Abstract: BACKGROUND: Multiple studies and observations have indicated a close relationship between inflammation, metabolites, and osteoporosis. However, it is still unclear whether there is a genetic causal effect between inflammation-related proteins and osteoporosis and whether metabolites play a mediating role in this process.
OBJECTIVE: To investigate the causal relationships between inflammation-related proteins and osteoporosis using Mendelian randomization method as well as the mediating effect of plasma metabolites in this process.
METHODS: Summary data from genome-wide association studies (GWAS) were used, with osteoporosis data sourced from the Fengenn database, and GWAS data on inflammation-related proteins and plasma metabolites obtained from published studies. The inverse-variance weighted method was primarily used to assess the exposure-outcome relationships. Bidirectional Mendelian randomization analyses were used to explore the causal relationships between inflammation-related proteins and osteoporosis, and two-step Mendelian randomization was used to discover potential mediating metabolites. Sensitivity analyses were then performed to further validate the robustness of the results. Cochran’s Q test was used to assess heterogeneity, and horizontal pleiotropy was evaluated using the MR-Egger intercept and MR-PRESSO methods.  
RESULTS AND CONCLUSION: The initial bidirectional Mendelian randomization analysis identified five inflammation-related proteins that showed a positive causal relationship with osteoporosis and no reverse causal relationship. Artemin (odds ratio [OR]=0.895, 95% confidence interval [CI]: 0.819-0.979, P=0.015) was negatively associated with osteoporosis, whereas chemokine (C-X-C Motif) ligand 1 (OR=1.100, 95% CI: 1.002-1.209, P=0.046), chemokine (C-X-C Motif) ligand 11 (OR=1.150, 95% CI: 1.043-1.268, P=0.005), interleukin 17C (OR=1.087, 95% CI: 1.004-1.176, P=0.040), and tumor necrosis factor-like weak inducer of apoptosis (OR=1.108, 95% CI: 1.002-1.226, P=0.046) were positively associated with osteoporosis. Sensitivity analyses indicated no heterogeneity or pleiotropy in these causal effects. Subsequently, we conducted a two-step Mendelian randomization to discover potential mediating metabolites. This study showed that 1-palmitoyl-gpc (16:0) increased the negative effect of Artemin on osteoporosis. 5α-androstan-3α,17β-diol monosulfate increased the risk of osteoporosis mediated by chemokine (C-X-C Motif) ligand 1 and chemokine (C-X-C Motif) ligand 11. The ratio of α-ketoglutarate to succinate led to an increased risk of osteoporosis mediated by chemokine (C-X-C Motif) ligand 11 and interleukin-17C. Spermidine and the ratio of proline to trans-4-hydroxyproline contributed to an increased risk of osteoporosis mediated by chemokine (C-X-C Motif) ligand 11. 12,13-DiHOME contributed to an increased risk of osteoporosis mediated by interleukin-17C. The sulfate level of piperine metabolite C16H19NO3(3) and adenosine 3’,5’-cyclic monophosphate contributed to an increased risk of osteoporosis mediated by tumor necrosis factor-like weak inducer of apoptosis. In conclusion, the above data indicate that some inflammation-related proteins can influence the risk of osteoporosis, both positively and negatively, and some of these effects are mediated by plasma metabolites. This provides new insights for future investigations into the occurrence and development mechanisms of osteoporosis.

中国组织工程研究杂志出版内容重点：组织构建；骨细胞；软骨细胞；细胞培养；成纤维细胞；血管内皮细胞；骨质疏松；组织工程

Key words: osteoporosis, inflammation-related proteins, plasma metabolites, immune response, Mendelian randomization, causality, reverse causation, two-way analysis, mediating effect, genome-wide association studies