31 Mar 2020

Researchers stunned the field last summer when they reported that phosphorylated tau turns up in peoples’ plasma decades before they show signs of dementia. The data bore the promise of a blood test for early Alzheimer’s disease—one that is simpler than measuring the notoriously finicky plasma Aβ, and more specific than neurofilament light. This data has now been published in two papers in the March 2 Nature Medicine. A third paper using a different immunoassay, and telling much the same story, is in press and will be presented this week at AAT-AD/PD.

In one of the Nature Medicine papers, researchers led by Oskar Hansson, Lund University, Sweden, reported that plasma phospho-181 tau distinguishes people with AD from healthy controls and from those with other neurodegenerative diseases. Plasma p-tau181 tracked with CSF p-tau181 and with tau PET, and it predicted further cognitive decline among people who were already mildly impaired. In the second paper, scientists led by Adam Boxer at University of California, San Francisco, described how the same assay differentiated people with AD from people with clinically diagnosed or autopsy-confirmed diseases on the FTLD spectrum. In this study plasma p-tau181 was much more specific than plasma neurofilament light, a promising marker for neurodegeneration. Both groups used a p-tau181 immunoassay developed by Jeffery Dage at Eli Lilly and Company, Indianapolis.

“The results of these two studies on plasma p-tau181 may support its use as a screening tool for asymptomatic AD or in people with mild cognitive impairment before diagnostic confirmation of AD by gold-standard tests (CSF or imaging),” wrote Randall Bateman and colleagues, Washington University, St. Louis, in an accompanying News & Views.

In 2018, Dage and colleagues had reported that in the Mayo Clinic Study of Aging, people with AD had more p-tau181 in their blood than did nondemented controls, and that it correlated with their PET measures of amyloid plaques and neurofibrillary tangles (Mielke et al., 2018). This raised many questions. When does p-tau181 start to show up in the blood? Does this happen in other neurodegenerative diseases? Can it predict dementia? As reported at AAIC in 2019, Hansson and colleagues set out to answer these questions using samples from the BIOFINDER longitudinal study (Aug 2019 conference news). 

P-tau Takes the Stage. Though present at 100-fold lower concentration in plasma (left) than CSF (right), p-tau181 in both fluids correlated with Braak stage as determined by tau PET. [Courtesy of Janelidze et al., Nature Medicine.]

Co-first authors Shorena Janelidze, Niklas Mattsson and colleagues tested 589 blood samples from three cohorts—two from the Swedish BioFINDER longitudinal study at Skåne University Hospital and the Hospital of Ängelholm, and one from the Arizona Study of Aging and Neurodegenerative Disorders Brain & Body Donation program at Banner Sun Health Institute.

Plasma and CSF p-tau181 correlated with each other in 515 volunteers from the two BioFINDER cohorts. In the first, which included tau PET, p-tau181 correlated with Braak stage in 174 of 182 people who were scanned (see image above). Breaking this down, 38 people who were cognitively unimpaired but positive for Aβ had plasma p-tau181 levels that indicated Braak I to IV, while 66 who had dementia or mild cognitive impairment due to AD had plasma p-tau181 levels that corelated with Braak stage III-VI.

The 344 people in the second BioFINDER cohort only had amyloid PET scans, as did 129 from cohort 1. In both, plasma p-tau181 correlated tightly with flutemetamol binding in the brain. In fact, plasma p-tau181 better correlated with amyloid than did the plasma Aβ42/Aβ40 ratio determined by Elecsys immunoassay, indicating that this particular phospho form of tau reflects ongoing Aβ pathology.

This fits with the idea that very early in AD, when Aβ toxicity first rears its head, neurons respond by activating kinases that phosphorylate tau at threonine 181. In keeping with this, Janelidze and colleagues found that plasma p-tau181 began to tick up in cohorts 1 and 2 when the standard uptake value ratio for flutemetamol in the neocortex reached 0.72 and 0.70, respectively. Remarkably, these SUVR values are even lower than the 0.74 cutoff for a positive amyloid test. “These findings show that plasma P-tau181 increases early in AD, around the timepoint of Aβ positivity, and support plasma p-tau181 as a possible early marker of AD,” wrote the authors.

Recently, researchers led by Bateman reported that p-tau181 and p-tau217 rise in the CSF of autosomal-dominant AD cases 19 and 21 years before symptom onset, respectively (March 2020 news). 

If p-tau181 rises specifically in response to Aβ toxicity, then it should be able to distinguish AD from primary tauopathies or other non-AD neurodegenerative diseases. To test this hypothesis, the authors compared 14 controls with 52 non-AD dementia cases in the first BioFINDER cohort, and to 47 autopsy-confirmed cases in the Banner cohort that spanned a range of non-AD diseases. They included 28 people with PD or parkinsonism, 17 with PDD, eight with PSP, six DLB, seven CBD, five FTD, three with MCI, two each with SD, VCI, or dementia with an unspecified etiology, and one each with ALS, MS, primary lateral sclerosis, vascular dementia, or vascular parkinsonism. Plasma p-tau181 distinguished AD from all other diseases with a sensitivity and specificity of 92 and 87 percent, respectively, and an area under the curve of 0.92. CSF p-tau181 performed only marginally better.

