Periodontitis is an inflammatory, destructive disease involving periodontal supporting tissues due to dysbiosis of the oral microbiota, causing interaction with the host immune defence system (Kinane et al., 2017). According to the analysis of the Global Burden of Disease Study 1990-2019, periodontitis, the sixth most prevalent disease in the world, has a prevalence of 50%, with 1.1 billion people suffering from severe periodontitis (Chen et al., 2021; Kassebaum et al., 2014). Loss of connective and bone tissue caused by periodontitis is the leading cause of tooth loss in adults (Pihlstrom et al., 2005). On the one hand, tooth loss and chewing dysfunction in patients with periodontitis may negatively affect their nutritional intake, quality of life, and even mental health (Reynolds and Duane, 2018). On the other hand, periodontal tissue inflammation may increase the body’s overall inflammatory burden, contributing to further deterioration of diseases such as diabetes and cardiovascular diseases (Preshaw and Bissett, 2019; Sanz et al., 2020). For this reason, much research has been done on the mechanisms of periodontitis over a long period. However, the exact pathogenesis of this disease remains unclear. It is widely recognized that a highly inflammatory state, triggered by ecological dysregulation of oral microorganisms and excessive host immune response to the microorganisms, is closely related to the pathogenesis of periodontitis (Kinane et al., 2007; Hajishengallis and Lamont, 2012; Roberts and Darveau, 2015). Notably, multiple forms of cell death have been found to play an essential role in developing inflammation and mediating immune responses (Weindel et al., 2022). Various forms of cell death have also emerged as a new hotspot for investigating the mechanisms of chronic inflammatory diseases such as periodontitis. Therefore, the investigation of the underlying mechanisms of periodontitis, especially the cell death and immuno-inflammatory regulatory mechanisms, will provide new insights into the pathogenesis of periodontitis.

Disulfidptosis is a new form of cell death, usually caused by abnormal accumulation of disulfides leading to disulfide stress, which disrupts the cytoskeleton (Liu et al., 2023). In addition, the reduced form of nicotinamide adenine dinucleotide phosphate (NADPH) provides the reducing ability to prevent disulfide stress critical for maintaining cell survival against disulfidptosis (Zheng et al., 2023). First, we hypothesized that the cell’s redox state is associated with the onset of disulfidptosis. Numerous studies have shown that periodontitis is an inflammatory disease caused by oxidative stress in which the oxidant and antioxidant systems are imbalanced. For example, excessive reactive oxygen species (ROS) are produced by overactivated polymorphonuclear neutrophils in dental plaque, and excessive ROS leads to an increased oxidant load, which results in oxidative stress in the tissues and triggers periodontal tissue destruction (Sczepanik et al., 2020). Second, studies have shown a significant correlation between the severity of periodontitis and serum levels of disulfides, which are significantly elevated in patients with periodontitis compared to the healthy population (Tayman et al., 2021). Another study also confirmed that thiols with reducing properties were significantly reduced in gingival tissues of periodontitis patients (Cerkezi et al., 2024), and that this reducing compound plays a crucial role in resistance to oxidative stress. Furthermore, disulfidptosis has been progressively studied in inflammatory diseases such as obligatory myelitis and ulcerative colitis (Li et al., 2024; Xiong et al., 2024). Thus, the state of oxidative stress in periodontitis and the changes in disulfide levels allow us to reasonably hypothesize that the development of periodontitis may be associated with disulfidptosis.

