Pittsburgh plasma p-tau217: Classification accuracies for autosomal dominant and sporadic Alzheimer's disease in the community

Abstract

Introduction: Most available phosphorylated tau (p-tau)217 immunoassays have similar performance. It is unclear if this is due to the use of the same antibody (the "ALZpath antibody"). We established and evaluated a novel p-tau217 assay that uses an alternative antibody and benchmarked the results against ALZpath-p-tau217.

Methods: After development and analytical validation of the University of Pittsburgh ("Pitt-p-tau217") method, clinical verification was performed in three independent cohorts (n = 363).

Results: Pitt-p-tau217 demonstrated high between-run stability, linearity, and specificity. Clinically, Pitt-p-tau217 differentiated neuropathologically confirmed PSEN1 mutation carriers from controls with area under the curve (AUC) = 0.94, and amyloid beta (Aβ) positron emission tomography (PET)-positive from Aβ PET-negative cognitively normal older adults with AUC up to 0.84, equivalent to ALZpath-p-tau217 results. Both Pitt-p-tau217 and ALZpath-p-tau217 were slightly elevated in tau PET-positive versus tau PET-negative participants. Between-assay correlations were up to 0.93.

Discussion: The new Pitt-p-tau217 assay exhibits high and reproducible classification accuracies for identifying individuals with biological evidence of Alzheimer's disease, equivalent to the widely used ALZpath-p-tau217.

Highlights: We designed and developed an alternative assay to quantify plasma phosphorylated tau (p-tau)217, aiming to enhance accuracy and enable early detection of Alzheimer's disease (AD). Comprehensive analytical and clinical validation demonstrated that the new p-tau217 assay is a valuable and affordable resource for investigating AD pathophysiology. The new p-tau217 assay showed similar performance to the established ALZpath assay in staging and monitoring early AD.

Keywords: Alzheimer's disease; autosomal dominant Alzheimer's disease; blood‐based biomarkers; phosphorylated tau217; population‐based studies.

FIGURE 1

Biochemical characterization of the p‐tau217–specific rabbit mAb. Each panel displays the results of sandwich enzyme‐linked immunosorbent assays, using either the p‐tau217–specific rabbit mAb or the anti‐tau mouse mAb, clone Tau5, for capture. Detection was performed using a biotinylated anti‐tau mouse mAb, clone Tau12‐the same antibody used for detection in the eventual Pittsburgh‐p‐tau217 assay. The assays tested varying concentrations of recombinant tau441 (non‐p‐tau441 [2N4R] isoform) or its GSK3beta‐phosphorylated variant (p‐tau441), either alone (A and B) or against a set concentration (0.1 µg/mL) of a synthetic peptide (C–E). These peptides, corresponding to tau441 amino acids 210 to 224 (SRTPSLPTPPTREPK), were linked to an N‐terminal cysteine via a peptide bond and phosphorylated at the residues specified in the illustrations in the figure. Panels (A,B) depict the binding profiles of the p‐tau217 and Tau5 antibodies to recombinant non‐p‐tau441 (A) and p‐tau441 (B). Panels (C–E) show the binding profiles of the p‐tau217 and Tau5 antibodies to p‐tau441 in the presence of synthetic peptides phosphorylated exclusively at threonine‐217 (C), peptides phosphorylated jointly at serine/threonine 210, 212, 214, and 217 (D), and non‐phosphorylated peptides (E). mAb, monoclonal antibody; p‐tau, phosphorylated tau.

FIGURE 2

P‐tau217 antibody stains NFTs in the middle temporal gyrus region of human brain tissue. Section of middle temporal gyrus from an AD case neuropathologically diagnosed as Braak stage VI, immunohistochemically stained with p‐tau217 antibody. A dense network of immunoreactive structures is present throughout the cortical laminae illustrated at low magnification in (A). The dashed red box in (A) delineates the area shown at higher magnification in panel (B). The dashed red boxes in (B) delineate areas illustrated at higher magnifications in panels (C, superficial lamina) and (D, deep lamina). The dashed red boxes in (C) and (D) delineate areas illustrated at higher magnifications in panels (E) and (F), respectively. Immunoreactive clusters of dystrophic neurites in neuritic plaques are most numerous in superficial lamina, while immunoreactive NFTs and neuropil threads are numerous throughout the cortical laminae. Scale bars = 500 µm (A), 200 µm (B), 100 µm (C, D), and 50 µm (E, F). AD, Alzheimer's disease; NFT, neurofibrillary tangle; p‐tau, phosphorylated tau.

