A blood test that tracks protein shape could help detect Alzheimer’s earlier, research suggests, by spotting disease-linked changes missed by current methods.
Current tests usually measure levels of two proteins, amyloid beta and phosphorylated tau, in the blood or spinal fluid, but these markers may not fully capture earlier biological changes linked to disease progression.
The new approach instead looks at how proteins are folded in the bloodstream.
John Yates, a professor at Scripps Research and senior author of the study, said: “Many neurodegenerative diseases are driven by changes in protein structure.
“The question was, are there structural changes in specific proteins that might be useful as predictive markers?”
He added: “Detecting markers of Alzheimer’s early is absolutely critical to developing effective therapeutics. If treatment can start before significant damage has been done, it may be possible to better preserve long-term memory.”
Proteins need to be folded into the right shape to work properly, and in Alzheimer’s that folding system, known as proteostasis, begins to break down.
Scientists at Scripps Research in the US analysed plasma samples from 520 people across three groups: cognitively normal adults, people with mild cognitive impairment, or MCI, which is often considered a precursor to Alzheimer’s, and patients diagnosed with the disease.
Using mass spectrometry, a technique that measures molecules in detail, to assess how exposed or buried specific protein sites were, and machine learning to identify patterns, the team found a clear trend. As Alzheimer’s progressed, certain blood proteins became less structurally “open”.
These structural changes gave a stronger signal for distinguishing disease stage than measuring protein levels alone.
Three proteins stood out from hundreds of candidates: C1QA, which is involved in immune signalling; clusterin, which is linked to protein folding and amyloid clearance; and apolipoprotein B, which helps carry fats in the bloodstream and plays a part in blood vessel health.
Casimir Bamberger, a senior scientist at Scripps Research, said: “The correlation was amazing. It was very surprising to find three lysine sites on three different proteins that correlate so highly with disease state.”
Structural differences at specific sites within these three proteins allowed the researchers to classify people as cognitively normal, MCI or Alzheimer’s with about 83 per cent overall accuracy.
In two-way comparisons, such as separating healthy people from those with MCI, accuracy exceeded 93 per cent.
The three-marker model performed consistently across independent groups of participants and remained accurate when tested on follow-up samples months later.
In repeat blood samples collected months apart, the panel classified disease status with about 86 per cent accuracy and reflected changes in diagnosis over time.
The structural score also closely matched cognitive test results and showed a more moderate link with MRI measures of brain shrinkage.
Together, the findings suggest that measuring protein structure in blood could provide complementary information to existing amyloid and tau tests.
By targeting structural changes linked to underlying disease biology, the approach could help distinguish disease stages, track progression and measure whether treatments are working.
The test will need larger validation studies with longer follow-up before it is ready for clinical use.
The researchers are also exploring whether the same structural profiling approach could be used in other conditions, including Parkinson’s disease and cancer.
