DNM3OS/miR-127-5p/CDH11, activates Wnt3a/β-catenin/LEF-1 pathway to form a positive feedback and aggravate spine facet joint osteoarthritis

Abstract

Spinal facet joint osteoarthritis (FJOA) is an OA disease with pathogenesis and progression uncovered. Our present study was performed to elucidate the role of DNM3OS on spinal FJOA. In this study, spine facet joint tissue of patients were collected. In vitro and in vivo models were constructed with SW1353 cells and rats. Hematoxylin and eosin (HE) staining, Safranin O-fast Green, Alcian blue staining, and Tolueine blue O (TBO) staining were employed for histology analyses. Quantitative PCR, western blotting, and Immunofluorescence were performed to evaluate the expression of genes. The levels of inflammatory cytokines were measured by enzyme-linked immunosorbent assay analysis. Cell Counting Kit-8 and flow cytometry were used for cell activity and apoptosis evaluation. The targeting sites between microRNA (miR)-127-5p and cadherin 11 (CDH11) were predicted TargetScan and miRbase database and confirmed by Dual-luciferase reporter assays. CHIP and EMS assay were employed to confirm the binding of LEF1and DNM3OS promoter. Our results showed that DNM3OS was found to upregulated, while miR-127-5p was downregulated in severe FJOA patients and inflammation-induced chondrosarcoma SW1353 cells. DNM3OS reduced cell activity, induced cell apoptosis and extracellular matrix (ECM) degradation by sponging miR-127-5p in vitro. miR-127-5p targeted CDH11 and inhibited wnt3a/β-catenin pathway to regulate OA in vitro. LEF1 promoted DNM3OS transcription to form a positively feedback in activated wnt3a/β-catenin pathway. In vivo rat model also confirmed that DNM3OS aggravated FJOA. In summary, DNM3OS/miR-127-5p/CDH11 enhanced Wnt3a/β-Catenin/LEF-1 pathway to form a positive feedback and aggravate spinal FJOA.

1. Introduction

Facet Joint Osteoarthritis (FJOA) is one of the degenerative spinal diseases that has significant morbidity in older adults, causing back and neck pain and economic burden [1]. Series treatments of physical therapies and weight management, diet and nutritional supplements, and medicines are applied to improve joint movement and reduce pain in FJOA [[2], [3], [4]]. Surgery is also one option when other treatments are ineffective. However, OA, including spinal FJOA is irreversible and cannot be cured, and the treatment strategy is relatively limited. Therefore, understanding the pathogenesis and progression of spinal FJOA has emerged as an important theme in disease prevention and treatment.

In recent years, the ceRNA network provided mounting evidence on underlying the pathogenesis and progression of OA. For examples, the dual-specificity phosphatase 1-specific ceRNA network have been considered as one diagnostic tool on assessing the OA progression [5]. Wu et al. [6] eluciated the pahtogenesis of OA by analyzing mRNA/lncRNA/circRNA expression and displaying the ceRNA network of lncRNAs. MicroRNA (miRNA) is one of the small non-encode RNA (ncRNA) that targets mRNA and is competitively bound by long ncRNA (lncRNAs). CeRNA XIST/miR-211/CXCR4 have been reported to promote the proliferation of OA chondrocytes and promote apoptosis [7]. Many studies have confirmed that miRNAs participate in the differentiation regulation of bone marrow mesenchymal stem cells (BMSCs) [[8], [9], [10]]. One of our previous studies revealed that miR-132-3p, miR-99a, miR-127-5p, and miR-125b-3p play essential roles in the tending induction of rat BMSCs [11]. However, they are barely studied on spinal FJOA. Of these miRNAs, miR-127-5p was demonstrated to promote early chondrogenic differentiation of MSCs by regulating gremlin 2 (GREM2) and was competitively bound by lncRNA DNM3 Opposite Strand/Antisense RNA (DNM3OS) in our preivous work [12]. Therefore, we wonder whether DNM3OS/miR-127-5p works in FJOA.

Cadherin 11 (CDH11) is one of integral membrane proteins and involved in Wnt/β-Catenin pathways. CDH11 binds with and stabilize β-Catenin, which prevents the degradation of β-Catenin by APC-Axin-GSK3β complex [13]. Extraceletic Wnt molecules bind to the receptor on the cell membrane, which is transformed by a signal to activate the β-catenin molecules in the cell. The β-Catenin molecule does not have a DNA binding site, but it can be combined with the transcription factors lymphoid enhancer binding factor 1 (LEF1) and TCF4 to adjust downstream target genes' transcription jointly [14]. Satriyo et al. [13]demonstrated that inhibited CDH11 deactivates the canonical Wnt/β-catenin signalling pathway and suppresses the cancer stem cell-like phenotype of triple negative breast cancer. Coincidentally, Wnt/β-Catenin is involved in the formation of bone and promotes the progression of OA [15,16]. For example, Dong et al. [17] revealed that miR127-3p released from BMSCs alleviates OA through regulating CDH11 mediated Wnt/β-Catenin pathways. Furthermore, it was precited by TargetScan that miR-127-5p might bind to CDH11 mRNA, which participated in Wnt/β-Catenin pathway. It meant that miR-127-5p might also regulate Wnt/β-Catenin pathways.

2. Method and materials

2.1. Patients

Patients with spinal FJOA were recruited from the Guangdong Second Provincial General Hospital in Guangzhou, China. The selection criteria and exclusion criteria could been seen in our previous research [18]. The FJOA severity was graded by the computed tomography (CT) and T2 magnetic resonance imaging (MRI) of patients according to the criteria of Weishaupt [19]. FJOA in Grade 2 was regarded as mild and FJOA in Grade 3 was regarded as severe. The spine facet joint tissue (L3-S1) and blood samples from spinal FJOA patients with mild (n = 10) or severe (n = 10) severity were obtained.

