Osteoarthritis (OA) is a joint disease that is pervasive in life. It is largely caused by cartilaginous injury and affects the whole joint tissue (Pereira et al., 2015). Nearly half of people over 65 suffer from OA.(Sakalauskienė and Jauniškienė, 2010; Glyn-Jones et al., 2015). Globally, the incidence and mortality of OA are increasing (Bijlsma et al., 2011). Arthrodynia, swelling, and inability to move freely are the main symptoms of OA and cause many adverse effects on people’s lives. Several risk factors (Prieto-Alhambra et al., 2014), including age, sex, obesity, genetics, and joint damage, have been linked to OA progression (Felson et al., 2000; Vincent, 2019; Abramoff and Caldera, 2020). Articular cartilage degeneration and secondary osteogenesis are the main pathological manifestations of OA (Burr and Gallant, 2012). The long-term development of OA will not only affect people’s behaviors and activities but also cause depression, anxiety, and other negative emotions (Litwic et al., 2013). To provide more perfect, targeted treatment for patients with OA, the progression of OA needs to be studied. The specific pathogenesis of OA may be related to metalloproteinases (Mehana et al., 2019), cytokines (Boehme and Rolauffs, 2018), signaling pathways (Rigoglou and Papavassiliou, 2013), and noncoding RNA (ncRNA) (Sondag and Haqqi, 2016).

ncRNA is a nonprotein-coding RNA that does not translate into proteins but participates in protein translation. At the RNA level, it can perform biological functions (Wu et al., 2019). microRNA (miRNA), long ncRNA (lncRNA), circular RNA (circRNA), ribosomal RNA (rRNA), transfer RNA (tRNA), small nuclear RNA (snRNA), small nucleolar RNA (snoRNA), small interfering RNA (siRNA), short hairpin RNA (shRNA) and Piwi-interactingRNA (piRNA) are the main ncRNAs(Chen et al., 2021). Studies have found that ncRNA is closely related to the occurrence of several diseases for the past few years (Esteller, 2011; Wang et al., 2019b). For example, promoter CpG methylation of two genes encoding members of the miR-200 family can easily lead to the occurrence and development of breast and colorectal cancer (Lim et al., 2013); miR-34b/c is a critical tumor suppressor. The methylation of miR-34b/c CpG island leads to the silence of miR-34b/c, thus increasing the incidence of tumors (Toyota et al., 2008); the decreased expression of miR-133 may induce myocardial hypertrophy by targeting the beta-1 adrenergic receptor pathway (Castaldi et al., 2014). Many studies have also found that miRNA, lncRNA, and circRNA are closely related to the course of OA, and play important regulatory roles in transcription, post-transcription, and post-translation (Li et al., 2019b; Zhang et al., 2021e). The interaction between lncRNA/circRNA, miRNA, and mRNA has attracted increasing attention. For example, lncRNA/circRNA can bind to miRNA, reduce the inhibitory effect of miRNA on mRNA, participate in regulating the progress of chondrocyte proliferation and apoptosis, extracellular matrix (ECM) degradation and inflammatory response in the progress of OA. Furthermore, lncRNA-p21 could induce chondrocyte apoptosis and slow the process of OA by binding to miR-451 and promoting the expression of downstream target gene mRNA (Tang et al., 2018a). This review describes the roles of miRNA, lncRNA, and circRNA in OA and the role of the lncRNA/circRNA–miRNA–mRNA axis in OA.

miRNA is a single-stranded RNA molecule with a length of about 20–24 nucleotides (Correia de Sousa et al., 2019). It belongs to one type of ncRNA and widely exists in eukaryotes to regulate the expression of other genes. miRNA regulates gene expression based on complete or incomplete pairing with mRNA. In most cases, the single-stranded miRNA in the complex is paired with the 3′UTR of the target mRNA in an incomplete complementary manner, blocking the translation of the gene and regulating gene expression. This process, called translation inhibition, is mainly found in animal cells. When the miRNA is completely complementary to the 3′UTR of the target mRNA, the mRNA in the complementary region would be specifically broken, eventually leading to gene silencing, and the process called post-transcriptional gene silencing, which will eventually lead to the degradation of target mRNA, mainly exists in plant cells (Liu et al., 2014a). The same gene can be regulated by multiple miRNAs, and multiple target genes can be regulated by the same miRNA (Iacona and Lutz, 2019). The formation and mechanism of miRNA are shown in Figure 1.

