Inflammatory bowel disease (IBD) and osteoporosis (OP) are often comorbid conditions, with IBD contributing to accelerated development of OP. To explore the molecular links, gene expression data from the GEO database were analyzed. The GSE126124 dataset (IBD) included 57 diseased and 21 normal tissue samples, while the GSE56815 dataset (OP) comprised 40 OP and 40 normal samples. Using the limma package, 9,039 genes were differentially expressed in IBD, and 378 genes were differentially expressed in OP with statistical significance (p < 0.05).
Weighted Gene Co-expression Network Analysis (WGCNA) was conducted on the IBD dataset, identifying modules of co-expressed genes significantly correlated with disease status. Intersection of differentially expressed genes and significant WGCNA modules yielded 4,417 genes associated with IBD. Subsequent overlap with OP-related differentially expressed genes resulted in 88 shared genes potentially involved in both diseases.
Functional enrichment analysis of these 88 genes was performed using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG). GO terms enriched included non-canonical NF-κB signaling, cell activation, receptor for advanced glycation endproducts (RAGE) binding, and protein kinase regulatory activity. KEGG pathways highlighted osteoclast differentiation and interleukin-17 (IL-17) signaling, processes relevant to bone metabolism and inflammation.
Given the significance of apoptosis in the pathophysiology of both IBD and OP, the 88 shared genes were cross-referenced with an apoptosis-related gene set from GeneCards (genes with relevance score ≥5). This intersection identified 12 apoptosis-associated genes. Lasso regression—a machine learning technique to refine variable selection—was applied separately on IBD and OP datasets.
The OP dataset lasso analysis identified 9 key genes (e.g., NFKBIA, HSPB1, S100A8, WNT1). The IBD data lasso analysis selected 5 genes (including S100A8, WNT1, MITF). Intersection of these lists yielded three genes common to both conditions: MITF, S100A8, and WNT1.
To assess their discriminative potential, the expression data from both datasets were combined and batch effect corrected. Differential expression analysis showed that S100A8 and WNT1 were significantly upregulated in disease groups. Receiver Operating Characteristic (ROC) curve analysis demonstrated that S100A8 and WNT1 had Area Under the Curve (AUC) values of approximately 0.695 and 0.684, respectively, indicating moderate diagnostic accuracy. MITF did not show significant differential expression or diagnostic utility.
Gene Set Enrichment Analysis revealed divergent functional profiles for these two genes. WNT1 was enriched in pathways related to cell cycle checkpoints, chromosome organization, and neuroactive ligand–receptor interactions. In contrast, S100A8 was associated with adaptive immune responses, inflammatory bowel disease pathways, and immune receptor recombination processes.
Pearson correlation analysis identified gene sets correlated with each key gene. WNT1 correlated with genes such as TSPAN32, POU3F3, and EGFL7, while S100A8 correlated with inflammatory and immune-related genes including MMP3, S100A9, CXCL6, and CXCR1, suggestive of their respective roles in regulatory and immune pathways bridging the two diseases.
This integrative bioinformatics study demonstrates that apoptosis-related genes WNT1 and S100A8 are crucial molecular links between inflammatory bowel disease and osteoporosis. Their involvement in cell cycle regulation and immune responses highlights potential therapeutic targets. These findings provide a molecular basis for the increased osteoporosis risk in IBD patients and suggest novel diagnostic biomarkers.
Further experimental validation in clinical samples and mechanistic studies are warranted to confirm their roles and to develop targeted interventions aimed at mitigating bone loss in IBD. Understanding these shared pathways may improve clinical management and reduce comorbidities associated with these chronic diseases.