RA is a severe chronic autoimmune disease that leads to joint inflammation and damage, as well as systemic impairment, which significantly impact patients’ health and quality of life. The early diagnosis and treatment of RA play a crucial role in delaying its progression, helping to prevent irreversible joint damage and disability (49).

The current diagnosis of RA is primarily based on the classification criteria of the American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) and certain serological markers, including rheumatoid factor (RF) and Anti-cyclic citrullinated peptide antibody (ACPA) (50). However, the limited sensitivity and specificity of these diagnostic markers often result in misdiagnosis and poor prognosis for RA patients (51). CircRNAs are ncRNAs that have been investigated extensively in RA following miRNAs and LncRNAs. Compared to traditional linear RNAs, circRNAs form a covalently closed loop structure which lacks 3’ PolyA tails and 5’ caps, making them effectively resistant to degradation by the ribonuclease R (RNase R) (52). Furthermore, circRNAs exhibit characteristics such as tissue specificity, conservation, and ubiquitous expression (53–56). Due to these characteristics, circRNAs hold greater potential as a diagnostic biomarker for diseases (48, 57–59). Several studies ( Table 6 ) had explored the circRNA expression profiles in peripheral blood mononuclear cells of RA patients by microarray assay and receiver operating characteristic (ROC) analysis, and identified circRNAs such as hsa_circ_0005008, hsa_circ_0140271, and hsa_circ_101328 exhibit specificity and sensitivity in RA, which indicates that circRNAs hold significant promise as a diagnostic biomarker in RA with broad application prospects (9, 67, 68).

Furthermore, circRNAs offer new perspectives and insights in research related to the pathogenesis and treatment of RA (69). Current studies are primarily focused on the competitive endogenous RNA (ceRNA) mechanism of circRNAs. CircRNAs can function by sponging miRNAs to reduce the number of miRNAs available to target mRNA, thus contributing to mRNA stability or protein expression, which in turn influences the onset and development of RA (70). CircRNAs act as ceRNAs in various cell types associated with RA, including macrophages, helper T cells, chondrocytes, fibroblast-like synoviocyte (FLS), and others, and fibroblast-like synoviocytes are the primary focus of research in this context ( Table 7 ). For example, CircRNA_09505 can promote AKT1 expression by acting as a miR-6089 sponge through the IκBα/NF-κB signaling pathway in macrophages, resulting in increased the production of TNF-α, IL-6, and IL-12, which cause joint inflammation and joint damage (77); CircNUP214 functioned as a miR-125a-3p sponge in RA, which promotes the expression of IL-17A and IL-23R in PBMCs and increases the proportion of Th17 cells, resulting in immune dysregulation and inflammatory responses (92); CircFADS2 promotes the expression of type II collagen, MMP-13, COX-2, and IL-6 via miR-498/mTOR axis, which leads to extracellular matrix (ECM) degradation, inflammation, and cell apoptosis in chondrocytes (72); Circ-PTTG1IP can promote TLR4 expression by acting as a miR-671-5p sponge, promoting the proliferation, migration, invasion and inflammatory response of FLSs in RA (89).

Based on the extensive research on circRNAs, researchers are further exploring the clinical applications of circRNAs as therapeutic targets. Compared to mRNA, miRNA, or lncRNA, circRNAs exhibit higher stability in the cytoplasm as therapeutic targets (111, 112). Additionally, circRNAs have lower immunogenicity of unmodified transcripts which will help to avoid activation of the innate immune responses that can decrease gene editing capabilities and disrupt protein production (113, 114). These two advantages can fully realize the potential of circRNAs in gene editing therapies and long-term protein expression regulation (115). Recent research has found that exosomes from circEDIL3-overexpressing synovial mesenchymal stem cells (SMSCs) can reduce VEGF expression by acting as a miR-485-3p sponge through the PIAS3/STAT3 signaling pathway in RA-FLS, and suppress inflammation-induced angiogenesis, which promotes pannus formation, ultimately ameliorated RA (91).

Consequently, circRNAs have started to exhibit the therapeutic potential in vitro. However, the application of circRNAs in clinical settings still requires further research, including how to construct more efficient and specific carriers for the delivery of circRNAs to specific cells or tissues. Additionally, it is important to clarify circRNAs’ pharmacokinetics, biodistribution in the body, and adverse reactions (116).