With the deepening of research, miRNAs have been discovered and studied increasingly, and they have become a potential target in disease prevention and treatment. miRNA has many functions roles in human diseases, such as regulating cell autophagy (Li et al., 2019h), epigenesis (Yao et al., 2019), glucose metabolism (Fu et al., 2015). Chen et al. (2018c) developed a computational model for disease association prediction to detect potential miRNA-disease associations accurately and efficiently. By studying three common human cancers (Zhang et al., 2021b), namely, colon cancer, esophagus cancer, and kidney cancer, many miRNAs were confirmed to be connected with the three kinds of cancer. In addition, many studies have proven that miRNA is related to the pathological processes of intervertebral disc degeneration (Shi et al., 2021b), muscle atrophy (Zhang et al., 2021a), and cardiovascular diseases (Liu et al., 2021a).

Currently, growing findings reveal that miRNA expression level changes exist in various tissues of patients with OA, leading to abnormal target gene expression. miRNA has many functions in OA, such as regulating cell autophagy and apoptosis (Yu et al., 2019b), inflammatory reaction (Sui et al., 2019), and cartilage degradation (Guo et al., 2020). Changes in miRNA expression levels in different tissues can be experimented with by gene sequencing. Gene sequencing is a new type of gene detection technology, which can analyze and determine the whole sequence of genes from blood or saliva to predict the possibility of suffering from various diseases and lock in individual diseased genes for early prevention and treatment. Zhou et al. (2020b) revealed 21 differentially expressed miRNAs in synovial tissues from OA patients compared with normal controls by gene sequencing technology. The expression levels of the first two DEmiRNAs(hsa-miR-17-5p and hsa-miR-20b-5p), which cover most of the DEmRNAs, were analyzed and found to be down-regulated in OA, which was also confirmed by qRT-PCR verification. Ntoumou et al. (2017) assessed differential miRNA expression by microarray analysis in the serum of patients with OA. Compared with the control group, 279 miRNAs were differentially expressed in OA. This study focused on analyzing and studying three differentially expressed miRNAs: hsa-miR-140-3p, hsa-miR-671-3p, and hsa-miR-33b-3p. We found that the expression of these three miRNAs was down-regulated in the serum of OA patients. Through serum microRNA array analysis and bioinformatics analysis, they determined that these three miRNAs were potential OA biomarkers involved in the metabolic processes of insulin and cholesterol. OA is a metabolic disease, and insulin resistance plays a vital role in metabolic syndrome. Therefore, the metabolic processes of insulin and cholesterol in the body are closely related to OA. In addition, based on RNA sequencing and miRNA analysis, Wu et al. (2021a) identified that miR-210-5p is highly enriched in the exosomes of OA sclerotic subchondral osteoblasts, triggering the expression of genes associated with catabolism in articular chondrocytes. Therefore, the abnormal up-regulation of miR-210-5p in exosomes could serve as a marker for OA. Notably, miRNA show obvious tissue specificity in different OA tissues. For example, the expression of miR-125b-5p in synovial fluid and chondrocytes is different in OA patients. Ge et al. (2017) found by PCR that miR-125b-5p in synovial fluid was significantly up-regulated in OA patients compared with normal subjects, promoting synovial cell apoptosis by targeting syvn1. Rasheed et al. (2019) treated chondrocytes with IL-1β to construct OA cell models and determined the expression of miR-125b-5p using Taqman analysis. They found that miR-125b-5p in chondrocytes was significantly down-regulated compared to healthy individuals and regulated inflammatory genes in OA chondrocytes by targeting TRAF6. Our appeal study found that expression levels of multiple miRNAs in the synovial membrane, cartilage, and subchondral bone were altered in OA patients compared to healthy individuals. In addition, even in the same tissue, if in different stages of development, the expression of miRNAs may be different. For example, in different stages of the knee joint cartilage of rats, Sun et al. (2011) used Solexa sequencing and RT-qPCR detection for the expression of miRNAs. They tested the miRNAs in the rat knee joint cartilage at the starting point, on Day 21 and Day 42, and found that the expression of miRNAs was different at each stage. Among them, 4 representative miRNAs were selected for further analysis. Compared with the initial stage, the expressions of aggrecan, colia1, and ColXa1 were up-regulated on day 21. The expression of ColXa1 was up-regulated on day 42, whereas those of aggrecan and colia1 were down-regulated. The expression of Sox9 showed minimal change during the three stages. Gabler et al. (2015) found that miRNA could control the differentiation of chondrocytes and regulate the occurrence of OA. During the development of human bone marrow mesenchymal stem cells (HMSCs), the expression of miRNA in different development stages is also different. By microarray analysis, the miR spectra of HMSCs in patients with OA at different development time points were measured. Among the 1,349 detected miRNAs, 553 were expressed in cartilage formation, they further performed miRNAs detection at 7, 14, 21, and 42 days after cartilage formation and found that their expression of miRNAs was also different. In summary, the expression of miRNAs in OA patients is different in different tissues and between different stages of development of the same tissue.

