Identification and analysis of differently expressed transcription factors in aristolochic acid nephropathy

Abstract

Background: Aristolochic acid nephropathy (AAN) is a rapidly progressive interstitial nephropathy caused by Aristolochic acid (AA). AAN is associated with the development of nephropathy and urothelial carcinoma. It is estimated that more than 100 million people worldwide are at risk of developing AAN. However, the underlying mechanisms driving renal deterioration in AAN remain poorly understood, and the treatment options are limited.

Methods: We obtained GSE27168 and GSE136276 series matrix data from the Gene Expression Omnibus (GEO) related to AAN. Using the R Studio environment, we applied the limma package and WGCNA package to identify co-differently expressed genes (co-DEGs). By GO/KEGG/GSVA analysis, we revealed common biological pathways. Subsequently, co-DEGs were subjected to the String database to construct a protein-protein interaction (PPI) network. The MCC algorithms implemented in the Cytohubba plugin were employed to identify hub genes. The hub genes were cross-referenced with the transcription factor (TF) database to identify hub TFs. Immune infiltration analysis was performed to identify key immune cell groups by utilizing CIBERSORT. The expressions of AAN-associated hub TFs were verified in vivo and in vitro. Finally, siRNA intervention was performed on the two TFs to verify their regulatory effect in AAN.

Results: Our analysis identified 88 co-DEGs through the "limma" and "WGCNA" R packages. A PPI network comprising 53 nodes and 34 edges was constructed with a confidence level >0.4. ATF3 and c-JUN were identified as hub TFs potentially linked to AAN. Additionally, expressions of ATF3 and c-JUN positively correlated with monocytes, basophils, and vessels, and negatively correlated with eosinophils and endothelial cells. We observed a significant increase in protein and mRNA levels of these two hub TFs. Furthermore, it was found that siRNA intervention targeting ATF3, but not c-JUN, alleviated cell damage induced by AA. The knockdown of ATF3 protects against oxidative stress and inflammation in the AAN cell model.

Conclusion: This study provides novel insights into the role of ATF3 in AAN. The comprehensive analysis sheds light on the molecular mechanisms and identifies potential biomarkers and drug targets for AAN treatment.

Keywords: ATF3; Aristolochic acid nephropathy; Transcription factor; c-JUN.

Fig. 1

Workflow: Limma (linear models for microarray data) was used to identify DEGs (differentially expressed genes) from the GEO (Gene Expression Omnibus Series) dataset. TF, transcription factor; GSVA, gene set variation analysis; AAN, aristolochic acid nephropathy.

Fig. 2

Identification of DEGs from the AAN model of HK-2 cells (GSE27168) and mice model (GSE136276). (A) Volcano plot of 24 h AAN cell model DEGs. (B) Heatmap of top 30 DEGs with highest P-Values in 24 h model. (C) Volcano plot of 48 h model DEGs. (D) Heatmap of top 30 DEGs with highest P-Values in 48 h model. (E) Volcano plot of mice model DEGs. (F) Heatmap of top 30 DEGs with highest P-Values in mice model. Color transitions from red to blue denote upregulation to downregulation.

Fig. 3

Identification of DEGs in the animal model. (A, B) Soft threshold established based on scale independence and mean connectivity. (C) Hierarchical clustering tree representing gene co-expression modules in distinct colors. (D) Heatmap illustrating module associations with AAN. The turquoise module demonstrated a significant correlation with AAN. Numbers above and below brackets indicate the correlation coefficient and P-value, respectively. (E) Venn diagram depicting the intersection of genes in turquoise module and DEGs.

Fig. 4

Identification of hub TFs (A) Venn diagram displaying the overlap of DEGs between cell and mice AAN models. (B) GO enrichment (BP, CC, MF) for up-regulated co-DEGs. (C) GO enrichment (BP, CC, MF) for down-regulated co-DEGs. (D) Construction of PPI network. (E) Hub genes are identified via the MCC algorithm. (F) GSVA-derived clustering heatmap of GSE136276’s differently expressed genes (DEGs). (G–I) ATF3 expression in cell and mice datasets. (J–L) c-JUN expression in cell and mice datasets. BP, biological process; CC, cellular component; MF, molecular function.

Fig. 5

Functional pathways showed by KEGG/Reactome Pathway Database.

Fig. 6

Immune cell infiltration analysis. (A) Heatmaps of the immune cell abundance in GSE136276; (B) Correlation matrixes of immune cells in the control group, red indicates negative correlations and higher values indicate higher correlations. Both horizontal and vertical axes demonstrate immune cell subtypes. (C) Correlation matrixes of immune cells in the control group; (D–H) The correlation analysis of ATF3 with Monocytes, basophils, and vessels. (I–L) The correlation analysis of c-JUN with Monocytes, basophils, and vessels.

Fig. 7

Validation of hub TFs in vivo. (A) Flow chart of AAN model construction. (B) Compared to controls, Scr levels in AAN mice, (C) Compared to controls, BUN levels in AAN mice. (D) Representative picture of HE staining. (E) The mRNA expression level of ATF3 in renal tissues of AAN mice. (F) Representative pictures of the ATF3 protein levels in renal tissue. (G) Densitometric analysis of ATF3 levels on western blots. (H) The mRNA expression level of c-JUN in renal tissues of AAN mice. (I) Representative pictures of c-JUN protein levels in renal tissues. (J) Densitometric analysis of c-JUN levels on western blots. (K) Immunohistochemical analysis of ATF3 and c-JUN levels in renal tissues. Scale bar: BUN, blood urea nitrogen, Scr, serum creatinine. Each result was repeated at least three times. Compared with the control group, *p < 0.05; **p < 0.01.

Fig. 8

Validation of hub TFs in vitro. (A) mRNA expression of ATF3 in AA-treated cells vs. controls. (B) Representative images of ATF3 protein levels in vitro AAN. (C) Densitometric analysis of western blot ATF3 levels. (D) mRNA expression of ATF3 and c-JUN under AA treatment compared to controls. (E) Representative images of c-JUN protein levels in vitro AAN. (F) Densitometric analysis of western blot c-JUN levels. (G–J) mRNA expression of GPX4, IL-6, IL-1β, and SMAD3 under AA treatment vs. controls. AA, aristolochic acid. Each result was replicated ⩾3 times. *p < 0.05; **p < 0.01 compared to control group.

Fig. 9

Knockdown of ATF3 attenuates oxidative stress and inflammation. (A) Proteins expression of ACSL4, IL-1β, IL-6, and cleaved Caspase-3 in in vitro AAN vs. controls. (B) The cell viability images of si-ATF3 and si-c-JUN in AA-treated cells. (C) Detection of reactive oxygen species under si-ATF3 treatment. (D) Protein expression of ACSL4, IL-1β and S100A8 after si-ATF3 treatment under AA treatment compared to controls. Each result was replicated ⩾3 times. *p < 0.05; **p < 0.01 compared to control group.