#2439 Decoding stromal remodeling associated with the anti-fibrotic effects of a ketogenic diet at single-cell resolution

Abstract Background and Aims Ketogenic diets are increasingly being explored as a potential treatment for various kidney diseases. Our previous work demonstrated the anti-fibrotic effects of a therapeutic ketogenic diet in an experimental mouse model of glomerulonephritis, which was associated with reduced levels of fibrosis and improved kidney function five weeks after ending the dietary intervention. However, the underlying mechanisms remain unclear. In this study, we conducted an in-depth analysis of the stromal cell compartment to better understand the anti-fibrotic molecular changes induced by the ketogenic diet that contribute to its beneficial effects. Method Mice were subjected to nephrotoxic serum nephritis (NTS) and treated with ketogenic diet (KD) starting 3 days after NTS induction. The control group was kept on standard chow (SC). On day 14 of NTS kidneys were harvested and analyzed by bulk (n = 6 vs. 5) and droplet-based single-nuclei RNA sequencing using the 10X Genomics Chromium platform (n = 3 vs. 3). The single-nuclei dataset was preprocessed using state-of-the-art bioinformatics tools, including SoupX, DoubletFinder, and scVI for integration. We performed detailed cell type annotation and subclustering, focusing on stromal cell populations. To assess extracellular matrix (ECM) dynamics, we leveraged a “matrisome” database of related gene sets, enabling an in-depth functional comparison across stromal subtypes. We conducted differential expression analysis and assessed cell fraction differences between KD and SC-fed mice. To evaluate the reproducibility of our identified stromal clusters, we integrated our dataset with an existing acute kidney injury time course dataset. Finally, we inferred dynamic gene expression changes and used our data to predict metabolic shifts and transcription factor activity changes (SCENIC) associated with the KD. Results Our final single-nuclei data included 70,000 cells, with over 1,600 cells annotated as stromal. KD-fed mice exhibited a significant reduction in gene expression across various collagens. To quantify ECM activity, we calculated a “matrisome” score for each cell, reflecting its overall ECM production. This score, combined with iterative subclustering and an extensive review of marker genes, enabled us to identify eight distinct stromal subtypes. These included various fibroblast and myofibroblast populations (characterized by higher ECM production), vascular smooth muscle cells, and mesangial cells. The relative abundance of these stromal subtypes differed significantly between KD and controls. Among these subtypes, we identified a population of Adamtsl1-expressing myofibroblasts, which exhibited the strongest reduction in collagen production in response to KD. This effect was specific to collagens, as proteoglycan expression remained unchanged between groups. A subtype of fibroblasts characterized by a reduced expression of ECM was enriched in KD mice. By integrating our data with previously published datasets, we confirmed the robustness of our identified stromal subclusters across different conditions. Finally, we inferred dynamic gene expression trajectories and identified metabolic changes and key transcription factors (regulons) associated with stromal remodeling in response to KD. Conclusion This study presents the first in-depth single-cell transcriptomics analysis characterizing the complex adaptations within the stromal cell compartment associated with the anti-fibrotic effects of a KD. We describe differences in fibroblast subtypes that may contribute to the improved long-term outcome observed in KD-treated animals. We provide a detailed characterization of transcription factor activity and gene expression changes underlying these stromal alterations. Given the growing interest in KD as a therapeutic approach in nephrology, our findings are particularly relevant and may help identify renal conditions where ketogenic interventions could slow kidney disease progression.