It is well known that many intracellular signaling pathways, such as nuclear factor-kappaB(NF-κB) and transforming growth factor β (TGF-β) played an vital roles in the pathogenesis of OA (Nishimura et al., 2020). In recent years, more studies discovered that miRNA can delay the pathological process of OA by promoting or inhibiting these pathways (Xu et al., 2016). NF-κB is an essential nuclear transcription factor in cells participating in the inflammatory and immune response of the body and apoptosis regulation (Lawrence, 2009). For example, as the 3′UTR of NF-κB contains the binding site of miR-143 and miR-124, when the DNA methylation degree of miR-143 and miR-124 promoters is reduced, the expression of miR-143 and miR-124 is up-regulated, and the transcription process is activated, thereby inhibiting the NF-κB signaling pathway, inhibiting apoptosis and delaying the progression of OA (Qiu et al., 2020). Similarly, When the expression levels of miR-34a and miR-181a were decreased, the expression of the BCL2 gene was increased, thereby limiting the term of NF-κ B translocation into the nucleus in OA Chondrocytes cultures and eventually reducing apoptosis and oxidative stressl (Cheleschi et al., 2019). The TGF-β signaling pathway is involved in many cellular processes in mature organisms and developing embryos, including cell growth, differentiation, apoptosis, dynamic cell balance, and other cellular functions. By promoting or inhibiting the TGF-β signaling pathway, we can regulate the cellular processes, thereby inducing or delaying the progression of OA (Shen et al., 2014). Hu et al. (2019b) established OA mouse models. QPCR and Western blot were used to compare the expression of miR-455-3p and PAK2 in the cartilage of healthy individuals and patients with OA, and the luciferase reporter gene was used to analyze the interaction between them. The results showed that miR-455-3p could inhibit the expression of pak 2, promote the TGF-β signaling pathway, and ultimately inhibit OA by directly targeting PAK2 3′UTR. In summary, various miRNAs are involved in regulating OA progression by handling a variety of intracellular signaling pathways. In addition, increasing evidence also emphasizes that changes in the expression of many miRNAs can also directly regulate the development of OA. The specific information of these miRNAs is listed in Table 1.

