Circular RNA circ_0000423 regulates cartilage ECM synthesis via circ_0000423/miRNA-27b-3p/MMP-13 axis in osteoarthritis

Expression of circRNAs in control and OA specimens

To evaluate the different circRNAs expression between control and OA tissues, Microarray analysis was performed. As shown in Figure 1A, 1B, Hierarchical clusters and scatter plots displayed the circRNAs expression between control and OA tissues, meanwhile, the significant differences between the two groups were shown in volcano plots filtering (Figure 1C). Based on the Microarray data analysis, 11 circRNAs were down-regulated and 101 circRNAs were up-regulated when compared to control tissues (Supplementary Material). According to the microarray data, we chose some circRNAs which showed significant changes among different circRNAs expressions to further confirm the significant changes among different circRNAs expressions with T-qPCR. The results showed that the expression of circ_0000423 was significantly increased in OA tissues (p < 0.05) (Figure 1E–1G).

Figure 1. Differential expression of circRNAs in normal and OA specimens. (A) Hierarchical clustering analysis of the differentially expressed circRNAs between 4 individuals normal and 4 individuals OA specimens; (B) Scatter plot of circRNAs expression between normal and OA specimens. The circRNAs below the bottom green line and above the top green line illustrate >2.0-fold changes between the normal and OA specimens. (C) Volcano plots of the differentially expressed circRNAs. The red point in the plot represents the differentially down or up-regulated circRNAs with statistical significance. (D–G) Expression of significant circRNAs in normal and OA specimens.

Expression of Circ_0000423 is up-regulated in OA tissues and IL-1β induced chondrocyte

To investigate the Circ_0000423 expression further, RT-qPCR was used to IL-1β induced chondrocytes. The previous results confirmed that Circ_0000423 was significantly increased in OA tissues (p < 0.05) (Figure 1D), and similar results was analyzed in IL-1β induced chondrocyte compared with the normal chondrocyte (Figure 2A, 2B). Sanger sequencing of the RT-qPCR products confirmed the back splice junction of Circ_0000423 (Figure 2C); the presence of Circ_0000423 was amplified by divergent primers from cDNA, but not from genomic DNA (Figure 2D). Furthermore, the RNase R was used to confirm the cyclic structure of Circ_0000423 (Figure 2E).

Figure 2. Expression of Circ_0000423 in cartilage and chondrocyte. (A) Expression of Circ_0000423 in chondrocyte with or without IL-1β induced at different time point; (B) Expression of Circ_0000423 in chondrocyte with or without IL-1β induced at different concentrations; (C) Sanger sequencing was used to illustrate the presence of Circ_0000423; (D) The RT-PCR was used to confirm the presence of Circ_0000423. Divergent primers amplified Circ_0000423 from cDNA, but not from genomic DNA; (E) The expression of Circ_0000423 and Line_0000423 mRNA treated with or without RNase R was measured by RT-qPCR; Data are presented as mean ± SEM. ^*P < 0.05 compared with the control group.

Circ_0000423 could be directly targeted by miRNA-27b-3p

According to the microarray data, some circRNAs were chosen to predict the targets. The results of gene co-expression networks showed that Circ_0000423 could be directly targeted by miRNA-27b-3p (Figure 3A). Moreover, the sequences of Circ_0000423 3′-UTR could match the miRNA-27b-3p target through bioinformatics algorithms analysis (Figure 3B). Meanwhile, Luciferase assay confirmed results which Circ_0000423 bind to miRNA-27b-3p. Circ_0000423-psiCHECK-2 vector and miRNA-27b-3p mimic were co-transfected into HEK-293 cells, the results of luciferase activity were significantly lower compared with the control group, however, after miRNA-27b-3p binding sites on Circ_0000423 were mutated, the results could be reversed (Figure 3C). These results demonstrate that Circ_0000423 could be directly targeted by miRNA-27b-3p and act as a miRNA-27b-3p sponge.

Figure 3. Circ_0000423 directly targeted by miRNA-27b-3p. (A) circRNAs-miRNA-mRNA network (B) Targeted miRNA-27b-3p relationship between Circ_0000423 and MMP-13; (C) Luciferase reporter gene assay confirmed the targeted relationship between Circ_00004234 and miRNA-27b-3p. Data are presented as mean ± SEM. ^*P < 0.05.

