Title : A nested case-control study of plasma-based extracellular vesicles biomarkers of Alzheimer’s disease
Abstract:
Determining a non-invasive, serum-based diagnostic panel for early diagnosis of Alzheimer’s disease (AD) played a significant role in the prevention and treatment of AD. Extracellular vesicles (EVs) can cross the blood–brain barrier and may provide a source for AD biomarkers. This study was carried out based on the AD high-risk population cohort in Zhejiang province enrolled from May 1st to June 30th, 2023. We investigated plasma-derived EV proteins for AD biomarkers from 25 AD patients and 25 healthy controls. The data collected included demographic characteristics such as gender, age, body mass index (BMI) and so on. Cognitive function scale information, such as MMSE, MoCA. Images detect information, such as positron emission tomography (PET), brain computed tomography (CT) or magnetic resonance imaging (MRI) scans. The EV research adherence to the most up-to-date guidelines outlined by the International Society for Extracellular Vesicles in the MISEV guidelines. EV purification was performed using EVlent™, the exosomes obtained by ultracentrifugation were extracted and diluted with 5 μL to 10 μL for sample detection. The results of transmission electron microscopy, nanoparticle tracking and western Blot analysis data were obtained. LC-MS/MS analysis was performed using timsTOF Pro mass spectrometry (Bruker) coupled with Nanoelute (Bruker). Ensemble-based feature selection and machine learning method was used to screen and verify biomarkers, ROC curve analysis was used to determine the performance. In this study, 4 EV biomarkers were found especially relevant for AD, including that PRKCSH (0.84, 0.82-0.90) which was up-regulated and ANGPT1 (0.90, 0.87-0.94), NCKAP1 (0.93, 0.90-0.99), FLJ78516 (0.90, 0.87-0.97) which were down-regulated. Based on the selected biomarkers to construct a risk diagnostic model for AD in the elderly, the AUC of the ROC was 0.87 (95% CI = 0.83–0.91), 0.87 (95% CI = 0.83–0.90), 0.88 (95% CI = 0.85–0.92) by LR, SVM, and RF, respectively. When the cutoff was 0.55, the accuracy determined by Youden’s index, and they were 0.55, 0.85, 0.91, and 0.89, respectively. Our study developed an approach including of EVs protein biomarkers, that could be used to distinguish AD from control candidates, thus providing an additional approach that can be used to complement classical diagnosis methods.