lncRNAs are ncRNAs with a length of more than 200 nucleotides that have little or no protein-coding potential, and account for more than 80% of total lncRNAs(Ponting et al., 2009). At first, lncRNA was considered the “noise” of genome transcription, with no biological function, and its mechanism of action was only in situ regulation, through recruitment and formation of chromatin modification complexes [such as IGF2RRNA antisense (AIR), XIST] to silence the transcription of neighboring genes. As more detection techniques were applied to RNA studies, such as microarray, RNA sequencing (RNA-seq), Northern blot, and real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR) (Zhu et al., 2013), more biological functions of lncRNAs gradually being discovered. Recent studies have discovered several mechanisms of action of lncRNA, which can interact with proteins, DNA, and RNA to regulate many biological processes (Zhu et al., 2013). For example, lncRNA MALAT1 acts on miR-150-5P and AKT3 to regulate cell proliferation and apoptosis (Zhang et al., 2019g), thus participating in the growth and development of the body and the pathological process of diseases (Kopp and Mendell, 2018) (Figure 2).

lncRNA is closely related to cell growth, differentiation, and senescence. In addition, lncRNA has a special relationship with some human diseases, such as cardiovascular diseases (Huang, 2018), nervous system diseases (Zhang et al., 2019f), and immune-mediated diseases (Zhou et al., 2018b). In the recently updated database of lncRNA-related diseases, more than 200,000 lncRNAs have been recorded in their association with diseases (Bao et al., 2019).

lncRNA can regulate chondrocyte proliferation and apoptosis, inflammatory response, and extracellular matrix degradation, and promote the repair and stability of articular cartilage. Recent studies have shown an essential relationship between some changes or disorders of lncRNAs and the occurrence and development of OA. There are many studies to detect the expression of lncRNA in OA patients. Yang et al. (2021a) examined the lncRNA profiles of patients with OA and healthy individuals by RNA sequencing. They found that 25 lncRNAs are differentially expressed in patients with OA compared with the control group. Through microarray analysis, Xing et al. (2014) detected the expression of lncRNA in KOA cartilage and normal cartilage and further verified it by real-time polymerase chain reaction (RT-PCR). They found that the expression of 121 lncRNAs in KOA is different from normal cartilage: 73 up-regulated lncRNAs and 48 down-regulated lncRNAs. Among the up-regulated lncRNAs, HOTAIR is the most up-regulated. Pearson et al. (2016) separated OA chondrocytes through collagenase digestion and analyzed lncRNA expression through RNA sequencing (RNAseq) and qPCR. Finally, 983 lncRNAs were identified in OA chondrocytes. A total of 125 differentially expressed lncRNAs were identified after interleukin-1B (IL-1B) stimulation. Through microarray and qPCR analysis, Liu et al. (2014b) compared the expression of lncRNA in OA cartilage and normal cartilage, and found 152 differentially expressed lncRNAs in OA cartilage. Compared with normal cartilage, 82 increased lncRNAs and 70 decreased lncRNAs were in OA cartilage. Using mRNA and lncRNA microarray analysis, Zhang et al. (2020a) found that 990 lncRNAs were different in OA chondrocytes compared with the control group: 666 up-regulated, 324 down-regulated. In addition, 546 mRNAs had a different expression: 419 up-regulated, 127 down-regulated. Six lncRNAs (ENST00000606283.1, ENST00000436872.1, ENST00000488584.1, ENST00000603682.1, XR-245446.2, and ENST00000605586.1) were tested by qPCR. The results were consistent with the test results. In summary, through the detection of lncRNA expression levels in the chondrocytes of OA patients and healthy individuals, we can finally find that there are differences in the expression of a variety of lncRNAs. In addition to the lncRNAs of appeal, several lncRNAs are closely related to the progress of OA, as shown in Table 2.

The circRNA molecule is in a closed-loop structure and is not affected by RNA exonuclease. They are mainly in the cytoplasm or stored in exosomes. They are stable and not easily degradable, and widely exist in many eukaryotes. circRNAs are formed by reverse splicing through nonclassical splicing. One model believes that in the transcription of pre-RNA, due to the partial folding of RNA, the originally nonadjacent exons are pulled closer, and exon jumping occurs, resulting in the formation of circular RNA intermediates in the region to be crossed. Moreover, ring RNA molecules composed of exons are formed by lasso splicing. Another model suggests that the reverse complementary sequence located in the intron region leads to intron region pairing mediated reverse splicing, resulting in the formation of circular RNA molecules (Chen and Yang, 2015). To date, the biological functions of circRNAs that have been discovered mainly include interactions with miRNAs(Cao et al., 2019a), binding of regulatory proteins (Zang et al., 2020), transcription of regulatory genes (Zhang, 2020), and coding functions (Lei et al., 2020) (Figure 3). For example, circRNA.33186 increased MMP-13 expression by interacting with miR-127-5p to regulate cell proliferation and apoptosis (Zhou et al., 2019b).

