Current kidney function parameters overestimate kidney tissue repair in reversible experimental kidney disease.

Although underlying mechanisms and the clinical course of kidney disease progression are well described, less is known about potential disease reversibility. Therefore, to analyze kidney recovery, we adapted a commonly used murine chronic kidney disease (CKD) model of 2,8- dihydroxyadenine (2,8-DHA) crystal-induced nephropathy to study disease recovery and efficacy of disease-modifying interventions. The recovery phase after CKD was characterized by improved kidney function after two weeks which remained stable thereafter. By contrast, even after eight weeks recovery, tubular injury and inflammation were only partially reduced, and fibrosis persisted. Deep-learning-based histologic analysis of 8,604 glomeruli and 596,614 tubular cross sections revealed numerous tubules had undergone either prominent dilation or complete atrophy, leading to atubular glomeruli and irreversible nephron loss. We confirmed these findings in a second CKD model, reversible unilateral ureteral obstruction, in which a rapid improvement of glomerular filtration rate during recovery also did not reflect the permanent histologic kidney injury. In 2,8-DHA nephropathy, increased drinking volume was highly effective in disease prevention. However, in therapeutic approaches, high fluid intake was only effective in moderate but not severe CKD and established tissue injury was again poorly reflective of kidney function parameters. The injury was particularly localized in the medulla, which is often not analyzed. Thus, recovery after crystal- or obstruction-induced CKD is characterized by ongoing tissue injury, fibrosis, and nephron loss, but not reflected by standard measures of kidney function. Hence, our data might aid in designing kidney recovery studies and suggest the need for biomarkers specifically monitoring intra-kidney tissue injury.