Abstract: BACKGROUND: Some studies have confirmed the changes in the function of immune cell subsets such as monocytes, T cells, B cells, and natural killer cells (NK cells) in patients with osteoarthritis, but the specific regulatory mechanisms are unclear.
Objective: To explore the causal relationship between plasma metabolite-mediated immune cells and hip osteoarthritis. 
METHODS: The Genome-Wide Association Studies (GWAS) data of 731 immune cells were used as the exposure, the GWAS data of hip osteoarthritis were used as the outcome, and 1 400 plasma metabolites were selected as mediating factors. The GWAS database is an important database for genetic association studies, maintained by international organizations with no country-specific affiliation. The inverse variance weighting method in the two-sample Mendelian randomization method was the main method, and the Bayesian weighted Mendelian randomization method was used to analyze the prior distribution, sample data and weights, which were then used to calculate the posterior distribution. The accuracy and reliability of the inverse variance weighting results were evaluated according to the posterior distribution, supplemented by MR-Egger, weighted median, simple model, and weighted mode methods. The pliotropy test and heterogeneity test were used to ensure the robustness of the process. The results of the inverse variance weighting method were used for subsequent mediating effect analysis. 
RESULTS AND CONCLUSION: (1) The inverse variance weighting method identified 4 immune cells strongly correlated with hip osteoarthritis, and 20 metabolites strongly associated with hip osteoarthritis, all of which had no reverse causal relationship. At the same time, the validation results of Bayesian weighted Mendelian randomization method showed that the posterior mean value was similar to the estimated value of the inverse variance weighting, and the posterior variance was relatively lower. One monocyte subtype (PDL-1 on CD14-CD16+) was finally screened out to have a causal relationship with hip osteoarthritis, with a total effect of -0.047 (odds ratio=0.954, 95% confidence interval: 0.926-0.983), and a mediating effect of -0.004 (odds ratio=0.939, 95% confidence interval: 0.902-0.978) mediated by alliin levels, accounting for 8.5% of the total effect. It was concluded that alliin is a protective factor in the progression of hip osteoarthritis, in which this metabolite plays a mediating role. (2) The large amount of data from international databases and European population analysis is of great significance to Chinese biomedicine, which can provide clues for research on the genetic susceptibility to similar diseases in the Chinese population, aiding in discovering the unique associations. The pharmacogenomic approaches used can be adapted to screen for drug response genes in the Chinese population, enhancing the precision of personalized medicine. Additionally, the advanced high-throughput technologies and statistical methods employed can be learned and applied to disease prevention and treatment research.

Key words: immune cells, plasma metabolites, hip osteoarthritis, two-sample Mendelian randomization analysis, mediation analysis, engineered tissue construction