Bipartite Network Projection allocates resources according to the known associations between different miRNAs and diseases, entirely using the similarity information of miRNA and diseases to predict various conditions accurately (Chen et al., 2018b). KATZ Measure is a graph-based calculation method, which converts the calculation of the similarity between lncRNA and diseases into the problem of similarity calculation between nodes in heterogeneous networks to predict the correlation between lncRNA and conditions. The integration of the two can recognize the association of circRNA with the disease (Chen, 2015). Through Bipartite Network Projection and KATZ Measure (Zhao et al., 2019a), many circRNAs related to diseases have been discovered, and circRNAs are involved in the diagnosis and treatment of atherosclerosis (Zhang et al., 2018a), cancer (Li et al., 2020f), cardiac hyperplasia (Li et al., 2020e), and other diseases. There are many experimental studies related to circRNA and diseases, and the main research types are cell experiments or animal experiments. Through these experiments, we have found multiple action mechanisms of circRNA on various conditions. For example, circRNA_100367 acts as a signaling molecule that regulates esophageal squamous cell carcinoma through the Wnt3 signaling pathway (Liu et al., 2019b); circRNA_0016624 regulates gene-based expression of interest in osteoporosis patients via sponge miR-98 (Yang et al., 2020d); circRNA_100395 mitigates the progression of breast cancer by directly targeting MAPK6 (Yu et al., 2020).

In addition, several circRNAs participate in the development of OA and the OA of the abnormal expression in various tissues. For example, Xiao et al. (2019a) used illumina sequencing platform to detect circRNA expression in patients with mild and severe KOA. In this paper, 197 differentially expressed circRNAs were identified. Among them, the up-regulation amplitude of Hg38_circ_0007474 is the largest, and the down-regulation amplitude of hg38_circ_0000118 is the largest. Further analysis of the three circRNAs selected from hsa_circ_0045714, hsa_circ_0005567, and hsa_circ_0002485 found that all three circRNAs can inhibit the function of the corresponding miRNA by serving as a sponge for miRNAs and indirectly promote its downstream process, thereby participating in the development of OA. Wang et al. (2019h) used microarray analysis to screen for circRNA expression in healthy and KOA articular cartilage. They found 1,380 circRNAs differentially expressed in the articular cartilage of knee joints of healthy individuals and patients with OA. Meanwhile, constructing a circRNA-miRNA network verified the ten most likely target genes related to circRNA. It was finally discovered that hsa_circ RNA_003231 might be involved in the occurrence and progression of OA. Zhou et al. (2018e) established OA models in interleukin-1β (IL1β)-treated mouse articular chondrocytes (MACs) to study the expression and function of circRNAs in OA using new sequencing methods and bioinformatic analysis. Compared with the control group, 255 circRNAs were differentially expressed in MACs treated with IL-1 β: 119 up-regulated, 136 down-regulated. Mmu-circRNA-30365 and Mmu-circRNA36866 were two substantially different circRNAs, and their specific expression changes in patients with OA and normal individuals were verified by QRT-PCR. Liu et al. (2016) analyzed circRNA expression between OA and normal cartilage samples by hierarchical clustering analysis and found that compared with normal cartilage, 71 circRNAs were differentially expressed (16 were increased, and 55 were decreased) in OA cartilage. In this study, we focused on the research of circRNA-CER. We found that this circRNA could compete with MMP13 for miR-136 and participate in the degradation of the extracellular matrix of chondrocytes. The above examples fully prove that the expression levels of circRNA in OA patients and healthy individuals are different, and these differentially expressed circRNA has a special relationship with the progression of OA.

