Two Blood Proteins Could Let Doctors Stage Alzheimer’s Disease With a Simple Draw

A single vial of blood may soon be able to tell clinicians not just whether Alzheimer’s disease is present in the brain, but how far it has progressed. A new study published in Nature Communications reports that two specific tau protein fragments circulating in blood plasma rise at distinct points along the disease’s biological timeline, and that combining them produces a plasma-based staging system that mirrors the gold-standard brain imaging measurements it was validated against. The work, led by researchers at Samsung Medical Center in Seoul together with collaborators in the United States, Sweden, and Europe, moves the field closer to a future where Alzheimer’s can be biologically staged, and prognostically sized up, without a PET scanner.

The study’s central insight concerns the difference between what the researchers call T1 and T2 biomarkers. Alzheimer’s pathology unfolds in a rough sequence: amyloid-beta plaques accumulate first, followed years later by the spread of tangled filaments of tau protein through the brain’s temporal and cortical regions. Biomarkers such as phosphorylated tau 217 and phosphorylated tau 181, which rise early and track the onset of amyloid pathology, behave like T1 markers in this framework. They are excellent for detecting that the disease process has begun, but they saturate relatively early and say comparatively little about how much fibrillar tau has since aggregated inside neurons. What has been missing is a reliable counterpart for the later, tau-driven phase of the disease.

Two candidate molecules fill that gap. The first is p-tau205, a tau fragment phosphorylated at residue 205, measured across multiple analytical platforms in the study. The second, and the more striking performer, is eMTBR-tau243, a fragment of tau containing the microtubule-binding region at residue 243 that is produced when tau is cleaved inside the brain and released into the fluid surrounding it. Because this endogenous cleavage event is intimately tied to the aggregation of tau into fibrils, the researchers reasoned that circulating levels of the fragment should reflect the later, tangle-rich stage of pathology rather than the early seeding phase.

Measuring a specific proteolytic fragment of tau in blood is technically formidable. Plasma contains vanishingly small quantities of brain-derived proteins, and nonspecific fragments generated during sample handling can masquerade as true biological signals. The team therefore used a dedicated liquid chromatography tandem mass spectrometry platform, LC-MS/MS, to quantify eMTBR-tau243, complementing it with immunoassays for the phosphorylated tau species developed with industry partners including C2N Diagnostics, Roche Diagnostics, and Eli Lilly. Mass spectrometry’s ability to target an exact peptide sequence gives the measurement a specificity that antibody-based assays can struggle to achieve for short tau fragments.

The clinical foundation of the study was a well-characterized Korean cohort in which every participant had undergone amyloid PET imaging and a subset had also received tau PET scans using the 18F-flortaucipir tracer. This imaging anchor matters enormously. Previous blood biomarker studies have often been forced to infer brain pathology from cognitive diagnosis alone, which conflates biology with symptoms. Here, the plasma measurements could be directly calibrated against the actual molecular contents of the living brain, stage by stage.

The results showed a clean temporal separation between biomarker classes. The T1 markers, including p-tau217, began rising during preclinical stages, when amyloid was accumulating but tau tangles and symptoms had not yet taken hold. In contrast, eMTBR-tau243 and p-tau205 increased later in the disease course, a pattern the authors interpret as consistent with the onset and spread of fibrillar tau aggregation. In other words, the two marker families appear to report on different chapters of the same story, and reading them together reveals where along that story a given patient currently sits.

Among all the markers tested, eMTBR-tau243 showed the strongest and most stage-dependent associations with regional tau PET signal and with cognitive performance. It outperformed the T1 biomarkers in predicting both cross-sectional cognition, meaning how well patients were functioning at the time of the blood draw, and longitudinal cognition, meaning how their thinking skills declined over subsequent follow-up. This prognostic power is precisely what early markers tend to lack, since a p-tau217 level that has plateaued cannot easily distinguish a patient whose tau spread will remain limited for years from one facing imminent deterioration.

The study’s most practical contribution is a proposed plasma-based tau staging system that integrates p-tau217 with eMTBR-tau243. When the researchers assigned participants to stages using only these two blood measurements and compared the assignments with staging derived from tau PET imaging, the two systems aligned closely. Critically, the plasma stage also predicted longitudinal cognitive trajectories, separating patients into groups whose memory and thinking declined at measurably different rates. A staging system that requires only a blood draw, yet tracks the brain’s tau burden and forecasts its cognitive consequences, would be transformative for a field in which PET imaging is expensive, available only at specialized centers, and impractical to repeat frequently.

The implications reach well beyond diagnosis. With anti-amyloid immunotherapies now approved and tau-targeted treatments advancing through clinical trials, clinicians increasingly need tools to decide which patients should receive which drugs, when to start them, and how to monitor response. A two-biomarker blood panel could, in principle, identify amyloid-positive patients whose tau pathology is still early, who are likeliest to benefit from disease-modifying therapy, and flag those whose advanced tau burden calls for different management. It could also make clinical trials faster and cheaper by replacing some imaging endpoints with repeated blood sampling, and it could extend access to biological staging to health systems and regions where PET infrastructure will never be routine.

Cautious interpretation is still warranted. The cohort was well characterized but drawn from a single country’s memory clinic network, and mass spectrometry platforms of this kind will need standardization, replication in independent and more diverse populations, and validation in primary care settings before entering everyday practice. Yet the conceptual advance is difficult to overstate. Alzheimer’s medicine has spent decades classifying patients by symptoms, which appear only after the underlying pathology has done much of its damage. The demonstration that two plasma fragments can reconstruct the disease’s internal staging, and predict what the future holds for a given brain, suggests that the era of staging Alzheimer’s by biology rather than by bedside examination, using nothing more invasive than a needle and a tube, may be closer than anyone dared expect.

Subject of Research: Plasma biomarkers for biological staging and prognosis of Alzheimer's disease

Article Title: Plasma eMTBR-tau243 and p-tau205 as T2 biomarkers for biological staging and prognosis in Alzheimer’s disease

Article References: Yim, S., Lee, E. H., Lee, J., Shin, D., Kang, H., Park, H. K., Kim, S., Braunstein, J. B., Verghese, P. B., West, T., Meyer, M. R., Day, T. A., Zetterberg, H., Gonzalez-Ortiz, F., Ashton, N. J., Bittner, T., Domenico, A. D., Kollmorgen, G., Jang, H., … Seo, S. W. (2026). Plasma eMTBR-tau243 and p-tau205 as T2 biomarkers for biological staging and prognosis in Alzheimer’s disease. Nature Communications. https://doi.org/10.1038/s41467-026-78582-x

Image Credits: AI Generated

DOI: 10.1038/s41467-026-78582-x

Keywords: Alzheimer's disease, plasma biomarkers, eMTBR-tau243, p-tau205, p-tau217, tau PET, amyloid PET, LC-MS/MS, biological staging, prognosis, neurodegeneration, blood test