Mendelian randomization reveals the causal links between miRNAs and rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease characterized by persistent synovial inflammation, progressive joint destruction, and diverse systemic manifestations, all contributing to significantly reduced quality of life and increased risks of disability and mortality in affected patients.Early diagnosis and prompt initiation of treatment are critical to halting disease progression, minimizing joint damage, and improving long-term prognosis.However, early-stage RA remains challenging to diagnose due to its clinical heterogeneity and overlapping symptoms with other arthritic disorders.Currently, the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria serve as the primary diagnostic criteria for RA, incorporating serological markers such as rheumatoid factor and anti-cyclic citrullinated peptide antibody.Despite their widespread clinical use, the sensitivity and specificity of these biomarkers remain suboptimal, often leading to delayed or missed diagnoses and suboptimal disease control.Consequently, there is an urgent need to identify novel and more reliable biomarkers for the early diagnosis and therapeutic evaluation of RA. [] ¹ [] ² [] ³ [] ⁴ [] ³

MicroRNAs (miRNAs), a class of small noncoding RNAs approximately consisting of 19 to 25 nucleotides, have emerged as critical post-transcriptional regulators of gene expression involved in immune regulation, inflammation, and autoimmunity.Due to their remarkable stability in various biofluids such as serum, plasma, and synovial fluid, miRNAs have attracted considerable interest as promising diagnostic, prognostic, and therapeutic biomarkers for RA.Previous studies employing microarray and RNA-seq technologies have revealed differential miRNA expression profiles in peripheral blood mononuclear cells and synovial tissues of RA patients, providing novel insights into disease pathogenesis.Nevertheless, most of these studies are observational in nature and cannot fully exclude the influence of confounding factors such as age, sex, smoking, and body mass index, nor can they establish definitive causal relationships between miRNA dysregulation and RA susceptibility or progression. [,] ^{5 6} [] ⁷ [–] ^{8 10}

Mendelian randomization (MR) is an analytical method that leverages genetic variants as instrumental variables (IVs) to estimate the causal effect of an exposure on an outcome,thereby complementing the limitations of conventional observational studies. In MR studies, genetic variants are randomly allocated at conception and are generally independent of environmental and behavioral confounders, thereby substantially reducing confounding bias and the risk of reverse causation.Therefore, MR analysis can be regarded as a natural tool approximating randomized controlled trials, providing a powerful approach for causal inference.In recent years, MR has been widely applied to explore the causal effects of various biomarkers, metabolites, and circulating molecules on complex diseases, including autoimmune disorders. [] ¹¹ [–] ^{12 14} [] ¹⁵

In this study, we conducted a 2-sample MR analysis by integrating data from large-scale genome-wide association studies (GWAS) and the largest cis-miRNA expression quantitative trait loci (cis-miR-eQTL) datasets. To our knowledge, this is the first large-scale 2-sample MR study to systematically assess the causal role of circulating miRNAs in RA using summary-level data from these resources. Furthermore, we performed comprehensive bioinformatics analyses, including target gene prediction and functional enrichment, to elucidate the potential biological mechanisms linking RA-associated miRNAs to disease pathogenesis. By identifying miRNAs with putative causal roles in RA development, our findings may contribute to the discovery of novel biomarkers and therapeutic targets, thereby advancing the field of precision medicine in RA management.

This study represents the first attempt to systematically investigate the causal effects of circulating miRNAs on RA risk using a 2-sample MR approach based on integrated GWAS and cis-miR-eQTL datasets.Our analysis identified 8 miRNAs with significant evidence supporting their potential causal roles in RA pathogenesis, including 4 risk-enhancing miRNAs (hsa-miR-31-5p, hsa-miR-130a-3p, hsa-miR-182-5p, and hsa-miR-183-3p) and 4 protective miRNAs (hsa-miR-139-5p, hsa-miR-152, hsa-miR-204-5p, and hsa-miR-598). Among these, hsa-miR-130a-3pand hsa-miR-204-5pdemonstrated the strongest evidence, supported by both MR analysis and prior experimental validation in circulating samples from RA patients. Considering that the cis-miR-eQTLs data employed in our MR analysis were derived from blood-based studies, we propose that only miRNAs with both MR-based causal inference and circulating-level experimental validation should be considered high-confidence risk or protective factors for RA. Although hsa-miR-31-5p,hsa-miR-139-5p,hsa-miR-152,and hsa-miR-182-5phave shown differential expression in RA fibroblast-like synoviocytes or mesenchymal stem cells, their lack of validation at the circulating-level limits their current classification as high-confidence biomarkers. Notably, hsa-miR-183-3p and hsa-miR-598 remain experimentally uncharacterized in RA, suggesting their potential value for future functional studies. [] ¹⁶ [] ³⁵ [,] ^{10 36} [,,] ^{9 47 48} [] ⁴⁹ [,] ^{50 51} [] ⁵²