Several studies have reported the functions and mechanisms of several circRNAs in OA, but relevant studies are few. Zhou et al. (2018d) established rat OA models, predicted the function of circRNA_ATP9b in rat knee chondrocytes through bioinformatic analysis, and finally found that circRNA_ATP9b regulated the degradation of extracellular matrix through sponge miR-138-5p, thereby controlling the progression of OA. Moreover, circRNA_ATP9b expression was increased, and miR-138-5p expression was down-regulated in IL-1β-induced chondrocytes. circRNA_ATP9b regulated the expression of related genes by targeting miR-138-5p. Li et al. (2017a) analyzed the dual-luciferase reporter genes and found that the transcriptional activity of miR-193b can be inhibited by overexpression of hsa_circ_0045714. Overexpression of hsa_circ_0045714 can also up-regulate the expression of insulin-like growth factor 1 receptor (IGF1R) because IGF1R is a crucial target gene of miR-193b. It is associated with cell proliferation and apoptosis. Further studies on the progression of circRNA in OA are presented in Table 3.

Studies have shown that lncRNA–miRNA–mRNA axis plays a vital control effect in the progression of several diseases, such as cardiovascular disease and cancer (He et al., 2018; Wang et al., 2019c). The mechanisms of interaction of lncRNAs, miRNAs, and mRNAs in various diseases are as follows: 1) The structure of most lncRNAs is similar to mRNAs, and miRNAs binding to mRNAs can reduce the expression of lncRNAs. lncRNA and miRNA compete to bind the 3′-UTR of target gene mRNA, thereby indirectly inhibiting the interaction between miRNA and mRNA. For example, in Alzheimer’s disease, the post-transcriptional regulation of BACE1 involves miR-485-5p, and the specific antisense transcription of BACE1 forms lncRNA-BacE1-As, which compete with lncRNA-Bace1-As to bind to the binding sites of related mRNAs (Faghihi et al., 2010). 2) lncRNAs sponge miRNAs as competitive endogenous RNAs (ceRNAs). lncRNA molecules contain miRNA binding sites, which can bind to miRNA, inhibit the interaction between miRNA and mRNA, improve the expression level of related mRNA, and regulate the expression of target genes. For example, Zhang et al. (2020l) constructed a complete mRNA-LncRNA-miRNA ceRNA regulatory network; lncRNAs ENST00000326237.3, ENST00000399702.5, and ENST00000463727.1 were found to regulate related genes through competitive binding of the same miRNA has-miR-1260a. Kong et al. (2019) demonstrated that lncRNA—CDC6 can further regulate CDC6 expression through direct uptake of miR-215 as a ceRNA. Luan et al.(Luan and Wang, 2018) found that in cervical cancer, XLOC_006390 may act as ceRNA and bind with miR-331-3p and miR-338-3p, thus regulating the expression of genes related to cervical cancer. 3) miRNAs mediate the degradation of lncRNAs. For example, miRNA-150 is the target gene for lncRNA CASC11 in human plasma, and increased concentrations of miRNA-150 decrease the activity of lncRNA CASC11(Luo et al., 2019b). 4) lncRNAs act as miRNAs precursors. For example, Tao et al. (2017) found that miR-869a and miR-160c could be clipped from lncRNAs npc83 and npc521. However, in OA, lncRNA mainly binds to miRNA as a competitive endogenous RNA (ceRNA), inhibiting its target genes’ expression and regulating OA’s progression by regulating cell proliferation, apoptosis, autophagy and extracellular matrix (ECM) degradation (Figure 4).

