KETOGENIC DIET ATTENUATES CHRONIC KIDNEY DISEASE PROGRESSION IN A MURINE MODEL OF GLOMERULONEPHRITIS VIA METABOLIC REPROGRAMMING OF FIBROBLASTS

Chronic kidney disease (CKD) represents a major global health burden, associated with high morbidity, mortality, and healthcare costs. Although novel therapeutics are being developed to improve outcomes in inflammatory kidney diseases such as glomerulonephritis (GN), many patients with GN still progress to CKD and eventually end-stage kidney disease (ESKD). Here, we explore the therapeutic potential and underlying mechanisms of dietary carbohydrate restriction (ketogenic diet, KD) in a mouse model of GN-to-CKD transition.

Mice were subjected to nephrotoxic serum nephritis (NTS) and treated with KD or standard chow (SC) starting three days post-induction for 18 days. Key parameters, including urinary albumin-to-creatinine ratio (ACR) and kidney histology, were evaluated at multiple time points. Long-term effects on kidney function were assessed five weeks after KD cessation using transdermal glomerular filtration rate (GFR) devices. Kidney fibrosis was evaluated with Sirius Red and Masson’s trichrome staining. Kidneys were further analyzed using NMR spectroscopy (metabolomics), bulk RNA-seq, and droplet-based single-nucleus RNA sequencing (snRNA-seq) in an integrated multi-omics approach. Computational analysis included subclustering of glomerular and stromal cells with evaluation of extracellular matrix (ECM) changes. Finally, unsupervised analyses of dynamical gene expression changes in key cell types were performed to infer pathway and transcription factor activity. 

KD significantly reduced albuminuria and the number of crescents, indicating an improved glomerular phenotype, but showed a trend toward increased tubular injury associated with tubular fat deposition. Mice switched back to SC after the initial KD phase showed reduced kidney fibrosis, preserved kidney function, and a restored tubular phenotype. The final snRNA-seq dataset comprised ~70,000 nuclei. Unsupervised analyses revealed significant adaptations within the stromal compartment. Despite higher inflammation and tubular injury, KD-fed mice exhibited reduced collagen gene expression. A “matrisome” score quantified ECM activity at single-cell level, revealing eight stromal subtypes, including fibroblasts, myofibroblasts, mesangial cells, and vascular smooth muscle cells, which varied in abundance between groups. A KD-enriched fibroblast subtype displayed low ECM expression, while Adamtsl1⁺ myofibroblasts showed the strongest KD-induced reduction in collagen production. Finally, we describe dynamical changes in transcription factor activity and energy metabolism linked to stromal remodeling.

Therapeutic KD was protective in a mouse model of GN, reducing albuminuria, preserving kidney function, and attenuating renal fibrosis. We provide the first single-cell analysis of stromal remodeling in experimental GN in response to KD, revealing metabolic fibroblast reprogramming as a central mechanism underlying its anti-fibrotic effects. These findings establish a mechanistic foundation for translational studies of ketogenic interventions in human kidney disease and may extend to other forms of chronic kidney injury.