FIGURE 3

P‐tau217 antibody stains NFTs in hippocampal region of human brain tissue. Section of hippocampus from an AD case neuropathologically diagnosed as Braak stage VI, immunohistochemically stained with p‐tau217 antibody. Numerous immunoreactive NFTs are present in the CA fields, immunoreactive dystrophic neurites in neuritic plaques are abundant in the dentate gyrus stratum moleculare, and immunoreactive neuropil threads are distributed throughout the CA fields and dentate gyrus illustrated in a low magnification composite in panel (A). The dashed red box in (A) delineates the area (field CA1 and dentate stratum moleculare) shown at higher magnification in panel (B). The dashed red boxes in (B) delineate areas illustrated at higher magnifications in panels (C) and (D). The dashed red boxes in (C) and (D) delineate areas illustrated at higher magnifications in panels (E) and (F), respectively. Immunoreactive NFTs embedded in a mesh of neuropil threads are the most prominent features of the CA1 area. Many immunoreactive neurons had a classic flame‐shaped tangle morphology, while others appeared to contain more homogeneous immunoreaction and morphology of intracellular tangles. AD, Alzheimer's disease; CA, Cornu Ammonis; NFT, neurofibrillary tangle; p‐tau, phosphorylated tau.

FIGURE 4

Technical validation of the newly developed novel plasma p‐tau217 assay. Dilution linearity of the p‐tau217 assay in pooled plasma samples (A). For each matrix, the plots show the measured AEB signal in three unique samples with variable levels of the biomarker. Three equal‐volume aliquots of each sample were prepared and measured diluted 2‐, 4‐, or 8‐fold with the assay diluent. Samples were run in duplicate. B, Percentage recovery of the p‐tau217 in pooled plasma sample spiked with assay calibrator of concentrations 5,10, and 20 pg/mL. Samples at each concentration were assayed in duplicate. The dashed line represents the acceptable recovery range between 80% and 120%. C, Assay specificity of the newly developed plasma p‐tau217 assay by competitive inhibition. The plot shows the results of competitive inhibition experiments of p‐tau217 concentrations in a pooled plasma sample that was either untreated or inhibited with the addition of different amounts (range: 0.1–200 ug/mL) of a synthetic peptide bearing phosphorylation at the threonine‐217 site. Data points represent mean concentration of plasma p‐tau217, measured in duplicate in two independent experiment runs, and error bars represent ± standard deviation. AEB, average enzyme per bead; p‐tau, phosphorylated tau.

FIGURE 5

Plasma p‐tau217 performance in distinguishing sporadic AD and ADAD from non‐AD controls. A, Correlations between Pitt and ALZpath p‐tau217 concentration. The correlation coefficient was calculated using Spearman rank‐based correlation. The red line indicates the least square regression line. B, Boxplot distributions of the p‐tau217 levels in ADAD participants with pathogenic PSEN1 mutations, sporadic AD participants, and control. P values were derived from post hoc tests after Kruskal–Wallis, with Bonferroni corrections for multiple comparisons. AD, Alzheimer's disease; ADAD, autosomal dominant Alzheimer's disease; p‐tau, phosphorylated tau.

FIGURE 6

Diagnostic utility of the Pittsburgh p‐tau217 assay in detecting Aβ PET positivity. A, Distribution of plasma p‐tau217 concentrations according to Aβ PET status in the MYHAT‐NI cohort at baseline. P values were determined using the Wilcoxon rank‐sum test. B, Plasma p‐tau217 levels in Aβ negative (CL < 15) to low‐burden Aβ (CL15‐25), and Aβ positive (CL > 25) in the MYHAT‐NI cohort at baseline. P values were determined from post hoc tests after Kruskal–Wallis, with Bonferroni corrections to account for multiple comparisons. C,D, AUC of Pittsburgh p‐tau217 assay measurements for distinguishing Aβ– from Aβ+ participants across the full cohort (C), as well as in cognitively normal participants (D) in the MYHAT‐NI and HCP cohorts. Aβ, amyloid beta; AUC, area under the curve; CL, Centiloid; HCP, Human Connectome Project; MYHAT‐NI, Monongahela‐Youghiogheny Healthy Aging Team‐Neuroimaging; PET, positron emission tomography; p‐tau, phosphorylated tau.

FIGURE 7

Diagnostic utility of the Pittsburgh p‐tau217 assay for tau PET positivity. A, Distribution of plasma p‐tau217 concentrations according to tau PET status in the MYHAT‐NI cohort at baseline. P value was determined using the Wilcoxon rank‐sum test. B, Plasma p‐tau217 levels across AT groups at baseline in the MYHAT‐NI cohort. P values were based on post hoc tests after Kruskal–Wallis, with Bonferroni corrections. C,D, AUC of Pittsburgh p‐tau217 assay measurements for distinguishing tau PET– from tau PET+ participants across the full cohort (C), as well as in cognitively normal participants (D) in the MYHAT‐NI cohort. A, amyloid; AUC, area under the curve; MYHAT‐NI, Monongahela‐Youghiogheny Healthy Aging Team‐Neuroimaging; PET, positron emission tomography; p‐tau, phosphorylated tau; T, tau.