Oxidative stress is characterized by the disruption of the intracellular redox balance, resulting in an excessive production of reactive oxygen species (ROS), leading to protein and DNA degradation, as well as lipid peroxidation (117). It plays a crucial role in accelerating the progression of RA and exacerbating clinical symptoms, which includes triggering immune cell activation, promoting synovial inflammation, inducing cell apoptosis and necrosis, increasing joint pain, and initiating joint damage (118–121). Research has revealed that the inflammatory synovial cavity in RA patients is a severely hypoxic environment, which serves as the primary site for oxidative stress and is the fundamental condition for the accumulation of ROS and mitochondrial damage (122). FLS play a pivotal role in the pathogenesis of RA (123). Under hypoxic conditions, FLS exhibit an increased frequency of mitochondrial DNA (mtDNA) mutations, leading to mitochondrial dysfunction, which results in the downregulation of mitochondrial respiration and a decrease in ATP production, along with an upregulation of glycolysis and the accumulation of ROS (124).

Oxidative stress directly activates redox-sensitive transcription factors such as NF-κB, Hypoxia-inducible factor-1α (HIF-1α), Activator Protein-1 (AP-1), and Nuclear factor erythroid2-related factor 2 (Nrf2), which exacerbate the inflammatory responses, oxidative stress, proliferation, invasion, and migration of FLS, induce the recruitment and activation of macrophages and neutrophils that contributing to oxidative stress and inflammatory responses, as well as directly or indirectly contribute to cartilage and bone damage (125–130).

ROS can induce the tyrosine phosphorylation of IκBα by activating Syk kinase, which leads to the dissociation, phosphorylation, and nuclear translocation of NF-κB p65 subunit, thereby stimulating the NF-κB pathway (131, 132). In RA, NF-κB is involved in the induction of various pro-inflammatory cytokines and chemokines, including TNF-α, IL-1β, IL-6, GM-CSF, and others, within monocytes, macrophages, and synovial cells (133). Consequently, these cytokines recruit additional immune cells from the immune system and activate the NF-κB factor, resulting in the initiation of an inflammatory loop (134–136). Furthermore, NF-κB also participates in the regulation of Cyclin D1, Matrix Metalloproteinases (MMPs), and Vascular Endothelial Growth Factor (VEGF), which contribute to the proliferation, migration, and invasion of FLS, as well as cartilage tissue damage (137–139).

HIF-1α is highly regulated by oxygen levels, and in the hypoxic environment of the synovium, HIF-1α accumulates and translocates to the nucleus, where it forms a heterodimer with HIF-1β, known as HIF-1 (140, 141). HIF-1 binds to the HRE promoter thus mediating VEGF expression, thereby leading to angiogenesis (142). Moreover, the upregulation of HIF-1α and VEGF leads to an increase in the expression of stromal cell–derived factor 1α (SDF1α)/CXCL12 mRNA in RA-FLS, which facilitates the recruitment of leukocytes to RA synovial tissue and triggers the release of matrix metalloproteinase 3 (MMP-3) by chondrocytes (143). Furthermore, through the overexpression of HIF-1α in FLS, the upregulation of MMP-1, MMP-3, and IL-8 expression was induced, indicating that HIF-1 can recruit neutrophils, exacerbate inflammation, and induce bone and cartilage damage (144). Studies had found that HIF-1α also plays a significant role in bone cells. Under hypoxic conditions, HIF-1α can upregulate receptor activator of nuclear factor‐κB ligand (RANKL) expression through the JAK2/STAT3 pathway, thereby inducing osteocyte‐mediated osteoclastogenesis (145). HIF-1α can also reduce osteoclast iron death by inhibiting osteoclast iron-starvation response and RANKL-induced ferritinophagy (146). Additionally, HIF-1α can increase the expression and release of angiopoietin-like 4 (ANGPTL4) in various cell types within the rheumatoid synovium, thereby promoting osteoclast-mediated bone resorption (147, 148).