To further elucidate the biological relevance of our findings, we focused subsequent bioinformatics analyses on hsa-miR-130a-3p and hsa-miR-204-5p. Our analyses revealed that hsa-miR-130a-3p potentially regulates 64 target genes enriched in the TGF-β, Hippo, and mTOR signaling pathways. Additionally, hsa-miR-130a-3p was found to interact with ceRNAs such as H19 and modulate key inflammatory and immune-related proteins including TNF, UBB, PPARG, and TGFBR1. Pharmacological assessment identified 20 druggable target genes and suggested therapeutic potential for agents such as resveratrol and flufenamic acid. Previous studies reported that hsa-miR-130a-3p is significantly elevated in the serum of RA patients, both pre-disease onset and in anti-citrullinated protein antibody-positive at-risk individuals, supporting its diagnostic and predictive biomarker potential.Moreover, mechanistic studies have shown that mechanical stress-induced downregulation of H19 in densely populated RA-FLSs leads to reduced miR-130a-3p levels and subsequent CDH11 upregulation, promoting synovial fibroblast invasiveness.Although the TGF-β,Hippo,and mTORpathways have been extensively studied in RA, direct regulatory links between hsa-miR-130a-3p and these pathways remain to be clarified. Notably, resveratrol has been reported to modulate the TGF-β pathway in RA models,raising the intriguing possibility that its therapeutic effects may involve miR-130a-3p-mediated mechanisms. Further experimental studies are warranted to explore this hypothesis. [] ³⁵ [] ⁵³ [–] ^{54 56} [,] ^{57 58} [,] ^{59 60} [] ³⁷

Similarly, hsa-miR-204-5p was identified as a putative regulator of 58 target genes enriched in AMPK, cGMP-PKG, and cAMP signaling pathways. This miRNA interacts with ceRNAs such as NEAT1 and NORAD and modulates critical RA-associated proteins including BCL2, SIRT1, and HMGA2. From a pharmacological perspective, hsa-miR-204-5p targets 20 druggable genes, with cilostazol, melatonin, and curcumin emerging as candidate therapeutic agents. Prior studies have reported that NORAD negatively regulates hsa-miR-204-5p, with elevated NORAD expression correlating with increased inflammatory markers such as TNF-α, IL-6, CRP, and ESR in RA patients.However, the downstream target genes of the NORAD/miR-204-5p axis remain largely undefined. Furthermore, NEAT1 has been shown to modulate inflammation in RA-FLSs by targeting hsa-miR-204-5p via the NF-κB signaling pathway,though the precise downstream mechanisms require further elucidation. The target prediction and enrichment results from our study provide a valuable reference for future functional validation of the ceRNA network centered on hsa-miR-204-5p and its role in RA pathogenesis. Notably, cilostazol, melatonin, and curcumin have demonstrated anti-inflammatory and immunomodulatory effects in RA models,suggesting potential avenues for miRNA-targeted therapeutic strategies. [] ³⁶ [] ⁶¹ [,,] ^{42 62 63}

While miRNAs have been extensively studied in the context of RA, most research has focused on synovial tissues and fibroblast-like synoviocytes, with relatively limited attention paid to circulating miRNAs.Current knowledge regarding the systemic regulatory roles, secretion dynamics, and diagnostic performance of circulating miRNAs in RA remains insufficient. Although several studies have reported altered circulating miRNA profiles in RA patients, the diagnostic sensitivity and specificity of these miRNAs remain suboptimal, with limited validation in large, multicenter cohorts. Moreover, the molecular mechanisms by which specific miRNAs modulate gene expression in peripheral blood mononuclear cells of RA patients are poorly understood.Emerging evidence also suggests that extracellular vesicles, including exosomes, play a crucial role in miRNA-mediated intercellular communication in RA. However, the mechanisms governing miRNA packaging, secretion, and functional transfer to target cells remain largely unexplored.Addressing these knowledge gaps will be essential for translating circulating miRNA biomarkers into clinical practice. In addition, investigating how existing anti-RA drugs modulate miRNA expression profiles may offer novel insights into drug repurposing and facilitate the development of miRNA-based therapeutic interventions. Furthermore, the advancement of targeted delivery systems for circulating miRNAs, such as lipid nanoparticles(LNPs) and engineered exosomes,presents a promising frontier for RA therapy.These strategies could enable precise modulation of disease-relevant miRNAs, thereby enhancing therapeutic efficacy while minimizing off-target effects. [] ⁶⁴ [] ⁶⁵ [,] ^{66 67} [] ⁶⁸ [] ⁶⁷ [] ⁶⁹

