Abstract: BACKGROUND: Rheumatoid arthritis is influenced by complex genetic and environmental factors. Although observational studies have found some correlation between plasma proteins and rheumatoid arthritis, the susceptibility to confounding and reverse causation makes it difficult to clarify whether these proteins are pathogenic factors of rheumatoid arthritis.
OBJECTIVE: To explore the potential of plasma proteins as biomarkers and therapeutic targets in rheumatoid arthritis through Mendelian randomization analysis of plasma proteins in the onset and progression of rheumatoid arthritis.
METHODS: A large-scale two-sample Mendelian randomization analysis was conducted to comprehensively assess the causal relationships between 1 553 circulating proteins and rheumatoid arthritis based on the Decode database (developed by Decode Genetics in Iceland, which contains genomic data from the Icelandic population), the MR-Base platform (developed by a team of researchers at the University of Oxford in the United Kingdom, specifically designed to provide genetic and phenotypic data for Mendelian randomization analyses), and the GWAS Catalog platform (developed by the European Institute of Bioinformatics, which provides data for genome wide association studies worldwide). The causal effects were estimated using the Wald ratio and inverse variance weighting methods, with Bonferroni correction applied to control for false positives caused by multiple testing. To ensure the robustness of the results, sensitivity analyses were performed to validate the positive causal relationship between circulating proteins and rheumatoid arthritis, and Bayesian colocalization and phenome scanning were used to exclude confounding effects and horizontal pleiotropy. Additionally, external validation was carried out using new plasma protein datasets to reduce the likelihood of false discoveries. Finally, small-molecule compounds associated with candidate proteins were identified using the Drug Signatures Database (DsigDB), and molecular docking was performed to predict the binding patterns and energies between proteins and compounds, identifying the most stable and likely binding molecules and mechanisms.
RESULTS AND CONCLUSION: (1) Sensitivity analyses, including Bayesian colocalization and phenome scanning, identified four plasma proteins with reliable causal relationships with rheumatoid arthritis: FCRL3, IL6R, ICOSLG, and TNFAIP3. Their genetic effects were estimated as follows: FCRL3 [odds ratio (OR)=1.12, 95% confidence interval (CI): 1.07–1.17], IL6R (OR=0.94, 95% CI: 0.91–0.96), ICOSLG (OR=2.42, 95% CI: 1.67–3.52), and TNFAIP3 (OR=2.19, 95% CI: 1.88–2.56). Furthermore, molecular docking analysis revealed that the small-molecule compound benzo[a]pyrene exhibited favorable binding with these candidate proteins, suggesting its potential as a therapeutic agent for rheumatoid arthritis. (2) This study provides a comprehensive analysis of the genetic causal relationships of FCRL3, IL6R, ICOSLG, and TNFAIP3 in rheumatoid arthritis. These proteins not only serve as potential molecular biomarkers for rheumatoid arthritis risk screening and disease prevention, but also offer key candidate targets for further understanding the pathogenic mechanisms of rheumatoid arthritis and developing targeted therapies. Although the study is based on European populations, its findings offer important insights for biomedical research in China. By incorporating Mendelian randomization methods to analyze genetic causality, future research on rheumatoid arthritis in the Chinese population could provide more accurate causal inferences, offering theoretical support for localized risk assessment and treatment strategies.

Key words: plasma proteins, rheumatoid arthritis, Mendelian randomization, Bayesian colocalization, phenome scanning, molecular docking, engineered tissue construction