Sijia Ma,1,2 Hongbing He,1,2 Xiaobin Ren1,2 

1Department of Periodontology, Kunming Medical University School and Hospital of Stomatology, Kunming, 650106, People’s Republic of China; 2Yunnan Key Laboratory of Stomatology, Kunming, 650106, People’s Republic of China

Correspondence: Hongbing He; Xiaobin Ren, Department of Periodontology, Kunming Medical University School and Hospital of Stomatology, Kunming, 650106, People’s Republic of China, Email hehongbing@kmmu.edu.cn; Renxiaobin6688@163.com

Background: Periodontitis represents an inflammatory disease with multiple contributing factors, affecting both oral and systemic health. The mechanisms linking mitochondrial dysfunction to immune responses in periodontitis remain unclear, limiting the development of individualized diagnostic and therapeutic approaches.
Objective:  This study aims to elucidate the roles of mitochondrial dysfunction and immune responses in the pathogenesis of periodontitis, identify distinct molecular subtypes, and discover robust diagnostic biomarkers to support precision medicine approaches.
Methods: Single-cell RNA sequencing and transcriptome data from periodontitis patients were analyzed to identify gene signatures linked to macrophages and mitochondria. Consensus clustering was applied to classify molecular subtypes. Potential biomarkers were identified using five machine learning algorithms and validated in clinical samples through qPCR and IHC.
Results: Four molecular subtypes were identified: quiescent, macrophage-dominant, mitochondria-dominant, and mixed, each exhibiting unique gene expression patterns. From  13 potential biomarkers, eight were shortlisted using machine learning, and five (BNIP3, FAHD1, UNG, CBR3, and SLC25A43) were validated in clinical samples. Among them, BNIP3, FAHD1, and UNG were significantly downregulated (p < 0.05).
Conclusion: This study identifies novel molecular subtypes and biomarkers that elucidate the interplay between immune responses and mitochondrial dysfunction in periodontitis. These findings provide insights into the disease’s heterogeneity and lay the foundation for developing non-invasive diagnostic tools and personalized therapeutic strategies.

Keywords: periodontitis, single-cell RNA sequencing, mitochondrial dysfunction, immune microenvironment, molecular subtypes, machine learning