Clearing amyloid plaques is one goal of Alzheimer’s treatment. Understanding what happens to a patient as those plaques disappear is a more complex measurement problem. Changes in amyloid, inflammation, and neuronal function may follow different paths, leaving researchers with a question: which blood proteins capture which aspects of the response?
A study published in The Lancet Neurology explores that question through plasma samples from patients receiving lecanemab in routine clinical care. Researchers at Washington University School of Medicine in St. Louis tracked a broad protein panel to examine how those signals changed over treatment.
“Anti-amyloid therapies are now part of routine care, but until now we have had a very narrow window into what is actually happening biologically in the patients receiving them,” notes Carlos Cruchaga, PhD, professor of psychiatry and co-senior author of the study.
Widening the View of Treatment Response
The team used Alamar’s NULISAseq CNS panel to measure 130 proteins spanning amyloid and tau pathology, inflammation, neurodegeneration, and synaptic function. Each analysis required 25 µL of plasma, allowing researchers to examine these processes together in a small sample.
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
The treatment cohort included 197 patients with early symptomatic Alzheimer’s disease. Blood draws and clinical assessments took place approximately every six months, with follow-up extending to two years. Repeated samples from 138 patients supported longitudinal analyses. The wider dataset, including untreated comparison groups, contained 2,385 samples from 1,967 participants that passed quality control.
Following protein signals across samples also required control of analytical variation. The researchers applied procedures to normalize plate and signal variability and used statistical adjustments for batch effects.
When Biomarkers Tell Different Stories
Across the panel, 34 proteins changed significantly with infusion number. Their trajectories resisted a single interpretation of improvement. Some, including phosphorylated tau and glial fibrillary acidic protein, moved toward levels in amyloid-negative participants without cognitive impairment. Amyloid β peptides moved toward and then beyond those reference levels. Other proteins moved further away, while many showed no significant change.
The distinction became more consequential when the team examined clinical outcomes. Proteins associated with amyloid clearance differed largely from those associated with subsequent cognitive decline. A marker that helps track plaque removal may therefore answer a different question from one that helps characterize a patient’s likely cognitive course.
Analyte selection added another layer. Brain-derived tau markers showed stronger differences between patient and control groups than their corresponding peripheral forms. Measuring “tau” alone does not capture that distinction; the molecular form under investigation shapes the information the assay provides.
“Our findings show that Alzheimer’s disease and neurodegeneration biomarkers have unique trajectories that capture specific biological process associated with treatment response,” Cruchaga added in the release.
From Protein Patterns to Useful Tests
These patterns identify candidates for further investigation. They do not yet establish a blood test that determines whether lecanemab benefits an individual patient. The treatment cohort came from one clinic and consisted mostly of White, non-Hispanic patients. The untreated reference groups lacked longitudinal samples for comparison, limiting conclusions about treatment effects.
The next challenge is to validate which measurements can support which decisions. For analytical scientists, the study makes that distinction tangible: a broad panel can uncover signals that move in different directions, but each candidate biomarker still needs a defined purpose and evidence that its measurement supports that purpose.
This article was produced under Separation Science's AI Editorial Guidelines.
