Plasma tau-species positive neuron-derived extracellular vesicles in progressive supranuclear palsy.

The diagnosis of progressive supranuclear palsy (PSP) remains challenging, particularly in differentiating it from Parkinson's disease (PD) at early stages. Circulating neuron-derived extracellular vesicles (NDEVs) providing a peripheral window into central nervous system pathology may serve as promising biomarkers. A total of 188 participants were recruited from 3 centers. A discovery cohort (40 PSP patients, 36 PD patients, and 31 healthy controls HCs) and a multicenter validation cohort (30 PSP patients, 27 PD patients, and 24 HCs) were established. NDEVs containing total tau, 4R tau, phosphorylated tau (ptau181, ptau217, ptau231 and ptau396) in plasma samples were analyzed using nano-scale flow cytometry. Multivariable logistic regression models were developed in the discovery cohort and strictly validated in the independent cohort using fixed model parameters. In the discovery cohort, the concentrations of tau, 4R tau, ptau181, and ptau217-containing NDEVs in PSP patients were significantly higher than those in HCs and PD patients, while ptau396-containing NDEVs showed elevation exclusively in PSP compared to HCs. (PSP vs. HCs: P<0.001 for tau, P<0.001 for 4R tau, P<0.001 for ptau181, P=0.008 for ptau217, P=0.009 for ptau396; PSP vs. PD: P<0.001 for tau, P<0.001for4R tau, P=0.029 for ptau181, P=0.003 for ptau217). An integrated model incorporating these biomarkers achieved an area under the curve (AUC) of 0.965 (90.0% sensitivity, 96.8% specificity) for distinguishing PSP from HCs, and an AUC of 0.963 (87.5% sensitivity, 94.4% specificity) for distinguishing PSP from PD. In the validation cohort, the concentrations of tau, 4R tau, ptau181, ptau217, and ptau396-containing NDEVs in plasma were significantly higher in PSP patients compared to HCs (PSP vs. HCs: P<0.001 for tau, 4R tau, and ptau217, P=0.03 for ptau181, P=0.017 for ptau396). Similarly, the concentrations of tau, 4R tau, ptau181, and ptau217 were higher in PSP patients than in PD patients (PSP vs. PD: P<0.001 for tau, 4R tau, and ptau217, P=0.025 for ptau181). The integrated model yielded an AUC of 0.971 for distinguishing PSP from HCs and 0.990 for distinguishing PSP from PD. Notably, in early-stage patients, the integrated model achieved an AUC of 0.987 in differentiating early-stage PSP from PD. These findings indicate that plasma tau-species-containing NDEVs are promising biomarkers for PSP, offering high sensitivity and specificity for distinguishing PSP from both HCs and PD, particularly in early disease stages. Further large-scale, longitudinal studies are warranted to fully validate these findings and explore their role in PSP pathophysiology and progression.