Exhausting circ_0136474 and Restoring miR-766-3p Attenuate Chondrocyte Oxidative Injury in IL-1β-Induced Osteoarthritis Progression Through Regulating DNMT3A

Circular RNA circ_0136474 is a new contributor of human osteoarthritis (OA) by suppressing chondrocyte proliferation. However, its role and mechanism in OA chondrocyte injury remain ill defined. Herein, we performed real-time quantitative PCR to detect RNA expression of circ_0136474, microRNA (miR)-766-3p, and DNA methyltransferase 3A (DNMT3A) and utilized Western blotting to measure protein expression of DNMT3A, matrix metalloproteinase-1 (MMP1), MMP13, collagen II, proliferating cell nuclear antigen (PCNA) and B cell lymphoma (Bcl)-2, and Bcl-2-associated X protein (Bax). Direct interaction between miR-766-3p and circ_0136474 or DNMT3A was confirmed by bioinformatics algorithms, dual-luciferase reporter assay, and RNA immunoprecipitation. Functional experiments including cell counting kit-8 assay, flow cytometry, and special assay kits were employed to measure oxidative injury in interleukin (IL)-1β-induced OA-like chondrocytes. First, IL-1β administration induced cell viability inhibition, collagen II suppression, and promotion of MMP1 and MMP13 in human chondrocyte CHON-001 cells. Expression of circ_0136474 and DNMT3A was upregulated, and miR-766-3p was downregulated in human OA cartilages and IL-1β-induced CHON-001 cells. Functionally, both blocking circ_0136474 and upregulating miR-766-3p could rescue cell viability and levels of PCNA, Bcl-2, reduced glutathione (GSH), and total superoxide dismutase (SOD), and attenuate apoptosis rate and levels of Bax, reactive oxygen species (ROS), and lipid peroxidation malondialdehyde (MDA). Mechanically, circ_0136474 served as miR-766-3p sponge to govern miR-766-3p-targeted DNMT3A expression. Accidently, restoring DNMT3A counteracted the miR-766-3p upregulation role, and silencing miR-766-3p weakened circ_0136474 knockdown effect in IL-1β-induced CHON-001 cells. In conclusion, exhausting circ_0136474 could mitigate OA chondrocyte oxidative injury through regulating miR-766-3p/DNMT3A axis.


INTRODUCTION
Osteoarthritis (OA) is a multifactorial chronic disease in the joints and is characterized by the degradation of articular cartilage, debilitating pain, and loss of mobility (Nguyen et al., 2017). The most prevalent risk factor of OA incidence is age (Loeser et al., 2016). Epigenetics such as non-coding RNAs and DNA methylation have emerged as a new and important area of research on OA etiology and therapy by interplaying with genetics (Grandi and Bhutani, 2020;Rice et al., 2020). Circular RNAs (circRNAs) are a new class of non-coding RNAs characterized by covalently closed loop structures produced by back-splicing event (Kristensen et al., 2019). Different expression profiles and functional analysis of circRNAs have been carried out in human OA cartilages (Li H. et al., 2019;Wang et al., 2019). Upregulated and downregulated circRNAs participate in the occurrence and progression of OA by regulating multiple pathological processes, such as extracellular matrix (ECM) degradation, inflammation, and apoptosis (Yu and Sun, 2018).
Moreover, oxidative stress and either inflammation or aging stress response are highly interdependent and interconnected (Haigis and Yankner, 2010;Biswas, 2016), thus serving as emerging players in OA pathogenesis and treatment (Minguzzi et al., 2018;Ansari et al., 2020). CircRNAs have been on the way forward for being promising biomarkers with therapeutic potential in OA . Furthermore, the competing endogenous RNA (ceRNA) network such as circRNA-microRNA (miRNA)-messenger RNA (mRNA) axis has been a possible mechanism and therapeutic target for OA (Kulcheski et al., 2016;Xiao et al., 2019). Circ_0136474 is a new contributor of human OA by suppressing chondrocyte proliferation ; however, its role and the ceRNA pathway in OA chondrocytes injury remain largely uncovered.

Cartilage Tissues
Osteoarthritis cartilage tissues were obtained from 33 OA patients who underwent total knee replacement surgery at The First People's Hospital of Lianyungang; another 33 normal cartilage tissues at knee joints were from trauma patients without previous history of OA or rheumatoid arthritis. Clinical characteristics of these subjects are listed in Table 1, and clinical samples were collected during January 2017 to December 2019. Informed consent was obtained from all tissue donors. The study was approved by the Ethics Committee of the First People's Hospital of Lianyungang and performed according to the Declaration of Helsinki.

