Integrated transcriptomic, machine learning, and network pharmacology analyses reveal shared inflammatory biomarkers...

Intervertebral disc degeneration (IDD) and osteoporosis (OP) frequently coexist, but their shared inflammatory molecular features remain unclear. We investigated whether the two disorders share an immune-extracellular matrix axis and sought candidate inflammation-associated biomarkers.

Public transcriptomic datasets GSE124272 and GSE35956 were analyzed separately using differential expression analysis and gene set enrichment analysis. Shared genes were prioritized using LASSO regression, random forest, and support vector machine algorithms, with model performance estimated by nested stratified cross-validation. Exploratory nomograms, network pharmacology, and molecular docking were performed. Expression of NCF2, CRH, and BTK was examined in TNF-α-stimulated immortalized human nucleus pulposus cells and osteoblasts.

Twenty shared differentially expressed genes were identified, with enrichment patterns involving immune, inflammatory, and metabolic processes. The candidate panel comprised BTK, NCF2, CRH, FCGR3A, and SERPINA3. Nested cross-validation yielded AUCs of 0.86 (95% CI, 0.78-0.93) for IDD and 0.84 (95% CI, 0.75-0.91) for OP. The nomograms were retained as theoretical classification models. TNF-α exposure increased NCF2 and decreased CRH expression in both cell models, while BTK fluorescence increased in nucleus pulposus cells. Docking generated exploratory compound-target predictions after relevance and safety screening.

The analyses identified candidate inflammation-associated molecular signatures shared by the two disorders. The cellular experiments demonstrated responsiveness to acute TNF-α exposure but did not establish biomarker function or disease causality. Independent clinical cohorts, primary cells from multiple donors, functional studies, and direct compound-target validation are required.