Identification of mitophagy-related genes with diagnostic value in acute rejection following kidney transplantation using bioinformatics analysis

Abstract

Acute rejection (AR) after kidney transplantation, is a common and serious complication that occurs when the recipient's immune system attacks the graft, and the specific genes and molecular mechanisms underlying the role of mitophagy are still unclear. This study integrated two transcriptomic datasets (GSE129166 and GSE25902) from the GEO database. Thirty differential mitophagy-related genes were identified by intersecting differentially expressed genes, module genes obtained through weighted gene co-expression network analysis and mitophagy-related genes. Functional enrichment analysis uncovered several biological processes and signaling pathways associated with these genes. Four candidate genes including CCND1, ZC3H15, RPL38, and ARPC4, were further identified through Random Forest and Support Vector Machine with recursive feature elimination. Internal, external datasets and a nomogram confirmed they could effectively predict AR. Moreover, these genes significantly correlated with the infiltration of multiple immune cells. Differential expressions of the four genes were also validated in patient's peripheral blood and AR mice. These four mitophagy-related genes may be novel biomarkers for predicting the occurrence and diagnosis of AR.

Keywords: Acute rejection; Gene signature; Immune cells; Kidney transplantation; Mitophagy.

Fig. 1

Identification of DEGs of the AR and normal groups. Volcano plot of all DEGs in the merged dataset (GSE129166 and GSE25902). Blue dots indicate upregulated genes, red dots represent downregulated genes, and black dots denote genes with no significant change in expression.

Fig. 2

WGCNA and DMRGs screening. (A) Clustering analysis of the samples. (B) Measurement of the soft threshold power based on the scale-free fit index (left) and mean connectivity (right). (C) Gene dendrogram with corresponding module colors identified by WGCNA. (D) Heatmap showing the module-trait relationships. (E) The scatter plot showing the correlation between AR and the red module. (F) The scatter plot indicating the association between AR and the yellow module. (G) The Venn diagram representing the intersection of DEGs, WGCNA module genes, and the mitophagy-related genes.

Fig. 3

Function enrichment analysis of DMRGs. (A) Dot plot of GO enrichment analysis for the significant BP, CC, and MF. (B) Dot plot of KEGG pathway enrichment analysis showing the significant pathway.

Fig. 4

Identification of candidate genes. (A) Random forest tree. (B) Gene importance ranking based on RF-RFE. (C) RF-RFE feature selection result plot. (D) SVM-RFE feature selection result plot. (E) Venn diagram of the intersection of the genes screened by the RF-RFE and SVM-RFE.

Fig. 5

Construction of a predictive nomogram model and expression levels of candidate genes. (A) Nomogram for assessing the risk of AR occurrence. (B) Calibration curve for the nomogram. (C) Box plot of the expression levels of four candidate genes in normal and AR groups. Data were analyzed using the Wilcoxon test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 6

Analysis of Immune Cell Infiltration. (A) Heatmap of immune cell infiltration in normal and AR groups. (B) Box plot of immune cell infiltration in normal and AR groups. Correlation analysis for CCND1 (C), ZC3H15 (D), RPL38 (E), and ARPC4 (F) and immune cell infiltration. Data were analyzed using the Wilcoxon test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, no significance.

Fig. 7

GSEA analysis of candidate genes. GSEA plot for (A) CCND1, (B) ZC3H15, (C) RPL38, and (D) ARPC4.

Fig. 8

Validation of the diagnostic value of candidate genes. ROC curves evaluating the diagnostic performance of CCND1, ZC3H15, RPL38, and ARPC4 in (A) the internal dataset (merged expression matrix of GSE129166 and GSE25902), (B) the external dataset GSE14328, and (C) the external dataset GSE1563.

Fig. 9

Consensus Clustering Analysis. (A) Consensus matrix for k = 2, illustrating the optimal clustering of samples. (B) UMAP plot showing the two-dimensional embedding of the re-clustered samples (Subgroup A and Subgroup B). (C) PCA plot demonstrating the separation of Subgroup A and Subgroup B along the first two principal components (PC1 and PC2). Heatmap (D) and box plot (E) of immune cell infiltration between the two re-clustered groups. Data were analyzed using the Wilcoxon test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, no significance.

Fig. 10

Differential analysis of candidate genes in clinical samples and animal models. (A) qRT-PCR analysis of the expression levels of four candidate genes in PBMCs from normal and AR patients. (B) Representative H&E staining of mouse kidney tissue from the normal and AR groups: (a-b) normal group at 100× and 400× magnification, (c-d) AR group at 100× and 400× magnification. (C) qRT-PCR analysis of four candidate genes in mouse kidney tissue from normal and AR groups. Data were analyzed using the unpaired two-tailed Student’s t-test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, no significance.