Identification of key antifibrotic targets FPR1, TAS2R5, and LRP2BP of valsartan in diabetic nephropathy: A transcriptomics-driven study integrating machine learning, molecular docking, and dynamics simulations

Abstract

Diabetic nephropathy (DN) is a major complication of diabetes and a leading cause of renal failure. While valsartan has been shown to alleviate DN clinically, its antifibrotic mechanisms require further investigation. This study used a transcriptomics-driven approach, integrating in vitro, Machine Learning, molecular docking, dynamics simulations and RT-qCPR to identify key antifibrotic targets. In vitro experiments demonstrated that valsartan combats fibrosis by reversing the mRNA expression levels of fibrosis markers. PCA, t-SNE and UMAP analyses suggest the effectiveness of valsartan in modifying gene expression patterns related to fibrosis. Differential expression analysis identified key fibrosis-related genes, while WGCNA highlighted DN-associated genes in human kidney samples, with 33 potential antifibrotic targets emerging from their intersection. To enhance the accuracy of key targets selection, multiple Machine Learning algorithms-LASSO, SVM-RFE, and XGBoost-were employed, refining the potential antifibrotic targets. Molecular docking and dynamics simulations confirmed strong interactions between valsartan and targets, with RT-qPCR validating their expression reversal. GSEA indicated involvement in RAS, AGE-RAGE, TGF-beta, and PI3K-Akt pathways, affecting oxidative phosphorylation and mitochondrial regulation. These findings provide insight into therapeutic mechanisms of valsartan and demonstrate the potential of transcriptomics-driven approaches in developing targeted DN treatments.

Keywords: Antifibrotic targets; Diabetic nephropathy; Machine learning; Transcriptomics analyses; Valsartan.