Abstract: BACKGROUND: Disturbances in bone metabolism have a significant association with ferroptosis in steroid-induced osteonecrosis of the femoral head (SONFH). Furthermore, the pathologic process of SONFH is characterized by the presence of cartilage damage and degeneration. However, the specific regulatory targets and the relationship between ferroptosis and cartilage concerning SONFH remain unclear.
OBJECTIVE: To employ bioinformatics and machine learning techniques to identify specific genes associated with ferroptosis that target cartilage and to investigate the correlation between ferroptosis and cartilage, thereby providing novel ideas and methodologies for the study and treatment of SONFH.
METHODS: Disease datasets pertinent to the study and ferroptosis-related genes were retrieved from the GEO and FerrDb databases. Subsequently, the disease datasets were normalized and differential analysis using the R language to identify ferroptosis-related differential genes (Fe-DEGs). We conducted Gene Ontology (GO) functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathway enrichment analysis of Fe-DEGs. Furthermore, ferroptosis-related signature genes were filtered based on the protein-protein interaction network of Fe-DEGs and machine learning methods. Finally, the rabbits were divided into normal and model groups. The normal group was given the same dose of saline to simulate the modeling drug, and the animal model of SONFH in rabbits was constructed by injection of modified horse serum combined with methylprednisolone. After successful modeling, the expression of signature gene was verified between different groups, and the phenotype of ferroptosis in cartilage was analyzed.
RESULTS AND CONCLUSION: Through the normalization and differential analysis of the dataset, a total of 1 315 differentially expressed genes were identified. Additionally, 379 ferroptosis-related genes were obtained from the FerrDb database. After intersecting both gene sets, 19 Fe-DEGs were obtained. The GO analysis revealed that Fe-DEGs were mainly involved in biological processes such as cell migration and cellular response to oxidative stress, cellular components such as kinase complexes, amino acid complexes, and cytoplasmic membranes, as well as molecular functions such as kinase activity, receptor activity, and protein binding. The KEGG analysis revealed that Fe-DEGs were mainly enriched in the FoxO signaling pathway, vascular endothelial growth factor signaling pathway, and FcγR-mediated phagocytosis. Constructing a protein-protein interaction network and using machine learning, we identified the ferroptosis-related signature gene, CA9. The gene set enrichment analysis of the signature gene CA9 revealed an upregulated expression in biological processes such as fatty acid metabolism and O-GlcNAc glycosylation modification, while being inhibited in terms of neural activity and ligand-receptor interactions. RT-PCR and western blot results showed that compared with the normal group, the expressions of ACSL4 and CA9 at mRNA and protein levels were significantly higher in the model group (P < 0.05), while the expressions of SLC7A11 and GPX4 at mRNA and protein levels were significantly lower in the model group (P < 0.05), coinciding with the expression levels of the signature genes in the dataset. These findings indicate that the cartilage of SONFH is closely related to ferroptosis, and targeting the signature gene may provide certain ideas and directions for the study and treatment of SONFH.

Key words: steroid-induced osteonecrosis of the femoral head, cartilage, ferroptosis, bioinformatics, signature gene, biomarker, CA9, ACSL4, SLC7A11, GPX4