Blood plasma phosphorylated-tau isoforms track CNS change in Alzheimer's disease

Abstract

Highly sensitive and specific plasma biomarkers for Alzheimer's disease (AD) have the potential to improve diagnostic accuracy in the clinic and facilitate research studies including enrollment in prevention and treatment trials. We recently reported CSF tau hyperphosphorylation, especially on T217, is an accurate predictor of β-amyloidosis at asymptomatic and symptomatic stages. In the current study, we determine by mass spectrometry the potential utility of plasma p-tau isoforms to detect AD pathology and investigate CSF and plasma tau isoforms' profile relationships. Plasma tau was truncated as previously described in CSF. CSF and plasma measures of p-tau-217 and p-tau-181 were correlated. No correlation was found between CSF and plasma on total-tau levels and pS202 measures. We found p-tau-217 and p-tau-181 were highly specific for amyloid plaque pathology in the discovery cohort (n = 36, AUROC = 0.99 and 0.98 respectively). In the validation cohort (n = 92), p-tau-217 measures were still specific to amyloid status (AUROC = 0.92), and p-tau-181 measures were less specific (AUROC = 0.75).

Graphical abstract

Figure 1.

Plasma tau truncation profile after chemical extraction and IP in the discovery cohort. Left: Plasma tau peptides concentration profiles obtained from the 36 individuals. Each line corresponds to the peptide profile from one participant. Right: Dots represents averaged tau peptides normalized concentration obtained from the overall cohort. Normalized concentration for each peptide is relative to the sum of the peptides concentrations measured in each participant. Bars represent SD. Label a indicates a decrease of 2N and 1N+2N peptide abundance consistent with 5/5/1 0N/1N/2N contribution in plasma tau. Label b indicates a decrease consistent with the presence of around 10% of phosphorylation on position 181. Phosphorylation on T181 induces a trypsin missed cleavage between residues 180 and –181. This contributes to a decrease of 175–180 and 181–190 peptides abundance proportional to the extent of phosphorylation on T181. Label c indicates a decrease consistent with tau truncation between residues 221 and 226. Cicognola et al. (2019) have reported CSF tau main cleavage occurring at residue 224. Label d indicates a decrease consistent with progressive C terminus degradation of plasma tau from residue 224 to microtubule binding region upstream region.

Figure 2.

Plasma tau and p-tau changes across groups in Tau SILK discovery cohort. (A) Measures of p-tau-217/T217 ratios in plasma and CSF are highly correlated in both entire cohort and CSF p-tau-217 positive subgroup. Spearman correlations and associated P value are shown. (B–E) Separation between amyloid-negative and -positive groups are calculated using AUROC. (B and C) Consistent with CSF measurement, plasma p-tau-217/T217 ratio and p-tau-217 level distinguish amyloid-negative from amyloid-positive groups regardless of the cognitive status. Amyloid-negative groups with high CSF p-tau-217 were also separated from other amyloid-negative groups. (D) Plasma tau level is not a biomarker for amyloid status and AD dementia. (E and F) Plasma p-tau-181/T181 ratio and p-tau-181 level increase in amyloid-positive groups. YNC, young normal controls; AMC, aged-matched controls.

Figure S1.

Spike and recovery experiment assessing assay performance using 4 ml of plasma volume. Plasma pool or 5% solution of recombinant HSA were spiked with increased volumes of AD or non-AD CSF pools. Plasma was extracted using the 4-ml protocol described in the Materials and methods. HSA samples were only immuno-purified (IP) before digestion and analysis.

Figure 3.

Plasma tau and p-tau changes across groups in Aβ SILK validation cohort. (A) Phosphorylation occupancies on T217 in plasma and CSF correlate. Spearman correlations and associated P value are shown. (B and C) As found in the discovery cohort, plasma p-tau-217/T217 ratio and p-tau-217 level distinguish amyloid-negative from amyloid-positive groups regardless of the cognitive status. Amyloid-negative with high CSF p-tau-217 were also separated from other amyloid-negative groups. Separations between groups are calculated using AUROC. (D) Plasma tau level is not a biomarker for amyloid status and AD dementia. (E and F) Plasma p-tau-181/T181 ratio and p-tau-181 level increase in amyloid-positive groups but are less accurate than p-tau-217 measures to detect abnormal tau phosphorylation. (G and H) Receiver operating characteristic curves discriminating amyloid-positive from amyloid participants using CSF and plasma tau measures. Corresponding AUCs are summarized in Table 2. AUC comparison between p-tau-217 and p-tau-181 measures is described in Table S3. (I) Plasma p-tau-217/T217 ratio associates with Aβ42/40 ratio measured in CSF.