Identification and validation of autophagy-related genes in Kawasaki disease

Abstract

Background: Kawasaki disease (KD) is a systemic vasculitis of unknown etiology affecting mainly children. Studies have shown that the pathogenesis of KD may be related to autophagy. Using bioinformatics analysis, we assessed the significance of autophagy-related genes (ARGs) in KD.

Methods: Common ARGs were identified from the GeneCards Database, the Molecular Signatures Database (MSigDB), and the Gene Expression Omnibus (GEO) database. ARGs were analyzed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis and protein-protein interaction (PPI) network analysis. Furthermore, related microRNAs (miRNAs), transcription factors (TFs), and drug interaction network were predicted. The immune cell infiltration of ARGs in tissues was explored. Finally, we used receiver operating characteristic (ROC) curves and quantitative real-time PCR (qRT-PCR) to validate the diagnostic value and expression levels of ARGs in KD.

Results: There were 20 ARGs in total. GO analysis showed that ARGs were mainly rich in autophagy, macro-autophagy, and GTPase activity. KEGG analysis showed that ARGs were mainly rich in autophagy-animal and the collecting duct acid secretion pathway. The expression of WIPI1, WDFY3, ATP6V0E2, RALB, ATP6V1C1, GBA, C9orf72, LRRK2, GNAI3, and PIK3CB is the focus of PPI network. A total of 72 related miRNAs and 130 related TFs were predicted by miRNA and TF targeting network analyses. Ten pairs of gene-drug interaction networks were also predicted; immune infiltration analysis showed that SH3GLB1, ATP6V0E2, PLEKHF1, RALB, KLHL3, and TSPO were closely related to CD8 + T cells and neutrophils. The ROC curve showed that ARGs had good diagnostic value in KD. qRT-PCR showed that WIPI1 and GBA were significantly upregulated.

Conclusion: Twenty potential ARGs were identified by bioinformatics analysis, and WIPI1 and GBA may be used as potential drug targets and biomarkers.

Keywords: Autophagy; Bioinformatics; GEO database; Kawasaki disease; qRT-PCR.

Fig. 1

Study flowchart and box diagram of data correction. A Study flowchart. B, C Boxplot of dataset GSE68004 before correction (B) and after correction (C). D, E Boxplot of dataset GSE73461 before correction (D) and after correction (E). KD samples are shown in blue, and control samples are shown in yellow

Fig. 2

Differential expression and Venn diagram. A, B Volcano plot showing the differential expression analysis results for datasets GSE68004 (A) and GSE73461 (B). C Heatmap showing the differential expression analysis results for datasets GSE68004 (C) and GSE73461 (D). E Venn diagram showing the numbers of ARGs obtained from the MSigDB database, the GeneCards database, and both. F Venn diagram showing the numbers of DEGs in dataset GSE68004, dataset GSE73461, and those in both. G Venn diagram showing the numbers of genes that were differentially expressed in both datasets, ARGs that were obtained from both databases, and their intersection, i.e., ARGs in both databases that were differentially expressed in both datasets. H Venn diagram showing the numbers of DEGs of dataset GSE68004, DEGs of dataset GSE73461, ARGs in the MSigDB dataset, and ARGs in the GeneCards database

Fig. 3

Functional enrichment analysis of ARGs in KD. A GO and KEGG enrichment bubble maps of ARGs in KD. B GO and KEGG enrichment circle maps of ARGs in KD. C Functional enrichment Circos maps of the GSE68004 dataset combined with log(fold change) values. D Functional enrichment Circos maps of the GSE73461 dataset combined with log(fold change) values

Fig. 4

PPI network and targeting network of ARGs in KD. A PPI network of ARGs in KD based on the STRING database. B Top 10 key genes of the PPI network. C Network of ARGs and targeted miRNAs. D Network of ARGs and targeted TFs. E Network of ARGs and targeted drugs. Blue oval dots represent ARGs; green rounded rectangles represent miRNAs targeted by ARGs; orange quadrangles represent TFs targeted by ARGs; and yellow hexagons represent gene-related drugs

Fig. 5

Diagnostic significance of ARGs in KD. A, B Boxplots of differential expression of ARGs in datasets GSE68004 (A) and GSE73461 (B). C–F ROC curves for genes SH3GLB1, GBA, RALB, DRAM1, and GNAI3 (C); TSPO, WIPI1, QSOX1, ATP6V0E2, and ATP6V1C1 (D); LRRK2, C9orf72, PIK3CB, PLEKHF1, and WDFY3 (E); and FBXL2, CAMKK2, KLHL3, EPAS1, and DEPP1 (F) in dataset GSE68004. (G–J). ROC curves for genes SH3GLB1, GBA, RALB, DRAM1, and GNAI3 (G); TSPO, WIPI1, QSOX1, ATP6V0E2, and ATP6V1C1 (H); LRRK2, C9orf72, PIK3CB, PLEKHF1, and WDFY3 (I); and FBXL2, CAMKK2, KLHL3, EPAS1, and DEPP1 (J) in dataset GSE73461. *P < 0.05, *P < 0.01, *P < 0.001

Fig. 6

Immune cell infiltration of GSE68004 and GSE73461. A, B Heatmap of the correlation between ARGs and immune cells in KD in datasets GSE68004 and GSE73461. C, D Heatmap of the correlation between immune cells in datasets GSE68004 and GSE73461

Fig. 7

Validation of the differential expression of potential diagnostic markers, WIPI1, WDFY3, ATP6V0E2, RALB, KLHL3, GBA, C9orf72, LRRK2, GNAI3, and PIK3CB, via qRT‑PCR. *P < 0.05, **P < 0.01; ns: not significant