Functions of Circ_0000423 in OA by targeting miRNA-27b-3p expression as a ceRNA

To analyze whether the functions of Circ_0000423 in OA were through targeting miRNA-27b-3p, we measured the mRNA and protein expression of MMP-13 and collagen II (COL-2) in chondrocytes by co-transfecting relative Circ_0000423 and miRNA-27b-3p, respectively. As shown in Figure 4, we overloaded the Circ_0000423 expression (p-circ) into chondrocytes, which decreased the expression of miRNA-27b-3p and collagen II, and this rate could be reversed by knocking down the Circ_0000423 expressions (si-circ1-2). However, the opposite results were found on the expression of MMP-13. In addition, co-transfection with miRNA-27b-3p mimic, the effects of p-circ_ 0000423 on the expression of MMP-13 and collagen II could be rescued. Besides, co-transfection with miRNA-27b-3p inhibitor, the effects of si-circ_0000423 on the expression of MMP-13 and collagen II were also shown to be rescued. These results demonstrate that Circ_0000423 could affect the expression of MMP-13 and collagen II by targeting miRNA-27b-3p expression as a ceRNA in OA.

Figure 4. Functions of Circ_0000423 in OA by targeting miRNA-27b-3p expression as a ceRNA. (A) The mRNA expression of Circ_0000423; (B) The mRNA expression of miRNA-27b-3p; (C) The mRNA expression of MMP-13; (D) The mRNA expression of COL-2; (E) The proteins expression of MMP-13 and COL-2; (F, G) The relative protein expression of COL-2 and MMP-13are demonstrated with bar graphs. (H) The relative protein expression of COL-2 and MMP-13 are demonstrated with Immunofluorescence analysis. (I, J) The relative fluorescence expression of COL-2 and MMP-13 are demonstrated with bar graphs. P-circ: overexpression of Circ_0000423; si-circ-1-3: siRNA against Circ_0000423; Data are presented as mean ± SEM. ^*P < 0.05 compared with the NC group, ^#P < 0.05 compared with the p-circ group, ^&P < 0.05 compared with the si-circ-1 group.

Circ_0000423 regulates OA progression in ACLT (anterior cruciate ligament transection) mice

To investigate cartilage surface histopathological changes after intra- Circ_0000423 articular injection in ACLT mice, we used the HE (Figure 5A) and Safranin O-fast green (Figure 5B). The ACLT groups showed more severe degenerative OA changes than the control group, including articular cartilage destruction and erosion, articular surface fibrillation, and osteophyte formation (Figure 5A, 5B). These modifications, however, were inhibited following shRNA-Circ 0000423 treatment. Moreover, the Osteoarthritis Research Society International (OARSI) grade scores results were consistent with histological analyses (Figure 5C). Meanwhile, immunohistochemistry analyses revealed MMP-13 and collagen II expressions in articular cartilage. The percentage of chondrocytes positive for MMP-13 was significantly lower in the shRNA-Circ_0000423 group compared to the ACLT, but it was much higher in the ACLT (Figure 5D–5G). Altogether, these results demonstrated that the shRNA-Circ_0000423 intra-articular injection can slow down OA progress by extracellular matrix (ECM) metabolic homeostasis maintenance and showed its therapeutic potential for OA.

Figure 5. Circ_0000423 regulates the progression of OA in ACLT mice. (A, B) Histological analysis of OA was measured by HE staining and Safranin O staining. (C) OARSI scores were used to measure the progression of OA. (D–F) Immunohistochemical analyses of MMP-13 and Collagen II in sagittal sections of the medial condyle. (E–G) Quantification of MMP-13 and Collagen II-positive cells. P values were computed vs. controls group^* or ACLT group^#; (^*, ^#) P < 0.05. (Scale bar = 50 μm).

