APOE4 triggers scarring in brain blood vessels that drives Alzheimer's, reversed in mice and a lab-grown human brain model
APOE4 triggers scarring in brain blood vessels that drives Alzheimer's, reversed in mice and a lab-grown human brain model
On September 24, researchers at the Icahn School of Medicine at Mount Sinai reported in Cell why the brain's blood vessels are particularly vulnerable in carriers of the APOE4 gene, the strongest genetic risk factor for Alzheimer's disease. In APOE4 carriers, pericytes, the support cells that wrap around blood vessels, turn into scar-forming cells that attract toxic amyloid to vessel walls. This transition proved to be reversible.
APOE4 is a variant of the gene encoding the brain's cholesterol transport protein. Two copies raise Alzheimer's risk up to 16-fold, and at least one copy is carried by 40–65% of everyone who has received this diagnosis. The brain's blood vessels are especially fragile in APOE4 carriers and accumulate amyloid, but which cell type and which signal are responsible was unknown. The only approved class of Alzheimer's drugs, anti-amyloid antibodies, carries a particular risk of brain hemorrhages and edema in APOE4 carriers.
Pericytes wrap around the brain's smallest blood vessels and maintain the blood-brain barrier, the filter that prevents unwanted substances in the blood from entering the brain. A gene-activity atlas of nearly 64,000 vascular cells from 220 individuals showed that APOE4 carriers have fewer pericytes and instead harbor cells with features of myofibroblasts, scar-forming cells that normally appear briefly during wound healing and then vanish. In APOE4 carriers these cells do not vanish; they persist along blood vessels throughout the brain, including before any Alzheimer's diagnosis. This was confirmed in postmortem human tissue, in aged mice carrying the APOE4 gene, and in miBrain, a lab-grown human brain tissue derived from stem cells.
Suspicion had fallen on fibronectin earlier: in 2024, a rare mutation that disables its gene reduced Alzheimer's risk in APOE4 carriers by 71% and delayed disease onset by more than three years. Which cell produces this protein in the brain's vasculature had remained unknown. Inside miBrain, the researchers replaced APOE4 pericytes with healthy ones from the same cell line carrying APOE3. The scar-forming cells disappeared, pericytes returned, and they proved to be the main source of fibronectin in the vasculature. Normally this protein strengthens tissues, but here it acted as a molecular adhesive for amyloid; knocking out its gene reduced amyloid accumulation.
The culprit proved to be TGF-β, a signaling protein that normally governs tissue repair. In APOE4 carriers its level is nearly indistinguishable from normal, but their pericytes carry far more receptors for this signal. The cells respond to ordinary background levels of TGF-β as though the signal were dangerously elevated, launching a scarring program where healthy cells remain quiescent.
This is why reversal worked. In mice and in miBrain, two different inhibitors of the TGF-β receptor restored pericytes to their normal appearance, reduced fibronectin by 24–41%, and reduced amyloid by 48–52%, even though treatment began after the damage had already formed. One of the inhibitors, galunisertib, has already been tested in humans: Eli Lilly evaluated it as an anticancer drug for liver cancer, and human safety data already exist. The vascular vulnerability conferred by APOE4 is also being approached from another direction: in a pilot study of 23 adults, rapamycin increased cerebral blood flow by more than 15% in carriers of this gene, and since December a trial of sirolimus in 225 women with APOE4 has been comparing it with placebo.
"Vascular damage in the brain is an active, potentially reversible biological process driven by the APOE4 gene, not simply a late consequence of Alzheimer's disease," says senior author Joel Blanchard.