2.2. In vitro model construction and transfections

As reported, the human chondrosarcoma (SW1353) cell line has been established as a human chondrocyte model for its similar arthritis-related biomarkers [20]. Therefore, our present study employed SW1353 cells to construct FJOA in vitro models. SW1353 cells were obtained from ATCC (American Type Culture Collection), cultured in Leibovitz's L-15 medium (L 15) containing 10 % fetal bovine serum (FBS), and stimulated by 10 ng/ml of IL-6 for 0, 12, 24, 36, 48, and 56 h (Inflammation group) and 1 μM of the Wnt/β-Catenin pathway inhibitors XAV-939. The control group was out of the treatments of IL-6 (Control group).

To explore the role of genes, SW1353 cells were subjected to transfections with vectors, two short-hairpin (sh) DNM3OS (sh-DNM3OS), overexpressed DNMOS, miR-127-5p mimics and inhibitor, overexpressed CDH11 (CDH11), sh-LEF1, and their negative control using Lipofectamine® 3000 (Invitrogen) in line with the manual's protocol. Vectors, shRNAs and siRNAs, mimics, inhibitors were obtained from Shanghai GenePharma Inc. (Shanghai, China). PcDNA™3.1 was bought from Invitrogen (USA).

2.3. In vivo model construction

Sh-DNM3OS lentivirus were obtained from Shanghai Genechem Co., Ltd (China) and miR-127-5p antagomir was obtained from Guangzhou Ribobio Co., Ltd (China). Twenty-four male Sprague-Dawley rats were provided by the Animal Center of the Chinese Academy of Sciences (Shanghai, China) at ten weeks old. The rats were housed in polypropylene cages in a room with a 12 h dark/light cycle; food and water were available ad libitum. Then, they were divided into four groups evenly, including Sham, spinal FJOA model (Model), spinal FJOA model injected with sh-DNM3OS lentivirus (Model + Lv-sh-DNM3OS), and spinal FJOA model injected with sh-DNM3OS lentivirus and antagomir (Model + Lv-sh-DNM3OS + antagomir). The FJOA model was constructed, followed by one previous report [18]. Consively, the rats were anesthetized with isoflurane, shaved, and inserted with A 21 G butterfly needle to the right of the middle of the L2 and L3 spinous processes. Then, the butterfly needle was slightly moved to slide into the L2/L3 facet joint. Finally, a 26 G needle microinjector was inserted into the butterfly needle, and the papain mixed solution was slowly injected. The left sides of the spinous processes were also treated with papain combined solution. Papain solution injection was performed once a week for 3 consecutive weeks. 1 × 10⁸ PFU lentivirus and 50 nmol antagomir were injected at the second week. At 4 weeks post-operation, all rats were anesthetized and sacrificed, then blood and joint cartilage tissue samples were collected. Similar to the clinical samples, the in vivo tissue samples were also sectioned and stained with hematoxylin and eosin (HE) staining, Alcian blue staining, and Tolueine blue O (TBO) staining.

2.4. Safranin O-fast Green, Alcian blue, and toluidine blue O staining staining

The cartilage tissues of patients and animals were fixed in formaldehyde, decalcified in ethylene diamine-tetra acetic acid (EDTA) solution, embedded in paraffin, and sectioned.

The clinical sections were used for Safranin O-fast Green [21] and Alcian blue staining [22]. Safranin O-Fast Green staining kit (DB0082, Solarbio, China) was obtained to observe the morphology of cartilage tissues under a light microscope. Briefly, the paraffin-embedded slices were dewaxed, rehydrated with degraded alcohol, and hydrated. Subsequently, the slices were stained with haematoxyline for 3min, hydrochloric acid (1 %), and ethanol for 15 s at room temperature. Then, the slices were washed and immersed in a 0.02 % Fast Green solution at 20 °C for 3 min, washed with 1 % glacial acetic acid, and stained in 0.1 % Safranin O at 20 °C for 3 min. Finally, the chondrocytes of cartilage tissues were observed under a light microscope (magnification, x50).

Alcian blue staining was conducted using a commercial kit (DB0063, Solarbio, China). In brief, the slices were decarculeried conventional, dehydrated diopine, washed with graded ethanol, and hydrated. Then, they were soaked in Alcian acidized solution for 3 min and Alcian dyeing liquid for 30min. Afterward, the slices were washed, dehydrated, and sealed for observation. The slices were photographed and analyzed under image processing program (ImageJ, U.S.A, 1.52 K) under an RGB scale to obtain signal intensity.

As for Toluidine Blue O stainingstaining (TBO), the sections were stained with toluidine blue solution for 20 s, rinsed in distilled water for 2 min, and washed with 95 % ethanol.

2.5. Enzyme-linked immunosorbent assay (ELISA) analysis

The inflammatory cytokines of tumor necrosis factor-alpha (TNF-α; #cat: ab181421 and ab236712; Abcam), interleukin 1β (IL-1β; #cat: ab46052 and ab255730; Abcam), and Interleukin 6 (IL-6; #cat: ab178013 and ab234570; Abcam) were measured by ELISA using the corresponding kits following the manufacturer's instructions. In brief, the samples were added with HRP-labeled antibody and incubated at 37 °C for 60 min. Then, 100 μl of TMB substrate was added, and incubated for 30 min at 25 °C in the dark. Finally, the Stop Solution was added into, and the absorbance at 405 nm was measured using an ELISA instrument (Thermo Fisher Scientific Inc.) after a 20 min incubation with the chromogenic agent.