AP-1 is typically a homodimer/heterodimer composed of c-Jun and c-Fos proteins. ROS can activate AP-1 in RA-FLS, leading to the upregulation of IL-1β, MMP-1, MMP-3, and MMP-9 expression (149, 150). Additionally, AP-1 promotes cPLA2 expression by binding to the cPLA2 promoter region, and subsequently increasing the production of eicosanoids like prostaglandin E₂ (PGE₂), which contributes to inflammation and exacerbates pain (151).

Nrf2 is a crucial factor in the RA antioxidant system, regulating the expression of multiple antioxidant genes (152–154). It is also an important target in current research on ncRNA regulation of oxidative stress in RA. In the cytoplasm, Kelch-like ECH-associated protein-1 (keap1) binds to Nrf2 and promotes its degradation, while ROS can directly activate Nrf2, leading to the dissociation of Nrf2 from Keap1 and its translocation into the cell nucleus, where it interacts with antioxidant response elements (ARE), such as heme oxygenase-1 (HO-1) and superoxide dismutase (SOD), to exert its antioxidant effects (155–157). Currently, multiple studies have revealed the involvement of ncRNAs in regulating Nrf2 expression in RA. For instance, overexpression of miR-30a-3p in RA-FLS of rat can downregulate the expression of Keap1 and cullin 3 (cul3), promoting the expression of Nrf2 and its downstream targets, HO-1 and NAD(P)H: Quinone Oxidoreductase 1 (NQO1), thus exerting antioxidant effects (158). Additionally, overexpression of LINC00638 in RA-FLS activates the Nrf2/HO-1 pathway, leading to the upregulation of SOD2 expression and a reduction in levels of ROS and reactive nitrogen species (RNS) (159).

NcRNAs have the capability to regulate the expression of multiple target genes. Currently, extensive and comprehensive studies have been conducted on ncRNAs in the context of synovial inflammation, vascular angiogenesis, and the destruction of bone and cartilage joints in RA. However, much remains to be discovered about oxidative stress in RA. Oxidative stress can expedite the progression of RA by intensifying the inflammatory loop and causing irreversible joint damage. Hence, the exploration of ncRNAs’ role in regulating oxidative stress in RA shows promise in offering new potential therapeutic approaches to manage the advancement of the disease.

FLSs are the predominant cells in synovial tissue, and in RA, they exhibit a high proliferative capacity and an aggressive pathogenic phenotype, invading joint cartilage and bone while secreting MMPs, resulting in joint damage (123, 160, 161). Consequently, it is important to target RA-FLS to suppress their proliferation, migration, and invasive capabilities to counteract joint damage.

Currently, several studies have revealed that ncRNAs play a role in regulating proteins associated with proliferation, migration, and invasion in RA-FLS. MiRNAs can bind to complementary base sequences on target mRNA molecules, resulting in decreased stability, degradation, or inhibition of the translation process of the target mRNA, ultimately reducing the expression level of the target protein (162). For example, miR-4423-3p inhibits the proliferation, migration and invasion of RA-FLS by downregulating the expression level of MMP13 (163); MiR-449a downregulates the expression of High Mobility Group Box 1 (HMGB1) to inhibit the proliferation, migration, and invasion of RA-FLS (164); MicroRNA-133 downregulates mesenchymal-epithelial transition factor (MET), Epidermal Growth Factor Receptor (EGFR), and Fascin Actin-Bundling Protein 1 (FSCN1) to suppress the proliferation, migration, and invasion of RA-FLS (165).

LncRNAs and circRNAs primarily act as sponges for miRNAs, indirectly regulating the expression of target gene mRNAs, thereby promoting the progression of RA (166, 167). Circ_0088194 regulates proliferation, migration, and invasion in RA-FLS through the miR-30a-3p/ADAM10 axis (94). Long non-coding RNA NR-133666 promotes proliferation and migration in RA-FLS through the miR-133c/MAPK1 axis (168). In addition, LncRNAs can also form complexes with proteins, collaboratively participating in the regulation of specific gene expression (166). The elevated expression of LncRNA HAFML in RA-FLS promotes the formation of the HAFML-HuR complex, facilitating the binding of HuR (Human antigen R) to APPL2 mRNA and enhancing its translation, which promotes the migration and invasion of RA-FLS (169). The decreased expression of LncRNA LERFS in RA-FLS leads to a reduction in the complex formation between LERFS and heterogeneous nuclear ribonucleoprotein Q (hnRNP Q), which reduces the binding of the complex to RhoA, Rac1, and Cdc42 mRNA, subsequently promoting the translation process and leading to the proliferation, migration, and invasion of RA-FLS (170).

