The umbilical cord mesenchymal stem cell-derived exosomal lncRNA H19 improves osteochondral activity through miR-29b-3p/FoxO3 axis

Abstract

Background: Our previous study revealed that the exosomal lncRNA H19 derived from umbilical cord mesenchymal stem cells (UMSCs) plays a pivotal role in osteochondral regeneration. In this study, we investigated whether the exosomal lncRNA H19 could act as a competing endogenous RNA (ceRNA) to potentiate osteochondral activity in chondrocytes.

Methods: Dual-luciferase reporter assay, RNA pull-down, RNA immunoprecipitation (RIP), and fluorescence in situ hybridization (FISH) were carried to verify the interaction between miR-29b-3p and both lncRNA H19 and the target mRNA FoxO3. Chondrocytes were treated with UMSC-derived exosomes, which highly expressing lncRNA H19 expression, followed by apoptosis, migration, senescence, and matrix secretion assessments. An in vivo SD rat cartilage defect model was carried out to explore the role and mechanism of lncRNA H19/miR-29b-3p.

Results: UMSCs were successfully identified, and exosomes were successfully extracted. Exosomes exhibited the ability to transfer lncRNA H19 to chondrocytes. Mechanistically, exosomal lncRNA H19 potentiated osteochondral activity by acting as a competing endogenous sponge of miR-29b-3p, and miR-29b-3p directly targeted FoxO3. Intra-articular injection of exosomes overexpressing lncRNA H19 could promote sustained cartilage repair; however, this effect could be undermined by miR-29b-3p agomir.

Conclusions: Our study revealed a significant role in the development of strategies against cartilage defects for UMSC-derived exosomes that overexpress lncRNA H19. Exosomal H19 was found to promote chondrocyte migration, matrix secretion, apoptosis suppression, as well as senescence suppression, both in vitro and in vivo. The specific mechanism lies in the fact that exosomal H19 acts as a ceRNA against miR-29b-3p to upregulate FoxO3 in chondrocytes.

Keywords: chondrocytes; exosomes; long noncoding RNA; umbilical cord mesenchymal stem cells.

FIGURE 1

The lncRNA H19 serves as a sponge for miR‐29b‐3p in chondrocytes. A, The subcellular localization of lncRNA H19 was verified by FISH analysis; scale bar: 200 μm. B, The binding sequences of lncRNA H19 and miR‐29b‐3p were predicted by bioinformatics analysis. C, The binding of lncRNA H19 and miR‐29b‐3p was confirmed by the dual‐luciferase reporter assay. *P < .05 compared with the miR‐NC group. D, Colocalization of lncRNA H19 and miR‐29b‐3p by FISH analysis; scale bar: 200 μm. E, The binding efficiency of lncRNA H19 and miR‐29b‐3p to Ago2 protein was detected by RIP. **P < .01 compared with the vector group. F, The relationship between lncRNA H19 and miR‐29b‐3p was tested by RNA pull‐down. *P < .05 compared with the Bio‐NC group. The data are shown as mean ± SD (n = 3)

FIGURE 2

UMSCs were successfully identified and exosomes were successfully extracted. A, Morphological observation of UMSCs (×100). B, UMSCs exhibited multiple osteogenic, adipogenic, and chondrogenic differentiation ability. C, The surface antigen expression of UMSCs was detected by flow cytometric analysis. D, Morphology of exosomes under TEM; scale bar: 200 nm. E, The diameter and concentration of exosomes by nanosight. F, The expressions of TSG101, CD63, CD81, and calnexin in exosomes were detected by Western blot analysis. The data are shown as mean ± SD (n = 3)

FIGURE 3

Exosomes derived from UMSCs could transfer lncRNA H19 to chondrocytes. A, The expression of lncRNA H19 in UMSCs and exosomes was detected by qRT‐PCR. **P < .01 compared with the control group. B, The expression of lncRNA H19 in exosomes treated with RNase A or Triton X‐100. **P < .01 compared with the control group. C, Uptake of PKH67‐labeled exosomes in chondrocytes was observed by confocal fluorescence microscopy; scale bar: 50 μm. The data are shown as mean ± SD (n = 3)

FIGURE 4

UMSC‐derived exosomal lncRNA H19 potentiates osteochondral activity by directly inhibiting miR‐29b‐3p. A, Expression of lncRNA H19 and miR‐29b‐3p in chondrocytes, which were cocultured with UMSCs or UMSC‐derived exosomes treated with si‐Rab27a, or MSC‐derived exosomes overexpressing H19, or MSCs transfected with sh‐H19. *P < .05 and **P < .01 compared with the control group; ^# P < .05 compared with the UMSCs. B and C, Chondrocytes were treated with H19‐Exos with or without miR‐29b mimics and stimulated by IL‐1β (10 ng/mL) challenge for 24 hours. Apoptosis was assessed by flow cytometry. *P < .05, **P < .01. D and E, Light microscopy images and quantitative analysis of chondrocytes stained with the senescent marker SA‐β‐Gal in monolayer culture; scale bar: 100 μm. *P < .05, **P < .01. F and G, Light microscopy images and quantitative analysis of scratch wound assays. H and I, Light microscopy images and number of transmitted cells in the transwell migration assay. *P < .05, **P < .01. J and K, The protein and mRNA level of genes associated with chondrocytes (MMP13, ADAMTS5, Runx2, Col II, aggrecan, and Sox9). L, Immunofluorescence for MMP13, Col II, and aggrecan. The data are shown as mean ± SD (n = 3)

FIGURE 5

FoxO3 was a direct target of miR‐29b‐3p. A, Binding sequences of miR‐29b‐3p and FoxO3 were predicted by bioinformatics analysis. B, Binding of miR‐29b‐3p and FoxO3 was confirmed by the dual‐luciferase reporter assay. *P < .05 compared with the miR‐NC group. C, miR‐29b‐3p overexpression reduced FoxO3 level, while miR‐29b‐3p inhibition increased FoxO3 level. D, FoxO3 overexpression increased the protein level of Col II, aggrecan, and Sox9 and decreased the level of MMP13, ADAMTS5, and Runx2. E, Chondrocytes were treated with miR‐29b mimics with or without overexpressing FoxO3 and stimulated by IL‐1β (10 ng/mL) challenge for 24 hours. Apoptosis was assessed by flow cytometry. F and G, Light microscopy images and quantitative analysis of chondrocytes stained with the senescent marker SA‐β‐Gal in monolayer culture; scale bar: 100 μm. H, Light microscopy images and number of transmitted cells in the transwell migration assay. I and J, Light microscopy images and quantitative analysis of scratch wound assays. K and L, Protein and mRNA level of genes associated with chondrocytes (MMP13, ADAMTS5, Runx2, Col II, aggrecan, and Sox9). M, Immunofluorescence for MMP13, Col II, and aggrecan. The data are shown as mean ± SD (n = 3; *P < .05, **P < .01)

FIGURE 6

Macroscopic and MRI images of the regenerated tissues at 4 and 8 weeks. A, Representative macroscopic and MRI images of the regenerated tissues at 4 and 8 weeks. B, ICRS macroscopic scores. C, T2 mapping scores. The data are shown as mean ± SD (n = 5; *P < .05, **P < .01)

FIGURE 7

Histologic evaluation of cartilage repair at 4 and 8 weeks. A, Staining results of hematoxylin and eosin (HE), toluidine blue (TB), Safranin‐O/fast green (Saf‐O), and immunohistochemical staining for type II and I collagens. B, Wakitani scores for the histological sections. The data are shown as mean ± SD (n = 5; **P < .01)

FIGURE 8

Schematic presentation of working hypothesis