2.6. Quantitative real-time PCR (QPCR)

The mRNA expression of DNM3OS, miR-127-5p, CDH11, LEF1 were evaluated by QPCR. The total RNA in tissues and cells was extracted by TRIzol reagent (Invitrogen, USA). The primers of DNM3OS, miR-127-5p, CDH11, LEF1, GAPDH and U6 were designed and synthesized by Sangon (Shanghai, China), which were showed in Table 1. Although DNM3OS had not been identified in rat, we found a similar sequence in chromosome 13 of rat gene, and found miR-127-5p binding sequence (CTTCAG) in its next 120bp. Therefore, rat DNM3OS primers were designed according to this similar sequence. The qPCR experiment were performed by using HiScript II One Step qRT-PCR SYBR Green Kit (#Q221-01) and miRNA Universal SYBR qPCR Master Mix (cat:#MQ101; Vazyme Biotech Co. Ltd., Nanjing, China), respectively. Then, the productions were conducted on an ABI 7900 system (Foster City, CA, USA) and calculated using the 2^–ΔΔCt method.

Table 1.

Primer pairs used for quantitative RT-PCR analysis.

Gene ID	Sequence (5′- 3′)
H-GAPDH F	TGTTCGTCATGGGTGTGAAC
H-GAPDH R	ATGGCATGGACTGTGGTCAT
H-DNM3OS F	GCACTGAATGCAGCAACAAT
H-DNM3OS R	TCTGTTTCGGCCATCACTTT
H-CDH11 F	GTATCCTCGAAGGACAACCCT
H-CDH11 R	GACATCGGTCAGTGTGATCGT
H-LEF1 F	TGCCAAATATGAATAACGACCCA
H-LEF1 R	GAGAAAAGTGCTCGTCACTGT
H–U6 F	CTCGCTTCGGCAGCACA
H–U6 R	AACGCTTCACGAATTTGCGT
H-miR-127-5p F	ACACTCCAGCTGGGCTGAAGCTCAGAGGGCTC
H-miR-127-5p R	CTCAACTGGTGTCGTGGA
R-GAPDH F	CCTCGTCTCATAGACAAGATGGT
R-GAPDH R	GGGTAGAGTCATACTGGAACATG
R-DNM3OS F	CTTGCTGTGCAGAACATCCG
R-DNM3OS R	TCAGGCTGGGTTGTCATGTG
R–U6 F	CTCGCTTCGGCAGCACA
R–U6 R	AACGCTTCACGAATTTGCGT
R-miR-127-5p F	ACACTCCAGCTGGGCTGAAGCTCAGAGGGCTC
R-miR-127-5p R	CTCAACTGGTGTCGTGGA

F: forward primer; R: reverse primer.

2.7. Western blotting analysis

The protein expression of the extracellular matrix (ECM)-related, included matrix metallopeptidase 13 (MMP13), collagen type II alpha 1 chain (Col2A1), Aggrecan (ACAN); cell apoptosis-related proteins included caspase3, BCL2, BCL2 Associated X (Bax); Wnt/β-Catenin included CDH11, Wnt3a, and β-Catenin were assessed by western blotting analysis. The proteins in cells and tissues were lysis by RIPA lysis buffer (#R0278, Sigma) and quantified by BCA Protein Assay Kit (Beyotime, Jiangsu, China). The standard sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) method was employed. In concisely, 20 μg protein was electrophoretically separated in 10 % sodium dodecyl sulfate separation gel under 100 V and concentration gel under 120 V. The protein on the gel was transferred onto the polyvinylidene fluoride (PVDF) membrane, which was washed with 25 mL TBST (Tween-20: TBS = 1:1000) for 5min, and then blocked by 5 % skimmed milk powder at 4 °C overnight. Then, the samples were incubated 1 h with primary antibodyMMP13 (#cat: ab39012, Abcam), Col2A1 (#cat: LS-C382807, LSBio), ACAN (#cat: ab3778, Abcam), caspase3 (#cat: ab184787, Abcam), Bcl2 (#cat: ab196495, Abcam), Bax (#cat: ab182733, Abcam), CDH11 (#cat: LS-B15836, LSBio), Wnt3a (#cat: LS-C313236, LSBio), β-catenin (#cat: ab32572, Abcam), and GAPDH (#cat: ab181602, Abcam) followed by three TBST washes (5 min/wash). The samples then oscillated with the secondary antibody of HRP-labeled goat anti-rabbit IgG (#cat: ab672, Abcam 1) and incubated at room temperature for 40min. The PVDF membrane was washed with 25 ml TBST (5 min each time). Put the membrane into the electrogenerated chemiluminescence (ECL) solution (ECL808-25, Biomiga, San Diego, CA) for 1min to act fully. Remove the excess liquid and cover the membrane with preservative film before being photographed and observed using the X-ray machine (36209ES01, Shanghai Qcbio Science and Technologies Inc., Shanghai, China). The net density value of the band was calculated by Image-Pro Plus 6.0 with the GAPDH as an internal reference band. All the experiments were tripled.

2.8. Cell proliferation assay

The cell activity was evaluated by Cell Counting Kit-8 (CCK-8, Solarbio, Beijing, China). 1 × 10⁴ SW1353 cells were seeded into a 96-well plate and cultured with IL-6 at 37 °C, then they were added with CCK-8 solution at 0, 12, 24, 36, 48, 56 h and incubated for an additional 2 h. Finally, the cell numbers were calculated under absorbance spectrometry at 450 nm.

2.9. Flow cytometry analysis

The Annexin V FITC/PI staining Apoptosis Detection kit (Sigma-Aldrich, Germany) was used to measure cellular apoptosis in cells. About 3 × 10⁴ cells/wells in 6-well plates were washed with cold PBS, then suspended and incubated with a binding buffer containing annexin V-FITC and PI stains. And finally, the cells were analyzed by flow cytometry (Agilent, USA) and NovoExpress software 14.1.

2.10. Targeting predicting and dual-luciferase reporter assay

The targeting sites of miR-127-5p and CDH11 were predicted by TargetScan and the miRbase database. The wild-type (WT) and mutant (MUT) binding sequences were synthesized by GenePharma (Shanghai, China) then inserted into the pGL3-promoter vectors (Thermo Fisher Scientific). Then, the vectors and miR-127-5p mimics were transfected into SW1353 cells. The Dual-Luciferase Reporter Assay System bought from Promega Corporation (Madison, WI, USA) was employed to detect luciferase activity.

