Abstract: BACKGROUND: Studies have shown that circulating inflammatory proteins and immune cells are associated with rheumatoid arthritis, but the causal relationship is unclear.
OBJECTIVE: To explore the causal relationships between circulating inflammatory proteins and rheumatoid arthritis mediated by immune cells.
METHODS: We downloaded data on circulating inflammatory proteins and immune cell phenotypes from the GWAS Catalog database (a publicly accessible database jointly established and maintained by the National Human Genome Research Institute and the European Bioinformatics Institute), and genome-wide association study data for rheumatoid arthritis from the FinnGen database (a genomics project resulting from collaboration between Finnish research institutions, biobanks, and international industry partners, also publicly accessible). A two-step Mendelian randomization analysis was then conducted. In the first step, inverse variance weighted method was used to evaluate the causal effects of 91 circulating inflammatory proteins and 731 immune cells on rheumatoid arthritis risk. MR-Egger, weighted median, weighted mode, simple mode, and sensitivity analyses were performed to ensure robustness and assess pleiotropy. The second step evaluated the mediating effects of identified immune cells on the relationship between circulating inflammatory proteins and rheumatoid arthritis.
RESULTS AND CONCLUSION: (1) Inverse variance weighted analysis showed that four circulating inflammatory proteins were significantly correlated with the risk of rheumatoid arthritis, with one circulating inflammatory protein being a risk factor for rheumatoid arthritis and three circulating inflammatory proteins being protective factors for rheumatoid arthritis. Forty-six kinds of immune cells were significantly associated with rheumatoid arthritis, comprising 20 risk factors and 26 protective factors. Reverse Mendelian randomization analysis found no causal associations between rheumatoid arthritis and four identified circulating inflammatory proteins. Sensitivity analyses revealed no significant heterogeneity or horizontal pleiotropy. Further Mediation analysis showed that CD19 on IgD- CD38br partially mediated the causal effect of interleukin-18 on rheumatoid arthritis (β=0.064, odds ratio=1.066, P=0.044), with a mediation effect of 0.004, accounting for 5.7% of the total effect, and a direct effect of 0.060. (2) The study findings reveal that circulating inflammatory proteins and immune cells have causal associations with rheumatoid arthritis, and demonstrate that CD19 on IgD- CD38br partially mediates the causal effect of interleukin-18 on rheumatoid arthritis. With the rich accumulation of data in international multi-omics research platforms and trans-ethnic cohorts, China can reference their multi-dimensional data integration frameworks to construct a dedicated database integrating epigenomics, proteomics, and metabolomics for the Chinese population, revealing molecular regulatory networks underlying complex diseases and aiding in disease subtyping and early diagnostic biomarker development. Learning from transnational collaborative research mechanisms, establishing multi-ethnic and multi-regional natural population cohorts in China will systematically dissect the impact of gene-environment interactions on health, providing a scientific basis for localized disease prevention strategies.

Key words: circulating inflammatory proteins, immune cells, rheumatoid arthritis, Mendelian randomization, inverse variance weighted, causality, mediation analysis, FinnGen database