In Boxer and colleagues’ hands, Dage’s p-tau181 assay was similarly discerning. First author Elisabeth Thijssen used plasma from 404 people in three cohorts: 301 from the UCSF Memory and Aging Center, 61 from the Advancing Research and Treatment for Frontotemporal Lobar Degeneration (ARTFL) study, and 42 from a Phase 1 trial of Lilly’s therapeutic antibody against pyroglutamate Aβ. Her primary analysis focused on the UCSF and ARTFL samples, which included 69 healthy controls, 47 people with MCI, 56 with AD, 39 CBD, 48 PSP, 50 bvFTD, 27 nfvPPA, and 26 svPPA.

Here, too, plasma p-tau181 was higher in AD than any of the FTLD disorders (see image above). An age-adjusted cutoff of 8.7 pg/mL distinguished AD from primary tauopathies with an AUC of 0.84. Blood p-tau181 distinguished logopenic variant PPA, a secondary tauopathy with Alzheimer’s-like proteopathy, from non-fluent variant and semantic variant PPAs, which are primary tauopathies. P-tau181 differentiated AD from other dementias better than did plasma neurofilament light, as NfL was higher in all diseases than in controls.

Autopsy-confirmed diagnoses were available for 82 participants, and they told a similar story. At a mean of 7.5 pg/mL, 15 AD cases had more p-tau181 in their plasma than 67 cases with FTLD, who averaged 2.1 pg/mL. Blood p-tau181 picked out AD versus FTLD with an AUC of 0.88. It distinguished AD even better from FTLD-TDP, with an AUC of 0.95, than from FTLD-tau, where the AUC was 0.86.

FTD mutations barely influenced plasma p-tau181. Among 61 people with MAPT, five with progranulin, and 10 with C9ORF72 variants, plasma levels were similar to each other and to controls. There was one exception: 14 MAPT mutation carriers who had mixed 3R/4R tangle pathology, p-tau181 levels were doubled, at 4.4 pg/mL. This supports the idea that the marker reflects protein pathology specific to the neurofibrillary tangles that form in AD.

As did Janelidze, Thijssen found that plasma p-tau181 correlated with CSF p-tau181, amyloid PET, flortaucipir PET, and with Braak stage. By contrast, plasma NfL correlated with none of these. Blood p-tau181 also correlated with gray-matter shrinkage in the medial temporal lobe, posterior cingulate, and precuneus in people with AD or MCI but not in FTLD. Plasma NfL correlated with gray-matter volume loss across the whole cohort, again in keeping with it being a general marker of neurodegeneration.

In both papers, plasma p-tau181 predicted progression. One BioFINDER cohort tracked 322 people for up to eight years with a mean of 4.9 years. Of these, 62 developed AD and 33 another dementia, including vascular, DLB, PDD, FTD, hydrocephalus, and CADASIL. The former had more plasma tau at baseline than the latter and than people who stayed cognitively sound.

Plasma p-tau181 was also higher at baseline in cognitively unimpaired people or people with MCI who were Aβ-positive and progressed to AD dementia. All told, upward of 1.8 pg/mL plasma p-tau181 came with an 11-fold higher risk of AD, and high p-tau181 meant a faster decline in people with AD and MCI. “P-tau181 is the main predictor of conversion to AD dementia,” said Hansson. “Adding NfL, Aβ42/40, or total tau did not help.” This opens up the possibility of using a simple blood test to predict and track dementia. “Plasma p-tau assays also have the potential to be used as outcome measures to monitor tau-isoform levels in clinical trials of AD,” wrote Bateman.—Tom Fagan

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

News Citations

1. Move Over CSF, P-Tau in Blood Also Tells Us There’s Plaque in the Brain 20 Aug 2019
2. Different CSF Phospho-Taus Match Distinct Changes in Brain Pathology 13 Mar 2020

Paper Citations

1. Mielke MM, Hagen CE, Xu J, Chai X, Vemuri P, Lowe VJ, Airey DC, Knopman DS, Roberts RO, Machulda MM, Jack CR Jr, Petersen RC, Dage JL. Plasma phospho-tau181 increases with Alzheimer's disease clinical severity and is associated with tau- and amyloid-positron emission tomography. Alzheimers Dement. 2018 Aug;14(8):989-997. Epub 2018 Apr 5 PubMed.

External Citations

1. Phase 1

Further Reading

No Available Further Reading