Integrative analysis of bone-formation associated genes and immune cell infiltration in osteoporosis, and the prediction of active ingredients in targeted traditional Chinese medicine

Objective

To explore the differential expression and mechanisms of bone formation-related genes in osteoporosis (OP) leveraging bioinformatics and machine learning methodologies, and to predict the active ingredients of targeted traditional Chinese medicine (TCM) herbs.

Methods

The Gene Expression Omnibus (GEO) and GeneCards databases were employed to conduct a comprehensive screening of genes and disease-associated loci pertinent to the pathogenesis of OP. The R package was utilized as the analytical tool for the identification of differentially expressed genes. Least absolute shrinkage and selection operator (LASSO) logistic regression analysis and support vector machine-recursive feature elimination (SVM-RFE) algorithm were employed in defining the genetic signature specific to OP. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses for the selected pivotal genes were conducted. The cell-type identification by estimating relative subsets of RNA transcripts (CIBERSORT) algorithm was leveraged to examine the infiltration patterns of immune cells, with Spearman’s rank correlation analysis utilized to assess the relationship between the expression levels of the genes and the presence of immune cells. Coremine Medical Database was used to screen out potential TCM herbs for the treatment of OP. Comparative Toxicogenomics Database (CTD) was employed for forecasting the TCM active ingredients targeting the key genes. AutoDock Vina 1.2.2 and GROMACS 2020 softwares were employed to conclude analysis results, facilitating the exploration of binding affinity and conformational dynamics between the TCM active ingredients and their biological targets.

Results

Ten genes were identified by intersecting the results from the GEO and GeneCards databases. Through the application of LASSO regression and SVM-RFE algorithm, four pivotal genes were selected: coat protein (CP), kallikrein 3 (KLK3), polymerase γ (POLG), and transient receptor potential vanilloid 4 (TRPV4). GO and KEGG pathway enrichment analyses revealed that these trait genes were predominantly engaged in the regulation of defense response activation, maintenance of cellular metal ion balance, and the production of chemokine ligand 5. These genes were notably associated with signaling pathways such as ferroptosis, porphyrin metabolism, and base excision repair. Immune infiltration analysis showed that key genes were highly correlated with immune cells. Macrophage M0, M1, M2, and resting dendritic cell were significantly different between groups, and there were significant differences between different groups (P < 0.05). The interaction counts of resveratrol, curcumin, and quercetin with KLK3 were 7, 3, and 2, respectively. It shows that the interactions of resveratrol, curcumin, and quercetin with KLK3 were substantial. Molecular docking and molecular dynamics simulations further confirmed the robust binding affinity of these bioactive compounds to the target genes.

Conclusion

Pivotal genes including CP, KLK3, POLG, and TRPV4, exhibited commendable significant prognostic value, and played a crucial role in the diagnostic assessment of OP. Resveratrol, curcumin, and quercetin, natural compounds found in TCM, showed promise in their potential to effectively modulate the bone-forming gene KLK3. This study provides a scientific basis for the interpretation of the pathogenesis of OP and the development of clinical drugs.