To detect the binding sequence of LEF1 and DNM3OS promoter, the pGL3-promoter vectors of segmented DNM3OS were constructed. To confirm the binding sequence of LEF1 and DNM3OS promoter, dual-fluorescent report gene carrier of DNM3OS WT and MUT sequences was established. Then, the dual-fluorescent report gene carrier were co-transfected into SW1353 with overexpressed LEF1 plasmid and luciferase activity was detected using the above methods.

2.11. Immunofluorescence (IF) analysis

The IF analysis was conducted referred to Zhao et al. [23]. Briefly, after the cell reached 60%–70 %, they were fixed by 4 % paraformaldehyde, permeabilized with 0.5 % TritonX-100, blocked with 1 % BSA, incubated with primary anti-β-catenin (#cat: ab32572, Abcam) and FITC conjugates goat anti-rabbit secondary antibody (Sigma, Darmstadt, DE). For nuclear/cytosolic localization, the cells were stained by 4’, 6-diamidino-2-phenylindole (DAPI). Finally, β-catenin expression was observed under a confocal laser microscope (Olympus Optical, Tokyo, Japan).

2.12. Chromatin immunoprecipitation (ChIP)

ChIP assay was performed on LEF1 and DNM3OS promoters according to a previously described method [24]. Briefly, cells were crosslinked with formaldehyde for 10 min and then lysed by lysis buffer. After sonication and incubation with anti-LEF1 and anti-IgG pre-conjugated with protein G agarose beads overnight. Enriched DNA was extracted and qPCR was performed.

2.13. Electrophoretic mobility shift assay (EMSA) and antibody-supershift assay

Complementary oligonucleotides probe targeting DNM3OS promoter were synthetized and 5′-biotinylated, competitive probe without 5′-biotinylated and mutant competitive probe were also synthetized. Then, the nuclear proteins were extracted and incubated with probes. For supershift assay, purified LEF1 protein and anti-LEF1 were added into the mixtures of nuclear extract and probes. Next, the complexes were subjected to electrophoresis, transferred to a nylon membrane and immobilized, detected by chemiluminescent substrate.

2.14. Statistical analysis

GraphPad Prism 9 (San Diego, CA, USA) was employed for the statistical analysis. Unpaired student's t-test and one-way ANONA analyses were performed depending on the groups. Data are exhibited as mean ± standard deviation (SD). P < 0.05 was considered a significant difference.

3. Results

3.1. DNM3OS/miR-127-5p and wnt/β-catenin dysregulated in clinical samples

We collected the spinal facet joint and blood samples from the mild and severe spinal FJOA patient. The representative CT and MRI images were showed in Fig. 1A. And the Safranin O-fast Green Staining (SF) and Alcian Blue, Staining (AB) images were showed in Fig. 1B, which demonstrated that the cartilaginous layer is thicker and chondrocytes were less in severe group than that in mild group. We then evaluated the inflammation factors of IL-1β, IL-6, and TNF-α in the blood samples of mild and severe groups. IL-1β, IL-6, and TNF-α were significantly more powerful in severe group than in the mild group (P < 0.001; Fig. 1C). To assess DNM3OS/miR-127-5p and Wnt/β-Catenin on Spinal FJOA, their expression levels were evaluated in cartilage tissues of mild and severe groups. LncRNA DNM3OS and miR-127-5p exhibited opposed expression trends in two groups, with DNM3OS significantly upregulated, and miR-127-5p downregulated considerably in a severe group compared with the mild group (P < 0.001; Fig. 1D). Compared with the expression levels of Wnt3a and β-catenin in mild group, their expression levels were significantly elevated in the severe group (Fig. 1E). These results suggested that DNM3OS/miR-127-5p and Wnt/β-Catenin pathway might regulate the Spinal FJOA.

Fig. 1.

DNM3OS/miR-127-5p and Wnt/β-Catenin dysregulated in spinal FJOA. (A) The representative CT and MRI images of mild and severe FJOA patients. (B) O-fast Green Staining (SF) and Alcian Blue, Staining (AB) images of the spinal facet joint tissue of patients. (C) ELISA analysis on evaluating the levels of IL-1β, IL-6, and TNF-α in blood samples. (D) qPCR analysis on assessing the expression levels of lncRNA DNM3OS and miR-127-5p in cartilage tissues. (E) Western blotting analysis on detecting the expression levels of Wnt3a and β-Catenin in cartilage tissues. N = 10. ***P < 0.001. FJOA, facet joint osteoarthritis.

3.2. DNM3OS/miR-127-5p and wnt/β-catenin dysregulated in vitro

In our present study, SW1353 cells stimulated with IL-6 were used for the construction of an OA in vitro model. The cell activity was significantly reduced in inflammation groups at 48 h and 56 h (P < 0.001; Fig. 2A). Flow cytometry on cell apoptosis showed that the apoptosis-cell ratio was significantly higher in the Inflammation group than in Control (P < 0.001; Fig. 2B). ECM degradation is one of the markers of OA, we then detected the expression of ECM-related proteins, including MMP13, Col2A1, and ACAN. As shown in Fig. 2C, MMP13 was significantly highly expressed, and Col2A1 and ACAN were significantly lowly expressed in the Inflammation group compared with the Control group. In addition, the cell apoptosis-related proteins of caspase3, Bcl2, and Bax also demonstrated that cell apoptosis ratio is more severe in the Inflammation group than in the Control group. These results revealed a destroyed cartilage microenvironment of cells in the Inflammation group, reflecting the success of in vitro model construction. Then, we assessed the expression of DNM3OS/miR-127-5p and Wnt/β-Catenin in vitro. We found that the expression trend in vitro was consistent with that in clinical samples (Fig. 2D and E) exhibited in Fig. 1D and E.

Fig. 2.