Ferroptosis is a type of iron-dependent regulatory cell death. It affects glutathione peroxidase, leading to decreased cellular antioxidant capacity, ROS accumulation, and lipid peroxidation, which results in oxidative cell death and tissue damage (Stockwell et al., 2017). Iron-dependent oxidative stress and lipid peroxidation are recognized as critical mechanisms of ferroptosis in periodontitis (Chen et al., 2022). Regarding the first aspect, iron concentration fluctuations correlate with the severity of periodontitis (Boyer et al., 2018). Periodontal pathogenic microorganisms sequester iron from iron-containing compounds such as ferritin and hemoglobin in infected periodontal tissues to support their growth and reproduction (Olczak et al., 2005). Excess free iron creates an iron overload, which triggers the Fenton reaction, producing substantial amounts of ROS (Lewis, 2010; Ke et al., 2017). Beyond ROS generated by iron overload, human immune defence cells, such as polymorphonuclear leukocytes, consume NADPH to produce ROS (Hirschfeld et al., 2017). The overproduction of ROS results in oxidative damage to proteins and lipid peroxidation in periodontal tissues, closely associated with the severity of periodontitis (Chiu et al., 2017). From the second perspective, by-products of lipid peroxidation are significantly elevated in patients with chronic periodontitis, and indicators such as glutathione peroxidase (GPX) and the glutathione/oxidized glutathione ratio are altered (Borges et al., 2007; Fentoğlu et al., 2015). Hypoxia in periodontal pockets leads to the high expression of hypoxia-inducible factor-1 (HIF-1), causing fatty acid deposition in the microenvironment (Ezzeddini et al., 2021). ROS attacks lipids, such as phospholipids, in periodontal tissues, resulting in oxidative damage to lipids, contributing to periodontitis’s progression. In addition, lipid peroxidation is a chain reaction that leads to biofilm damage and altered fluidity (Halliwell and Chirico, 1993). In summary, the correlation between the development of periodontitis and ferroptosis is well-supported.

Multiple forms of cell death can coexist and interact within the context of periodontitis. The roles of cell death modalities such as autophagy, pyroptosis, and necroptosis in periodontitis have been extensively studied. However, the interplay between ferroptosis and disulfidptosis in periodontitis remains underexplored. Our findings suggest that an imbalance in the redox state not only negatively impacts the reduction state preventing disulfide stress, thereby linking to disulfidptosis, but is also closely associated with the mechanism of ferroptosis. Further exploration is required to learn the specific mechanisms of disulfidptosis and ferroptosis in periodontitis. In our study, we developed transcriptional biomarkers related to periodontitis that integrate disulfidptosis and ferroptosis. These biomarkers provide new insight into periodontitis pathogenesis’s molecular mechanisms and help explore innovative diagnostic methods for periodontitis.

Periodontitis, an inflammatory disease, is predominantly caused by plaque microorganisms and results in irreversible damage to periodontal tissues. Multiple regulatory cell deaths, such as cuproptosis, autophagy, pyroptosis, and necroptosis, have been extensively studied in periodontitis (Song et al., 2017b). Yet, the roles of disulfidptosis and ferroptosis in this disease remain to be clarified. Both ferroptosis and disulfidptosis seem to be closely linked to the redox state. Therefore, our study utilized public databases and clinical samples to associate disulfidptosis- and ferroptosis-related transcripts with the pathogenesis of periodontitis through bioinformatics analysis and experimental validation. This approach aimed to identify potential critical transcripts for disulfidptosis- and ferroptosis-related transcripts in periodontal tissues, offering new insights into the pathological role of DFR in periodontitis and its interplay with the immune microenvironment. In this research, we identified 27 disulfidptosis- and ferroptosis-related transcripts with dysregulated expression in periodontitis, six transcripts that are tissue biomarkers for periodontitis, the top three transcripts with the best classification, established correlations between these transcriptional biomarkers and the immune microenvironment, developed a classification model based on the six transcriptional biomarkers, and delineated two disulfidptosis- and ferroptosis-related expression patterns in periodontitis.

