Integrated Network Pharmacology and Single-Cell Transcriptomics Reveal Transketolase as a Potential Target for the...

Acute kidney injury (AKI) lacks targeted pharmacological interventions. While the DanShen-DaHuang (DS-DH) herb pair shows clinical potential for AKI treatment, and our prior study has validated its nephroprotective efficacy in a cisplatin-induced murine model, its specific molecular targets within the renal microenvironment remain undefined. In this study, we integrated network pharmacology and weighted gene co-expression network analysis (WGCNA) to screen AKI-related targets of the DS-DH pair. A multi-algorithmic machine learning pipeline (including LASSO, Boruta, Random Forest, GBM, XGBoost, and Decision Trees) was utilized to calculate feature importance scores and rank core genes. Subsequently, single-cell RNA sequencing (scRNA-seq) data (GSE197266) were analyzed for transcriptomic mapping, pseudotime trajectory, and cell-cell communication. Finally, molecular docking evaluated theoretical binding affinities. After database screening, a total of 603 drug-disease intersecting targets were obtained. Subsequently, 917 module genes significantly associated with AKI were identified by WGCNA, and 62 core candidate genes were determined after intersecting with the above targets. Multi-algorithm machine learning ranked the importance of the 62 targets, with transketolase (TKT) ranking the highest. To elucidate the mechanism of TKT in AKI, scRNA-seq analysis was performed on 77,593 high-quality cells. The results showed that Tkt was specifically enriched in renal macrophages, with the highest expression in the M2-polarized subset. Pseudotime analysis further revealed that Tkt expression dynamics were highly synchronized with the differentiation trajectory of M2 macrophages and positively correlated with the repair markers Arg1 and Mrc1. Cell-cell communication analysis predicted that Tkt⁺ M2 macrophages act as active communication hubs via the Spp1 and Mif signaling axes. Molecular docking validated the favorable binding affinity between core DS-DH compounds and the TKT active pocket. This computational framework predicts that the DS-DH herb pair might mitigate AKI by potentially targeting TKT, a metabolic enzyme closely associated with macrophage M2 polarization. By prioritizing targets via multi-algorithmic scoring, we provide a data-driven rationale and candidate targets for future experimental validation.