Mitochondrial dysfunction in peri-implantitis: bioinformatics and machine learning analysis with in vivo experiment

This study aimed to identify mitochondria-related hub genes in peri-implantitis and to detect their expression in a Sprague-Dawley (SD) rat model. Human peri-implantitis tissue datasets (GSE223924, GSE33774, and GSE106090) were obtained from the GEO database and cross-referenced with the MitoCarta3.0 database to identify mitochondria-related differentially expressed genes (MitoDEGs). Functional characterization was performed through protein-protein interaction (PPI) network analysis, Gene Ontology (GO) enrichment, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. Hub genes were further selected using least absolute shrinkage and selection operator (LASSO) regression and the Boruta algorithm. Their discriminative capacity was assessed via receiver operating characteristic (ROC) curve analysis. Finally, a peri-implantitis model was established in SD rats, and hub gene expression was detected by quantitative real-time polymerase chain reaction (qRT-PCR). A total of 115 MitoDEGs were identified, among which 80 genes formed the core interaction network. Machine learning methods identified five hub genes (TSPO, THEM5, SARDH, COX4I2, and ACSM1), all of which exhibited favorable discriminative ability in both the training and testing sets. Animal experiments confirmed that, compared with the healthy group, the expression patterns of the hub genes in peri-implantitis tissues were consistent with the bioinformatics results. This study offers novel insights into the molecular mechanisms of peri-implantitis and presents potential targets for therapeutic intervention.