Brain lipidomics for region-specific biomarker discovery in neurodegenerative diseases

Background: Alzheimer's disease (AD), Parkinson's disease (PD), and Huntington's disease (HD) are progressive neurodegenerative diseases (NDs) characterized by chronic neuronal loss. The lack of effective treatments highlights the urgent need for reliable lipid biomarkers to enable diagnosis and monitor disease progression. Previous lipidomic investigations of altered lipid metabolism have focused on a single disease type, limiting cross-disease comparisons. Methods: = 24). All groups include three males and three females, with brain tissues from four cortical regions sacrificed from each individual. Results: A total of 243 lipid molecular species spanning five major classes were annotated, revealing distinct disease-specific lipidomic profiles that differentiated HV from the AD, HD, and PD groups via multivariate analysis. Sphingomyelins and oxidized phosphatidylserine [PS (16:1/24:0;O1)] were significantly increased, while lysophosphatidylcholines (LPC 18:2, LPC 17:2) were decreased in the AD group relative to HV. HD exhibited elevated PS (O-17:0/22:6) and ω-6 fatty acid esterified cholesteryl esters (CE 18:2, CE 20:4), alongside decreased essential neuronal lipids such as phosphatidylinositols (PI). The PD lipidome alterations closely resembled those of HD, indicating partially overlapping disruptions in brain lipid metabolism. Receiver operating characteristic analysis identified PS (16:1/24:0;O1), PS (O-17:0/22:6), and PI (18:1/18:1) as potential discriminatory biomarkers with strong diagnostic performance. Regional heatmap analysis revealed significant lipid perturbations were observed in the parietal and occipital lobes across all NDs. Conclusion: This study provides a comprehensive overview of disease- and region-specific alterations in the brain lipidome of AD, HD, and PD. The identified lipid species-PS (16:1/24:0;O1), PS (O-17:0/22:6), and PI (18:1/18:1)-may serve as promising candidate biomarkers for NDs diagnosis and warrant further mechanistic and longitudinal validation with large data set.