Exploration of purine metabolism-related genes in the development of osteoporosis by integrated analyses and experiments.

OBJECTIVE: Metabolomics analyses suggest abnormal purine metabolism during the development of osteoporosis. This study aimed to investigate the role of purine metabolism-related genes in osteoporosis. METHODS: Three microarray datasets for osteoporosis were used in this study. Purine metabolism activity score was quantified by GSVA. Differential expression analysis and weighted Gene Co-expression Network Analysis (WGCNA) were employed to screen dysregulated purine metabolism activity-related genes in osteoporosis. Three machine learning algorithms were employed to screen potential biomarkers, and a predictive nomogram based on biomarkers was then established. Drugs that target the biomarkers were predicted based on the DGIDB database. An ovariectomized rat model causing osteoporosis was established to validate the expression of biomarkers and the purine metabolism in vivo. RESULTS: Purine metabolism activity of osteoporosis samples was significantly different from controls. Based on differential analysis and WGCNA, 122 dysregulated purine metabolism-related genes were found, and they were mainly involved in immune inflammation-related functions. PDPK1 and FOXO3 were identified as potential biomarkers, and the nomogram based on PDPK1 and FOXO3 showed well performance to predict the onset risk of osteoporosis, with a clinical net benefit greater than PDPK1 and FOXO3 each alone. Two approved drugs, resveratrol and celecoxib, were predicted to target FOXO3 and PDPK1, respectively. Expression of FOXO3 and PDPK1 was enhanced in osteoporosis both in bioinformatics analysis and in vivo experiments. Rat model showed higher serum levels of xanthine and hypoxanthine as well as high level of uric acid. CONCLUSION: FOXO3 and PDPK1 might be biomarkers and targets for the prevention and treatment of osteoporosis.