DNM3OS/miR-127-5p and Wnt/β-Catenin dysregulated in vitro. (A) CCK-8 assay on selecting the optimal treatment time of IL-6. (B) Flow cytometry on detecting the cell apoptosis situation in Control and Inflammation groups. (C) Western blotting analysis on detecting the expression levels of ECM and cell apoptosis-related proteins. (D) qPCR analysis on assessing the expression levels of lncRNA DNM3OS and miR-127-5p. (E) Western blotting analysis on detecting the expression levels of CDH11, Wnt3a, and β-catenin. N = 3. ***P < 0.001. FJOA, facet joint osteoarthritis. ECM, extracellular matrix.

3.3. Knockdown of DNM3OS extenuated spinal FJOA by promoting cell viability, inhibiting cell apoptosis, and accelerating ECM accumulation, while DNM3OS overexpression had the opposite effect

To evaluate the role of DNM3OS in vitro, the two sh-DNM3OS plasmids (sh-DNM3OS-1 and sh-DNM3OS-2) and DNM3OS-overexpressed plasmid (DNM3OS), as well as vector were separately transfected into SW1353 cells, then expression levels of DNM3OS and miR-127-5p were assessed. QRT-PCR assay showed that DNM3OS was significantly decreased in sh-DNM3OS-1, sh-DNM3OS-2, and overexpressed in DNM3OS (P < 0.001; Fig. 3A). miR-127-5p exhibited opposed expressed trends with DNM3OS, which is significantly highly expressed in sh-DNM3OS-1 and sh-DNM3OS-2 groups, but lowly expressed in DNM3OS overexpression group (P < 0.001; Fig. 3A). The cell viability at 56 h was highest in sh-DNM3OS-1 and sh-DNM3OS-2 groups, followed by vector, and lowest in DNM3OS group (Fig. 3B). The ECM accumulation was promoted by sh-DNM3OS-1, sh-DNM3OS-2 plasmids, and inhibited by DNM3OS-overexpressed plasmid. Moreover, cell apoptosis-induced proteins caspase3 and Bcl2 were promoted, but cell apoptosis-inhibited protein Bax was inhibited by sh-DNM3OS-1, sh-DNM3OS-2, while DNM3OS overexpression had the opposite effect, and the expression of apoptosis-inhibited protein Bcl-2 was induced by sh-DNM3OS and suppressed by DNM3OS (Fig. 3C). Flow cytometry experiment results were also consistent with Western blot assay (P < 0.001; Fig. 3D). As for the Wnt/β-Catenin pathway, western blotting analysis showed that knockdown of DNM3OS (sh-DNM3OS) inhibited Wnt3a and β-Catenin expression, while DNM3OS overexpression promoted their expression (Fig. 3E). These results demonstrated that knockdown of DNM3OS extenuated OA by promoting cell viability, inhibiting cell apoptosis, and accelerating ECM accumulation, while DNM3OS overexpression had the opposite effect.

Fig. 3.

Knockdown of DNM3OS extenuated spinal FJOA by promoting cell viability, inhibiting cell apoptosis, and accelerating ECM accumulation, while DNM3OS overexpression had the opposite effect. (A) The expression levels of DNM3OS and miR-127-5p were evaluated in transfections by qPCR. (B) CCK-8 assay on evaluating the cell viability in transfections at different time points. (C) Western blotting analysis on detecting the expression levels of ECM, cell apoptosis-related proteins in transfections. (D) Flow cytometry on detecting the apoptosis cell ratio in transfections. (E) Western blotting analysis on detecting the expression levels of wnt3a and β-catenin in transfections. N = 3. ***P < 0.001 vs. vector. FJOA, facet joint osteoarthritis. ECM, extracellular matrix.

3.4. DNM3OS aggravated spinal FJOA by sponging miR-127-5p in vitro

To confirm the role of DNM3OS/miR-127-5p, the transfections of sh-DNM3OS and miR-127-5p inhibitor were constructed in SW1353 cells. Firstly, the expression levels of DNM3OS and miR-127-5p were detected to evaluate the transfections, and we found that DNM3OS was significantly downregulated in sh-DNM3OS and upregulated in miR-127-5p inhibitor. Moreover, at the combination transfection of sh-DNM3OS + miR-127-5p inhibitor, DNM3OS expression was significantly decreased compared to that in the miR-127-5p inhibitor group (P < 0.01, P < 0.001; Fig. 4A). The expression of miR-127-5p was significantly upregulated in sh-DNM3OS and downregulated in the miR-127-5p inhibitor group (P < 0.001; Fig. 4A). Similarly, sh-DNM3OS promotes the expression of miR-127-5p decreased by inhibitor, which revealed that sh-DNM3OS and miR-127-5p inhibitor altered the level of miR-127-5p and DNM3OS mutually. The qPCR results support the sponging role of miR-127-5p for DNM3OS. The cell activity results showed that the cells in the sh-DNM3OS group have the highest activity, followed by sh-DNM3OS + miR-127-5p inhibitor, and lowest in the miR-127-5p inhibitor group (Fig. 4B). At this point, the apoptosis cell ratio was significantly decreased in sh-DNM3OS and increased in miR-127-5p inhibitor (P < 0.001; Fig. 4C). Moreover, sh-DNM3OS and miR-127-5p inhibitors have counteracted regulation of cell apoptosis. This balances regulation between sh-DNM3OS and miR-127-5p inhibitor was also exhibited in regulating ECM and cell apoptosis-related proteins, in which knockdown of DNM3OS (sh-DNM3OS) could promote the ECM accumulation and decreased cell apoptosis by sponging miR-127-5p in vitro (Fig. 4D). The western blotting analysis showed that knockdown of DNM3OS (sh-DNM3OS) inhibited Wnt/β-Catenin expression in a counteract way with miR-127-5p (Fig. 4E). These results demonstrated that DNM3OS aggravated spinal FJOA by sponging miR-127-5p in vitro.

Fig. 4.

