Identification of Key Pyroptosis-Related Genes and Distinct Pyroptosis-Related Clusters in Periodontitis

Abstract

Aim: This study aims to identify pyroptosis-related genes (PRGs), their functional immune characteristics, and distinct pyroptosis-related clusters in periodontitis.

Methods: Differentially expressed (DE)-PRGs were determined by merging the expression profiles of GSE10334, GSE16134, and PRGs obtained from previous literatures and Molecular Signatures Database (MSigDB). Least absolute shrinkage and selection operator (LASSO) regression was applied to screen the prognostic PRGs and develop a prognostic model. Consensus clustering was applied to determine the pyroptosis-related clusters. Functional analysis and single-sample gene set enrichment analysis (ssGSEA) were performed to explore the biological characteristics and immune activities of the clusters. The hub pyroptosis-related modules were defined using weighted correlation network analysis (WGCNA).

Results: Of the 26 periodontitis-related DE-PRGs, the highest positive relevance was for High-Mobility Group Box 1 (HMGB1) and SR-Related CTD Associated Factor 11 (SCAF11). A 14-PRG-based signature was developed through the LASSO model. In addition, three pyroptosis-related clusters were obtained based on the 14 prognostic PRGs. Caspase 3 (CASP3), Granzyme B (GZMB), Interleukin 1 Alpha (IL1A), IL1Beta (B), IL6, Phospholipase C Gamma 1 (PLCG1) and PYD And CARD Domain Containing (PYCARD) were dysregulated in the three clusters. Distinct biological functions and immune activities, including human leukocyte antigen (HLA) gene expression, immune cell infiltration, and immune pathway activities, were identified in the three pyroptosis-related clusters of periodontitis. Furthermore, the pink module associated with endoplasmic stress-related functions was found to be correlated with cluster 2 and was suggested as the hub pyroptosis-related module.

Conclusion: The study identified 14 key pyroptosis-related genes, three distinct pyroptosis-related clusters, and one pyroptosis-related gene module describing several molecular aspects of pyroptosis in the pathogenesis and immune micro-environment regulation of periodontitis and also highlighted functional heterogeneity in pyroptosis-related mechanisms.

Keywords: immune microenvironment; inflammasome; periodontitis; prognostic; pyroptosis.

Figure 1

Study workflow. Step1: 26 differentially expressed (DE)-pyroptosis-related genes (PRGs) were determined by merging the expression profiles of GSE10334 and 37 obtained PRGs. The 26 DE-PRGs were further validated in another periodontitis dataset (GSE16134). Protein-protein interaction (PPI) network and correlation analysis were conducted to investigate the interaction and correlation between the 26 DE-PRGs; Step2: The prognostic value of the 26 DE-PRGs in periodontitis (Training set) was assessed by univariate logistic regression analysis, and least absolute shrinkage and selection operator (LASSO) regression was applied to screen 14 prognostic PRGs from the 26 DE-PRGs. Then, a 14-PRGs signature model was developed based on multivariate logistic regression analysis. The risk scores and receiver operating characteristic (ROC) curves were applied for the model validation; Step3: Consensus clustering was applied based on the 14 signatures, and three pyroptosis-related clusters were determined. Differential expression levels of the 14 PRGs were compared among the three clusters. Functional analysis was applied to compare the Gene Ontology (GO) terms, including Biological Processes (BP), Cellular Components (CC), and Molecular Functions (MF), as well as Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways among the three clusters. Single-sample gene set enrichment analysis (ssGSEA) were applied to compare the immune activities among the three clusters. Step4: Weighted correlation network analysis (WGCNA) identified the pink module as the hub pyroptosis-related module based on the module-trait relationship. Functional analysis was conducted for the pink module, in order to explore the role of pyroptosis in periodontitis

Figure 2

Identification of differentially expressed pyroptosis-related genes (DE-PRGs) in periodontitis. (A) Heat map of 26 DE-PRGs in the periodontitis dataset (GSE10334 and GSE16134). (B, C) The box plots show the different expression levels of the 26 PRGs between disease (periodontitis) and healthy samples in the GSE10334 dataset (B) and GSE16134 dataset (C). *P < 0.05, **P < 0.01, ***P < 0.001. (D) Volcano plot of the 26 DE-PRGs in the GSE10334 dataset. (E) Protein–protein interaction (PPI) network of the 26 DE-PRGs. (F) Correlations of the 26 DE-PRGs in the periodontitis samples Red, positive correlation; Green, negative correlation. The color depth and the size reflect the strength of the relevance. The strongest positive and the strongest negative correlation were displayed in scatter plots.

Figure 3

Construction and validation of a prognostic model. (A) Univariate logistic regression analysis for the 26 differentially expressed pyroptosis-related genes (PRGs) in periodontitis samples (GSE10334) with P < 0.05. (B, C) A total of 14 prognostic genes were selected among 26 periodontitis-related PRGs with least absolute shrinkage and selection operator (LASSO) model, selecting the optimum parameter (λ) as 14 with 10-fold cross-validation. (D) Multivariate logistic regression for the 14 prognostic PRGs. (E) Risk scores of the disease (periodontitis) and healthy samples in the GSE10334 dataset (P < 2.2E-16). (F) Receiver operating characteristic (ROC) curve of the GSE10334 dataset (training set) and the GSE16134 dataset (test set). The area under the curve (AUC) value of the train set and the test set is 0.940 and 0.933.

Figure 4

Identification of the pyroptosis-related clusters in periodontitis. (A–C) Consensus clustering for the 183 periodontitis samples in GSE10334 based on the prognostic pyroptosis-related genes (PRGs). Three clusters were classified according to the consensus matrix (A), consensus index of cumulative distribution function (CDF), and CDF delta area curve (C) for k = 3 by increasing the index from 2 to 9. (D) The principal component analysis (PCA) shows a different distribution of the three clusters. (E, F) The expressions of the 14 prognostic PRGs in the three clusters are shown in the heat map (E) and box plots (F). *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5

Distinct biological functions identified among the three pyroptosis-related clusters. (A–C) Comparison of the GO-BPs (A), GO-CCs (B), and GO-MFs (C) enriched in the three clusters. GO, Gene Ontology; BP, biological processes; CC, cellular components; MF, molecular functions. (D) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enriched in the three clusters. The terms with p-value <0.05 were shown as triangle points (TRUE), and others were displayed as round points (FALSE).

Figure 6

Distinct immune characteristics underlined in the three pyroptosis-related clusters. (A–C) Human leukocyte antigen (HLA) gene expression (A), immune pathway activities (B), and immune cell infiltration (C) of the three pyroptosis-related clusters. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7

Identification of hub pyroptosis-related gene modules by performing weighted gene correlation network analysis (WGCNA). (A) Scale independence and mean connectivity analysis for various soft threshold powers, where the optimum power = 9. (B) Network heat map plot of topological overlap for all genes. Each row and column correspond to a gene. The color row underneath the dendrogram indicated the module assignment, of which 13 modules were identified according to the dynamic tree cut. (C) The module–trait relationships between the 13 modules and three clusters are shown in the heat map. The most significant relationship was between the pink module eigengene and cluster 2 (Cor = 0.36, P = 7e-07). (D) The scatter plots of module membership in pink module vs. gene significance in cluster 2 (Cor=0.41, P=7.7e-25). Cor, correlation..

Figure 8

Biological functions behind the pink module. (A–C) GO-BPs (A), GO-CCs (B), and GO-MFs (C) enriched in the pink module. GO, Gene Ontology; BP, biological processes; CC, cellular components; MF, molecular functions. (D) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enriched in the pink module. The terms with p-value <0.05 are visualized.