Initially, this study identified 27 differentially expressed disulfidptosis- and ferroptosis-related transcripts. GO and KEGG analyses revealed that these transcripts are functionally related to the intrinsic apoptotic signaling pathway, fatty acid metabolic process, and oxidoreductase activity. Enrichment was observed in the NADPH oxidase complex, mitochondria, and other cellular components. NADPH plays an essential part in both disulfidptosis and ferroptosis. The reduced state of NADPH counteracts disulfide stress and prevents cellular disulfidptosis (Zheng et al., 2023). Furthermore, the increased oxidation of NADPH, disruption of cellular redox homeostasis, and altered lipid metabolism constitute important biochemical and metabolic characteristics of ferroptosis (Yang and Stockwell, 2016; Lin et al. 2016). Additionally, morphological characteristics of ferroptosis encompass mitochondrial contraction, enhanced mitochondrial membrane density, and the rupture of cellular and mitochondrial membranes (Xie et al., 2016). This observation aligns with our study’s findings, providing additional evidence that disulfidptosis and ferroptosis could play a role in the development of periodontitis.

Secondly, our study identified six transcripts that are tissue biomarkers for periodontitis (ALOX12B, BNIP3, CEBPG, LURAP1L, RGS4, and TFAP2C) potentially significant for periodontitis using three machine learning methods: LASSO, SVM, and RF. LURAP1L and RGS4 showed significant overexpression, whereas ALOX12B, BNIP3, CEBPG, and TFAP2C were notably underexpressed in periodontitis. CEBPG, also known as C/EBPγ, belongs to the C/EBP family. Since CEBPG is challenging in forming stable homodimers, it usually forms heterodimers with other family members and represses its transcriptional activity. C/EBPγ forms a heterodimer with ATF4 to alleviate various stress responses such as oxidative stress, etc., and acts as a new antioxidant modulator to regulate redox homeostasis (Huggins et al., 2015). TFAP2C is part of the TFAP2 family (transcription factor activating protein 2). The study showed that TFAP2C is a master regulator of periodontitis (Sawle et al., 2016). BNIP3 protein belongs to the Bcl-2 protein superfamily and promotes mitochondrial autophagy (Zhang and Ney, 2009). Mitochondrial autophagy is critical in osteogenesis (Pei et al., 2018). Exposure to inflammatory cytokines inhibits mitochondrial autophagy, suppressing osteogenesis in periodontal ligament stem cells (PDLSCs) (Lin et al., 2023). Specifically, exposure to inflammatory cytokines inhibits mitochondrial autophagy, suppressing osteogenesis in periodontal ligament stem cells (PDLSCs) (Lin et al., 2023). ALOX12B, from the lipoxygenase (LOX) family, catalyzes the formation of hydroperoxides from polyunsaturated fatty acids (Mashima and Okuyama, 2015). ALOX12B is expressed in dermal epithelial cells, and its metabolites are essential for skin barrier protection (Mashima and Okuyama, 2015). RGS4, or Regulator of G-protein Signaling 4, is recognized as a biomarker of ferroptosis (Dixon et al., 2014). Most studies on these transcripts are bioinformatics-based (Xu et al., 2023), and further investigation is needed to elucidate their role in periodontitis. Identifying these periodontitis biomarkers offers new avenues for early monitoring, risk assessment, and disease progression treatment.

Moreover, a classification model was constructed based on the six transcriptional biomarkers, demonstrating higher AUC values than each transcript. This suggests the combined analysis of these six transcriptional biomarkers is more effective in distinguishing periodontitis gingival tissues from healthy ones. Additionally, our study constructed a nomogram model to visualize the contribution of the degree of transcript expression to the disease risk, providing new insights into the prediction of periodontitis risk from a molecular perspective. An external dataset (GSE10334) validated the efficacy of the classification model and nomogram models’ efficacy. We screened the top three transcripts (CEBPG, TFAP2C, BNIP3) with optimal classification ability among these six transcriptional biomarkers and verified their expression, classification efficacy, and correlations through clinical experiments. We collected 32 gingival samples (16 gingival samples from periodontitis patients and 16 gingival samples from healthy patients) at Jilin University Stomatological Hospital. We recorded the information of these subjects, including gender, age, PD, AL GI, and PI. First, qRT-PCR using these gingival tissues helped us to verify the expression levels of the top three transcripts (CEBPG, TFAP2C, BNIP3) with optimal classification ability. The expression levels of BNIP3 and TFAP2C aligned with the analytical outcomes, exhibiting a notably reduced expression compared to the control group. However, the expression of CEBPG showed no significant deviation from the controls. This lack of significant difference may be ascribed to the limited size of the sample group. Second, These AUCs were determined using single-transcript models. Therefore, these three could be correlated. We performed a correlation analysis of these three transcripts using the expression levels of the transcripts obtained by qRT-PCR. The results confirmed our hypothesis that these three could be correlated. Further, the expression of these three transcripts was dysregulated in periodontitis patients compared to gingival tissues from healthy populations. So, is there a correlation between their expression and the severity of periodontitis? Correlating patient clinical information with transcript expression is one of the highlights of this study. Our results showed no significant correlation between these three transcripts’ expression levels and clinical indicators of periodontitis. After excluding the interference of age and gender (Supplementary Material S5), we initially hypothesized that these three transcripts expression levels may be involved only in the onset of periodontitis and may not be significantly involved in the development and progression of periodontitis. However, our speculation still lacks sufficient evidence, and further studies are needed to elucidate this point.

