Periodontitis is a chronic inflammatory periodontal disease resulting in destroyed periodontal tissue. Many studies have found that the host's inflammatory immune responses are involved in the risk of periodontal tissue damage. In this study, we aim to identify potential biomarkers and therapeutic targets related to immunity in periodontitis.
GSE16134 and GSE10334 were downloaded from the Gene Expression Omnibus (GEO) database, and the immune-related genes were obtained from the Immunology Database and Analysis Portal (ImmPort). After the differentially expressed immune-related genes (DE-IRGs) were identified, enrichment analysis was performed. Two machine learning methods, the least absolute shrinkage and selector operation (LASSO) logistic regression and the support vector machine-recursive feature elimination (SVM-RFE), were used to screen out potential markers for the diagnosis of periodontitis. The CIBERSORT algorithm and LM22 matrix were used to analyze the percentage of infiltrating immune cells in periodontitis. Finally, the potential drug targets for the selected immune-related marker genes were predicted using relevant databases.
A total of 7 genes (CD19, CXCR4, FABP4, FOS, IGHD, IL2RG, and PPBP) were upregulated in periodontitis samples. The area under the receiver operating characteristic curve (AUC) value of only one gene for distinguishing periodontitis from healthy samples ranged from 0.724 to 0.894. The prediction ability of the combined risk score of these 7 DE-IRGs was improved (AUC = 0.955). Naïve B cells, neutrophils, plasma cells, and activated memory CD4 T cells were significantly enriched in periodontitis samples, and 25 drugs targeting 4 DE-IRGs were predicted.
We developed a diagnostic model based on seven IRGs for periodontitis. The possible drugs targeting IRGs may provide new ideas for periodontitis treatment.