A toxin made by a common gut bacterium needs one specific doorway to break into colon cells, and researchers just found it, along with a possible way to block it.
Quick Take
- Johns Hopkins scientists identified a protein called claudin-4 as the receptor a gut bacteria toxin uses to invade colon cells
- The finding, published in Nature in April 2026, came from a genome-wide CRISPR screen testing thousands of genes
- A lab-made decoy protein blocked the toxin and protected mice from tissue damage in early tests
- Prior research already linked this toxin to colorectal cancer, but the exact invasion method stayed unknown for 15 years
- Scientists say more work is needed to confirm the decoy can prevent cancer in humans
A Bacterium Linked to Cancer for 15 Years
Bacteroides fragilis lives in nearly everyone’s gut without causing trouble. But a toxic strain called enterotoxigenic Bacteroides fragilis makes a poison known as BFT. Researchers have tied this toxin to colorectal cancer for more than a decade. Exactly how it broke into colon cells to start the damage stayed a mystery. Johns Hopkins Medicine called it a 15-year puzzle finally solved.
Scientists at Johns Hopkins ran a genome-wide CRISPR screen, a method that checks thousands of genes at once, to find what BFT needs to grab onto a cell. Their answer, published in Nature in April 2026, pointed to a protein called claudin-4 as the toxin’s receptor. Without it, the toxin has nothing to latch onto.
When the research team removed claudin-4 from lab-grown cells, the toxin lost its grip entirely. It could no longer bind to the cells at all, according to a Newsweek account of the findings. That single result confirmed claudin-4 wasn’t just linked to the toxin’s behavior. It was the actual entry point the bacterium’s weapon depended on.
How the Toxin Picks the Lock on Colon Cells
Once BFT grabs claudin-4, it does more than attach. The toxin uses that grip to slice apart E-cadherin, a protein that holds colon cells tightly packed together. That cut weakens the colon’s inner lining. Gaps open up, inflammation moves in, and nearby cells start multiplying faster than normal, a chain reaction researchers believe helps set the stage for cancer.
The discovery matters because earlier research already showed something was going on. One meta-analysis of colorectal cancer patients found the toxic bacteria strain showed up far more often in tumors than in healthy tissue, with more than double the odds of being present. That pattern raised suspicion for years. It just didn’t explain the mechanism behind it, until the claudin-4 connection filled in the missing step.
A Decoy Protein Offers a Possible Shield
To test their discovery, the Johns Hopkins team built a soluble decoy, a free-floating piece of protein copied from claudin-4’s structure. Instead of reaching real colon cells, the toxin latched onto the decoy, essentially tricked into attacking a target that could not be damaged. The toxin wasted its attack on a harmless stand-in.
In mice, that decoy strategy worked. Researchers reported it “successfully protected” the animals from toxin-induced tissue damage. That is a meaningful early result, not a finished cure. The team has not yet shown the same decoy stops colorectal cancer from forming, in mice or in people, and says more research is needed before it could ever reach patients.
Scientists finally discover how a common gut bacterium can trigger colon cancer. Bacteroides fragilis toxin damages the colon lining, drives inflammation and promotes tumors, Johns Hopkins found. https://t.co/RLey3dmIpa
— Drew Grimaldi (@Grimillionaire) October 31, 2026
The claudin-4 discovery hands scientists something they lacked before: a specific molecular target instead of a loose statistical pattern. Years of studies had already flagged this toxic bacteria strain as a suspect in colorectal cancer. Now there’s an actual mechanism to chase, test, and potentially interrupt, turning a long-standing correlation into a concrete lead for prevention research.
Sources:
sciencedaily.com, hopkinsmedicine.org, newsweek.com, pmc.ncbi.nlm.nih.gov