Thirdly, Immune cells, integral to the immune system, are closely linked to the development of periodontitis (Kinane et al., 2007; Kinane and George, 2009). Our study investigated the differences in immune cell infiltration between healthy individuals and periodontitis patients and the correlation between six transcriptional biomarkers and immune cells. Significant alterations in immune cell ratios were observed between healthy individuals and periodontitis patients, with increased levels of naive B cells, plasma cells, native CD4⁺ T cells, activated memory CD4⁺ T cells, γδT cells, and neutrophils in periodontitis tissues. Conversely, the infiltration levels of Tfh and Tregs were significantly reduced. This is consistent with existing studies (Li et al., 2020). Furthermore, Spearman correlation analysis revealed correlations between six transcriptional biomarkers and immune cells. Among the six transcriptional biomarkers, the top three transcripts with the best classification (BNIP3, CEBPG, and TFAP2C) exhibited a closer association with immune cells. Specifically, dendritic cells, M1 macrophages, and resting mast cells showed positive correlations with these transcripts, while plasma cells demonstrated negative correlations. This suggests that the top three transcripts with the best classification possess strong classification efficacy and may have an essential role in influencing the pathogenesis of periodontitis through immune functions. In the array of cells infiltrated by periodontitis, B cells, and plasma cells predominantly occupy the inflammatory tissue (Zouali, 2017). B cells have a dual role in PD, promoting bacterial clearance and the destructive effects of inflammation, bone resorption, and matrix lysis (Berglundh, Donati, and Zitzmann, 2007). Plasma cells, which develop from B cells, are characterized by their high number and density, marking a critical distinction between advanced periodontitis and long-term gingivitis without attachment or bone loss (Thorbert-Mros et al., 2015). Naive CD4⁺ T cells become activated and differentiate into various subtypes upon interaction with antigenic complexes (Luckheeram et al., 2012). Among the subtypes, CD4⁺ T cells are the primary source of IL-17, a cytokine closely associated with periapical bone loss (Luckheeram et al., 2012). An imbalance between Th17 cells and Tregs plays a critical role in periodontitis, with Th17 cells related to inflammation and destruction of periodontal tissues (Monasterio et al., 2018), and Tregs implicated in reducing bone resorption in periodontitis (Cafferata et al., 2021). γδT cells are more prevalent in inflamed gingiva than in healthy individuals, consistent with our study (Lundqvist et al., 1994; Kawahara et al., 1995). Previous evidence suggests a potential role for γδT cells in periodontitis. Neutrophils have increasingly been recognized for their irreplaceable role in periodontitis (Herrero-Cervera et al., 2022). Besides releasing inflammatory mediators with both pro-inflammatory and anti-inflammatory effects, neutrophils also release toxic substances like ROS and collagenase, which contribute to connective tissue damage and the initiation of bone resorption (Lee et al., 1995; Chapple and Matthews, 2007). The correlation of disulfidptosis- and ferroptosis-related transcriptional biomarkers, especially the top three transcripts with the best classification, with various immune cells, disulfidptosis and ferroptosis were associated with immune responses. Still, the exact mechanism needs to be further investigated.

