Small extracellular vesicle-mediated miR-320e transmission promotes osteogenesis in OPLL by targeting TAK1

Abstract

Ossification of the posterior longitudinal ligament (OPLL) is an emerging spinal disease caused by heterotopic ossification of the posterior longitudinal ligament. The pathological mechanism is poorly understood, which hinders the development of nonsurgical treatments. Here, we set out to explore the function and mechanism of small extracellular vesicles (sEVs) in OPLL. Global miRNA sequencings are performed on sEVs derived from ligament cells of normal and OPLL patients, and we have showed that miR-320e is abundantly expressed in OPLL-derived sEVs compare to other sEVs. Treatment with either sEVs or miR-320e significantly promote the osteoblastic differentiation of normal longitudinal ligament cells and mesenchymal stem cells and inhibit the osteoclastic differentiation of monocytes. Through a mechanistic study, we find that TAK1 is a downstream target of miR-320e, and we further validate these findings in vivo using OPLL model mice. Together, our data demonstrate that OPLL ligament cells secrete ossification-promoting sEVs that contribute to the development of ossification through the miR-320e/TAK1 axis.

Fig. 1. Small extracellular vesicles (sEVs) are vital to the development of OPLL.

A Scheme of GW4869 injection in ttw mice to examine the effect of sEVs inhibition on OPLL development. B Computed tomography showing the occupation of ossified ligament tissue in the spinal canal of ttw mice. The red dashed line indicates the area of spinal canal. C Typical CT image of cervical OPLL patient (left panel with red dashed square indicate OPLL in the spine canal), and postoperative X-ray image showing the resection of the vertebrate and the ossified ligament tissue (left panel with blue dashed square). The workflow showing how sEVs are collected and purified from OPLL and normal PLL ligament cell supernatant was shown in the right panel. D The Nanoparticle Tracking Analysis (NTA) showing the particle size/concentration (left panel) and particle size/relative intensity (right panel) plot of the collected sEVs. Note that most detected particles were around 100 nm in size. E The Transmission Electron Microscopy (TEM) analysis showing the typical image of sEV collected in the experiment was shown. The scale bar represents 100 nm, n = 6 biologically independent repeats with similar results. F Western blot analysis confirmed the extracellular vesicle marker expression in ligament cells (Cell group), EV-free supernatant (EFS group) and sEV group, n = 3 biologically independent repeats with similar results. G sEV uptake analysis was performed and analyzed using immunofluorescence microscopy. sEVs were labeled with membrane-specific dye PKH67 and purified before incubation with ligament cells for 6 h. Cells were washed, fixed and counterstained with DAPI (upper panel). G Non-labeled sEV and EFS were incubated with ligament cells for 6 h, and after wash, the cells were fixed and immunocytochemistry analysis was performed to show cellular expression of CD81 and CD63 using specific antibodies (lower panel). EFS represented the collected supernatant that free of sEVs. The scale bar represents 100 μm. H Comparison of the sEV size and amount data between PLL and OPLL ligament cell derived sEVs recorded from NTA analysis. EFS represented the collected supernatant that free of sEVs. Data were presented as mean ± SD, n = 6 biologically independent samples, **p = 0.01, t test two tailed; ns not significant, p = 0.0692. Source data are provided as a Source Data file.

Fig. 2. Micronome analysis identified OPLL-sEV specific microRNAs.

