Decoding vascular calcification-neutrophil signature in pathogenesis of chronic IgA nephropathy: evidence from artificial intelligence-driven multi-omics and in vitro validation.

BACKGROUND: Dysregulation of neutrophil and consistent vascular calcification(VC) contributes to the pathogenesis and progression of chronic IgA nephropathy(C-IgAN). However, its integrated mechanism in IgAN has not yet been elucidated.

METHODS: We integrated GSVA, CIBERSORT and WGCNA algorithms in C-IgAN patient bulk profiles(GSE35487 and GSE93798) downloaded from GEO database for identified VC and neutrophil (VN)-related shared DEGs. Next, explainable and systemic machine learning algorithms in C-IgAN training and validation bulk profiles identified VN-associated diagnostic model and hub gene. In addition, based on VN-associated shared gene signature, we performed NMF analysis for identification of VN-related molecular subgroups, and then we also investigated the molecular and immune patterns between subgroups. Besides, we also investigated hub gene, VC and neutrophil molecular landscapes at C-IgAN patient single-cell level via advanced analytical frameworks, such as sctenifoldknk and Bayes-prism algorithms. Novelty, deep learning algorithm (DrugRefLector) and molecular docking based on C-IgAN identified optimal therapeutic agent targeting hub gene. Finally, in vitro assays examined the expression pattern of hub gene.

RESULTS: Our study first traced dynamic VN patterns in C-IgAN and identified VN-associated diagnostic and molecular subgroups. Besides, CLEC4D can be considered as neutrophil-distributed hub gene involved in VC and C-IgAN pathogenesis, which also can be considered as a drug target of BRD-K27184429 in C-IgAN.

CONCLUSION: Our study first identified VN-related patterns for C-IgAN patients, which provides novel idea into clinical applications.