Comprehensive machine learning and experimental verification reveal the mechanism of action of autophagy-related genes FIZ1 and FBXO21 in acute kidney injury.

BACKGROUND: Acute kidney injury (AKI) is a serious disease with a high incidence and easy induction. The search for innovative biomarkers and treatment methods is of great significance for improving the prognosis of patients. Autophagy is closely related to the occurrence and development of AKI. This study aims to explore the role of autophagy-related genes (ARGs) as potential biomarkers and therapeutic targets in AKI. METHODS: In this study, the gene microarray data of the GEO dataset were used to explore the molecular profile of AKI, and three machine learning algorithms were used to screen autophagy-related feature genes. To further validate the reliability of the screening results, we constructed a cisplatin-induced AKI rat model to validate potential biomarkers of machine learning screening. RESULTS: Machine learning analysis identified 17 differentially expressed ARGs and selected the core genes FIZ1 and FBXO21, with area under curve (AUC) values both exceeding 0.7 (95% CI [0.706-0.899]). Immune analysis revealed that the number of Mast cells resting significantly decreased in AKI samples compared to normal samples (P < 0.05). Electron microscopy observations of the cisplatin-induced AKI rat model indicated thickening of the basement membrane, fusion of foot processes, and swelling and rupture of mitochondria in the model group, suggesting a correlation between AKI and mitochondrial autophagy; Western blot results indicated a significant increase in the expression of FIZ1 and a significant decrease in FBXO21 in the AKI group (P < 0.01). The results of IHC staining were also consistent with those of Western blot results. CONCLUSION: This study highlights the significant role of ARGs in AKI and identifies FIZ1 and FBXO21 as promising biomarkers with high diagnostic potential, offering new insights into the molecular mechanisms underlying AKI.