Elucidating common biomarkers and pathways of osteoporosis and aortic valve calcification: insights into new therapeutic targets

Abstract

Background: Osteoporosis and aortic valve calcification, prevalent in the elderly, have unclear common mechanisms. This study aims to uncover them through bioinformatics analysis.

Methods: Microarray data from GEO was analyzed for osteoporosis and aortic valve calcification. Differential expression analysis identified co-expressed genes. SVM-RFE and random forest selected key genes. GO and KEGG enrichment analyses were performed. Immunoinfiltration and GSEA analyses were subsequently performed. NetworkAnalyst analyzed microRNAs/TFs. HERB predicted drugs, and molecular docking assessed targeting potential.

Results: Thirteen genes linked to osteoporosis and aortic valve calcification were identified. TNFSF11, KYNU, and HLA-DMB emerged as key genes. miRNAs, TFs, and drug predictions offered therapeutic insights. Molecular docking suggested 17-beta-estradiol and vitamin D3 as potential treatments.

Conclusion: The study clarifies shared mechanisms of osteoporosis and aortic valve calcification, identifies biomarkers, and highlights TNFSF11, KYNU, and HLA-DMB. It also suggests 17-beta-estradiol and vitamin D3 as potential effective treatments.

Keywords: Aortic valve calcification; Bioinformatics analysis; Machine learning; Molecular mechanisms; Osteoporosis.

Fig. 1

Study flowchart of the whole procedures.

Fig. 2

Identification of DEGs and common genes in the OP GSE56815 and AVC GSE51472 datasets. A. Volcano plots of DEGs in AVC.B. Volcano plots of DEGs in OP.C. Wayne plots of intersecting genes of DEGs of the two diseases in OP and AVC.D. Heatmaps of common genes of the two diseases in the AVC dataset. E. Heatmaps of common genes of the two diseases in the OP dataset.

Fig. 3

Enrichment analysis of shared genes for both OP and AVC diseases. A. PPI network of 13 shared genes obtained by GeneMANIA construction. B, C. GO, KEGG enrichment analysis of 13 shared genes.

Fig. 4

Screening candidate diagnostic biomarkers for OP vs. AVC using machine learning. A, D. Screening biomarkers for AVC and OP using SVM-RFE algorithm, respectively. B, E. Importance of genes sorted by MeanDecreaseAccuracy value for screening biomarkers for AVC and OP using Randomized Survival Forest Algorithm. C. Wayne plots showing the importance of the AVC 6 genes common to both algorithms. F. Wayne diagram showing 8 genes common to both algorithms for OP.

Fig. 5

Expression of common genes in the two diseases. (A) Wayne plots showing the three common genes obtained by two algorithms for the two diseases. (B) Expression level of HLA-DMB in the two diseases. (C) Expression level of TNFSF11 in the two diseases. (D) Expression level of KYNU in the two diseases. (E, F, G) The ROC curves of the three genes, namely, HLA-DMB, TNFSF11 and KYNU ROC curves of the three genes in the OP dataset.

Fig. 6

Composition of immune cells in OP and AVC. (A) Infiltrating immune cells in the AVC group are plotted as stacked bars. (B) Heat map of correlation between immune cells in the AVC group. (C) Box line diagram of immune cell composition in the AVC group. (D) Infiltrating immune cells in the OP group are plotted as stacked bars. (E) Heat map of correlation between immune cells in the OP group. (F) Box line diagram of immune cell composition in the OP group. line diagram.

Fig. 7

GSEA of the expression levels of three genes, HLA-DMB, TNFSF11 and KYNU, in the OP dataset. (A, B) KEGG analysis by GSEA. (C) GO: BP analysis by GSEA. (D) GO: MF analysis by GSEA. (E) GO: CC analysis by GSEA. (F) Reactome pathway analysis by GSE. (A) Reactome pathway analysis. (G) Demonstration of GSEA enrichment results by ridge diagram.

Fig. 8

Network diagram of common genes miRNA, TF and small molecule drug prediction. (A) Network diagram of common gene KYNU with miRNA. (B) Network diagram of common genes HLA-DMB, TNFSF11 and KYNU with TF. (C) Network diagram of two genes, TNFSF11 and KYNU, with small molecule drugs.

Fig. 9

A. Molecular docking of TNFSF11 with vitamin D3 and KYNU with 17-beta-estradiol visualised. (A) Molecular docking of TNFSF11 with vitamin D3 visualised, producing 1 hydrogen bond, this amino acid is ARG-39. (B) Molecular docking of KYNU with 17-beta-estradiol visualised to produce 2 hydrogen bonds, ARG-313 and MET-316.