Silencing of Circ_0000423 suppresses bone resorption induced by ACLT in vivo

In the OA initial period, bone loss is crucial for bone remodeling [14]. To illustrate Circ_0000423 effects on OA subchondral bone remodeling, it was intra-articular injected shRNA-Circ_0000423 in mice after ACLT. When compared to the Sham group, bone resorption caused by increased osteolysis in the tibia subchondral bone was significantly higher in the ACLT-induced OA group (Figure 6A, 6B). However, after shRNA-Circ_0000423 treatment, bone resorption decreased and led to increasing bone mass in the subchondral bone and joint bone integrity. This indicated that shRNA-Circ_0000423 treatment can reduce osteolysis in vivo. Meanwhile, the tibial plateau subchondral bone volume (BV) in the ACLT (0.15 ± 0.46 mm3) was lower compared to the Sham group (0.39 ± 0.05 mm3; p < 0.05). However, after shRNA-Circ_0000423 treatment, BV significantly increased (0.34 ± 0.55 mm3, p < 0.05). Moreover, other bone-related parameters (Tb. N, BV/TV, Tb. Sp and Tb. Th) further confirmed the shRNA-Circ_0000423 bone destruction prevention effect in vivo (Figure 6C).

Figure 6. Silencing of Circ_0000423 suppresses bone resorption induced by ACLT in vivo. (A) Three-dimensional micro-CT images of frontal views of the knee joints at 8 weeks after the operation. (B) Sagittal views of medial compartment subchondral bone. (C) Quantitative analysis of BV, BV/TV, Tb.Sp, Tb.N and Tb.Th. P-values were computed vs. controls group^* or ACLT group^#; (^*, ^#) P < 0.05. (Scale bar = 1 mm).

Clinical specimens

OA cartilage was obtained from knee joints undergoing total knee arthroplasty operation (n = 4, age 62–70 years), and normal articular cartilage specimens were isolated from patients with femoral neck fractures (n = 4; ages 60–68 years) with no history of OA, who received total hip arthroplasty surgery. All of the patients gave their informed permission. The Ethics Committee on Human Experimentation at Sun Yat-Sen University’s First Affiliated Hospital in China approved the study (IRB: 2011011). Table 1 summarizes the patient information.

Microarray and bioinformatics analysis

Specimens were analyzed by the Arraystar Human circRNA Array V2 analysis, and then total RNA from each sample was analyzed by the Nanodrop ND-1000. The sample preparation and microarray hybridization were performed according to Arraystar’s standard protocols [43, 44]. In brief, total RNAs were digested with RNase R (Epicentre, Inc.) to remove linear RNAs and enrich circular RNAs. Then, the enriched circular RNAs were measured by Arraystar Human circRNA Array V2 (8x15K, Arraystar). The Agilent Scanner G2505C was used to scan the array images, which were then extracted using Agilent Feature Extraction software (version 11.0.1.1). The R software limma package was used to do quantile normalization and further data processing. The circRNA data, as well as common target mRNAs and miRNAs of the circRNAs, were used to create the mRNA-miRNA-circRNA network. The circRNA/miRNA/mRNA interaction was predicted with Arraystar’s homemade miRNA target prediction software according to miRanda [11] and TargetScan [45].

Primary chondrocyte isolation and cell culture

Primary chondrocyte isolation was isolated from articular cartilage tissues [12]. In brief, the articular cartilage tissues were sliced into small pieces less than 1 mm³ and then stirred to sequential digestion in pronase (90 min) and collagenase P (12 h) at 37°C. Primary articular chondrocytes were maintained in DMEM/F12 (Gibco Life Technology, USA) contained with 5% FBS, 100U/ml penicillin, and 100U/ml streptomycin at 37°C in 5% CO₂ humidified atmosphere. Every two days, the medium was changed. The following investigations used first-passage chondrocytes. The chondrocytes were treated with IL-1β (PeproTech, China) according to the experiment design.

