Jevgenij Raskatov kept finding his own experiments cited against him. His lab at UC Santa Cruz had shown, back in 2020, that a short peptide called P3 clumps into oligomers and fibrils, and does it faster than amyloid beta does. Yet papers kept appearing in peer-reviewed journals that described P3 as innocuous and non-amyloidogenic, with his work listed among the supporting references. In the perspective article he published in ChemBioChem on 27 February 2026 (open access, volume 27, issue 5, e202500912), he puts the complaint in parentheses and italics: we showed the diametric opposite of that. One ACS journal, Chemistry of Materials, has since issued a correction rewriting the offending sentence, replacing "is considered innocuous" with a line conceding that P3's potential amyloidogenic toxicity remains debated.
The chemistry is simple enough to follow at a kitchen table. Amyloid precursor protein, APP, sits in the membranes of brain cells. Cut it with beta-secretase and then gamma-secretase and you get amyloid beta, the 40 and 42 amino acid peptides that have absorbed the field, and the money, for decades. Cut it instead with alpha-secretase, which snips after residue 16 in amyloid beta numbering, and then gamma-secretase, and you get a shorter fragment: P3, also written as amyloid beta 17-40 and 17-42. Because that route skips the peptide everyone was watching, it was named the non-amyloidogenic pathway, and its product inherited the reputation of its pathway. Raskatov proposes renaming it amyloid alpha, or Aa, on the grounds that a peptide which forms amyloid should not be called non-amyloidogenic.
He is not arguing alone. In Nature Communications on 27 February 2025, exactly a year before his perspective appeared, John Viles's group at Queen Mary University of London reported that P3 peptides form amyloid fibrils rapidly, with kinetics dominated by secondary nucleation, that electron microscopy shows them assembling into ring-shaped annular oligomers, and that those assemblies disrupt membranes, let calcium flood into cells, form pore-like channels and kill cells in culture. Crucially, when P3 was mixed with full-length amyloid beta, the two cross-seeded each other. The same paper notes that P3 is not an exotic laboratory construct: it is a common APP cleavage product, and it is found in the diffuse amyloid deposits of Alzheimer's brains and of brains affected by Down syndrome.
What would that change for someone living with a diagnosis? Nothing today, and that boundary matters. But it changes the model, and models decide what gets targeted. Raskatov's argument is that if amyloid alpha and amyloid beta can co-aggregate, binary assemblies might form at lower peptide concentrations than amyloid beta alone requires, which would place them earlier in the disease timeline, closer to initiation than to the plaques visible at autopsy. If so, a drug aimed only at amyloid beta is being asked to do a job with half the picture. The direction of the effect is genuinely unsettled, which is the most honest thing in the whole story: in a separate study in the Journal of the American Chemical Society, Raskatov's own group reported the first cryo-electron microscopy structure of an amyloid alpha fibril and found that amyloid alpha nudged amyloid beta into aggregates that were less toxic than pure amyloid beta aggregates. A modulator can push either way.
The context he sets this against is the familiar one. Current Alzheimer's treatment, as his introduction lists it, runs to cholinesterase inhibitors, an NMDA receptor antagonist, and the recently approved antibodies that clear amyloid beta, which carry a risk of bleeding and swelling in the brain. UC Santa Cruz's announcement of the paper puts the scale of the effort at more than 400 clinical trials, most aimed at amyloid beta, most of them failed or modestly effective. Raskatov's own summary of progress is blunt: extremely slow. His remedy is not a drug but a revision, an amyloid cascade hypothesis that includes P3 as a possible contributing actor rather than a bystander.
It is worth being precise about what kind of document this is. A perspective article is an argument, not an experiment: it was received in November 2025, accepted in January 2026, and it reviews work done in test tubes and in cultured cells. No one has shown that P3 drives cognitive decline in a living person, and no therapy targets it. What has happened is that a peptide which spent thirty years classified as nothing has been reclassified, at least in the chemistry literature, as something unresolved.