Identification of Differentially Expressed Genes Related to UFMylation Modifications in Periodontitis: Integrated Insights from Transcriptomics and Clinical Experiments

Technical Summary: Identification of Differentially Expressed Genes Related to UFMylation Modifications in Periodontitis

Problem Statement
Periodontitis is a prevalent chronic inflammatory oral disease characterized by irreversible destruction of periodontal supporting structures, driven by oral microbiota dysbiosis and aberrant host immune responses. While the role of post-translational modifications (PTMs) in inflammatory signaling is recognized, the specific contribution of UFMylation—a ubiquitin-like modification regulating protein stability and immune signaling—to the pathogenesis of periodontitis remains poorly understood. There is a critical knowledge gap regarding the expression patterns, functional roles, and molecular mechanisms of UFMylation-related genes (UFMRGs) in periodontal tissue.

Methodology
The study employed an integrated approach combining transcriptomic bioinformatics with clinical experimental validation:

1. Data Acquisition: Three Gene Expression Omnibus (GEO) datasets were utilized: GSE16134 (training set, 310 samples), GSE10334 (validation set, 247 samples), and GSE223924 (validation set, 20 samples). A list of 1,910 UFMRGs was retrieved from the GeneCards database.
2. Differential Expression & Screening: Differentially expressed genes (DEGs) were identified in the training set using the limma package (criteria: P < 0.05, |log2FC| > 0.5). Candidate genes were defined as the intersection between DEGs and UFMRGs.
3. Machine Learning Feature Selection: Three algorithms—LASSO, SVM-RFE, and Boruta—were applied to the candidate genes to screen for robust feature genes. Overlapping genes from these algorithms were selected as key candidates.
4. Model Construction & Validation: A diagnostic nomogram was constructed using multivariable logistic regression. Model performance was evaluated via ROC curves, Decision Curve Analysis (DCA), and calibration plots.
5. Functional & Network Analysis:
   • Enrichment: GO and KEGG analyses were performed on candidate genes.
   • Immune Infiltration: ssGSEA was used to quantify 28 immune cell types and correlate them with key genes.
   • Regulatory Networks: ceRNA (lncRNA-miRNA-mRNA) and transcription factor (TF)-mRNA networks were constructed.
   • Drug Prediction: Potential therapeutic drugs were predicted using the DSigDB database, followed by molecular docking (CB-DOCK2) to assess binding affinities.
6. Experimental Validation: Quantitative real-time PCR (qRT-PCR) was performed on 10 periodontitis and 10 healthy control tissue samples to verify the expression levels of the identified key genes.

Key Contributions and Results

• Identification of Key Genes: From an initial intersection of 58 UFMylation-related DEGs, machine learning integration identified C4A, GANAB, and TUBA4A as the core diagnostic genes. These genes demonstrated consistent expression trends and Area Under the Curve (AUC) values > 0.7 across all three datasets.
• Diagnostic Performance: The constructed nomogram model based on these three genes exhibited high diagnostic accuracy, achieving an AUC of 0.948 (95% CI: 0.914–0.981) in the periodontitis dataset.
• Functional Insights:
  • Pathways: Enrichment analysis linked the genes to protein processing in the endoplasmic reticulum (ER), N-glycan biosynthesis, and immune pathways (e.g., complement activation).
  • Immune Correlation: C4A correlated positively with activated B cells and myeloid-derived suppressor cells; GANAB with monocytes and plasmacytoid dendritic cells; and TUBA4A showed negative correlations with immature B cells and activated CD4+ T cells.
  • Regulatory Networks: A ceRNA network comprising 139 nodes (3 mRNAs, 99 miRNAs, 37 lncRNAs) and a TF network with 173 factors were established, highlighting complex upstream regulation, particularly for GANAB.
• Therapeutic Potential: Molecular docking identified vinblastine (binding to TUBA4A, -10.0 kcal/mol), fenbendazole (binding to GANAB, -7.5 kcal/mol), and isotretinoin (binding to C4A, -6.9 kcal/mol) as having favorable binding affinities.
• Experimental Confirmation: qRT-PCR results confirmed the bioinformatics predictions: C4A and GANAB were significantly upregulated in periodontitis tissues, while TUBA4A was significantly downregulated.

Significance and Claims
The authors claim this study provides the first systematic multi-dimensional investigation of UFMylation-related genes in periodontitis. The primary significance lies in:

1. Novel Biomarkers: Establishing C4A, GANAB, and TUBA4A as reliable UFMylation-related diagnostic key genes for periodontitis.
2. Mechanistic Insight: Elucidating the role of UFMylation in periodontitis through the lens of ER stress (GANAB), complement activation (C4A), and cytoskeletal remodeling (TUBA4A), suggesting a synergistic relationship between UFMylation dysregulation and immune microenvironment disorders.
3. Clinical Utility: Providing a high-accuracy diagnostic tool (the nomogram) and identifying potential therapeutic targets and drug candidates.

The authors maintain a modest tone regarding clinical application, acknowledging limitations such as the small sample size for qRT-PCR validation and the need for further preclinical evaluation (cell/animal models) to confirm drug efficacy and the direct UFMylation status of the identified proteins. They emphasize that while molecular docking suggests binding, further experimental verification is required.