The brain's most intimate boundary — the blood-brain barrier — has long been known to falter in Alzheimer's disease, but the mechanism behind that failure has remained elusive. Now, researchers at Columbia University have traced a significant portion of the damage to fibronectin, a protein the body deploys to heal wounds, which in the presence of the APOE4 gene variant accumulates around brain blood vessels and quietly dismantles the seal meant to protect the mind. The discovery, published in Nature Aging, does not yet offer a cure, but it illuminates a new corridor in the architecture of Alzh
Scientists Identify Fibronectin as Key Driver of Blood-Brain Barrier Breakdown in Alzheimer's
The repair crew leaving behind so much scaffolding the door can no longer close
Mark
So fibronectin is normally a good thing—it helps the body heal. What changes in Alzheimer's?
Mimi
In Alzheimer's, especially in people with APOE4, the brain's support cells start overproducing it. It piles up around blood vessels and disrupts the signals that keep them sealed.
Luke
How do we know it's actually causing the damage and not just showing up after the fact?
Mimi
They tested it directly in zebrafish. They made brain cells produce extra fibronectin, and the barrier leaked. The protein alone was enough.
Mark
And they found a way to fix it?
Mimi
Multiple ways, actually. Reducing fibronectin levels lowered leakage. Blocking the pathway it uses to suppress protective signals restored those signals. Adding growth factors downstream also worked.
Luke
But none of this has been tested in humans yet, right?
Mimi
Correct. It's all preclinical—zebrafish, mice, human cells in dishes. No one has shown it preserves memory or cognition in animals.
Mark
What about that genetic variant you mentioned—the one that lowers risk by 71 percent?
Mimi
That's the promising part. It's a rare variant that reduces fibronectin, and people who carry it show no signs of harm. It suggests you could safely lower fibronectin without eliminating it entirely.
Luke
So the goal isn't to eliminate fibronectin, which the body needs for healing.
Mimi
Exactly. It's to selectively reduce the pathological buildup at brain blood vessels. An antibody or small-molecule drug might do that.
Mark
How long until we see a treatment?
Mimi
That's unknown. They need to show it works in animal cognition studies, establish safety, and make sure it reaches the brain without disrupting fibronectin's other roles.
Luke
And the human findings so far are correlational—they found more fibronectin in Alzheimer's brains, but that's not the same as proving it causes the disease.
Mimi
Right. It's a mechanistic discovery, not a clinical one. But it opens a new door.
Le Pouls
- The blood-brain barrier — the brain's primary defense against toxins and pathogens — begins leaking in Alzheimer's patients, sometimes before memory loss even appears, accelerating neurodegeneration in ways current treatments do not address.
- Fibronectin, a protein normally deployed for wound repair, is overproduced by brain support cells in APOE4 carriers and accumulates around blood vessels, actively suppressing the chemical signals that keep the barrier sealed.
- Researchers confirmed fibronectin's causal role by engineering zebrafish to overproduce it — the barrier leaked from the protein alone — then successfully reversed the damage through multiple interventions targeting different points in the pathway.
- A rare natural gene variant that reduces fibronectin levels cuts Alzheimer's risk in APOE4 carriers by 71 percent with no observed harm, suggesting that precision reduction — not elimination — of the protein is a viable and potentially safe therapeutic goal.
- The findings are preclinical and have not yet demonstrated cognitive benefits in animals, meaning antibody, small-molecule, or gene-based treatments targeting fibronectin remain promising but years from clinical application.
The brain's most intimate boundary — the blood-brain barrier — has long been known to falter in Alzheimer's disease, but the mechanism behind that failure has remained elusive. Now, researchers at Columbia University have traced a significant portion of the damage to fibronectin, a protein the body deploys to heal wounds, which in the presence of the APOE4 gene variant accumulates around brain blood vessels and quietly dismantles the seal meant to protect the mind. The discovery, published in Nature Aging, does not yet offer a cure, but it illuminates a new corridor in the architecture of Alzheimer's — one that leads not to plaques and tangles, but to the vessels themselves, and to the possibility that what the body sends to repair can, in excess, become the source of harm.
The brain is famously selective about what it permits entry. A tightly sealed barrier of cells lining its blood vessels keeps toxins out and the internal environment stable. In Alzheimer's disease, that barrier begins to fail — and a study published in Nature Aging, led by neuroscientist Çağhan Kızıl at Columbia University, has identified an unexpected driver of that failure: fibronectin, a protein the body ordinarily uses to scaffold tissue repair.
In people carrying the APOE4 gene variant — the strongest known genetic risk factor for late-onset Alzheimer's — a cascade of inflammation and amyloid-beta accumulation prompts astrocytes, the star-shaped support cells surrounding brain blood vessels, to overproduce fibronectin. The protein piles up around the vessels, suppressing the chemical signals that hold the barrier together. Gaps open. Blood-borne material leaks into brain tissue, feeding inflammation and accelerating neurodegeneration.
To confirm fibronectin was causing the damage rather than merely appearing at damaged sites, the team engineered zebrafish to overproduce the human protein. The barrier leaked from fibronectin alone. Working backward through the pathway, researchers then showed they could reverse the damage — by reducing fibronectin levels, by blocking the pathway through which it suppressed protective signals, and by reintroducing growth factors that restored the barrier's tight junctions.