Evidently, non-coding RNAs (ncRNAs) can effectively modulate the proliferation, migration, and invasion capacities of RA-FLSs through various pathways and targets, thus contributing to the reduction of irreversible damage to joint cartilage and bone (167). Moreover, drugs specifically tailored to target the biological functions of RA-FLS can be locally administered, enabling higher drug concentrations and prolonged therapeutic durations, which enhances treatment efficacy while reducing systemic drug distribution, thereby minimizing adverse effects on other tissues and organs (116, 171). This comprehensive therapeutic strategy represents a promising avenue for RA management.

MiRNAs are small, single-stranded transcripts of about 18-24 nucleotides in length. They do not directly encode proteins, instead, they regulate gene expression at both the transcriptional and translational levels (172, 173).

Several studies have indicated abnormal expression of miRNAs in blood or tissue samples from RA patients, suggesting their potential role in the pathogenesis of RA. As listed in Table 5 , miR-146a and miR-155 had been consistently identified through multiple studies as being overexpressed in PBMCs (21, 23, 174–176). However, isolating PBMCs and fresh plasma is not a common practice in routine clinical procedures, and there is a possibility of significant alteration in miRNA biomarker level caused by minor variations in cell counts or hemolysis during the isolation of cell-free plasma or blood-derived immune cells (such as PBMCs) (177, 178). Based on these two points, further research has revealed a strong correlation between whole blood samples and PBMC expression levels of miR-155 and miR-146 in healthy controls and RA patients (179). This discovery offers a new perspective on the application of miRNAs as clinical biomarkers in RA, facilitating the use of miRNA-based diagnostics in routine clinical practice by using easily accessible samples, such as whole blood. Moreover, recent evidence has demonstrated a promising platform for detecting certain cancers through profiling miRNAs derived from whole blood (180, 181). This significantly broadens the scope for the application of miRNAs as clinical biomarkers.

Additionally, miRNAs are also aberrantly expressed in RA-FLSs and regulate the inflammatory responses, proliferation, migration, and invasion of RA-FLSs. For example, miR-21 promotes the proliferation of RA-FLS by mediating NF-κB nuclear translocation (25); overexpression of miR-27a inhibits migration and invasion of RA-FLS and simultaneously decreases the expression of MMP2, MMP9, and MMP13 via the FSTL1/TLR4/NF-κB axis (33); miR-34a can be transported to RA-FLS via extracellular vesicles (EVs), which reduced inflammation, proliferation, and resistance to apoptosis of RA-FLS by inhibiting the cyclin I/ATM/ATR/p53 signaling pathway (30). MiRNAs can also regulate immune cells and the differentiation of bone cells to modulate the progression of RA. For instance, miR-146a enhances a pro-inflammatory phenotype in Tregs by upregulating the expression of STAT1 (182); miR-182 had been identified as a critical positive regulator of osteoclastogenesis, which is associated with the Foxo3, Maml1 and PKR/IFN-β signaling pathways (26, 27).

LncRNAs are transcripts with a length larger than 200 nucleotides that regulate gene expression through various mechanisms, including epigenetic regulation, transcriptional control, and post-transcriptional regulation. The regulatory mechanisms of lncRNAs mainly encompass (183, 184): (1) LncRNAs can act as sponges for miRNAs, indirectly regulating the expression of target gene mRNAs; (2) LncRNAs can interact with proteins, influencing the binding of proteins to chromatin DNA, altering chromatin structure, and affecting the activity of transcription factors, thus regulating gene transcription; (3) LncRNAs recruit chromatin-modifying enzymes (such as methyltransferases and deacetylases) to specific gene regions, altering histone modification patterns, and modulating chromatin status and gene expression; (4) LncRNAs can form complexes with transcription factors or other RNAs, serving as part of ribonucleoprotein complexes, recruiting other proteins or RNAs, thereby influencing the formation and function of transcription complexes; (5) LncRNAs can regulate mRNA processing, splicing, stability, and translation in the post-transcriptional phase, affecting the final products of gene expression. Currently, the dysregulation of lncRNAs is also observed in the PBMCs, FLS, and immune cells of RA patients, suggesting that lncRNAs might influence the progression of RA by impacting the function of these cells.

