Cerebrospinal fluid tau is an established Alzheimer’s disease biomarker, but efforts to develop a blood test for tau have been hampered by exceedingly low concentrations of the protein there. At least, until now, claims a paper published in Alzheimer’s and Dementia on November 9. Using a highly sensitive detection platform combined with antibodies trained to the N-terminus of tau, researchers led by Dominic Walsh at Brigham and Women’s Hospital in Boston developed a plasma tau test that accurately distinguished between AD cases and controls. The findings point to distinct forms of tau in the brain versus blood, and suggest that the minimally invasive blood test could prove useful for diagnosis and screening in clinical studies.
- Researchers developed ultrasensitive tau detection assays that pick up multiple forms of tau in plasma and CSF.
- Tau in plasma is different from tau in CSF.
- People with AD had more of an N-terminal fragment of tau in plasma than controls.
“The results are exciting because they suggest that tau, like Aβ, may be tracked in the blood to provide insight into the CNS processing of tau,” commented Randall Bateman of Washington University in St. Louis, who was not involved in the study (Aug 2018 news).
Most widely used tests for CSF tau employ antibodies specific for the protein’s mid-section. However, tau exists in many forms, including N-terminal fragments that lack part or all of tau’s mid-region (Kanmert et al., 2015; Hu et al., 2018). Researchers also recently reported that in CSF, full-length tau is virtually undetectable, while N-terminal fragments appear to be secreted by neurons, an active process that increases in the presence of Aβ pathology (Mar 2018 news). Attempts to detect plasma tau have yielded unclear results (Sep 2017 news). Walsh and colleagues sought to develop assays to detect N-terminal fragments of tau for both CSF and blood.
Using four different pairs of antibodies specific for different epitopes spanning the length of tau, the researchers developed enzyme-linked immunoabsorbent assays (ELISAs) to detect the N-terminus, mid-region, and full-length (FL) tau. They first tested these assays using CSF samples from the Harvard Brain Aging study (HABS) and University College London. They separated the samples into three groups based on both cognitive and biomarker data: cognitively normal people who tested negative for CSF Aβ42 and tau via standard assays, people with mild cognitive impairment (MCI) who tested positive for these biomarkers, or people diagnosed with AD who were biomarker-positive.
As would be expected, their mid-region tau test performed similarly to the standard INNOTEST assay used to select the cohort. Their two N-terminal assays—dubbed NT1 and NT2—also detected high levels of tau. Both the mid-region and N-terminal assays discriminated between controls and people with MCI-AD or AD, although separation between cases and controls was 25 percent higher with the mid-region test than the N-terminal ones. The researchers next transferred their assays to the ultrasensitive Single Molecule Array platform. Using Simoa, they were able to detect full-length (FL) tau in the CSF as well, which accounted for less than 5 percent of tau there.
What about in the blood? Using blood samples from the same participants who donated CSF, the researchers detected tau with the FL, NT1, and NT2 assays. Surprisingly, FL-tau was the predominant form in blood, but only the NT1 tau assay discriminated between controls and people with AD or MCI-AD. Using a cutoff of 3.07pg/mL, this assay accurately predicted AD or MCI-AD cases 95 percent of the time, and excluded 92 percent of controls.
The researchers validated their findings in samples from the UCSD ARDC cohort. Using the same cutoff, the test was less sensitive and specific in that cohort, predicting 84 percent of cases and ruling out 65 percent of controls. However, on average, the UCSD cohort had lower concentrations of plasma tau than the discovery cohort. Lowering the cutoff to 2.88pg/mL, the NT1 test detected 70 percent of AD cases and excluded nearly 80 percent of controls. In both cohorts, levels of NT1 tau did not significantly differ between people with AD and those with MCI-AD, suggesting plasma NT1 tau does not predict disease progression.
“These are really exciting findings that point to the potential importance of the N-terminal portion of tau,” commented Celeste Karch of Washington University in St. Louis. “It will be interesting moving forward to understand the source of blood tau and the biological processes that lead to N-terminal tau fragments changing as a function of disease.”
Walsh said it is still unclear why there are different forms of tau in blood and brain, or even what the main source of blood tau is. He speculated that peripheral neurons might contribute.
Walsh told Alzforum that the researchers are further validating their tau assays on plasma samples from longitudinal cohorts, including the Religious Orders Study and Rush Memory and Aging Project (ROSMAP), a cohort with longitudinal cognitive, biomarker, and autopsy data.
Walsh envisions the test being used for an initial screening for AD, especially for people who live far afield from medical centers equipped with PET imaging equipment or physicians skilled in CSF extraction. A plasma test might also prove useful in clinical studies. Walsh and colleagues are in talks with the Alzheimer’s Disease Drug Foundation and, after further validation, plan to try the test in prevention trials such as A4.
David Brody, of the Uniformed Services University of the Health Sciences in Bethesda, Maryland, believes the test could prove useful beyond AD. “We are all looking forward to seeing how these blood tests perform in larger cohorts, in longitudinal studies, in therapeutic trials, and in other diseases characterized by tau pathology,” he wrote. “It would be especially interesting if different diseases that can be hard to distinguish clinically (e.g., AD, CTE, FTD) have different patterns of tau fragments present in blood.”—Jessica Shugart
- With Sudden Progress, Blood Aβ Rivals PET at Detecting Amyloid
- Isotope Labeling Links Tau Production to Aβ Burden
- Bloodborne Tau: Foggy Window into the Brain for TBI, Dementia
- Kanmert D, Cantlon A, Muratore CR, Jin M, O'Malley TT, Lee G, Young-Pearse TL, Selkoe DJ, Walsh DM. C-Terminally Truncated Forms of Tau, But Not Full-Length Tau or Its C-Terminal Fragments, Are Released from Neurons Independently of Cell Death. J Neurosci. 2015 Jul 29;35(30):10851-65. PubMed.
- Hu NW, Corbett GT, Moore S, Klyubin I, O'Malley TT, Walsh DM, Livesey FJ, Rowan MJ. Extracellular Forms of Aβ and Tau from iPSC Models of Alzheimer's Disease Disrupt Synaptic Plasticity. Cell Rep. 2018 May 15;23(7):1932-1938. PubMed.
- Chen Z, Mengel D, Keshavan A, Rissman RA, Billinton A, Perkinton M, Percival-Alwyn J, Schultz A, Properzi M, Johnson K, Selkoe DJ, Sperling RA, Patel P, Zetterberg H, Galasko D, Schott JM, Walsh DM. Learnings about the complexity of extracellular tau aid development of a blood-based screen for Alzheimer's disease. Alzheimers Dement. 2019 Mar;15(3):487-496. Epub 2018 Nov 9 PubMed.