Abstract
KEY POINTS: Gene expression profiling of tubuloids derived from human fetal kidneys, adult nephrectomies, and kidney organoids reveals common and unique signatures. Induced pluripotent stem cell–derived kidney organoids express higher levels of proximal tubule markers than tubuloids derived from the adult kidney. Tubuloids provide a new capacity to model chronic stressors and repetitive kidney injury. BACKGROUND: Epithelial kidney organoids (tubuloids) made from kidney biopsies, urine, or induced pluripotent stem cell (iPSC)–derived kidney organoids offer new opportunities in experimental and clinical nephrology. Yet, we have limited knowledge of how tubuloid models differ from each other, from iPSC-derived kidney organoids and from the human kidney. New insight is required to guide model selection for studies in kidney physiology and disease. METHODS: Tubuloids were generated from adult nephrectomy samples (adult tubuloids n=3), iPSC-derived kidney organoids (iTubuloids n=3), and for the first time, from human fetal kidneys (fetal tubuloids n=3). Kidney organoid and tubuloid models were compared with each other and to adult human kidney using bulk RNA sequencing. As a proof of principle study, the potential to investigate the tubular response to repeated hypoxic insults was examined in iTubuloids. RESULTS: Expression signatures of proximal and distal tubules were stronger in adult kidneys than any organoid or tubuloid model. iPSC-derived kidney organoids expressed proximal tubule markers at higher levels than any tubuloid culture, despite adult tubuloids being derived from mature kidneys. Collecting duct signatures were enriched in adult and fetal tubuloids. Adult tubuloids showed stronger signatures of aging and inflammation, while fetal tubuloids had enhanced ureteric tip progenitor signatures. Over 80 genes linked to inherited disorders were expressed in all tubuloid cultures, while an additional 54 were expressed at higher levels in adult, fetal, or iTubuloids. iTubuloids subject to a single hypoxic injury effectively recovered by the end of the passage, while cultures exposed to hypoxia over three passages expressed markers of maladaptive repair. CONCLUSIONS: This study provides new transcriptome-wide reference data to aid in the selection and optimization of disease modeling for the human kidney. It defines common and unique opportunities to model inherited disorders in adult, fetal and iTubuloid models and illustrates new potential to model repetitive injury in long-lived tubuloid cultures.