Table 5 shows the important LncRNAs associated with the pathogenesis of RA. LncRNA GAS5 can be upregulated by Tanshinone IIA, thereby enhancing the expression of caspase-3/caspase-9 and suppressing the PI3K/AKT signaling pathway, resulting in the apoptosis of RA-FLS (185). This demonstrates the functional capacity of LncRNAs as potential therapeutic targets and their effectiveness as biomarkers to evaluate treatment responses in RA patients. Shui X (39) exploring the dysregulated lncRNAs in PBMCs and differentiated Th17 cells of RA patients and observed an upregulation of LncRNA NEAT1. Additionally, knockdown of lncRNA NEAT1 inhibits Th17/CD4+ T cell differentiation through reducing the STAT3 protein level (39). LncRNA H19 expression was found to be higher in PBMCs of RA patients compared to healthy volunteers (186). Overexpression of H19 in PMA-induced THP-1 cells (a common model used to study macrophage activity modulation) upregulates KDM6A and various M1 macrophage-associated factors (particularly CCL8, CXCL9, CXCL10, and CXCL11) which promote M1 macrophage activity and macrophage migration (186). LncRNA H19 can also aggravate TNF-α-induced inflammatory injury via TAK1 pathway in MH7A cells (41). LncRNA Hotair exhibited notably high expression levels in the PBMCs and serum exosomes of RA patients, promoting the migration of activated macrophages (187). Meanwhile, markedly lower levels of LncRNA Hotair were detected in differentiated osteoclasts and RA-FLS, and upregulating LncRNA Hotair can decrease the activation of MMP-2 and MMP-13 in both differentiated osteoclasts and RA-FLS (187).

In general, whether it’s miRNAs, lncRNAs, or circRNAs, their roles in the pathogenesis of RA are quite similar and can be summarized in the following three points: (1) ncRNAs regulates the inflammatory response of immune cells in the synovium (188); (2) the expression of ncRNAs influences the biological behavior of RA-FLSs, including affecting inflammation, proliferation, migration, and invasion functions; (3) ncRNAs participate in bone metabolism in RA, promoting the secretion of MMPs by RA-FLSs to disrupt the bone matrix and regulating the activation of osteoclasts. During the process in which non-coding RNA exerts its influence, miRNA, being relatively short, exhibits high specificity when binding to target genes. However, this characteristic also limits its diversity in molecular interactions. On the other hand, lncRNA, usually longer and forming stable secondary or tertiary structures, remains relatively stable in cells, resisting degradation by RNase R. Its more complex structure and diverse functional domains allow it to interact with multiple molecules. CircRNAs, share similar functions with lncRNAs, have a circular structure that provides an advantage in resisting degradation compared to linear RNA. While circRNAs and lncRNAs have become hot topics in research due to their structural and functional advantages, it is undeniable that the regulatory roles played by circRNA, lncRNA, and miRNA at the cellular and molecular levels should not be overlooked. Further clarification of the regulatory network constituted by circRNA, lncRNA, and miRNA will be a major research direction in the future (10).

In addition to understanding the role of ncRNAs in the pathogenesis of RA, exploring their application in the diagnosis and treatment of RA is equally crucial. The future application of ncRNAs in RA might develop in several aspects: (1) ncRNA as a disease biomarker for RA diagnosis; (2) ncRNA as a predictor of complications in RA patients, such as cardiovascular complications (189); (3) ncRNA as a biomarker for assessing the treatment response of RA patients (190); (4) the use of engineered exosomes containing ncRNA for RA treatment (13). Exploration in these areas is essential for advancing our comprehension of RA pathogenesis and enhancing diagnostic and treatment approaches.