The therapeutic challenge is specificity. Fibronectin is essential for normal healing, so eliminating it entirely would cause harm. But nature has already suggested a solution: a rare loss-of-function variant in the FN1 gene lowers Alzheimer's risk in APOE4 carriers by 71 percent, with no apparent harm to those who carry it. Kızıl sees this as evidence that selectively reducing fibronectin's pathological accumulation at brain vessels — through a future antibody, small molecule, or gene-based treatment — is biologically achievable.
Significant work remains. The interventions have not yet been shown to preserve cognition in animals, and the human data is correlational. But the discovery reframes the problem meaningfully: in Alzheimer's, part of the damage may come not from what the brain fails to build, but from what the body's own repair crew leaves behind.
The brain is famously selective about what it lets through. Oxygen and nutrients cross the blood-brain barrier freely. Toxins, pathogens, and most other substances traveling in the bloodstream do not. This protective seal is maintained by tightly connected cells lining the brain's blood vessels—a biological checkpoint that keeps the brain's internal environment stable and safe. In Alzheimer's disease, that checkpoint begins to fail. The barrier leaks. And researchers have now identified a surprising culprit: a protein normally tasked with healing wounds.
Fibronectin is part of the body's repair machinery. When tissue is injured, fibronectin acts as temporary scaffolding, giving cells something to grip while they rebuild. But in Alzheimer's disease, particularly in people carrying the APOE4 gene variant, fibronectin accumulates around brain blood vessels in dangerous amounts. A study published in Nature Aging, led by neuroscientist Çağhan Kızıl at Columbia University, traces how this happens and shows that the damage can be reversed in experimental models.
The chain of events begins with APOE4, the strongest known genetic risk factor for late-onset Alzheimer's. Carrying one copy of the variant increases disease risk; two copies increase it further. APOE4 does not guarantee Alzheimer's will develop, but it has long been associated with early breakdown of the blood-brain barrier. How it caused that breakdown remained unclear until now. The new research suggests that APOE4, combined with inflammation and amyloid-beta accumulation, triggers astrocytes—star-shaped support cells surrounding brain blood vessels—to overproduce fibronectin. The protein piles up around the vessels, disrupting the chemical signals that keep the barrier sealed. Without those messages, the connections weaken. Gaps open. Material from the blood leaks into brain tissue.
To confirm fibronectin was actually causing the damage rather than simply appearing at damaged sites, the researchers conducted a direct test. They engineered zebrafish brain cells to produce excess human fibronectin. The protein alone was sufficient to make the barrier leak. Fibronectin was not a bystander. It was driving the breakdown. The team then worked backward, testing interventions at multiple points in the pathway. Reducing fibronectin levels lowered leakage in zebrafish exposed to amyloid-beta. Blocking the cellular pathway through which fibronectin suppressed protective signals restored those messages and strengthened the barrier. In another experiment, they added growth factors downstream of the damaged pathway—HB-EGF and IGF-1—which restored the tight junctions holding the barrier together.
These findings matter because blood-brain barrier breakdown can occur early in Alzheimer's, sometimes before memory problems appear. Once the seal weakens, inflammation increases and toxic substances accumulate in brain tissue, accelerating neurodegeneration. Most Alzheimer's research and current treatments target amyloid-beta and tau, the proteins that form plaques and tangles. This work points to a different target: the blood vessels themselves and the proteins disrupting their integrity. "The most direct and most translatable finding is that reducing fibronectin itself rescues the barrier defects," Kızıl told ScienceAlert.
The therapeutic challenge is precision. The body needs fibronectin for normal wound healing and blood-vessel maintenance. Eliminating it entirely would cause harm. But an earlier genetic discovery suggests a path forward. Researchers identified a rare loss-of-function variant in the FN1 gene—the gene that codes for fibronectin—that lowers Alzheimer's risk in APOE4 carriers by 71 percent, with no evidence of harm in people who carry it. "Nature has already run a version of this experiment," Kızıl noted. This suggests that selectively reducing fibronectin's pathological buildup at brain blood vessels, without eliminating the protein's essential functions elsewhere, is possible. A future antibody or small-molecule drug might clear excess fibronectin from around brain vessels. A gene-based treatment mimicking the protective variant's effect in the brain is another possibility.
But significant work remains before any of this reaches patients. The interventions have not been shown to preserve memory or cognition in animals, let alone treat Alzheimer's in people. The human findings are correlational and require confirmation in larger groups. Researchers must demonstrate that targeting fibronectin improves cognition in animal models, establish its safety profile, and ensure that any treatment reaches the brain without disrupting fibronectin's necessary roles in the rest of the body. The discovery is mechanistic and preclinical—a new suspect identified, a pathway illuminated, but not yet a cure. Still, it reframes the problem. In Alzheimer's, part of the trouble may come from the repair crew leaving behind so much scaffolding that the door can no longer close.
Citations marquantes
The most direct and most translatable finding is that reducing fibronectin itself rescues the barrier defects.— Çağhan Kızıl, Columbia University neuroscientist
Nature has already run a version of this experiment with a rare protective genetic variant that lowers Alzheimer's risk without causing harm.— Çağhan Kızıl
Envie de l'histoire complète ?
Lire l'original sur ScienceAlert ↗
Health & Medicine