BALTIMORE
For fifteen years, scientists knew that a toxin produced by a common gut bacterium was somehow linked to colorectal cancer. What they didn't know was how the toxin actually got inside colon cells in the first place to do its damage. That black box has now been pried open, and the answer, published in the journal Nature, is both simpler and more unsettling than anyone expected: the toxin needs a very specific lock to pick, and scientists have finally identified it.
The discovery, led by a multi-institutional team anchored at the Johns Hopkins Kimmel Cancer Center and the Bloomberg~Kimmel Institute for Cancer Immunotherapy, working alongside collaborators at Harvard Medical School and the Molecular Biology Institute of Barcelona, pinpoints a protein called claudin-4 as the essential receptor that a bacterial toxin known as BFT — short for B. fragilis toxin — must bind to before it can breach the protective lining of the colon. Once bound, the toxin sets off a chain reaction that damages the gut's barrier, triggers inflammation, and appears to nudge cells toward the kind of runaway proliferation that defines cancer.
It's the kind of finding that sounds narrow and technical on paper but carries genuinely big stakes: colorectal cancer is one of the most common cancers worldwide, rates in adults under 50 have been climbing for reasons researchers still don't fully understand, and this is the clearest mechanistic explanation yet for one of the specific ways a gut microbe may be helping to switch it on.
THE BACTERIUM HIDING IN PLAIN SIGHT
Enterotoxigenic Bacteroides fragilis is not some exotic pathogen lurking in a lab freezer. It's a strain of bacteria that lives in the guts of a meaningful chunk of the human population, often without causing any obvious symptoms at all. For years, epidemiological studies had linked colonization by this particular toxin-producing strain to higher rates of colorectal cancer, and lab studies in mice had shown that exposing the colon to the toxin could kick off tumor formation. But knowing that a toxin is associated with cancer and knowing exactly how it does its damage at the molecular level are two very different things — and for a decade and a half, the second part remained frustratingly out of reach.
The toxin itself, BFT, is classified as a metalloprotease — a type of enzyme that works by cutting other proteins. Researchers already knew that BFT was capable of severing a critical protein called E-cadherin, which normally acts like molecular glue holding neighboring colon cells tightly together to keep the gut's lining sealed and intact. Cut enough E-cadherin, and that seal starts to fail: the barrier between what's inside your gut and the rest of your body gets leaky, immune cells rush in to deal with the breach, chronic inflammation sets in, and the surrounding cells are pushed into overdrive trying to repair the damage — the kind of persistent cellular stress and division that can eventually tip into malignancy.
What nobody could explain was how BFT physically got close enough to E-cadherin to cut it in the first place. E-cadherin isn't just sitting exposed on the surface of a cell waiting to be attacked; the toxin needed some kind of docking point, a way to grab onto the outside of the cell before it could reach its target. That docking point had eluded researchers for fifteen years.
THE CRISPR SCREEN THAT CRACKED IT OPEN
The breakthrough came from a genome-wide CRISPR knockout screen — a technique that allows researchers to systematically disable one gene at a time across thousands of cells and then observe which of those disabled genes changes how the cells respond to a given challenge. In this case, the team exposed colon epithelial cells, each missing a different gene, to the BFT toxin and watched for which cells became resistant to its effects.
One result stood out immediately. When the gene for claudin-4 — a protein normally involved in forming the tight junctions that seal cells together — was knocked out, the toxin simply couldn't get a grip. Without claudin-4 present, BFT failed to bind to the cell surface at all, and without binding, it couldn't go on to cleave E-cadherin or trigger the downstream inflammatory cascade.
"We've made several attempts over time to identify the receptor, so this is an exciting moment," said Cynthia Sears, M.D., the Bloomberg~Kimmel Professor of Cancer Immunotherapy and professor of medicine at Johns Hopkins, who served as senior author on the study. The team had circled this problem for years using other approaches without success; it took the brute-force, unbiased power of a genome-wide screen to finally surface the answer clearly.
Maxwell White, a researcher involved in the work, described the moment the data came in: "It took a while to get the assay working and validate the approach, but once we were able to do the screen, claudin-4 was a clear, resounding top hit." In a field where ambiguous or conflicting results are the norm, a "clear, resounding top hit" is the kind of language researchers use only when the data leaves little room for doubt.