Protein Extraction and Western Blotting
Total protein from tissues and CHON-001 cells was extracted from tissues and cells by radioimmunoprecipitation assay (RIPA; Beyotime, Shanghai, China) containing phenylmethylsulfonyl fluoride (PMSF; Beyotime). Next, protein samples were subjected to standard Western blotting as previously described . Briefly, protein separation was performed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and protein transferring was conducted onto polyvinylidene difluoride membrane (Millipore, Bedford, MA, United States). Membranes were blocked by 5% non-fat milk, incubated with special primary antibodies and horseradish peroxidaseconjugated secondary antibodies, and last developed by enhanced chemiluminescence kit (Millipore). The antibodies are summarized in Table 3. The relative protein expression was normalized to β-actin and presented as fold change of that in the control group.

Flow Cytometry
Apoptosis of CHON-001 cells after IL-1β treatment was detected by the fluorescein isothiocyanate (FITC) Annexin V Apoptosis Detection Kit (BD Biosciences, San Jose, CA, United States). Cells were reaped with the scraper and washed with cold phosphate-buffer saline; then 1 × 10 5 cells were re-suspended in 1 × Annexin V Binding Buffer and added with 5 µl of FITC annexin V and 5 µl of propidium iodide (PI) staining solution for 15 min at room temperature free from light. Eventually, these staining cells were diluted in 400 µl of 1 × annexin V binding buffer and analyzed on a flow cytometer (BD Biosciences). The percentage of early apoptotic cells and late apoptotic cells was considered as apoptosis rate (%).
In addition, GSH, MDA, and SOD levels in cell extracts were severally tested by GSH and Oxidized Glutathione Disulfide (GSSG) Assay Kit (Beyotime), Lipid Peroxidation MDA Assay Kit (Beyotime), and Total SOD Assay Kit with WST-8 (Beyotime) according to the manufacturer's instructions.

Dual-Luciferase Reporter Assay and RNA Immunoprecipitation
The wild type circ_0136474 (WT-circ_0136474) and its mutant derivative lacking the miR-766-3p binding site (MUT-circ_0136474) were synthesized and inserted into pGL3 luciferase report vectors (containing Firefly; Promega, Madison, WI, United States) downstream of the luciferase-coding region.
Similarly, pGL3-WT-DNMT3A 3 UTR vector and pGL3-MUT-DNMT3A 3 UTR vector were constructed. Then CHON-001 cells were co-transfected with the abovementioned recombinant vectors and miR-766-3p mimic or miR-NC mimic, as well as pRL-TK luciferase report vector (containing Renilla; Promega). A dual-luciferase reporter assay system (Promega) was applied to analyze the luciferase activities at 48-h post-transfection with normalization to Renilla. RIP assay was performed using EZ-Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore); during this experiment, anti-argonaute 2 (anti-AGO2) or anti-IgG (negative control) were used to label magnetic beads, which were then incubated with cell extract of CHON-001 cells overnight at 4 • C. The RNA-protein complex bound to the beads was lysed in Norgen Total RNA Isolation Plus Micro Kits (Norgen Biotek) and qPCR.

Statistical Analysis
All experiments were repeated at least three times, and measured data were expressed as mean ± standard error of the mean in one representative experiment. Data were analyzed, and comparisons were measured on GraphPad Prism 7 (GraphPad, La Jolla, CA, United States). P < 0.05 was considered to be of statistical significance.

Circ_0136474 Was an Upregulated circRNA in Osteoarthritis Patients
Compared with normal cartilages, circ_0136474 was highly expressed in human OA cartilages with a back-splicing event of a junction site from exons 12-21 of ASH2L (Figures 1A,B). Moreover, its expression was little affected by RNase R treatment, whereas the host gene ASH2L mRNA expression was dramatically reduced by RNase R (Figure 1C). Circ_0136474 level was dominantly discovered in the cytoplasm of CHON-001 cells, which was paralleled with GAPDH and contrary to U6 ( Figure 1D). These data showed that circ_0136474 was an abnormally upregulated circRNA in OA.

IL-1β Induced Cell Viability Inhibition and Extracellular Matrix Degradation in CHON-001 Cells
Interleukin-1β administration induced cell viability inhibition in CHON-001 cells in a certain of concentration-and timedependent manner (Figures 2A,B); 10 ng/ml of IL-1β caused about 50% cell viability inhibition at 24 h (Figure 2B), and accompanied with high expression of matrix metalloproteinase-1 (MMP1) and MMP13 and low expression of collagen II ( Figure 2C). These data demonstrated that IL-1β induced cell viability inhibition and ECM degradation in CHON-001 cells.
OA was a disease of the cartilage pericellular matrix (Guilak et al., 2018), and thus, these data suggested a success of OA cell model in chondrocytes.