A Transwell assay followed by qRT-PCR detecting the expression changes of osteogenic genes in Mesenchymal stem cells (MSCs) co-cultured with or without PLL and OPLL cells. The scheme was shown in the left panel, n = 6 biologically independent samples, gene expression in each group was compared to MSC alone group using two tailed t test, **p = 0.0001; ns not significant, p = 0.999. B qRT-PCR analysis detecting the expression changes of osteogenic genes in MSCs treated with OPLL (OPLL-sEV) or PLL (PLL-sEV) derived sEVs, n = 6 biologically independent samples, gene expression in each group was compared to control group using two tailed t test, **p = 0.0001; ns not significant, p = 0.999. High through-put microRNA (miRNAs) sequencing was performed in OPLL and PLL derived sEVs, and the scatter plot displayed the differentially expressed miRNAs between OPLL and PLL derived sEVs (C). The abundance of expressed miRNAs is shown in percentage in the pie chart of OPLL derived sEVs (D) and PLL derived sEVs (E), while data of OPLL (F) and PLL cells (G) were also displayed. Note that significant differences were found in top 15 abundant miRNAs between OPLL or PLL cell and derived sEVs. The top 50 differentially expressed miRNAs were shown between OPLL derived sEVs and PLL derived sEVs (H) or OPLL cells and PLL cells (I). The expression fold change of the top three differentially expressed miRNAs were listed in (I), as they are not found in the cellular level of top 50 differentially expressed miRNAs. All qPCR data were presented as mean ± SD, and GAPDH level were detected and served as internal reference. Source data are provided as a Source Data file.

Fig. 3. miR-320e is highly expressed in OPLL-sEV and upregulated in OPLL.

A Comparison of the miRNA expression levels of miR-320 family in OPLL and PLL cell derived from High through-put sequencing data (n = 3, two tailed t test). TPM represents transcripts per kilobase million in the sequencing data. B qRT-PCR analysis detecting the expression level of miR-320e in OPLL (n = 14) and PLL (n = 12) tissues (two tailed t test). C qRT-PCR analysis detecting the expression level of miR-320e in osteogenic induced OPLL and PLL cells at different time point. Data showing respective microRNA expression fold changes compared to that of non-induced PLL cells (day 1) were presented (n = 6, two tailed t test). D qRT-PCR analysis detecting the expression level of miR-320e in osteogenic induced OPLL and PLL cell derived sEVs at different time point. Data showing respective microRNA expression fold changes compared to that of non-induced PLL cells derived sEV (NC) were presented (n = 6, two tailed t test). E Transwell assay followed by qRT-PCR analysis detecting the expression level of miR-320e in the co-cultured MSC or PLL cells (n = 6, two tailed t test). F qRT-PCR analysis detecting the expression level of miR-320e in sEV or EFS treated PLL cells (n = 6, two tailed t test). EFS represents with the supernatant of sEV collection that is free of sEVs. U6 level were detected and served as internal reference. All data were presented as the mean ± SD. *p < 0.05, **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.

Fig. 4. miR-320e promote the osteoblastogenesis of posterior longitudinal ligament cells.

The osteogenic properties of PLL cells are analyzed using alizarin red staining (A) or alkaline phosphatase staining (B) after osteogenic induction for 21 days, n = 6. The colorimetric quantification is shown in the right panels, respectively (compared to NC group using two tailed t test). NC group represents transfecting scramble control miRNA mimics. The quantification of expression of ossification related genes were detected using either qRT-PCR (C, n = 6, compared to NC group using two tailed t test) or Western blot (D, n = 3 biologically independent repeats with similar results) under the same condition. Alizarin red staining (E) or alkaline phosphatase staining (F) were used to analysis osteogenic properties of miR-320e inhibition (320e-Inh) and miR-10a-3p inhibition (10a-3p-Inh) in OPLL cells after osteogenic induction for 21 days, n = 6. The colorimetric quantification is shown in the right panels, respectively (compared to NC group using two tailed t test). NC group represents transfecting scramble control miRNA mimics. Ossification-related genes are assessed by real-time PCR (G, n = 6, compared to NC group using two tailed t test) and Western Blot (H, n = 3 biologically independent repeats with similar results) after osteo-induction for 21 days under the same conditions respectively. GAPDH level were detected and served as internal reference. All data were presented as the mean ± SD. *p < 0.05, **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.

Fig. 5. miR-320e inhibits the osteoclastogenesis of monocytes.