BUILDING A DECOY TO FIGHT BACK
Identifying the lock the toxin uses to get in is one thing. Actually doing something useful with that knowledge is another — and this is where the study takes its most promising turn. Once the team confirmed claudin-4 was the toxin's essential entry point, they set out to build something to jam that lock before the toxin ever reached a real cell.
Their solution: a soluble decoy protein engineered to display the same claudin-4 sequences the toxin normally targets, but untethered from any actual cell. Released into the environment around colon cells, the decoy acts like a set of fake locks scattered everywhere, giving the BFT toxin something to latch onto that leads nowhere. In experiments, the toxin bound to these decoys instead of binding to the real claudin-4 receptors sitting on the surface of colon cells — meaning the toxin got tricked into wasting itself on a dummy target rather than doing damage to living tissue.
That's a meaningfully different kind of intervention than most existing approaches to gut bacteria and cancer risk, which tend to focus on broadly wiping out or suppressing bacterial populations with antibiotics — a blunt strategy that also kills off beneficial microbes and can have its own long-term consequences for gut health. A decoy-based approach, by contrast, wouldn't need to eliminate B. fragilis from the gut at all. It would simply neutralize the one toxin responsible for the damage, potentially leaving the rest of a person's microbiome undisturbed.
WHY THIS MATTERS BEYOND ONE BACTERIUM
Colorectal cancer remains one of the most diagnosed and deadliest cancers globally, and doctors have spent the last decade increasingly alarmed by a stubborn rise in cases among adults younger than 50 — patients who fall well outside the age range typically targeted by routine colonoscopy screening. Researchers have long suspected that shifts in diet, environment, and the gut microbiome play some role in that trend, but concrete, mechanistic evidence connecting a specific microbial toxin to a specific molecular pathway toward cancer has been rare. This study provides exactly that kind of concrete link for at least one contributing factor.
It also opens a broader conceptual door. If one toxin from one bacterial species relies on a single, druggable receptor to cause damage, it raises the question of whether other cancer-associated microbes and their toxins might operate through similarly specific, targetable mechanisms — mechanisms that could, in principle, be blocked the same way, with decoys or receptor-blocking drugs rather than broad antimicrobial treatment.
EXPERT PERSPECTIVE
Cancer researchers not involved in the study have pointed to the CRISPR screening approach itself as a template other labs are likely to borrow, given how cleanly it isolated a single gene responsible for an entire toxic cascade out of the tens of thousands of genes in the human genome. The finding also lands amid a broader wave of research into the gut microbiome's role in cancer more generally; coverage of the study has noted it as one of several recent lines of evidence pointing to specific bacterial products, rather than the microbiome in vague aggregate, as drivers of disease — a shift that makes the biology far more tractable for drug developers than "the microbiome" as a whole ever was.
WHAT THIS MEANS FOR YOU
This isn't a treatment you can walk into a clinic and ask for tomorrow — the decoy protein has so far been tested in cell and animal models, and turning a laboratory proof-of-concept into an approved therapy typically takes years of further development, safety testing, and clinical trials. But the discovery matters for anyone who has watched colorectal cancer rates creep upward, especially among younger adults, and wondered whether science actually understood why. It's also a reminder that a healthy-seeming gut can still harbor bacteria capable of contributing to disease over years or decades, reinforcing why screening guidelines and awareness of family history matter regardless of age.
For patients and families affected by colorectal cancer, the practical takeaway is more hope than immediate treatment: a specific, targetable mechanism now exists where none was understood before, and a working decoy-based countermeasure already exists in early form. That's the kind of foundation new drugs get built on.
THE BOTTOM LINE
Fifteen years is a long time to chase a single molecular mystery, but the payoff here is the kind that could ripple well beyond one toxin and one bacterium. By identifying claudin-4 as the lock BFT needs to pick, and by proving a decoy can jam that lock in the lab, researchers haven't just closed a long-standing question — they've handed drug developers a concrete, specific target to aim at, in a cancer where new tools are badly needed.