Exhausting circ_0136474 Mitigated Interleukin-1β-Induced Apoptosis and Oxidative Stress in Chondrocytes
Expression of circ_0136474 was increased in 10 ng/ml of IL-1β-induced CHON-001 cells (Figure 3A), and si-circ_0136474 pre-transfection could decrease this high level of circ_0136474 ( Figure 3B). The inhibition of cell viability in IL-1β-disposed CHON-001 cells was saved with circ_0136474 exhaustion via transfection, as evidenced by the higher OD values during 72 h ( Figure 3C). Apoptosis rate of CHON-001 cells was highly induced by IL-1β stimulation, and it could be partially attenuated by pre-transfecting si-circ_0136474 ( Figure 3D). Molecularly, accompanied with cell viability promotion and apoptosis inhibition, elevated PCNA and Bcl-2 levels and attenuated Bax level were discovered in circ_0136474silenced CHON-001 cells under IL-1β stress ( Figure 3E). Besides, oxidative stress-related factors ROS and MDA were upregulated, while GSH and SOD were downregulated in CHON-001 cells in response to IL-1β treatment (Figures 3F-I); surprisingly, silencing circ_0136464 could overall counteract these oxidative stress levels (Figures 3F-I). These results revealed that exhausting circ_0136474 was responsible to attenuate apoptosis and oxidative stress in IL-1β-induced OA model in human chondrocytes.

Circ_0136474 Could Target miR-766-3p in Human Chondrocytes
According to the prediction of circinteractome, we hypothesized that there might be a novel target binding interaction between circ_0136474 and miR-766-3p. To further validate this hypothesis, pGL3-WT-circ_0136474 vector and pGL3-MUT-circ_0136474 vector were constructed ( Figure 4A). As a result, miR-766-3p overexpression via mimic transfection resulted in a loss of luciferase activity of pGL3-WT-circ_0136474 vector in CHON-001 cells, whereas there was no deficit of that with the pGL3-MUT-circ_0136474 vector (Figures 4B,C). Furthermore, circ_0136474 and miR-766-3p were co-enriched in AGO2 RIP with normalization to IgG RIP ( Figure 4D). In OA, miR-766-3p expression was lower in human OA cartilages and CHON-001 chondrocytes under IL-1β (10 ng/ml) stress (Figures 4E,F).
Collectively, miR-766-3p could be targeted by circ_0136474 and downregulated in human OA.

