Prediction and analysis of osteoarthritis hub genes with bioinformatics.

Abstract

Osteoarthritis (OA) is the most common type of arthritis. OA can cause joint pain, stiffness, and loss of function. The pathogenesis of OA is not completely clear. Moreover, there is no effective treatment, and clinical management is limited to symptomatic relief or joint surgery. This study utilized bioinformatics to analyze normal and OA articular cartilage samples to find biomarkers and therapeutic targets for OA. The GSE169077 gene chip dataset was downloaded from the public gene chip data platform of the National Biotechnology Information Center. The dataset included 6 samples of OA tissues and 5 samples of healthy cartilage tissues. Differentially expressed genes (DEGs) were screened using the R language "limma" function package under the threshold of log2[fold change (FC)] ≥2 and a P value <0.05. The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) signal pathways of the target genes were enriched and analyzed using the database for annotation, visualization, and integrated discovery (DAVID), and a protein-protein interaction (PPI) network was further constructed using the search tool for the retrieval of interacting genes/proteins (STRING) database. The coexpression relationship of the genes in the module was visualized and screened with Cytoscape. A total of 27 DEGs were identified, including 9 downregulated genes and 18 upregulated genes. GO signal pathway enrichment analysis showed involvement in hypoxic response, fibrous collagen trimer, and extracellular matrix structural components. KEGG analysis demonstrated associations with protein digestion and absorption, extracellular matrix receptor interaction, and the peroxisome proliferator-activated receptor signal pathway, among several other pathways. A PPI network was obtained through STRING analysis, and the results were imported into Cytoscape software. The 27 DEGs were sequenced by the cytoHubba plug-in by various calculation methods, and 5 hub genes (COL1A1, COL1A2, POSTN, BMP1, and MMP13) were finally selected. These genes were analyzed by PPI again and annotated with GO and KEGG in different colors. Bioinformatics technology effectively identified differential genes in the knee cartilage tissue of healthy controls and patients with OA, providing opportunities to further explore the mechanism and treatment of OA on a transcriptional level.