Abstract

Ischemia–reperfusion injury (IRI) limits graft function and long-term outcomes after kidney transplantation. Proximal tubular (PT) cells are highly mitochondria-rich and metabolically active, making them vulnerable to ischemic and oxidative stress. However, the molecular mechanisms linking mitochondrial dysfunction to graft function remain incompletely understood. We integrated single-nucleus RNA sequencing, bioinformatics analyses, and experimental validation to identify mitochondria-associated genes in PT cells related to graft injury and vulnerability in the transplant setting. Mitochondria-related differentially expressed genes (Mito-DEGs) were used to construct machine learning models for delayed graft function (DGF) risk stratification. The candidate gene HAO2 was further evaluated by immunohistochemistry (IHC) in kidney transplant biopsies, a murine IRI model, and H₂O₂-treated HK-2 cells. Human biopsy samples were obtained at or prior to reperfusion and may not fully capture the extent of post-transplant IRI. Both overexpression and siRNA-mediated knockdown were performed to assess its function. Cell viability, apoptosis, and mitochondrial function were assessed using standard assays. PT cells from acute kidney injury samples exhibited mitochondrial dysfunction and metabolic impairment. Donor kidney clustering suggested heterogeneity in DGF risk associated with mitochondrial bioenergetic capacity. An eight-gene Mito-DEG signature demonstrated moderate performance in stratifying DGF risk. IHC analysis demonstrated that HAO2 expression was reduced in DGF recipient biopsies. Consistently, HAO2 was downregulated in murine IRI kidneys and injured HK-2 cells. Functionally, HAO2 overexpression alleviated oxidative stress, apoptosis, and mitochondrial dysfunction, whereas HAO2 knockdown exerted opposite effects and further aggravated H₂O₂-induced injury.