A New Genetic Risk Score for Alzheimer's Disease Directly Matches Microscopic Damage to the Brain's Cellular Recycling Machinery
A New Genetic Risk Score for Alzheimer's Disease Directly Matches Microscopic Damage to the Brain's Cellular Recycling Machinery
A team at the University of Washington combined 14 known genetic risk variants for Alzheimer's that act in endosomes and lysosomes (the cell's waste-recycling machinery) into a single score. The score predicted diagnosis and disease severity across two independent donor cohorts. On postmortem cortical sections, individuals with high scores had more damaged endosomes in neurons and more lysosomal aggregates in microglia (the brain's immune cells), regardless of disease stage.
Genome-wide association studies link a DNA variant to a gene indirectly, by chromosomal proximity rather than by a demonstrated cellular defect. Since 2009, this approach has identified dozens of candidate genes for Alzheimer's. Among the first to be found, just after APOE (the best-known Alzheimer's risk gene), were CLU, PICALM, and BIN1. Some of these shared a common denominator: they function in endosomes and lysosomes. In cells lacking one such gene, endosomes were already swollen, and a similar picture had been observed in people with mild cognitive impairment and Down syndrome before amyloid plaques appeared. These observations led Suman Jayadev's laboratory at the University of Washington to combine the pathway's risk variants into a single score and build a bridge between diagnostic statistics and actual organelle damage.
On September 11, a new version of the preprint was published testing this idea in humans. The score, built from 14 variants across 12 genes, was validated in 293 donors from the team's own cohort and 27,399 donors from the national sequencing project ADSP: in both groups, it predicted diagnosis and severity of postmortem-confirmed pathology.
The authors then looked inside the cell itself.
On postmortem cortical sections from individuals with high risk scores, neurons contained more early endosomes; these endosomes were larger and clustered more tightly around the cell nucleus. In microglia, lysosomal aggregates were more abundant. The damage was present both in donors without a diagnosed disease and in those whose brains already carried its signs.
RNA sequencing of 150,000 individual cell nuclei showed that a high risk score shifts gene expression in several brain cell types in the same direction: DNA damage response programs are upregulated, while programs that maintain protein homeostasis and energy metabolism are downregulated.
The direction of the effect changed with disease stage: at early stages, a high risk score amplified synaptic and endolysosomal programs in neurons; at late stages, the same programs were suppressed at the same score. For BIN1 and CLU, this reversal is visible in the genes' own expression levels and coincides with a similar switch previously described independently for one of the CLU risk variants.
The same laboratory had previously shown similar logic for a single gene in this pathway, SORL1: neurons lacking it upregulated synaptic genes more actively, yet synaptic function itself weakened. The present work extends that finding from a single gene to the entire pathway and to actual human brain tissue.
The study design remains postmortem: a cortical section captures only a single moment in a deceased individual. The authors propose that proving whether genetic burden itself damages the organelles will require future longitudinal studies with biomarkers and experiments with direct genetic manipulation of cells.