Development of a classification model and an immune-related network based on ferroptosis in periodontitis

Abstract

Background and objectives: Periodontitis is an immunoinflammatory disease characterized by irreversible periodontal attachment loss and bone destruction. Ferroptosis is a kind of immunogenic cell death that depends on the participation of iron ions and is involved in various inflammatory and immune processes. However, information regarding the relationship between ferroptosis and immunomodulation processes in periodontitis is extremely limited. The purpose of this study was to investigate the correlation between ferroptosis and immune responses in periodontitis.

Methods: Gene expression profiles of gingivae were collected from the Gene Expression Omnibus data portal. After detecting differentially expressed ferroptosis-related genes (FRGs), we used univariate logistic regression analysis followed by logistic least absolute shrinkage and selection operator (LASSO) regression to establish a ferroptosis-related classification model in an attempt to accurately distinguish periodontitis gingival tissues from healthy samples. The infiltration level of immunocytes in periodontitis was then assessed through single-sample gene-set enrichment analysis. Subsequently, we screened out immune-related genes by weighted correlation network analysis and protein-protein interaction (PPI) analysis and constructed an immune-related network based on FRGs and immune-related genes.

Results: A total of 24 differentially expressed FRGs were detected, and an 8-FRG combined signature constituted the classification model. The established model showed outstanding discriminating ability according to the results of receiver operating characteristic (ROC) curve analysis. In addition, the periodontitis samples had a higher degree of immunocyte infiltration. Activated B cells had the strongest positive correlation while macrophages had a strong negative correlation with certain FRGs, and we found that XBP1, ALOX5 and their interacting genes might be crucial genes in the immune-related network.

Conclusions: The FRG-based classification model had a satisfactory determination ability, which could bring new insights into the pathogenesis of periodontitis. Those genes in the immune-related network, especially hub genes along with XBP1 and ALOX5, would have the potential to serve as promising targets of immunomodulatory treatments for periodontitis.

Keywords: ferroptosis; immune infiltration; periodontitis; weighted correlation network analysis.