FOXO1 Switch Lets Bacteria Open Body's Barriers

The body's first defense is a wall of cells, and bacteria have a faster way through it than anyone had documented. In a study published in Nature Communications on November 6, 2026, Stella Alimperti of Georgetown University and Dana T. Graves of Penn Dental Medicine show that bacteria trigger the protein FOXO1 to enter the cell nucleus, and that this single move pulls the barrier's sealing proteins out of position within hours, leaving the wall leaky.

The barriers at stake line the mouth, the gut and the lungs. Their cells are stitched together by proteins, chiefly E-cadherin and beta-catenin, that form the adherens junctions between neighbors. For decades the working model held that bacteria broke those stitches by chewing the proteins up or by switching off the genes that make them. The three-year study, run at Georgetown and the University of Pennsylvania, found a third route that does neither.

Within a few hours of bacterial exposure, E-cadherin and beta-catenin hadn't disappeared. They had relocated, abandoning the cell membrane for the interior, so the links between cells weakened while every part of the machinery stayed intact. The trigger was FOXO1, a stress-sensing transcription factor. On bacterial challenge it moves into the nucleus, binds beta-catenin more strongly and raises phosphorylation of beta-catenin at the Tyr654 site, which destabilizes the beta-catenin-E-cadherin complex that holds the junction together.

The causal tests run in both directions. Forcing FOXO1 to stay in the nucleus was sufficient on its own to replicate the bacterial effect and fail the barrier, with no bacteria present. Keeping FOXO1 in the cytoplasm, or deleting the Foxo1 gene from epithelial cells, preserved membrane E-cadherin and beta-catenin and rescued barrier function despite bacterial challenge. The team confirmed the sequence in a 3D microfluidic device and in mice exposed to oral bacteria, supported by National Institutes of Health grants including R01DE031046.

Why a mouth protein matters beyond the mouth is the downstream map. "The oral cavity is a gateway," Graves said. "When the barrier breaks down, it provides an opportunity for bacteria to enter the blood stream." Bacteria entering the blood from the mouth have been tied to rheumatoid arthritis flare-ups, higher blood glucose in diabetes, worsened dementia and cardiovascular disease, the researchers note. A barrier that fails by relocation, rather than destruction, is in principle a barrier that can be switched back, which is the therapeutic opening the paper points to. No treatment exists yet; the finding is a mechanism, demonstrated in models and animals, and the authors present it as a roadmap rather than a therapy.

The study also reframes what a barrier drug would have to do. A treatment aimed at this pathway wouldn't need to kill bacteria or rebuild junctions from scratch. It would need to keep FOXO1 out of the nucleus, or keep it from binding beta-catenin, so the junctions never receive the relocation signal. The rescue experiments in the paper are, in effect, a proof of that concept inside a model system: hold FOXO1 in the cytoplasm and the wall holds. Turning that into a medicine that works in a human mouth, gut or lung is a separate, unsolved problem, and the researchers don't claim otherwise.