Lipidomics reveals potential biomarkers and pathophysiological insights in the progression of diabetic kidney disease

• Published in: Metabolism open Vol. 25; p. 100354
• Main Authors: Guo, Xiaozhen, Zhang, Zixuan, Li, Cuina, Li, Xueling, Cao, Yutang, Wang, Yangyang, Li, Jiaqi, Wang, Yibin, Wang, Kanglong, Liu, Yameng, Xie, Cen, Zhong, Yifei
• Format: Journal Article
• Language: English
• Published: England Elsevier Inc 01.03.2025; Elsevier
• Subjects: Biomarker; Diabetes kidney disease; Diabetes mellitus; Endocrinology and Metabolism; Lipidomics; Lysophosphatidylethanolamines; Original Research Paper; FFA; PS; ALT; DM; AUC; HDL-C; LASSO; UHPLC; DKD; CKD-EPI; SCr; ROC; PLA2; Lipidomics; PKC; Aclycar; Ctrl; CVD; Cer; MAG; RF; SVM-RFE; LacCer; AMPK; LPA; LPC; ESRD; LPE; Biomarker; LDL-C; BUN; SM; UACR; HbA1c; eGFR; AST; CE; OR; Diabetes mellitus; CI; Q-TOF mass spectrometer; PLS-DA; PP2A; DAG; TC; DKD-A; PA; PC; DKD-E; PE; Lysophosphatidylethanolamines; PG; Diabetes kidney disease; PI; TAG; Cardiovascular disease; Random forest; Chronic Kidney Disease Epidemiology Collaboration; Advanced stage of DKD; Lysophosphatidylcholine; Phosphatidylinositol; Total cholesterol; Ceramide; Early stage of DKD; Lysophosphatidylethanolamine; Urinary albumin-to-creatinine ratio; High-density lipoprotein cholesterol; Free fatty acid; Quadrupole time-of-flight mass spectrometer; Estimated glomerular filtration rate; Low-density lipoprotein cholesterol; Serum creatinine; Lysophosphatidic acid; Partial least squares-discriminant analysis; Phosphatidylcholine; Sphingomyelin; Protein phosphatase 2A; Ultra performance liquid chromatography; End-stage renal disease; Lactosylceramide; Protein kinase C; Diacylglycerol; Triacylglycerol; Aspartate transaminase; Phosphatidic acid; Cholesterol ester; Phospholipase A2; Aclycarnitine; Support vector machine-recursive feature elimination; Least absolute shrinkage and selection operator; Glycosylated hemoglobin, type A1c; AMP-activated protein kinase; Receiver operating characteristic curve; Area under ROC curve; Control; Phosphatidylglycerol; Phosphatidylserine; Diabetic kidney disease; Phosphatidylethanolamine; Confidence interval; Blood urea nitrogen; Odds ratio; Alanine aminotransferase; Monoacylglycerol
• ISSN: 2589-9368; 2589-9368
• DOI: 10.1016/j.metop.2025.100354

Diabetic kidney disease (DKD) is the leading cause of end-stage renal disease, affecting over 30 % of diabetes mellitus (DM) patients. Early detection of DKD in DM patients can enable timely preventive therapies, and potentially delay disease progression. Since the kidney relies on fatty acid oxidation for energy, dysregulated lipid metabolism has been implicated in proximal tubular cell damage and DKD pathogenesis. This study aimed to identify lipid alterations during DKD development and potential biomarkers differentiating DKD from DM. lipidomics analysis was performed on serum collected from 55 patients with DM, 21 with early DKD stage and 32 with advanced DKD, and 22 healthy subjects. Associations between lipids and DKD risk were evaluated by logistic regression. Lipid profiling revealed elevated levels of certain lysophosphatidylethanolamines (LPEs), phosphatidylethanolamines (PEs), ceramides (Cers), and diacylglycerols (DAGs) in the DM-DKD transition, while most LPEs, lysophosphatidylcholines (LPCs), along with several monoacylglycerol (MAG) and triacylglycerols (TAGs), increased further from DKD-E to DKD-A. Logistic regression indicated positive associations between LPCs, LPEs, PEs, and DAGs with DKD risk, with most LPEs correlating significantly with urinary albumin-to-creatinine ratio (UACR) and inversely with estimated glomerular filtration rate (eGFR). A machine-learning-derived biomarker panel, Lipid9, consisting of LPC(18:2), LPC(20:5), LPE (16:0), LPE (18:0), LPE (18:1), LPE (24:0), PE (34:1), PE (34:2), and PE (36:2), accurately distinguished DKD (AUC: 0.78, 95 % CI 0.68–0.86) from DM. Incorporating two clinical indexes, serum creatinine and blood urea nitrogen, the Lipid9-SCB model further improved DKD detection (AUC: 0.83, 95 % CI 0.75–0.90) from DM, and was notably more sensitive for identifying DKD-E (AUC: 0.79, 95 % CI 0.67–0.91). This study deciphers the lipid signature in DKD progression, and suggests the Lipid9-SCB panel as a promising tool for early DKD detection in DM patients.