Plasma tau levels in Alzheimer's disease

Compared with patients with mild cognitive impairment (MCI) or controls, P <0.001. b Compared with patients with MCI, P = 0.001. c Compared with controls, P = 0.02. AD, Alzheimer’s disease;


Introduction
Eff orts to fi nd reliable blood biomarkers for Alzheimer's disease (AD) in a highly warranted clinical laboratory test have met with little success. Th ere is no clear change in plasma β-amyloid in AD, and assays for the axonal injury marker tau have been hampered by a lack of analytical sensitivity for accurate measurement in blood samples [1]. Here, the results of a novel ultra-sensitive assay for tau in peripheral blood are reported.

Findings
We have developed an ultra-sensitive assay for tau in peripheral blood [2]. In brief, the assay is based on digital array technology [3] and uses the Tau5 monoclonal antibody for capture (Covance, Princeton, NJ, USA) and HT7 and BT2 monoclonal antibodies for detection (Pierce, now part of Th ermo Fisher Scientifi c Inc., Waltham, MA, USA). Th is combination reacts with both normal and phosphorylated tau with epitopes in the mid-region of the molecule, making the assay sensitive to all known tau isoforms. Th e calibrator was recombinant tau 381 (EMD Millipore Corporation, Billerica, MA, USA). To minimize matrix eff ects, all samples were diluted 1:4 in phosphatebuff ered saline with 2% bovine serum albumin diluent prior to assay. Th e limit of detection of the assay, which requires 30 μL of plasma, is 0.02 pg/mL [2], which is more than 1,000-fold more sensitive than conventional immunoassays.
Here, we assess the association of plasma tau levels with AD in a cross-sectional study of 54 patients with AD dementia [4], 75 patients with mild cognitive impairment (MCI) [5], and 25 cognitively normal controls ( Table 1). All participants were recruited at the specialized memory clinic at Skåne University Hospital in Malmö, Sweden, and underwent extensive clinical evalu ation, including cerebrospinal fl uid (CSF) sampling by lumbar puncture, in addition to venipuncture and collec tion of blood in ethylenediaminetetraacetic acid (EDTA) tubes for plasma preparation by centrifugation within 15 minutes from sampling. Plasma samples were ali quoted into cryo tubes and stored at −80°C pending analysis, which was performed on one occasion by using one batch of reagents with an average coeffi cient of variation of 9.7% for triplicate measurements of each sample. Th e patients with MCI were cognitively stable for an average of 101 months (n = 36) or developed AD dementia (n = 35) or other types of dementias -vascular dementia (n = 3) and semantic dementia (n = 1) -during follow-up. Th e study was approved by the regional ethics committee at Lund University and complied with the Declaration of Helsinki. Informed consent was obtained from all study participants.
Tau levels in plasma were signifi cantly higher in AD patients compared with both controls and MCI patients (Figure 1a). MCI patients who developed AD during follow-up had tau levels similar to those of patients with stable MCI and cognitively normal controls (Figure 1b). Th ere was no correlation between tau levels in plasma and CSF in any diagnostic group (Figure 1c).
Th e results of this study have several important implications. First, plasma tau levels are elevated in AD but with overlapping ranges across diagnostic groups.  Th is overlap diminishes the utility of plasma tau as a diagnostic test. However, further studies are needed to evaluate plasma tau as a fi rst-in-line screening tool (for example, in the primary care setting and perhaps together with other markers in a biomarker panel). Second, normal plasma tau levels in the MCI stage of AD suggest that plasma tau is a late marker, requiring substantial axonal injury before increasing to abnormal levels. In this context, other neurodegenerative diseases (for example, Creutzfeldt-Jakob disease) as well as acute conditions (for example, stroke and brain trauma) should be tested. Th ird, the lack of correlation of tau levels in plasma and CSF suggests that steady-state concentrations of tau in these two body fl uids are diff erentially regulated. In our earlier study of patients with hypoxic brain injury following cardiac arrest, tau was rapidly (within 24 hours) cleared from blood in patients with good neurological outcome [2], indicating potent clearance mechanisms for this marker in the bloodstream. Th is may obscure any correlation with CSF tau levels, which stay elevated for weeks following an acute neurological insult [6].

Competing interests
DW and JR, who are employees of Quanterix Corporation (Lexington, MA, USA), and HZ and KB are listed as inventors on a US patent application for plasma tau as a brain injury marker. The other authors declare that they have no competing interests. This article is part of a series on Peripheral Biomarkers, edited by Douglas Galasko. Other articles in this series can be found at http://alzres.com/series/biomarkers