments was established in cultured primary BMSCs to validate the relationship between miR-455 and SOX11. Elevation of SOX9 and ACAN expression levels by miR-455 mimics was rescued by restoration of SOX11 expression. In comparison, inhibition of SOX9, ACAN, and FOXO1 expression by SOX11 overexpression was rescued by miR-455 mimics. D Upregulation of SOX9 and ACAN expression levels by SOX11 siRNA was abolished by silencing of FOXO1 expression. In comparison, upregulation of FOXO1 target genes (p27 and Cathepsin L) expression by FOXO1 siRNA was abolished by silencing of SOX11 expression.

Fig. 6. Injection of T3-EV-conjugated composite hydrogel for cartilage regeneration in vivo.

A Rat T3-EV were derived for T3-EV hydrogel formation and delivery. T3-EVs were derived for embedding into composite BMSC hydrogel produced with a mixture of gelatin, fibrinogen, and HA. Cartilage defect injury was created, and hydrogel injection was performed in situ to deliver T3-EV hydrogel for defect repair. B Before transplantation, T3-EV hydrogel was cross-linked by the addition of thrombin to further maintain the shape fidelity of the hydrogel. (a) The cross-linked hydrogel demonstrated good shape fidelity and b–d) good distribution of EVs internalized within BMSCs (b: PKH26-labeled EVs stained red; c: nucleus stained blue; d: merged image with b and c). C Cell viability >95% was demonstrated for 7 days with live/dead assay for the T3-EV hydrogel. D Histological assessment of cartilage repair with T3-EV hydrogel gel at 24 weeks. Neo-cartilage formation in different groups was compared with tissue integrity (HE staining in the 1st row) and GAG deposition with Toluidine blue and Safranin-O staining (second and third row). E–G Histological grading of repaired cartilage in different groups (n = 6 for each) over 24 weeks. F Articular cartilage in both un-EV-gel and T3-EV-gel groups showed declined Mankin score and G higher ICRS histological score compared to the control group in the femoral condyle (FC) and tibial plateau (TP) over the 24 weeks in vivo. H Immunostaining of chondrogenic markers SOX9 (first row), ACAN (second row), and MMP13 (third row) were also conducted and observed (red for the protein, blue for the nucleus) in the generated cartilage in different groups. I Immunostaining of miR-455 (FISH), SOX11, and FOXO1 was also performed and observed (red for the protein, blue for the nucleus) to examine the miR-455/SOX11/FOXO axis in the generated cartilage in different groups. *P < 0.05, **P < 0.01, ***P < 0.001, NS not significant. #p < 0.05 compared to the un-EV gel group.

Fig. 7. Intra-articular T3-EV injection reversed OA progression by regulating the miR-455/SOX11/FOXO signaling axis.

A To determine whether T3-EV transplantation could reduce or reverse the progression of OA, intra-articular injection of T3-EV labeled with Dir was performed for rats with DMM surgery. B PKH26-labeled EVs (red) to detect their internalization by BMSC in vitro. C Intra-articular delivery of the Dir-label T3-EVs (left: red) were monitored with in vivo fluorescence imaging to evaluate the near-infrared imaging effect and distribution of EVs for 12 weeks, showing an ideal intra-articular delivery effect (right). D–F Histological assessments of joint cartilage with D HE (left column), TB staining (right column), and E, F immunostaining for ACAN, MMP13, miR-455, SOX11, and FOXO1 in different groups. G–J Quantification and comparison of histological grade for OA progression in different groups (n = 8 for each). Joint destruction severity was determined with OARSI score, osteophyte formation, subchondral bone plate thickness, and synovial inflammation as previously reported^–. Saline: OA model group with saline injection; un-EV: only untreated EVs were injected for OA treatment; T3-EV: TGFβ3-treated EVs were injected for OA treatment. *P < 0.05, **P < 0.01, ***P < 0.001, NS not significant. #p < 0.05 compared to the un-EV gel group.