In conclusion, our study provides novel evidence supporting the causal roles of specific circulating miRNAs in RA pathogenesis and highlights potential molecular targets and pharmacological agents for future therapeutic development. These findings lay a theoretical foundation for further mechanistic investigations and experimental validations, which are necessary to fully realize the clinical utility of miRNA-based diagnostics and therapeutics in RA.

Despite offering valuable insights into the causal roles of circulating miRNAs in RA and elucidating their potential regulatory mechanisms, this study has several limitations that should be acknowledged. First, the genetic data used for both the exposure (cis-miRNA expression quantitative trait loci) and outcome (RA GWAS summary statistics) were derived exclusively from individuals of European ancestry. As a result, the generalizability of our findings to populations with different genetic backgrounds remains uncertain. Future studies should incorporate multi-ethnic cohorts to validate the observed associations. Second, due to the lack of individual-level data, we were unable to perform stratified analyses based on important demographic and clinical variables such as sex, ethnicity, age, and socioeconomic status. This limitation may obscure potential effect modifiers and population-specific risk profiles. Third, current cis-miR-eQTL resources are primarily derived from blood-based samples, with limited data available for other disease-relevant tissues such as synovial tissue. The absence of synovial cis-miR-eQTL data prevented us from assessing tissue-specific miRNA regulatory effects, which may play a critical role in RA pathogenesis. Fourth, while our bioinformatics analyses identified multiple target genes, signaling pathways, and druggable genes potentially regulated by hsa-miR-130a-3p and hsa-miR-204-5p, these mechanistic findings remain theoretical. Experimental validation, including in vitro and in vivo functional assays, is necessary to confirm the biological relevance and therapeutic potential of these miRNAs in RA. Finally, although 2 of the 8 miRNAs (hsa-miR-130a-3p and hsa-miR-204-5p) have been previously validated in clinical RA cohorts at the circulating level, the remaining miRNAs identified in this study lack experimental confirmation. Future prospective studies and well-designed functional experiments are warranted to address these gaps.

In summary, this study systematically investigated the causal relationships between circulating miRNAs and RA risk using a 2-sample MR framework. We identified 8 miRNAs with significant causal evidence, including 4 risk-enhancing and 4 protective miRNAs. Notably, the roles of hsa-miR-130a-3p as a risk factor and hsa-miR-204-5p as a protective factor were not only supported by MR analysis but also corroborated by previous clinical validation studies, thereby strengthening the credibility of our findings. Furthermore, through comprehensive bioinformatics analyses, we explored the potential molecular mechanisms, regulatory networks, and druggable targets associated with these key miRNAs. These findings provide novel theoretical insights into the pathogenesis of RA and highlight promising directions for the development of miRNA-based diagnostic tools and therapeutic strategies. Our study lays the groundwork for future experimental and clinical research aimed at translating these findings into clinical applications, ultimately contributing to precision medicine and improved patient outcomes in RA management.

The authors thank FinnGen and Huan et al (¹⁶) for data sharing. Their contributions significantly facilitated the progress of this study. https://doi.org/10.1038/ncomms7601↗

Zehong Wei, Junping Yang. Conceptualization:

Zehong Wei, DongXu Chen, LianFa Li. Data curation:

Zehong Wei, DongXu Chen, Junping Yang, Ying Wang. Funding acquisition:

Zehong Wei, Ying Wang. Investigation:

Zehong Wei. Methodology:

DongXu Chen, Junping Yang. Project administration:

Zehong Wei, LianFa Li. Software:

LianFa Li, Junping Yang, Ying Wang. Supervision:

DongXu Chen, LianFa Li, Junping Yang, Ying Wang. Validation:

Zehong Wei. Visualization:

Zehong Wei. Writing – original draft:

Zehong Wei, Junping Yang, Ying Wang. Writing – review & editing: