Identification of hub cuproptosis related genes and immune cell infiltration characteristics in periodontitis

Abstract

Introduction: Periodontitis is an inflammatory disease and its molecular mechanisms is not clear. A recently discovered cell death pathway called cuproptosis, may related to the disease.

Methods: The datasets GSE10334 of human periodontitis and control were retrieved from the Gene Expression Omnibus database (GEO) for analysis.Following the use of two machine learning algorithms, least absolute shrinkage and selection operator (LASSO) and support vector machine-recursive feature removal (SVM-RFE) were used to find CRG-based signature. Then the Receiver operating characteristic (ROC) curves was used to evaluate the gene signature's discriminatory ability. The CIBERSORT deconvolution algorithm was used to study the link between hub genes and distinct types of immune cells. Next, the association of the CRGs with immune cells in periodontitis and relevant clusters of cuproptosis were found. The link between various clusters was ascertained by the GSVA and CIBERSORT deconvolution algorithm. Finally, An external dataset (GSE16134) was used to confirm the diagnosis capacity of the identified biomarkers. In addition, clinical samples were examined using qRT-PCR and immunohistochemistry to verifiy the expression of genes related to cuprotosis in periodontitis and the signature may better predict the periodontitis.

Results: 15 periodontitis-related DE-CRGs were found,then 11-CRG-based signature was found by using of LASSO and SVM-RFE. ROC curves also supported the value of signature. CIBERSORT results of immune cell signature in periodontitis showed that signature genes is a crucial component of the immune response.The relevant clusters of cuproptosis found that the NFE2L2, SLC31A1, FDX1,LIAS, DLD, DLAT, and DBT showed a highest expression levels in Cluster2 ,while the NLRP3, MTF1, and DLST displayed the lowest level in Cluster 2 but the highest level in Cluster1. The GSVA results also showed that the 11 cuproptosis diagnostic gene may regulate the periodontitis by affecting immune cells. The external dataset (GSE16134) confirm the diagnosis capacity of the identified biomarkers, and clinical samples examined by qRT-PCR and immunohistochemistry also verified that these cuprotosis related signiture genes in periodontitis may better predict the periodontitis.

Conclusion: These findings have important implications for the cuproptosis and periodontitis, and highlight further research is needed to better understand the mechanisms underlying this relationship between the cuproptosis and periodontitis.

Keywords: cuproptosis; hub; immune; infiltration; periodontitis.

Figure 1

Identification of differentially expressed cuproptosis-related genes (DE-CRGs) in periodontitis and controls. (A) Heat map of 15 DE-CRGs in the periodontitis; (B) The box plots showed 15 CRGs differernt were expressed between periodontitis and healthy samples, while PDHA1 and ATP7B showed no difference; (C) Protein–protein interaction (PPI) network of the 15 DE-CRGs; (D) Correlations of the 15 DE-CRGs 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. *P < 0.05, ***P < 0.001.

Figure 2

Screening for key feature genes by LASSO and SVM-RFE algorithms. (A, B) Feature genes identified using the least absolute shrinkage and selection operator (LASSO) logistic regression algorithm. Covariates are selected using the regularization parameter λ. (C, D) Support vector machine-recursive feature elimination (SVM-RFE) algorithm to screen feature genes. (E) Venn diagram demonstrating overlapping key feature genes screened by LASSO and SVM-RFE. (F) ROC curve.

Figure 3

CIBERSORT immune infiltration analysis. (A) Percentage of immune cells. (B) Immune cell difference between periodontitis and normal control. (C) Correlation between 11 DE-CRGs and 22 immune cells. *P <0.05, **P < 0.01, ***P < 0.001.

Figure 4

Identification of the cuproptosis-related clusters in periodontitis. Consensus clustering for the 183 periodontitis samples in GSE10334 based on the diagnosistic cuproptosis-related genes (PRGs). Three clusters were classified according to the consensus matrix (A), consensus index of cumulative distribution function (CDF) (B), 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. The expressions of the 11 diagnosistic CRGs in the three clusters are shown in the heat map (E) and box plots (F). **P < 0.01, ***P < 0.001.

Figure 5

GSVA analysis among three cuproptosis clusters in the merged cohort. (A) GSVA analysis between cuproptosis cluster 1 and cluster 2. (B) GSVA analysis between cuproptosis cluster 1 and cluster 3. (C) GSVA analysis between cuproptosis cluster 2 and cluster 3. (D) CIBERSORT analysis between three distinct cuproptosis clusters based on immune-related gene expression. The asterisk symbol indicates the statistical p-value *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 6

Nomogram predicting the risk in periodontitis. (A) Nomogram for patients with periodontitis (B) The calibration curve of nomogram; (C) The Decision curve analysis of nomogram; (D) ROC curve, AUC of 0.908.

Figure 7

Expression analysis of 11 DE-CRGs in periodontitis and controls. (A–K) 11 DE-CRGs mRNA differences between the periodontitis and control groups. (L) Immunohistochemical staining (M) Histograms of quantitative immunohistochemical staining results, *p < 0.05; **p < 0.01; ***p < 0.001.