Single–Cell and Spatial Transcriptomics Identify a Candidate COL6A3+ Stromal State and an SPP1‐Associated Signaling Network in Renal Fibrosis in Chronic Kidney Disease

Renal fibrosis is a common pathological manifestation in chronic kidney disease (CKD) and a major route to irreversible renal function loss, involving persistent remodeling of stromal and inflammatory compartments. However, disease‐associated cell states within fibrotic tissue, their spatial organization, and intercellular signals remain incompletely defined. To identify cellular and spatial features of fibrotic remodeling, we integrated scRNA‐seq data from GSE183277 and GSE211785, bulk transcriptomic data from GSE66494, and spatial transcriptomic data from GSE183456 across several renal disease states. Ro/e analysis, Scissor, pathway and functional scoring, trajectory inference, machine‐learning–based spatial classification, CellTrek mapping, differential co‐localization, and CellChat were used to connect disease‐associated cell states with their spatial context and potential intercellular signaling networks. The integrated single‐cell dataset contained approximately 280,000 high‐quality cells representing 13 major renal cell classes. Stromal and myeloid compartments exhibited the most evident disease‐associated remodeling. Reclustering of stromal cells identified seven transcriptional states, among which c2_COL6A3 was enriched in Scissor + cells and displayed myofibroblast‐like, extracellular matrix (ECM)‐producing, and profibrotic features. Myeloid reclustering further identified inflammatory states, including Neu_S100A8. Comparison of 12 machine‐learning models selected SVM for spatial classification of ECM‐associated regions, identifying focal ECM‐positive areas in Visium sections. CellTrek‐based mapping suggested that ECM‐positive and ECM‐negative regions differed in their inferred patterns of cellular co‐localization. CellChat analysis further connected proximal tubular cells (PTCs) with the COL6A3+ stromal state through an SPP1 signaling network involving candidate integrin and CD44 receptors, while virtual COL6A3 perturbation predicted regulatory changes in a limited set of downstream genes. Together, these findings characterize an inflammation‐associated and spatially organized fibrotic microenvironment in which injury‐associated PTCs, activated COL6A3+ stromal cells, and inflammatory myeloid cells remodel together within ECM‐rich regions, with SPP1 representing a potential tubular–stromal signaling link. This integrated multiomic framework identifies a candidate COL6A3+ stromal state and associated signaling programs, providing a framework for subsequent mechanistic and clinical validation.