DNM3OS aggravated spinal FJOA by sponging miR-127-5p in vitro. (A) The expression levels of DNM3OS and miR-127-5p were evaluated in transfections by qPCR. (B) CCK-8 assay on selecting the optimal treatment time point of IL-6 in transfections. (C) Flow cytometry on detecting the apoptosis cell ratio in transfections. (D, E) Western blotting analysis on detecting the expression levels of ECM, cell apoptosis-related proteins, and proteins involved in Wnt/β-Catenin in transfections. (F) qPCR analysis on assessing the expression levels of CDH11 in transfections. (G) Bioinformatics analysis on revealing the miR-127-5p binding sites of CDH11. (H) Dual-luciferase reporter assay on confirming the targeting role of miR-127-5p on CDH11. N = 3. ***P < 0.001 vs. vector. ###P < 0.001 vs. inhibitor. FJOA, facet joint osteoarthritis. ECM, extracellular matrix.

We then explore the mechanism by which DNM3OS/miR-127-5p regulated OA and Wnt/β-Catenin pathway. It was precited by TargetScan that miR-127-5p might bind to CDH11, which participated in Wnt/β-Catenin pathway. Firstly, we assessed the expression of CDH11 was inhibited by sh-DNM3OS and promoted by miR-127-5p inhibitor (Fig. 4E and F). The binding sites of miR-127-5p on CDH11 mRNA predicted by TargetScan was showed in Fig. 4G. Dual-luciferase reporter assay also confirmed the binding between miR-127-5p and CDH11 mRNA. The ratio of luciferase activity was significantly decreased by miR-127-5p mimics in WT group compared with NC, while mimics had no effect in MUT group (Fig. 4H). These results confirmed that CDH11 is the target gene of miR-127-5p, and CDH11 might play a role in OA regulated by DNM3OS/miR-127-5p.

3.5. miR-127-5p targets CDH11 and regulate spinal FJOA in vitro

After verifying the binding between miR-127-5p and CDH11 mRNA, their regulation on spinal FJOA was also evaluated in vitro experiment. Overexpressed miR-127-5p and CDH11 cells were constructed. As showed in CCK-8 assay, the cell activity was promoted by miR-127-5p mimics and inhibited by overexpressed CDH11 plasmid (P < 0.001; Fig. 5A). miR-127-5p mimics suppressed cell apoptosis and Wnt/β-Catenin pathway, induced ECM accumulation, while CDH11 overexpression had the opposite effect (P < 0.001; Fig. 5B–E). In addition, the expression level of CDH11 was considerably reduced in mimics, and raised in overexpressed CDH11 groups, which altered the effect of mimics in both western blotting and qPCR analyses (Fig. 5D and F). These results indicated that miR-127-5p targets CDH11 and regulates spinal FJOA in vitro. Moreover, qPCR analyses indicated that miR-127-5p and CDH11 were regulated by each other, and DNM3OS was also regulated by miR-127-5p and CDH11, which implied that there might be a feedback loop in DNM3OS/miR-127-5p/CDH11 (Fig. 5F). Taken together, our results confirmed that DNM3OS was overexpressed in spinal FJOA, which inhibited miR-127-5p, and promoted the expression level of CDH11 through the ceRNA network, thus activating Wnt/β-Catenin pathway and aggravating the progress of spinal FJOA.

Fig. 5.

miR-127-5p targets CDH11 and regulates spinal FJOA in vitro. (A) CCK-8 assay on selecting the optimal treatment time point of IL-6 in transfections. (B) Flow cytometry on detecting the apoptosis cell ratio in transfections. (C, D) Western blotting analysis on detecting the expression levels of ECM, cell apoptosis-related proteins, and proteins involved in Wnt/β-Catenin in transfections. (E) IF research on evaluating the expression of CDH11 in transfections. (F) qPCR analysis on assessing the expression levels of DNM3OS, miR-127-5p, and CDH11 in transfections. N = 3. ***P < 0.001 vs. vector. ###P < 0.001 vs. CDH11. FJOA, facet joint osteoarthritis. ECM, extracellular matrix.

3.6. LEF1 promotes DNM3OS transcription

As mentioned above, there might be a feedback loop in DNM3OS/miR-127-5p/CDH11, we then wonder whether Wnt/β-Catenin pathway regulated DNM3OS expression in turn. We first detect the expression of DNM3OS in cells transfected with LEF1 overexpression plasmid, and found that LEF1 overexpression promoted the expression of DNM3OS (Fig. 6A). PROMO dataset also predicted LEF1 as a potential transcription factor of DNM3OS. We then performed CHIP-qPCR and EMSA analysis to check whether LEF1 bind to DNM3OS promoter. EMSA and antibody-supershift analysis revealed that LEF1 interacted with the core promoter sequence of DNM3OS in vitro (Fig. 6B). The ChIP-qPCR results showed that the enrichment folds on the promoter of DNM3OS using anti-LEF1 was several times that of anti-IgG (P < 0.001; Fig. 6C). Next, segmented DNM3OS promoter was inserted into dual-luciferase reporter pGL3 to detect the binding sequence of LEF1 on DNM3OS promoter. The results showed that those reporter with higher luciferase activity all contained −1166 to −1159 sequence of DNM3OS promoter, which was predicted as the binding sequence of LEF1 by PROMO dataset (Fig. 6D). To confirm the binding sequence, dual-luciferase reporter containing wild type or mutant type sequence was co-transfected into cells with LEF1 overexpression plasmid or sitimulated with IL-6 (Fig. 6E). It was found that both LEF1 overexpression or IL-6 sitimulation promoted the luciferase activity in WT reporter, but not in MUW reporter (Fig. 6F).

Fig. 6.

