APC and ZBTB2 May Mediate M2 Macrophage Infiltration to Promote the Development of Renal Fibrosis: A Bioinformatics Analysis

Abstract

Background and Purpose: The continuous accumulation of M2 macrophages may potentially contribute to the development of kidney fibrosis in chronic kidney disease (CKD). The purpose of this study was to analyze the infiltration of M2 macrophages in uremic patients and to seek new strategies to slow down the progression of renal fibrosis. Methods: We conducted a comprehensive search for expression data pertaining to uremic samples within the Gene Expression Omnibus (GEO) database, encompassing the time frame from 2010 to 2022. Control and uremic differentially expressed genes (DEGs) were identified. Immune cell infiltration was investigated by CIBERSORT and modules associated with M2 macrophage infiltration were identified by weighted gene coexpression network analysis (WGCNA). Consistent genes were identified using the least absolute shrinkage and selection operator (LASSO) and selection and visualization of the most relevant features (SVM-RFE) methods to search for overlapping genes. Receiver operating characteristic (ROC) curves were examined for the diagnostic value of candidate genes. Quantitative real-time PCR (qPCR) examined the expression levels of candidate genes obtained from uremic patients in M2 macrophage. Results: A total of 1298 DEGs were identified within the GSE37171 dataset. Significant enrichment of DEGs was observed in 20 biological processes (BP), 19 cellular components (CC), 6 molecular functions (MF), and 70 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. CIBERSORT analysis observed a significant increase in B-cell memory, dendritic cell activation, M0, M1, M2, and plasma cell numbers in uremic samples. We identified the 10 most interrelated genes. In particular, adenomatous polyposis coli (APC) and zinc finger and BTB structural domain 2 (ZBTB2) were adversely associated with the infiltration of M2 macrophages. Importantly, the expression levels of APC and ZBTB2 were far lower in M2 macrophages from uremic patients than those in healthy individuals. Conclusion: The development of renal fibrosis may be the result of M2 macrophage infiltration promoted by APC and ZBTB2.

Keywords: biomarker; chronic kidney disease; immune infiltration; renal fibrosis; uremia.

Figure 1

Workflow of this study.

Figure 2

Analyses of the transcriptome profiles of both control and uremia samples. (a) Volcano plot showing DEGs between control and uremic groups. (b) Heat map showing 30 DEGs with the largest up or downregulation. (c) Bar graph showing the biological processes, cellular components, and molecular functions of DEGs enrichment. (d) Bar graph showing the KEGG pathway of DEGs enrichment.

Figure 3

Landscape of the immune infiltration in control and uremia samples. (a) Heat map showing the immune infiltration in all samples analyzed by the CIBERSORT method. (b) Comparison of immune cell infiltration in control and uremic samples. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001, and ns = not significant.

Figure 4

Identification related modules. (a) Evaluation of modules associated with immune infiltration in uremia. The correlation heat map demonstrates the correlation between modules and immune cells of different infiltrations. Each row represents a color-coded module and each column indicates one type of immune cell infiltrating into the tissue. The numbers for each cell represent the correlation coefficient and p value. (b) Scatter plot showing the relationship between the associated genes of M2 macrophages and the module members of MElightgreen. (c) Scatter plot showing the relationship between the associated genes of M2 macrophages and the module members of MEgreen.

Figure 5

Identification of 10 hub genes. (a) Wayne diagram showing the 85 potential genes shared by DEGs and MElightgreen and MEgreen modules. (b) Cytoscape visualization showing the network diagram of protein-protein interactions. (c) Network diagram of hub gene junctions generated by cytoHubba plugin. (d) Correlation analysis of 10 pivotal genes between the two groups.

Figure 6

Identification of candidate genes by LASSO regression model and SVM-RFE. (a) Cross-validation to determine the best adjusted parameter log (lambda) for LASSO regression analysis. (b) LASSO coefficients of potential genes. (c, d) ROC curve analysis of the training and testing sets. (e) Venn diagram illustrating that there are three overlapping candidate genes in LASSO and SVM-RFE methods.

Figure 7

Validation of the obtained genes. (a–c) Expression levels of the three diagnostic biomarker candidate genes in the GSE37171 dataset in the control and uremic samples. (d) ROC curves to assess the accuracy of the three predictive biomarkers. 0: control samples; 1: uremic samples. (e, f) Expression levels of APC and ZBTB2 in human peripheral blood mononuclear cells were detected by qPCR. ^∗p < 0.05.