Identification and validation of a classifier based on hub aging-related genes and aging subtypes correlation with immune microenvironment for periodontitis

Background: Periodontitis (PD), an age-related disease, is characterized by inflammatory periodontal tissue loss, and with the general aging of the global population, the burden of PD is becoming a major health concern. Nevertheless, the mechanism underlying this phenomenon remains indistinct. We aimed to develop a classification model for PD and explore the relationship between aging subtypes and the immune microenvironment for PD based on bioinformatics analysis.

Materials and methods: The PD-related datasets were acquired from the Gene Expression Omnibus (GEO) database, and aging-related genes (ARGs) were obtained from the Human Aging Genomic Resources (HAGR). Four machine learning algorithms were applied to screen out the hub ARGs. Then, an artificial neural network (ANN) model was constructed and the accuracy of the model was validated by receiver operating characteristic (ROC) curve analysis. The clinical effect of the model was evaluated by decision curve analysis (DCA). Consensus clustering was employed to determine the aging expression subtypes. A series of bioinformatics analyses were performed to explore the PD immune microenvironment and its subtypes. The hub aging-related modules were defined using weighted correlation network analysis (WGCNA).

Results: Twenty-seven differentially expressed ARGs were dysregulated and a classifier based on four hub ARGs (BLM, FOS, IGFBP3, and PDGFRB) was constructed to diagnose PD with excellent accuracy. Subsequently, the mRNA levels of the hub ARGs were validated by quantitative real-time PCR (qRT-PCR). Based on differentially expressed ARGs, two aging-related subtypes were identified. Distinct biological functions and immune characteristics including infiltrating immunocytes, immunological reaction gene sets, the human leukocyte antigen (HLA) gene, and immune checkpoints were revealed between the subtypes. Additionally, the black module correlated with subtype-1 was manifested as the hub aging-related module and its latent functions were identified.

Conclusion: Our findings highlight the critical implications of aging-related genes in modulating the immune microenvironment. Four hub ARGs (BLM, FOS, IGFBP3, and PDGFRB) formed a classification model, and accompanied findings revealed the essential role of aging in the immune microenvironment for PD, providing fresh inspiration for PD etiopathogenesis and potential immunotherapy.