LEF1 regulates DNM3OS promoters. (A) The segmented plasmid of the DNM3OS promoter was constructed. (B) Dual-luciferase reporter assay on evaluating the binding sites. (C) Schematic of the DNM3OS-promoter showing the position of the LEF1 binding sites and the locations of ChIP-qPCR fragments. (D) Dual-luciferase reporter assay on evaluating the ratio of luciferase activity of WT and MUT-DNM3OS. (E) ChIP-qPCR on assessing the enrichment of DNM3OS promoter in si-NC and si-LEF1. (F) EMSA and antibody-supershift analysis on revealing the interaction between LEF1 and DNM3OS. N = 3. ***P < 0.001.

3.7. LEF1 positively feedback DNM3OS transcription

After confirming the targeting role of LEF1 on DNM3OS, further functional analysis on OA was performed in vitro. As shown in Fig. 7A, the cell activity increased by sh-LEF1, but further decreased by DNM3OS overexpression (P < 0.001). Not only in cell activity, the alterative regulation between sh-LEF1 and DNM3OS was also observed in cell apoptosis (Fig. 7B), ECM- and cell apoptosis-related protein expression (Fig. 7C). The Wnt/β-Catenin pathway was activated by overexpressed DNM3OS and inhibited by sh-LEF1 (Fig. 7D and E). CDH11 expression was consistent with that of the Wnt/β-Catenin pathway in transfections. LEF1 expression evaluated by western blotting, IF, and qPCR showed that overexpressed DNM3OS have no effects on elevating the expression of LEF1. While, DNM3OS/miR-127-5p was significantly influenced by sh-LEF1 (Fig. 7F), indicating a positive feedback regulation of LEF1 on DNM3OS/miR-127-5p (P < 0.001; Fig. 7D–F). These results demonstrated that LEF1 positive feedback promotes DNM3OS transcription in spinal FJOA in vitro.

Fig. 7.

LEF1 positively feedback DNM3OS transcription in spinal FJOA in vitro. (A, B) CCK-8 assay and flow cytometry analyses on evaluating the cell activity and cell apoptosis in transfection. (C, D) Western blotting analysis on detecting the expression levels of ECM, cell apoptosis-related proteins, and proteins involved in Wnt/β-Catenin in transfections. (E) IF analysis on evaluating the expression of CDH11 in transfections. (F) qPCR analysis on assessing the expression levels of DNM3OS, miR-127-5p, and CDH11 in transfections. N = 3. ***P < 0.001 vs. vector. ###P < 0.001 vs. sh-LEF1. FJOA, facet joint osteoarthritis. ECM, extracellular matrix.

3.8. Overexpressed CDH11 aggravates OA by enhancing the wnt/β-catenin pathway

XAV-939 stabilizes AXIN by inhibiting the agglomeration of ADP-nucleic sugar-based enzymes Tankyrase 1 and Tankyrase 2, and stimulated β-Catenin degradation [25,26]. Therefore, the Wnt/β-Catenin pathway inhibitors XAV-939 were employed to assess the CDH11/Wnt regulation mechanism. XAV-939 counteracts CDH11 by promoting cell activity, inhibiting cell apoptosis, and ECM accumulation (Fig. 8A–C). The western blotting, IF, and qPCR analyses demonstrated that XAV-939 downregulated the expression of β-Catenin and CDH11 (Fig. 8D–F). Moreover, DNM3OS was also significantly downregulated by XAV-939, which can be altered by CDH11 (P < 0.001; Fig. 8F). These results demonstrated that overexpressed CDH11could activate Wnt/β-Catenin pathway to aggravate OA. Taking the results together, we could conclude that upregulated CDH11 activate Wnt/β-Catenin pathway, which cooperate with LEF1 to regulate the transcription of DNM3OS, promoting the positive feedback of DNM3OS, therefore, accelerating the progression of spinal FJOA.

Fig. 8.

Overexpressed CDH11 in FJOA can activate the Wnt/β-Catenin pathway. (A, B) CCK-8 assay and flow cytometry analyses on evaluating the cell activity and cell apoptosis in transfection. (C, D) Western blotting research on detecting the expression levels of ECM, cell apoptosis-related proteins, and proteins involved in Wnt/β-Catenin in transfections. (E) IF research on evaluating the expression of LEF1 in transfections. (F) QPCR analysis on assessing the expression levels of CDH11 and DNM3OS in transfections. N = 3. ***P < 0.001 vs. vector. ###P < 0.001 vs. CDH11. FJOA, facet joint osteoarthritis. ECM, extracellular matrix.

3.9. DNM3OS/miR-127-5p/CDH11 ceRNA network promotes spinal FJOA in vivo

The in vivo rat model was divided into four groups: Sham, OA, OA + Lv-sh-DNM3OS, and OA + Lv-sh-DNM3OS + antagomir. As observed in Fig. 9A, the cartilage in OA group were seriously damaged, which was relieved by sh-DNM3OS, but further aggravated by miR-127-5p antagomir. Moreover, the OARSI score in each group was consistent with sections staining. WB assay showed that OA group had less ECM accumulation and more cell apoptosis, sh-DNM3OS relieved the change, while antagomir had the opposite effect (Fig. 9B). The inflammation cytokines levels increased in OA group, decreased by sh-DNM3OS and further elevated by antagomir (Fig. 9C). Additionally, the expression levels of DNM3OS, miR-127-5p, Wnt/β-Catenin, CDH11, and LEF1 in vivo were consistent with that in vitro, in which Wnt/β-Catenin pathway was inhibited by sh-DNM3OS, but promoted by antagomir (Fig. 9D and E).

Fig. 9.

DNM3OS/miR-127-5p/CDH11 ceRNA network promotes spinal FJOA in vivo. (A) HE staining, Alcian blue staining (AB), and Toluidine Blue O staining staining (TBO) on cartilage tissues. (B) Western blotting analysis on detecting the expression levels of ECM, cell apoptosis-related proteins in transfections. (C) ELISA analysis on evaluating the levels of IL-1β, IL-6, and TNF-α. (D) qPCR analysis on assessing the expression levels of DNM3OS and miR-127-5p in transfections. (E) Western blotting research on detecting the expression levels of Wnt/β-Catenin. HE, hematoxylin, and eosin. TBO, Tolueine blue O. N = 6. **P < 0.01, ***P < 0.001 vs. sham. #P < 0.05, ###P < 0.001 vs. OA. $P < 0.05, $$P < 0.01, $$$P < 0.001 vs. OA + Lv-sh-DNM3OS. FJOA, facet joint osteoarthritis. ECM, extracellular matrix.

