Identification of Cuproptosis-Related Genes and Immunological Roles in Steroid-Induced Osteonecrosis of the Femoral Head.

OBJECTIVE

Steroid-induced Osteonecrosis of the Femoral Head (SONFH) is a disabling disorder arising from chronic glucocorticoid exposure. Cuproptosis, a copper- dependent form of cell death, has been implicated in various diseases, but its role in SONFH remains unclear. This study aimed to characterize cuproptosis-related molecular subtypes and develop a predictive model for SONFH.

METHODS

Using the GSE123568 cohort (30 SONFH cases, 10 controls), we analyzed Differentially Expressed Cuproptosis-related Genes (DE-CRGs) and immune infiltration via CIBERSORT. Unsupervised clustering and Weighted Gene Coexpression Network Analysis (WGCNA) identified molecular subtypes and module genes. Three machine learning models (Random Forest (RF), Generalized Linear Models (GLMs), Support Vector Machines (SVMs)) were compared; the RF model was validated using a nomogram, calibration, decision curve analysis, and the GSE74089 cohort. Target genes were confirmed by immunohistochemistry on clinical femoral head tissue.

RESULTS

NLRP3, NFE2L2, and PDHB were differentially expressed in SONFH and confirmed by immunohistochemistry. Two molecular clusters with distinct immune infiltration profiles were identified. The RF model performed best (AUC = 0.889), and five predictive genes-KCNE3, SOD1, ACAP1, ALDH2, and ITGAL-were selected. The nomogram demonstrated high predictive accuracy in external validation.

DISCUSSION: These findings reveal the potential role of cuproptosis-related genes in SONFH pathogenesis and immune regulation, providing biomarkers for molecular subtyping and risk prediction.

CONCLUSION

NLRP3, NFE2L2, and PDHB are key modulators of SONFH pathogenesis, and the 5-gene RF-based nomogram offers a novel, effective diagnostic approach for clinical application.