Osteoclast differentiation was induced by treating the human monocyte cells with 30 ng/ml M-CSF and 100 ng/ml RANKL. Cells were fixed and stained for Tartrate-resistant acid phosphatase (TRAP) activities at days 20 in miR-320e and miR-10a-3p overexpression (A) and inhibition groups (B), n = 6, the scale bars represent 500 μm. The quantification of osteoclast precursors is shown in the right panel, respectively (compared to NC group using two tailed t test). The quantification of expression of osteoclastogenesis related genes were detected using either qRT-PCR (C, n = 6, compared to NC group using two tailed t test) or Western blot (D, n = 3 biologically independent repeats with similar results) in in miR-320e and miR-10a-3p overexpressed monocytes. Similarly, the quantification of expression of osteoclastogenesis related genes were detected using either qRT-PCR (E, n = 6, compared to NC group using two tailed t test) or Western blot (F, n = 3 biologically independent repeats with similar results) in in miR-320e and miR-10a-3p inhibited monocytes. GAPDH level were detected and served as internal reference. All data were presented as the mean ± SD. *p < 0.05, **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.

Fig. 6. TAK1 is targeted by miR-320e.

A qRT-PCR analysis detecting the mRNA expression levels of miR-320e candidate targets in PLL and OPLL cells, n = 6, two tailed t test. B qRT-PCR analysis detecting the mRNA expression levels of miR-320e predicted targets after miR-320e or miR-10a-3p overexpression in PLL cells, n = 6, two tailed t test. Here miR-10a-3p is served as a negative miRNA control that are not correlated with the candidates. C MiRanda prediction of miR-320e binding motif in TAK1 3’UTR. Note that the sites were picked according to the binding energy and rodent conservation. D Dual luciferase reporter assay detecting the activities of firefly luciferase generated by respective 3’UTR bearing plasmids after miR-320e or miR-10a-3p overexpression in HEK293T cells (n = 6, two tailed t test). E Western Blot analysis showing the protein levels of TAK1 or FOXO3 after miR-320e or miR-10a-3p overexpression in PLL cells. Here, FOXO3 is served as a negative target control, which miR-320e had no function on it, n = 3 biologically independent repeats with similar results. F mRNA levels of TAK1 and FOXO3 after miR-320e or miR-10a-3p inhibition in OPLL cells using qRT-PCR analysis, n = 6, two tailed t test. G Western Blot analysis showing the protein levels of TAK1 or FOXO3 after inhibition of miR-320e or miR-10a-3p in OPLL cells, n = 3 biologically independent repeats with similar results. GAPDH level were detected and served as internal reference. All data were presented as the mean ± SD. *p < 0.05, **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.

Fig. 7. TAK1 is essential for miR-320e to regulate osteoblastogenesis and osteoclastogenesis.

A In situ hybridization and histochemistry is used to analysis the expression level of TAK1 in PLL and OPLL patients’ tissue (n = 6, two tailed t test, the scale bars represent 500 μm). B qRT-PCR analysis showing the mRNA expression level of osteogenic genes after knockdown of TAK1 using small interference RNAs in osteogenic induced PLL cells, n = 6, two tailed t test. The siNC represents transfecting scramble control siRNAs which serve as control group. Alizarin red staining (C) or alkaline phosphatase staining (D) was used to analysis osteogenic properties of respective treatment in miR-320e inhibited and osteogenic induced OPLL cells after various treatment, both n = 6, two tailed t test. The colorimetric quantification is shown in the right panels, respectively. 5Z-7-oxozeaenol (a specific inhibitor of TAK1) is used at 20 nM to inhibit TAK1 activities. The relative RNA level (E, n = 6, two tailed t test) and protein level (F, n = 3 biologically independent repeats with similar results) of osteogenic genes in miR-320e inhibited and osteogenic induced OPLL cells after various treatment are shown. G Osteoclast induction combined with TRAP staining was used to evaluate the function of miR-320e inhibition and TAK1 inhibition in monocytes, the scale bars represent 500 μm. The quantification of osteoclast precursors is shown in the right panel, n = 6, two tailed t test. The quantification of expression of osteoclastogenesis related genes were detected using qRT-PCR (H) under the same treatment in monocytes, n = 6, two tailed t test. GAPDH level were detected and served as internal reference. All data were presented as the mean ± SD. **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.

