2023, Vol. 27 ›› Issue (36): 5898-5904

PTGS2 and STAT3 in steroid-induced osteonecrosis of the femoral head: ferroptosis-related potential diagnostic biomarkers

Liang Xuezhen1, 2, Luo Di2, Li Jiacheng1, Wen Mingtao1, Zhang Jian1, Xu Bo1, 2, Li Gang1, 2   

Abstract: BACKGROUND: Previous experimental studies have shown that ferroptosis might be involved in the pathological process of steroid-induced osteonecrosis of the femoral head, but the pathogenesis of ferroptosis in steroid-induced osteonecrosis of the femoral head remains unclear.  
OBJECTIVE: To analyze the key potential ferroptosis-related genes involved in steroid-induced osteonecrosis of the femoral head to further clarify the biological mechanism of ferroptosis in steroid-induced osteonecrosis of the femoral head by bioinformatics.
METHODS: The GSE123568 mRNA expression profile dataset, including 10 non-steroid-induced osteonecrosis of the femoral head (following steroid administration) samples and 30 steroid-induced osteonecrosis of the femoral head samples, was downloaded from the Gene Expression Omnibus (GEO) database. Ferroptosis-related genes were obtained from the Human Ferroptosis Database (FerrDb). The ferroptosis-related genes involved in steroid-induced osteonecrosis of the femoral head were screened by intersecting the GSE123568 dataset with the set of ferroptosis genes. The differentially expressed ferroptosis-related genes involved in steroid-induced osteonecrosis of the femoral head were identified with R software. In addition, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses of the differentially expressed ferroptosis-related genes involved in steroid-induced osteonecrosis of the femoral head were conducted by “clusterProfiler” R package. Spearman correlations between the expression levels of the differentially expressed ferroptosis-related genes involved in steroid-induced osteonecrosis of the femoral head were confirmed with “corrplot” R package. Moreover, the protein-protein interaction (PPI) network was analyzed by using the Search Tool for the Retrieval of Interacting Genes (STRING); significant gene cluster modules were identified with the MCODE Cytoscape plugin, and hub genes among the differentially expressed ferroptosis-related genes involved in steroid-induced osteonecrosis of the femoral head were screened by CytoHubba plugin. The ROC curves of the final specifically expressed hub genes were analyzed by “pROC” R package. Finally, the expression levels of the hub genes of the differentially expressed ferroptosis-related genes involved in steroid-induced osteonecrosis of the femoral head were validated by using the GSE74089 dataset.  
RESULTS AND CONCLUSION: (1) A total of 30 differentially expressed ferroptosis-related genes were identified between the peripheral blood samples of steroid-induced osteonecrosis of the femoral head patients and non-steroid-induced osteonecrosis of the femoral head patients based on the defined criteria of adjusted P value <0.01 and |log2 FC| >log21.5, including 20 upregulated genes and 10 downregulated genes. (2) The GO and KEGG pathway enrichment analyses revealed that these differentially expressed ferroptosis-related genes were particularly enriched in oxidative stress, hypoxia-inducible factor-1 signaling pathway and ferroptosis. Spearman correlation analysis revealed significant correlations among the differentially expressed ferroptosis-related genes. (3) The PPI results demonstrated that the differentially expressed ferroptosis-related genes interacted with each other. Two significant gene cluster modules were identified through the MCODE plugin, and seven hub genes were identified by using the intersecting results within the MCC, MNC, DMNC, Degree and EPC algorithms of CytoHubba. (4) ROC curve suggested that compared to non-steroid-induced osteonecrosis of the femoral head samples, these seven overlapped ferroptosis-related hub genes had higher diagnostic value in the steroid-induced osteonecrosis of the femoral head samples. (5) The GSE74089 dataset showed that PTGS2 and STAT3 were significantly upregulated in the hip cartilage specimens, which was consistent with the GSE123568 dataset. (6) It is concluded that thirty potential ferroptosis-related genes were identified via bioinformatics analysis. The PPI network was constructed to select the Hub genes in which PTGS2 and STAT3 might serve as potential diagnostic biomarkers because they regulated ferroptosis. These results provide a target and new insight for further exploring the ferroptosis-related action mechanism and diagnosis of steroid-induced osteonecrosis of the femoral head.

Key words: steroid-induced osteonecrosis of the femoral head, ferroptosis, bioinformatics, oxidative stress, biomarker