DISCUSSION
Oxidative stress and ROS overproduction played an important role in the pathogenesis of OA, and IL-1β could induce oxidative damage in both articular chondrocytes and synovial fibroblasts Bao et al., 2020). Furthermore, oxidative stress, epigenetic changes, including miRNA regulation, were suggested as potential therapeutic targets in OA (Portal-Nunez et al., 2016). However, the role of circRNAs and long non-coding RNAs that could function as an miRNA sponge was still unknown in oxidative stress injury in osteoarthritic chondrocytes. Thus, we focused on investigating circ_0136474 dysregulation, role, and ceRNA mechanism in IL-1β-induced apoptosis and oxidative stress.
In this study, we observed an upregulation of circ_0136474 in human OA cartilage tissues and chondrocytes under IL-1β stimulation, and this finding was consistent with that in primary OA chondrocytes isolated from patients . IL-1β could induce an OA cell model in vitro, as evidenced by cell viability inhibition and ECM degradation in IL-1β-exposed CHON-001 cells. Functionally, blocking circ_0136474 could neutralize oxidative stress and apoptosis in IL-1β-induced chondrocytes by scavenging ROS and MDA, promoting cell proliferation and activating GSH and SOD. Among these (10 ng/ml)-induced CHON-001 cells pretransfected with DNMT3A vector or empty pcDNA vector (vector), miR-766-3p/NC or miR-766-3p combined with DNMT3A or vector, si-circ_0136474/NC or si-circ_0136474 combined with anti-miR-677-3p or anti-miR-NC. *P < 0.05 was determined by unpaired t-test, ordinary one-way ANOVA or two-way ANOVA.
chondroprotective effects of circ_0136474 knockdown, cell proliferation promotion and apoptosis suppression had been declared by ; however, the anti-oxidative stress role of exhausting circ_0136474 might be a new finding in OA pathogenesis and treatment. Besides, expression of circ_0136474 was predominantly detected in the cytoplasm instead of the nucleus in human chondrocytes, suggesting a latent potential of circ_0136474 as an endogenous miRNA sponge. Furthermore, we tested this hypothesis and validated that miR-766-3p was a novel target for circ_0136474, except for the previously reported miR-127-5p .
According to this research, miR-766-3p was abnormally downregulated in human OA cartilages and IL-1β-induced OA model in chondrocytes. On one hand, lower miR-766-3p could abate the chondroprotective effect of circ_0136474 deficiency in IL-1β-induced chondrocytes; on the other hand, high miR-766-3p attenuated IL-1β-induced cell proliferation inhibition, apoptosis, and oxidative stress, implying that miR-766-3p overexpression could mimic the chondroprotective effect of circ_0136474 deficiency. Furthermore, inhibiting circ_0136474 and re-expressing miR-766-3p might also be therapeutic approaches to antagonize OA progression. Additionally, the suppressive role of miR-766-3p in IL-1β-induced apoptosis and ECM degradation had been previously demonstrated (Li et al., 2020). Besides, the antioxidant property of miR-766-3p, it was also reported that a higher ROS was discovered to be paralleled with downregulation of a panel of five miRNAs including miR-766-3p in patients with low intrafollicular fluid melatonin (Khan et al., 2020). Even though miR-766-3p was considered as an anti-inflammatory miRNA in IL-1β-blamed and tumor necrosis factor-α (TNF-α)-blamed synoviocyte fibroblasts FIGURE 7 | Restoring miR-766-3p attenuated IL-1β-induced oxidative injury in chondrocytes, and this effect was abrogated by DNMT3A upregulation. CHON-001 cells were transfected with miR-766-3p/NC or miR-766-3p combined with DNMT3A or vector prior to treatment of IL-1β (10 ng/ml) for 24 h. (A) CCK-8 monitored the OD values of inoculated cells during 72 h. (B) FCM tested the apoptosis rate (%). (C) Western blotting examined the relative protein expression of PCNA, Bcl-2, and Bax. (D-G) Special assay kits determined the levels of ROS, GSH, MDA, and SOD. *P < 0.05 was determined by ordinary one-way ANOVA or two-way ANOVA. (Hayakawa et al., 2019), inflammatory response had not been further confirmed in this study.
We identified DNMT3A as a novel functional target gene for miR-766-3p. In addition, DNMT3B was confirmed to be directly regulated by miR-766-3p several years ago (Afgar et al., 2016), and DNA methylation of some tumor suppressors such as DKK2 could be decreased by transfecting miR-766-3p-expressing viruses. Notably, miR-766-3p was obviously downregulated in human tumors (You et al., 2018;Zhang et al., 2020), and its expression could be regulated by DNA methylation (Chen et al., 2017). DNMT3A expression was elevated in OA cartilages and OA-like chondrocytes, which supported the previous data (Wu et al., 2017;Ma et al., 2019). This high DNMT3A could functionally alleviate miR-766-3p overexpression-mediated chondroprotective effect under IL-1β stimulation by promoting apoptosis and oxidative stress levels. Notably, Vrtacnik et al. (2018) presently indicated that gene expression of epigenetic enzymes including DNMT3A was affected by oxidative stress and hypoxia in vitro and in patients with postmenopausal osteoporosis and OA. Moreover, the correlation between DNMT3A and oxidative stress had also been wildly spread in diverse cells (Maugeri et al., 2018;El Henafy et al., 2020;Jung et al., 2020). Apart from modulating proliferation, apoptosis, and oxidative stress, DNMT3A was previously demonstrated to be implicated in ECM destruction in OA (Wu et al., 2017). In addition, Wu et al. (Chen et al., 2017) claimed that the promoter of miR-766-3p was highly methylated in renal cell carcinoma tissues, suggesting that DNA methylation could, in turn, control miR-766-3p expression as well. Plus, PPARγ played a key role in OA development, and its promoter hypermethylation was attributed to DNMT3A and DNMT1 . These finding suggested a possible regulatory back loop of miR-766-3p/DNMT3A in OA chondrocytes; however, this hypothesis was left to be further deciphered, as well as the discovery of DNMT3A downstream targets.

DATA AVAILABILITY STATEMENT
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The First People's Hospital of Lianyungang. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
HZ designed the study and wrote the manuscript. SZ and XS analyzed the data. XM and SJ performed the experiments. LY and YD summarized the data. All authors contributed to this study, and read and approved the manuscript.