Multi-Omics discovery and clinical validation of IGFBP2, B2M, and CST3 as a serum biomarker panel for diabetic kidney disease progression.

Diabetic kidney disease (DKD) is a major microvascular complication of diabetes. Current diagnostic markers (e.g., UACR, eGFR) exhibit limitations in sensitivity and specificity, underscoring the need for novel approaches. We integrated single-cell RNA sequencing (scRNA-seq) of kidney tissues from 30 individuals (18 healthy, 12 DKD) with serum proteomics data. Cellular heterogeneity and dysregulated pathways were analyzed using Seurat and pathway enrichment analyses. Intercellular communication networks were deciphered via tensor-cell2cell. A LASSO model identified hub genes, followed by robust validation across independent bulk RNA-seq cohorts (Nephrectomy and Nephroseq). Integration with proteomics prioritized three candidate biomarkers-IGFBP2, B2M, and CST3- which were further assessed in murine DKD models (STZ/HFD-induced) and clinical serum samples (n = 139). Drug-target interactions were predicted using ChEMBL and validated by molecular docking. (1) scRNA-seq revealed aberrant activation of immune pathways and enhanced tubule cell repair in DKD. (2) Cell-cell communication analysis identified 43 hub genes, with 8 genes showing consistent upregulation in glomerular and tubular compartments. (3) Integration of proteomics and transcriptomics pinpointed three serum biomarkers-IGFBP2, B2M, CST3. (4) Longitudinal validation in STZ-induced murine models (n = 16) and human clinical samples (n = 139) confirmed progressive, stage-dependent elevation of all three biomarkers, with late-stage DKD showing the most pronounced elevation versus controls (p < 0.001). (5) Molecular docking predicted high-affinity binding of pramlintide acetate to CST3/IGFBP2 and rivipansel to B2M. Through a multi-omics approach, we identified IGFBP2/B2M/CST3 as a non-invasive biomarker panel for DKD progression, highlighting their roles as both diagnostic markers and therapeutic targets.