Cerebrospinal fluid phospho-tau T217 outperforms T181 as a biomarker for the differential diagnosis of Alzheimer's disease and PET amyloid-positive patient identification

Abstract

Background: Cerebrospinal fluid biomarker profiles characterized by decreased amyloid-beta peptide levels and increased total and phosphorylated tau levels at threonine 181 (pT181) are currently used to discriminate between Alzheimer's disease and other neurodegenerative diseases. However, these changes are not entirely specific to Alzheimer's disease, and it is noteworthy that other phosphorylated isoforms of tau, possibly more specific for the disease process, have been described in the brain parenchyma of patients. The precise detection of these isoforms in biological fluids remains however a challenge.

Methods: In the present study, we used the latest quantitative mass spectrometry approach, which achieves a sensitive detection in cerebrospinal fluid biomarker of two phosphorylated tau isoforms, pT181 and pT217, and first analyzed a cohort of probable Alzheimer's disease patients and patients with other neurological disorders, including tauopathies, and a set of cognitively normal controls. We then checked the validity of our results on a second cohort comprising cognitively normal individuals and patients with mild cognitive impairments and AD stratified in terms of their amyloid status based on PiB-PET imaging methods.

Results: In the first cohort, pT217 but not pT181 differentiated between Alzheimer's disease patients and those with other neurodegenerative diseases and control subjects much more specificity and sensitivity than pT181. T217 phosphorylation was increased by 6.0-fold in patients with Alzheimer's disease whereas T181 phosphorylation was only increased by 1.3-fold, when compared with control subjects. These results were confirmed in the case of a second cohort, in which the pT217 cerebrospinal fluid levels marked out amyloid-positive patients with a sensitivity and a specificity of more than 90% (AUC 0.961; CI 0.874 to 0.995). The pT217 concentrations were also highly correlated with the PiB-PET values (correlation coefficient 0.72; P < 0.001).

Conclusions: Increased cerebrospinal fluid pT217 levels, more than those of pT181, are highly specific biomarkers for detecting both the preclinical and advanced forms of Alzheimer's disease. This finding should greatly improve the diagnosis of Alzheimer's disease, along with the correlations found to exist between pT217 levels and PiB-PET data. It also suggests that pT217 is a promising potential target for therapeutic applications and that a link exists between amyloid and tau pathology.

Keywords: Alzheimer’s disease; Cerebrospinal fluid; Tau proteins.

Fig. 1

CSF tau and p-tau levels in the Montpellier AD cohort. CSF concentration of total tau (E_tau) and pT181 (E_pT181) measured by ELISA (in pg/mL) in the NAD and AD population (a, b). CSF concentration of pT181 and pT217 measured by quantitative mass spectrometry (MS) (in fmol/mL) in the NAD and AD population (c, d) and in the diseases included in the cohort (e, f). ①, ②, and ③ indicate participants with high pT181 level but normal pT217 level. The red arrow indicates participants with mixed dementia having higher MS_p217 level than controls. Differences between NAD and AD populations are statically significant (see Table 1). Abbreviations: AD Alzheimer disease, NAD non-Alzheimer disease, FTLD frontotemporal lobar degeneration, LBD Lewy body dementia, PSP progressive supranuclear palsy, ACIH adult chronic idiopathic hydrocephalus

Fig. 2

MS_pT181 and MS_pT217 plotted against their corresponding unmodified counterparts. CSF concentration of MS_pT181 (a) and MS_pT217 (b) were plotted on their corresponding unmodified counterparts in the Montpellier cohort. Linear regression was computed in the AD and NAD populations (rho Spearman’s correlation coefficient and P value are indicated). Note that for MS_pT217, AD and NAD regression lines have slopes that are significantly different (slope ANCOVA comparison P < 0.001). The sample from the patient with brain metastasis (arrow) is clearly an outliner with low concentration of phosphorylated peptides and high concentration of non-phosphorylated peptides

Fig. 3

Site occupancy of phosphorylation on T181 and T217 in the two cohorts. The percentage of T181 and T217 phosphorylation corresponding to the amount of the phosphorylated peptide divided by the sum of the phosphorylated and non-phosphorylated peptide was plotted in the two cohorts. Significant differences between NAD and AD or amyloid (−) and (+) populations for the two peptides were observed (see also Sup Table 2)

Fig. 4

CSF p-tau in the WUSTL amyloid-positive cohort. CSF concentration of pT181 (E_pT181) was measured by ELISA (in pg/mL) (a), MS_pT181 (b), and MS_pT217 (c) were measured by quantitative MS (in fmol/mL) in amyloid (−) and (+) populations. The ROC curves of the detection of the amyloid (+) patients for E_pT181, MS_pT181, and MS_pT217 were plotted (d) (see SupTable 3 for statistical differences between curves). The MS_pT217 values (e) and percentage of T217 phosphorylation (% MS_pT217, f) were plotted against their corresponding PiB-PET values in the WUSTL cohort composed of amyloid (−) and (+) patients. Linear regression and correlation coefficients are indicated