There are many examples where lncRNA functions as a binding of ceRNA to miRNA in OA. For example, Zhang et al. (2020m) took IL -1β-induced OA chondrocytes as the research object to study the molecular mechanism of LINC00511 in regulating OA. The study found that the expression of LINC00511 was up-regulated, and the lncRNA could be used as a sponge of miR-150-5p and combined with 3′-UTR of transcription factor inhibit the proliferation of chondrocytes, promote apoptosis and degradation of ECM, and finally regulate OA. Liu et al. (2018) established an OA chondrocyte model induced by IL -1β and an OA mouse model caused by collagenase. The experiments were performed in vivo and in vitro at two levels, and the cell state was examined by the CCK-8 method and flow cytometry. Studies have found that KLF3-AS1, as a ceRNA interacting with miR-206, promotes the expression of GIT1 and then promotes the proliferation of chondrocytes and inhibits apoptosis, ultimately alleviating the progression of OA. Likewise, Tian et al. (2020) studied the relationship between SNHG7, miR-34a-5p, and SYVN1 in human chondrocytes. It has been found that in OA tissues, SNHG7 is down-regulated, and SNHG7 can regulate SYVN1 by sponging miR-34a-5p, thereby promoting cell proliferation and inhibiting apoptosis and autophagy. In addition, studies have found that lncRNA XIST is up-regulated in OA articular cartilage. Like a sponge, XIST regulates the target proteins miR-211, miR-17-5p, miR-149-5p, and miR-27b-3p, thereby promoting the proliferation and apoptosis of chondrocytes and finally inducing OA (Li et al., 2018b; Zhu et al., 2021). These results suggest that lncRNAs can act as miRNA sponges in the interaction of lncRNAs, miRNAs, and mRNA in OA.

At present, the incidence of OA is very high, and its pathogenesis is still unclear. Studying the specific pathological process and molecular pathway of OA is of great clinical significance (Duan et al., 2020). First, ncRNA can be used to diagnose OA. The expression of many ncRNAs between patients with OA and normal individuals have remarkable differences, which can be seen in humans and animals. For example, Huang et al. (2019c) showed that miRNA-204 and miRNA-211 are decreased in OA, resulting in Runx2 accumulation in multiple types of joint cells and elevated OA markers, and leading to total joint degeneration. Second, several ncRNAs are associated with the prognosis of OA. Rousseau et al. (2020) took the miRNAs in the serum of female patients with KOA as the research objects. He first made a preliminary screening of the research objects through next-generation sequencing and then further analyzed the research objects through RT-QPCR. He found that miR-146A-5p is up-regulated in patients with mild OA, and the prognosis of OA caused by the up-regulation of miRNA is relatively good. In addition, the increase of miR-186-5p in an individual means that the individual might have the imaging changes of OA in the past 4 years, which could be prevented in advance to avoid the occurrence of OA as much as possible. Finally, several ncRNAs can be used for the treatment of OA. Several new drugs can be developed to promote or inhibit several ncRNAs, or change the pathway of action of ncRNA to treat OA. For example, miR-93 is down-regulated in mice with OA and lipopolysaccharide-treated chondrocytes, and acts directly on TLR4 to exert biological effects. miR-93 regulates OA by inhibiting the TLR4/NF-κB pathway, lipopolysaccharide-induced inflammation, and apoptosis. In patients with OA and down-regulation of miR-93, corresponding drugs can be developed to promote its up-regulation and inhibit the aggravation of OA (Ding et al., 2019). These studies indicate that ncRNA has great potential for clinical use in OA. At present, most of the tissue comes from cartilage and is found in the knee joint, and the chondrocytes are cultured to construct the OA cell model. Further research is needed, and more clinical trials must be explored to find biomarkers associated with OA while developing the immense potential of ncRNA.

In recent years, ncRNAs have become one of the most widely studied fields in the development of OA. However, the studies on the regulation of miRNA, lncRNA, and circRNA in diseases and their use as indicators for diagnosis or treatment of OA are still in the early stages, and the mechanism of action ofOA, which may involve multiple signaling pathways, is still unclear. This study reviews theinteractions between lncRNA/circRNA and miRNA in OA. Through high-throughput sequencingtechnologies such as microarray analysis and RNA sequencing, the findings reveal that a large number of miRNA, lncRNA, and circRNA are dysregulated in patients with OA, and the clinical trials related to ncRNA and OA are summarized. The present research progress of ncRNA in the prevention, diagnosis, and treatment of OA is illustrated, which provides a basis for the treatment of OA by ncRNA in the future.