Plasmid construction

The construction of circRNA over-expression and Luciferase-hsa_circ_0000423 plasmids were synthesized by Obio Technology (Shanghai, China). The construct method was carious out according to previously described [46]. The resulting construct (pcDNA3.1(+)-S-hsa_circ_0000423) was verified by direct sequencing with the hsa_circ_0000423-F (CGCAAATGGGCGGTAGGCGTG) and hsa_circ_ 0000423-R (TAGAAGGCACAGTCGAGG) primers. The resulting construct (pMIR-REPORT Luciferase-hsa_circ_0000423(WT)) was verified by direct sequencing with the hsa_circ_0000423-F(GCTCAT AGGCCGGCATAGACGCGTCCCTGTAAGACCAGTAATAAATTTCT) and hsa_circ_0000423-R (AAA TAAAAGATCCTTTATTAAGCTTCTGGGTGGCCTTTTCCAACT) primers. The resulting construct (pMIR-REPORT Luciferase-hsa_circ_0000423(MUT)) was verified by direct sequencing with the hsa_circ_ 0000423-F(AAAAATTAATCAGTTTTGAGTGGCA GCGGCGATGCTACTCTTTTGAACTGTCAAGTT) and hsa_circ_0000423-R(GAGTAGCATCGCCGCTG CCACTCAAAACTGATTAATTTTTGTGATATATAATAAGTGG) primers.

RNA transfection

Chondrocytes transfection was used by Lipofectamine 2000 (Gibco Life Technology, USA) based on the manufacturer’s instructions [47]. PcDNA3.1-Circ_0000423, si-Circ_0000423-1, and si-Circ_0000423-2 (small interfering RNAs for Circ_0000423) were synthesized by Obio Technology (Shanghai, China); miR-27b-3p mimics, miR-27b-3p inhibitor, and negative control (NC) were purchased from RiboBio Technology (Shanghai, China).

Quantitative real-time polymerase chain reaction (qRT-PCR)

Trizol reagent was used to extract total RNA from chondrocytes and cartilage tissues (Invitrogen, USA). To digest RNaseR, the samples were heated to 37°C for 25 minutes and digested with RNaseR (2 U/mg, Epicenter). For circRNA and mRNA analysis, the total RNA was reversed via the First cDNA Synthesis Kit (Takara, Japan), and RT-PCR was carried outperformed using the SYBR green system (Toyobo, Japan). For miRNA analysis, total RNA was reversed using the miRNeasy Mini Kit (QIAGEN, CA, USA), and RT-PCR was performed using the YBR Premix Ex TaqTM II (Takara, Japan). The circRNA and miRNA expressions were normalized against U6 expression and the mRNA expression was normalized against GAPDH expression. Data were quantified by the 2^−ΔΔCT methods. Primer sequences are displayed in Table 2.

Western blot analysis

Western blot experiments were carious out according to previously described [48]. To conclude, total protein was extracted with lysis buffer (Beyotime, China) and measured with a BCA Protein Assay Kit (Beyotime, China). Then, 50 μg of protein was separated on 10% SDS-PAGE and transferred onto 0.2 μm polyvinylidene difluoride membranes (Millipore, USA). The membranes were incubated with primary antibodies against MMP-13 (1:1000 dilutions, Abcam, UK) and collagen II (1:1000 dilution, Abcam, UK) overnight at 4°C after blocking with 5% non-fat milk for 30 min. GAPDH was used as a control. Then, the membranes were incubated with respective secondary antibodies conjugated with HRP (1:1000, cell signaling technology) for 60 min at room temperature. Band intensity was captured using chemiluminescence imaging reagents (Bio-Rad, USA) were used to measure Band intensity, and ImageJ software (National Institutes of Health, Bethesda, MD) was used to analyze the data.

Immunofluorescence

Immunofluorescence experiments were carious out according to previously described [49]. Chondrocytes from various treatment groups were washed twice in PBS before being used to make cell smears. The cells were fixed for 30 minutes at room temperature with 4 percent paraformaldehyde, washed twice with PBS, and then treated for 15 minutes at room temperature with 0.1% Triton-X 100 to render them permeable. After washing the permeate, Collagen II (1:200 dilution, Abcam, UK) and MMP-13 antibody (1:200 dilution, Abcam, UK) were added and the reaction was incubated overnight at 4°C. After washing out the first antibody, a fluorescently labeled secondary antibody (1:200, Abcam, UK) was added and incubated at room temperature for 1 hour before being washed three times with PBS. After adding DAPI staining and washing out the staining solution the cells were observed under the fluorescence microscope (Olympus DP80, Japan).