4. Discussion

Spinal FJOA is a type of spinal OA that is rarely explored in terms of ceRNA networks. Through these experiments, we can draw the following conclusions (Fig. 10): (1) DNM3OS is overexpressed in spinal FJOA, which inhibits miR-127-5p, and promotes the expression level of CDH11 through the ceRNA network, thus aggravating the progress of spinal FJOA. (2) Overexpressed CDH11 enhanced the Wnt/β-Catenin pathway, which cooperate with LEF1 to induce the transcription of DNM3OS, promoting the positive feedback of DNM3OS, therefore, accelerating the progression of spinal FJOA.

Fig. 10.

The mechanism flow chart of DNM3OS/miR-127-5p/CDH11 on spinal FJOA. We can observe that DNM3OS/miR-127-5p/CDH11 positively adjust β-Catenin/LEF-1 complex, thus promoting spine facet joint osteoarthritis.

Our present study verified the ceRNA network of DNM3OS/miR-127-5p/CDH11 in spinal FJOA for the first time. Although DNM3OS is rarely explored in OA, miR-127-5p has been found to lowly express in spinal FJOA by previous researches. For examples, miR-127-5p was downregulated in OA, which was regulated by circular RNA circ_0128846, thus promoting the progression of OA [27]. Li et al. [28] revealed that miR-127-5p was downregulated in OA, competitively bind to MMP-13 and circ_0136474, and suppressed the cell proliferation. CDH11, a protein located in integral membrane and stabilized β-Catenin, was predicted as one of the miR-127-5p target genes by TargetScan database. Considering the contributing role of Wnt/β-Catenin pathway on OA, we speculated that miR-127-5p might regulate Wnt/β-Catenin pathway by targeting to CDH11 mRNA, therefore paly a role on FJOA. In our present study, we found that miR-127-5p bound to and degraded CDC5L mRNA, which was competitively interfered by lncRNA DNM3OS, and the highly expressed DNM3OS aggravated spinal FJOA by promoting ECM degradation and cell apoptosis. The microenvironment of ECM changes promotes the decomposition metabolism of cartilage cells, eventually leading to cartilage destruction [29]. Therefore, DNM3OS/miR-127-5p/CDH11 might regulate the ECM degradation and cell apoptosis in spinal FJOA integration.

After confirmation of DNM3OS/miR-127-5p/CDH11 on spinal FJOA, we further investigated how it acted. We found that CDH11 was highly expressed in spinal FJOA, which was consistent with that in OA [17,30]. Vaamonde-Garcia et al. [30] demonstrated that the expression level of CDH11 was significantly increased in OA. The upregulated CDH11 activates the β-catenin/LEF1 complex, which regulates DNM3OS regulation. Moreover, this regulation was positive feedback. The function of β-catenin/LEF1 complexes has been studied in various diseases, including OA [31]. Excessive wnt/β-catenin pathway has been confirmed to induce OA [32]. As the co-activator of wnt/β-catenin pathway, LEF1 is also found to overexpress in OA tissue [33]. Chen et al. [31] demonstrated that CircMYO10/miR-370-3p/RUVBL1 axis promotes OA progression by regulating to encourage chromatin remodeling and thus enhances the transcriptional activity of the β-catenin/LEF1 complex. Our second hypothesis was also verified. Combined with the first hypothesis, our present study revealed that DNM3OS/miR-127-5p/CDH11 positively adjusts Wnt/β-Catenin/LEF-1 complex, which promotes ECM degradation and cell apoptosis, thus accelerating spinal FJOA. As discussed above, most of the regulation in spinal FJOA was similar to that in OA, which might make the insights into whether this mechanism is also available in other kinds of OA needed to be verified and might be very important.

5. Conclusions

In summary, we found that DNM3OS/miR-127-5p/CDH11 enhanced Wnt3a/β-Catenin pathway, which cooperated with LEF1 to promoted DNM3OS transcription, forming a positive feedback to aggravate spinal FJOA.

Institutional review board statement

Our study was approved by the Ethical Committee of Guangdong Second Provincial General Hospital (AE-202004-3).

Informed consent statement

Informed consent was obtained from all individual participants included in the study.

Declaration of competing interest

The authors declared that they have not competing interest in this research.

Funding information

This study is supported by Science and Technology Projects in Jieyang Science and Technology Bureau (No. ylxm037), Science and Technology Projects in Guangzhou (No. 202102020485), Doctoral Workstation Foundation of Guangdong Second Provincial General Hospital (No. 2019BSGZ005), The Youth Scientific Research foundation of Guangdong Second Provincial General Hospital (No. YQ2017-011), Doctoral Workstation Foundation of Puning City TCM Hospital (No. BSGZZ2021001), Scientific Research Project of Guangdong Traditional Chinese Medicine Bureau (No. 20221474) and Zhongshan City Social Welfare and Basic Research Project (Medical and Health) (No. 2022B3014).

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article.

CRediT authorship contribution statement

Jing Wang: Writing – original draft, Validation, Supervision, Software, Methodology, Investigation. Zhenyu Yang: Writing – original draft, Validation, Supervision, Software, Methodology, Data curation. Xiuming He: Software, Investigation, Data curation. Yeyang Wang: Validation, Software, Data curation. Dixin Luo: Software, Methodology, Data curation. Wangyang Xu: Validation, Software, Data curation. Hongtao Zhang: Validation. Xiaozhong Zhou: Writing – review & editing, Validation, Resources, Project administration, Methodology, Funding acquisition, Conceptualization.

Appendix A. Supplementary data

The following is the Supplementary data to this article: