Alzheimer's disease may not be one disease at all. A protein-level survey of 2,279 people with neurodegenerative conditions, published in Cell, identified three distinct molecular subtypes of Alzheimer's, four subtypes of Lewy body dementia and four of frontotemporal degeneration. The resource behind these findings, called the pan-neurodegeneration atlas, or PanNDA, was built by a multi-institute collaboration led by scientists at St. Jude Children's Research Hospital.
PanNDA maps the "proteomes" of six major neurodegenerative diseases - Alzheimer's, Parkinson's, vascular dementia, Lewy body dementia, progressive supranuclear palsy and frontotemporal degeneration with TDP-43 pathology - cataloguing not just which proteins are present but how they are chemically modified and which other proteins they bind. That matters because these diseases typically begin with proteins misfolding or piling up, and the damage does not stop there: faulty proteins drag down their binding partners and the signaling pathways connected to them. Co-corresponding authors Junmin Peng of St. Jude and Bin Zhang of the Icahn School of Medicine at Mount Sinai combined several proteomic techniques to trace that ripple effect.
The analysis turned up changes unique to individual diseases as well as changes shared across them, which helps explain why conditions with different names can look so similar in the clinic. The subtype findings carry the most immediate practical weight: the team reported more than 20 proteins that could act as biomarkers to sort Alzheimer's patients into its three subtypes. If those markers hold up, they would give clinicians a way to distinguish patients who currently receive the same diagnosis and, potentially, the same treatment approach.
Sorting a broad diagnosis into biologically defined groups is the kind of step that has reshaped cancer treatment, and it may explain part of why some dementia drug trials produce mixed results across patient populations. PanNDA is framed by its creators as a resource for the wider research community rather than a finished diagnostic tool, with the biomarker candidates still requiring validation before they reach clinical use.