Dual-luciferase assays

The linear sequence of Circ_0000423 was cloned into the psiCHECK-2 vector (Promega, China), and the potential binding sites of miR-27b-3p were mutated. pcDNA3.1 empty vector (EV) with either Circ_0000423 WT or Circ_0000423 MUT was co-transfected into HEK-293 cells with miR-27b-3p mimics or negative control (NC) based on the manufacturer’s instructions of Lipofectamine 2000 (Gibco Life Technology, USA). The cell lysates were collected 48 hours after transfection. Renilla luciferase activity was measured using the Dual Glo Luciferase Assay System (Promega).

OA mouse model establishment by anterior cruciate ligament transection (ACLT)

Ten weeks-old male C57 BL/6 mice were randomly assigned into four groups. Each group consisted of 8 mice: Sham, ACLT, AAV shRNA-NC and AAV shRNA-Circ_0000423. The AAV shRNA-NC and AAV shRNA-Circ_0000423 were constructed by HANBIO (Shanghai, China). Then, 3% pentobarbital sodium (35 mg/kg) was injected intraperitoneally for general anesthesia. The process was carried out in an aseptic condition. The mice were acclimated for one week before the operation. We used ACLT to create the mouse OA model, as previously described [50]. The AAV shRNA-N Cand AAV shRNA-Circ_0000423 groups were intra-articularly injected with 10 μL of AAV shRNA-NC and AAV shRNA-Circ_0000423 through the medial parapatellar area at one week after the operation. The Sham and ACLT groups were intra-articularly injected with the same volume of vehicle. After 8 weeks, all mice were put to death. We harvested knee joints and fixated them in 4% paraformaldehyde for follow-up experiments. Animal experiments were carried out with the approval of the Institutional Ethics Committee of Sun-Yat University’s First Affiliated Hospital. All animal studies were carried out following the Institutional Animal Ethics Committee (IAEC) guidelines for the use of laboratory animals.

Evaluations with micro-CT scanning

Specimens were placed in Eppendorf tubes containing 70% ethanol and scanned using micro-CT (Skyscan 1172, Bruker, Belgium) as previously described [50, 51]. The entire tibial plateaus subchondral bone was set as a region of interest (ROI). We analyzed three-dimensional anatomical structures, including bone volume (BV), bone volume/total tissue volume (BV/TV), trabecular number (Tb. N), thickness (Tb. Th), and separation (Tb. Sp).

Histological observation and immunohistochemistry

Tissue samples were transferred to 10% EDTA solution for complete decalcification, dehydration, and paraffin embedding. Sagittal planes were cut into a 5-m-thick slice of the medial compartment of the knee joint and stained with hematoxylin and eosin (H&E) and Safranin O-fast green (S&O). The Osteoarthritis Research Society International (OARSI) scoring system was applied as previously described [52].

Immunohistochemistry was used to identify tissue proteins. We sealed the paraffin sections with a serum first and then hatched them with rabbit polyclonal antibody working solution against collagen II or anti-MMP-13(gb12021, gb11247; all 1:100 dilution; Servicebio, PR China) antibodies. They were then incubated with the secondary antibody working solution, which resulted in the formation of the chromogenic reagent color. The number of positively stained cells per specimen was calculated, and three continuous specimens groups were taken from each mice group.

Statistical analysis

At least three biological replicates were used in each experiment. All data were displayed as mean ± standard deviation (SD). Student’s t-test or One-way ANOVA was applied to identify differences between groups. P < 0.05 was considered statistically significant. Data analysis was performed with SPSS Statistics 18.0 software (IBM Corp, NY, USA).

Ethics approval and consent to participate

The research plan approval was obtained from The Ethics Committee on Human Experimentation at The First Affiliated Hospital of Sun Yat-Sen University, China (IRB: [2020]113-1), and complied with The Declaration of Helsinki (2000). All participants signed the informed consent. Animal experiments were performed with the approval of the Institutional Ethics Committee of The First Affiliated Hospital of Sun-Yat University.

Availability of data and materials

The data used to support the findings of this study are available from the corresponding author upon request.