APOE ε4 is the strongest common genetic risk factor for Alzheimer’s disease, and roughly one in five people carries at least one copy. Carrying it raises the likelihood of developing the disease, but it does not guarantee it. Some carriers remain mentally sharp well into their 80s, and scientists have long wondered what protects them.

A team at Columbia University now reports a biological mechanism that may help explain that resilience. The study, published in Nature Aging on September 11, 2026, centers on a protein called fibronectin and on the blood-brain barrier, the protective lining of the brain’s blood vessels.

The genetic clue

The new work builds on a study the same group published two years ago. It found that some people who carry APOE ε4 also carry a rare protective mutation in the fibronectin gene, FN1. Fibronectin helps support tissue structure, but in people with APOE ε4, excess amounts can build up around blood vessels in the brain and may compromise the blood-brain barrier. The rare FN1 mutation appears to limit that buildup and the damage linked to it.

That raised a question the new study sets out to answer: why is excess fibronectin harmful in the first place? In the words of corresponding author Caghan Kizil, the study moves “from a genetic clue to a disease mechanism.”

Why the blood-brain barrier matters

The blood-brain barrier is made of tightly connected cells around the brain’s blood vessels. It regulates which substances can pass from the bloodstream into brain tissue and helps maintain the conditions needed to clear waste. When it becomes damaged or too permeable, substances that are normally kept out can get in more easily, which may contribute to inflammation and other processes involved in neurological disease.

What the researchers did and found

The team combined many approaches: human brain tissue and cerebrospinal fluid analyzed with genetic and epigenetic methods, single-cell sequencing, spatial imaging, human stem-cell models, three-dimensional blood vessel cultures, and experiments in zebrafish and mice. Across these methods the findings pointed the same way.

  • APOE ε4 drives fibronectin buildup. Mice engineered to carry human APOE ε4 had nearly twice as much fibronectin in the brain as mice carrying APOE ε3, another common version of the gene, along with signs of a leaky blood-brain barrier.
  • Fibronectin does more than accompany the damage. Raising fibronectin specifically in astrocytes, the support cells that help maintain the barrier, was enough to make it more permeable. Lowering fibronectin improved barrier function.
  • A signaling pathway is disrupted. Excess fibronectin activates abnormal signals through integrins, proteins that help cells sense their surroundings, and an enzyme called focal adhesion kinase. These signals disturb a network of growth factors (VEGF, HBEGF and IGF1) that astrocytes and blood vessel cells use to coordinate and keep the barrier intact.
  • Human samples show a similar pattern. In brain tissue and cerebrospinal fluid, higher fibronectin was associated with inflammation in astrocytes, and genetic analyses linked FN1 to blood vessel disease and to VEGFA, a growth factor important for healthy brain vessels.

Possible routes to treatment

The mechanism suggests three possible strategies: preventing harmful amounts of fibronectin from accumulating around brain blood vessels, blocking the abnormal signals that excess fibronectin triggers, or restoring the growth factor signaling that keeps the barrier healthy. The researchers are investigating these in laboratory experiments.

Turning this into a therapy will not be simple. Fibronectin has essential roles throughout the body, including supporting tissue structure and helping damaged tissue heal, so any treatment would need to target its harmful accumulation or effects without disrupting those functions. The team also suggests the work could help identify the best time to intervene and provide ways to measure whether a treatment is protecting the brain.

How to read these results

The strongest causal evidence, that raising fibronectin opens the barrier and lowering it helps, comes from cell and animal experiments. The human data show associations consistent with the same process, which strengthens the case but does not prove that the mechanism operates identically in people. No therapy based on this work has been tested in humans, and the researchers describe the potential treatments as still under laboratory investigation. For now this is a well-supported mechanism and a promising target, not a treatment or a way to prevent Alzheimer’s.

One more piece of context: according to the press release, Kizil and two co-authors, Richard Mayeux and Badri Vardarajan, are founders and shareholders of KVM Therapeutics and are inventors on patent applications related to FN1-targeted therapeutic approaches. That does not invalidate the findings, but it is relevant when weighing optimism about future drugs.

This post is a summary based on the Columbia University Irving Medical Center press release. For full methods, statistics and limitations, see the original paper.

Sources