The Alzheimer's field spent 30 years on cholesterol. The protection was in DNA repair.

Alzheimer's genetics has told the same story for three decades: a single gene, a lipid-transport problem, a drug target that never quite worked. The Buck Institute's new study, led by Lisa Ellerby, breaks that frame. In neurons engineered to differ only at the APOE locus, the protective variant APOE2 activates DNA-repair programs. The risky variant APOE4 does the opposite. The two differ by two amino acids.

Wire summaries will tell readers that APOE2 carriers live longer and resist dementia. That is the population fact. The mechanistic fact — what APOE2 is actually doing in neurons — goes beyond cholesterol transport, the role the field has studied since the 1990s.

When the protective signal is repair, not clearance, the drug target shifts from lipid metabolism to the DNA-damage response. Stress tests using radiation and doxorubicin bore it out: APOE2 neurons showed lower senescence markers and better-preserved nuclear architecture under stress. Older APOE2 knock-in mice carried the same signature in their hippocampi.

The most consequential finding: when Ellerby's team added recombinant APOE2 protein to APOE4 neurons, the DNA-damage signal dropped. The protection is not locked into people born with the variant. It is a function the protein performs. If that holds in follow-up studies, drug developers would have a new target — though the finding comes from iPSC neurons and mouse models, so therapeutic development is still prospective. Two amino acids, three decades, one reframe.