We are excited to share our latest study, published today in Nature Aging, identifying astrocyte-derived fibronectin (FN1) as a key mechanistic driver of blood-brain barrier (BBB) dysfunction in Alzheimer’s disease — and a proximal mediator of the risk conferred by the APOE ε4 allele.
Building on our earlier work showing that a rare, protective loss-of-function genetic variant in FN1 reduces Alzheimer’s risk by 71% in APOE4 carriers, we set out to understand how fibronectin causes damage. Integrating postmortem human brain tissue, cerebrospinal fluid and plasma proteomics, single-nucleus transcriptomics, iPSC-derived human astrocyte and 3D vascular (VAMP) models, and in vivo zebrafish and mouse models, we show that:
- APOE4, amyloid-β42, and inflammatory signaling (TNF/NF-κB) drive astrocytes to overproduce and abnormally deposit fibronectin at the gliovascular interface.
- Excess fibronectin engages integrin/focal adhesion kinase (FAK) signaling in astrocytes, suppressing the VEGF → HB-EGF → IGF-1 signaling cascade required for BBB integrity.
- Reducing fibronectin, or restoring this growth factor axis directly, rescues BBB function across our models.
Together, this positions fibronectin as a therapeutic target for the vascular component of Alzheimer’s disease — an early disease mechanism distinct from, but interacting with, amyloid and tau pathology.
Read the full paper (open access): https://doi.org/10.1038/s43587-026-01204-0
This work reflects a major collaborative effort with co-first authors Prabesh Bhattarai and Elanur Yilmaz, and colleagues across Columbia University and our collaborating institutions.
Kizil Lab 