Fig. 8. miR-320e is needed for OPLL derived sEVs to modulate osteoblastogenesis and osteoclastogenesis.

Transwell assay followed by qRT-PCR analysis detecting the expression changes of osteogenic related genes (A) or osteoclastogenic related genes (B) in PLL cells (A) or Monocyte (B) co-cultured with PLL cells or OPLL cells treated with miR-320e inhibitor or NC, n = 6, two-way ANOVA. C scheme of generation of miR-320e inhibited sEVs. The efficiency of inhibition is detected using qRT-PCR analysis in PLL cells shown in the right panel, n = 6, two tailed t test. D The effect of miR-320e inhibited sEVs on TAK1 expression detected by qRT-PCR (n = 6, two tailed t test) or Western blot (n = 3). OPLL-sEV-Inh represents miR-320e inhibited OPLL derived sEVs. The osteoblastogenic properties were examined by alizarin red staining (E) or alkaline phosphatase staining (F) and quantification in osteogenic induced PLL cells, n = 6, two tailed t test. G The osteoclastogenic properties were examined by osteoclast induction combined with TRAP staining and quantification, n = 6, two tailed t test, the scale bars represent 500 μm. H The relative protein level of osteoblastogenesis or osteoclastogenesis related genes were detected using Western blot (n = 3 biologically independent repeats with similar results). GAPDH level or U6 level were detected and served as internal reference. All data were presented as the mean ± SD. **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.

Fig. 9. miR-320e promoted heterotopic bone formation in vivo.

A Scheme for heterotopic bone formation assay procedure. Small EVs (sEV) were treated with OPLL cells prior to implantation. B Representative reconstructed three-dimensional micro-CT images of implanted bio-scaffold after 6 weeks, n = 6 for each group. C Bone analysis showing the BV (bone volume)/TV (tissue volume) and BMD (bone marrow density) of cultured bone constructs (n = 6, one-way ANOVA). D H&E staining and immunohistochemical staining of implanted bio-scaffold after 6 weeks in respective groups, n = 6 for each group. The scale bars represent 500 μm. E Quantification of OCN and TAK1 expression in the immunohistochemical staining of implanted bio-scaffold after 6 weeks (n = 6, one-way ANOVA). All data were presented as the mean ± SD. *p < 0.05, **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.

Fig. 10. sEV derived miR-320e promoted OPLL development in ttw mice.

A Experimental outline of in vivo OPLL formation assay using ttw mice injected with OPLL-sEVs or miR-320e inhibited sEV. Injections were performed once every 2 days for one week. Mice were sacrificed upon death or until 18 weeks after last injection, spine samples were harvested and limb functions were observed everyday. B Immunohistochemistry analysis showing the distribution of human CD81 expression in OPLL-sEV injected spine tissue (n = 3). Note that in the Blank group, no cellular CD81 expression were found. While in OPLL-sEV injected group, the ligament cells displayed strong expression of human CD81. The scale bars represent 500 μm. C micro-CT images of spine harvested spines from ttw mice in various groups. The occupation percentage of the ossified mass in the spinal canal of ttw mice were compared (lower right panel, n = 6, one-way ANOVA). D H&E staining, TRAP staining and immunohistochemistry analysis showing the expression of TAK1 and OCN in the spine samples from treated ttw mice were analyzed (n = 6). AF stands for annulus fibrosus, NP stands for nucleus pulposus and L stands for posterior longitudinal ligament. The scale bars represent 800 μm. E The quantification of TRAP positive, TAK1 positive and OCN positive ligament cells in the posterior longitudinal ligament region were shown and compared, n = 6 for each marker tested, one-way ANOVA. F The conclusive illustrations depict the potential mechanism uncovered in this study. All data were presented as the mean ± SD. *p < 0.05, **p < 0.01. Detailed statistical data and source data are provided in a Source Data file.