Fourthly, using unsupervised consistent clustering analysis, two periodontitis subclusters (C1 and C2) were identified based on differing expression patterns of the six characterized DFRs. In the C1 subcluster, ALOX12B, BNIP3, CEBPG, and TFAP2C showed higher expression, while LURAP1L and RGS4 were more expressed in the C2 subcluster. Additionally, these subclusters exhibited distinct immune infiltration profiles and functional pathways. Immune cells that were more infiltrated in periodontitis were more infiltrated in the C2 than in the C1. In addition, the results of GSEA and GSVA showed that C2 was more enriched in immune-related signaling pathways. Cytokines, critical peptide mediators for cell signaling and communication, perform various functions, including modulating immune and inflammatory responses (Ramadan et al., 2020). Chemokines coordinate leukocyte recruitment and activation, thus attracting macrophages, neutrophils, and lymphocytes to inflammation sites (Ramadan et al., 2020). The inflammatory process involves a complex network of cytokines and chemokines, crucial in mediating inflammation in periodontal tissue (Ramadan et al., 2020). Hematopoietic cell lineage encompasses various blood cells differentiating from hematopoietic stem cells, serving as the foundational source of immune cells. Toll-like receptors (TLRs) recognize pathogens and play a pivotal part in activating the host’s innate immune response and adaptive immunity against periodontal disease bacteria (Song et al., 2017a). In summary, the C2 subcluster demonstrates greater immune relevance than the C1 subcluster. This clustering approach, widely utilized in oncology research, has significantly contributed to the clinical management of tumors and clinical decision-making. Identifying various disulfidptosis- and ferroptosis-related subtypes enhances understanding of the role of disulfidptosis and ferroptosis in periodontitis and offers new insights into prognostic assessment and personalized treatment. Owing to limited clinical data in the public database, further investigation of this molecular-level typing about clinical symptoms and prognosis of periodontitis patients is necessary, potentially paving the way for its future application in clinical decision-making and treatment.

This study represents the first instance of linking disulfidptosis to ferroptosis and exploring their pathophysiological roles in periodontitis and their interactions with the immune microenvironment. This work has led to several novel discoveries that may inform future research on the relationship between cell death mechanisms and periodontitis. However, the current study does have some limitations. Although the expression of disulfidptosis- and ferroptosis-related transcripts with the best classification efficiency was validated in a clinical trial, further validation with a larger sample size is needed. Additionally, the absence of comprehensive clinical data, including symptoms and prognostic information in public databases, limited our ability to probe further the association between disulfidptosis- and ferroptosis-related transcripts -based periodontitis subtypes and clinical features. It is essential to elucidate the precise molecular pathways involved in disulfidptosis and ferroptosis within the context of periodontitis in the future to advance understanding in this field.

Our work integrates disulfidptosis and ferroptosis for the first time, utilizing bioinformatics approaches to investigate their role in periodontitis. We constructed a classification model incorporating six disulfidptosis- and ferroptosis-related transcriptional biomarkers, and consistency clustering analysis yielded two disulfidptosis- and ferroptosis-related subtypes enhances understanding of the role of disulfidptosis and ferroptosis in periodontitis and offers new insights into prognostic assessment and personalized treatment modification patterns. Our study demonstrated that six transcripts that are tissue biomarkers for periodontitis, particularly the top three transcripts with the best classification, exhibit correlations with multiple infiltrating immune cells. This information indicates that disulfidptosis and ferroptosis could affect periodontitis by eliciting immune responses. Additionally, the top three transcripts with the best classification we discovered may be involved only in the onset of periodontitis and may not be significantly involved in the development and progression of periodontitis. This will offer new insights into the mechanisms of disulfidptosis and ferroptosis in both the onset and development of periodontitis.