01.
Funding theme · Near-application

Plasma Biomarkers: p-tau217 & Aβ42/40

Markers in the blood that reveal Alzheimer-related changes — long before memory begins to fade.

The question

Can Alzheimer pathology be reliably detected in blood while people are still cognitively unimpaired — opening a window for prevention before symptoms emerge?

State of the science

p-tau217 — the dynamic early marker

A tau protein that responds as soon as the first amyloid deposits form.

Tau phosphorylated at threonine 217 (p-tau217) is not a measure of existing tau fibrils but a dynamic readout of the early amyloid-β-triggered hyperphosphorylation of soluble tau. It begins rising in the preclinical phase, when amyloid is accumulating but cognition is intact, and later tracks the tau-PET signal closely — acting simultaneously as an early state marker of amyloid and a staging marker of downstream tau pathology.

The Aβ42/40 ratio and the small-dynamic-range problem

Shows how much of the aggregation-prone amyloid is already being trapped in the brain.

The ratio of amyloid-β 42 to 40 reflects the relative depletion of aggregation-prone Aβ42 as it is incorporated into plaques, correcting for individual differences in total production and pre-analytical loss. Its weakness is a small dynamic range: the difference between amyloid-positive and -negative plasma is only about 8–15%, so even ~10% measurement imprecision misclassifies a large share of cases.

%p-tau217 — the phosphorylated-to-non-phosphorylated ratio

A clever normalization that makes the signal more robust against noise.

Expressing p-tau217 as the ratio of phosphorylated to non-phosphorylated tau-217 (%p-tau217) normalizes for individual differences in total tau production and reduces confounding. This ratio shows the disease's largest fold-changes — roughly six-fold for CSF T217 phosphorylation, and about 280% of control means in plasma — far greater separation than Aβ42/40. Caveat: because the non-phospho peptide is shared with total tau, rapid neurodegeneration or acute neuronal injury can skew the ratio.

GFAP & NfL — astrocytic reactivity and neurodegeneration

Two additional blood markers capturing inflammation and nerve-cell damage.

GFAP (glial fibrillary acidic protein) is released during reactive astrogliosis, rises early, and shows relative specificity for the preclinical Alzheimer phase among neurodegenerative diseases. NfL (neurofilament light chain) is a sensitive but non-specific marker of neuroaxonal injury, elevated across many neurodegenerative conditions. Both add orthogonal biology to amyloid and tau.

The A/T/N biological staging concept

A framework that combines the markers into a biological picture of the disease.

The field conceptually organizes markers by Amyloid (A), Tau (T) and Neurodegeneration (N): core amyloid via Aβ42/40, tau via p-tau217, neurodegeneration via NfL. GFAP is increasingly added as a non-core marker of neuroinflammation that does not map cleanly onto a single A/T/N node.

Pre-analytics & confounders

Factors such as kidney function or body weight can shift the readings.

Reduced kidney function, BMI, age and comorbidities measurably shift marker concentrations; falling eGFR can substantially raise the optimal p-tau217 cutoff for amyloid positivity. Add pre-analytical handling, lot-to-lot variability and the absence of cross-platform standardization — cutoffs from different assays are currently not interchangeable.

What we fund

We support research that moves these markers from diagnostics in already-symptomatic patients toward presymptomatic detection in cognitively unimpaired and demographically diverse populations: validating cutoffs in low-prevalence settings, assay standardization and reference materials, systematic confounder correction, and two-step screening algorithms (blood test as triage, followed by confirmation).

Evidence & limits

Near-application but unsettled. p-tau217 is the most mature blood marker, yet screening healthy people in low-prevalence settings yields many false positives. A positive marker indicates existing pathology — not a diagnosis, and not certainty of future disease. Disclosing risk to symptom-free people is ethically sensitive and should involve counseling. Many validation cohorts lack diversity; generalizability remains an open question.

back