This Gut Bacterium Starved a Cancer-Linked E. coli Out of Inflamed Guts

If you live with ulcerative colitis, you have almost certainly been told that your gut bacteria matter, and then left without anything concrete to do about it. A new mouse study out of the Université de Montréal puts an unusually specific claim on the table: a single ordinary gut commensal, Bifidobacterium pseudolongum, can hold down a strain of E. coli that carries a cancer-linked toxin gene cluster, and calm the inflammation that strain makes worse. The mechanism traced back to one small molecule, and it was not the one this field usually credits.

Key Takeaways

• In mice colonized with pks+ E. coli and given a colitis-inducing chemical, daily oral B. pseudolongum limited the pathogen's expansion and reduced weight loss, colon shortening, and inflammatory markers in stool.
• The protection tracked with acetate, not the more famous butyrate, and with activation of PPAR-γ, a receptor in the gut lining that switches off inflammatory signalling.
• pks+ E. coli makes colibactin, a DNA-damaging toxin repeatedly tied to colorectal cancer, so suppressing it inside an already inflamed colon is worth paying attention to.
• This is animal work with 9 to 13 mice per group and one bacterial strain. Nothing here has been tested in people.

Bifidobacterium pseudolongum limited pks+ E. coli expansion in colitic mice

Oral administration of Bifidobacterium pseudolongum stopped a genotoxic E. coli strain from expanding in the guts of mice with acute colitis, and made their disease measurably milder. The team, led by Manuela M. Santos at the Institut du cancer de Montréal, worked with E. coli NC101, a mouse adherent-invasive strain that carries the pks island (a cluster of genes that lets bacteria manufacture the DNA-damaging toxin colibactin). They gave C57BL/6 mice antibiotics to strip away the resident microbes that normally crowd such a strain out, colonized them with NC101, then induced colitis with dextran sulfate sodium.

Colonized mice got sicker than uncolonized ones: more weight loss, higher disease activity scores, shorter colons, worse crypt damage and ulceration on histology, and higher stool levels of the inflammation markers LCN-2, IL-6, and TNF-α. That gap established the model. The question was whether anything could close it.

Colibactin-producing E. coli is a live concern for people with IBD

Anyone managing ulcerative colitis or Crohn's disease has a practical stake in this, because pks+ E. coli is not a laboratory curiosity. These strains are found more often in the guts of people with inflammatory bowel disease and in colorectal tumour tissue, and large prospective cohort data has linked pks island-carrying E. coli to colorectal cancer risk. Colibactin itself alkylates DNA and leaves a recognizable mutational signature in colon cells.

The uncomfortable part is the feedback loop: inflammation creates exactly the conditions in which E. coli and its relatives thrive, so a flare hands the pathobiont an advantage and the pathobiont then deepens the flare. Breaking that loop with a resident microbe rather than an antibiotic is appealing precisely because antibiotics flatten the beneficial community along with the target.

The study moved from a correlation to a direct causal test

What makes this paper more than a survey is that the researchers did not stop at the association they found. They started with putrescine, a polyamine their own lab had previously shown to limit pks+ E. coli and tumour development in the colon. Adding putrescine to the drinking water during colitis suppressed NC101 growth, kept the mice heavier, lowered disease activity scores, preserved colon length, and cut stool LCN-2, IL-6, and TNF-α.

Then they asked what putrescine had changed. Sequencing showed it raised microbial diversity and shifted community structure, with Actinobacteriota rising, Proteobacteria and Bacteroidota falling, and both Enterococcaceae and Enterobacteriaceae significantly reduced. The Actinobacteriota increase was driven mostly by bifidobacteria, and one sequence variant annotated as B. pseudolongum stood out: it was negatively correlated with NC101 colonization and with colonic TNF-α, and positively correlated with colon length.

At that point most papers would write a discussion paragraph and stop. Instead the team took a reference strain, B. pseudolongum DSM 20099 (98% identity across the V5–V6 region with the variant they had spotted), and gave it to colonized mice by daily oral gavage during colitis. It reproduced the effect: NC101 growth inhibited, less weight loss, lower disease activity, less colon shortening, and lower LCN-2, IL-6, and TNF-α in stool. Group sizes ran from 9 to 13 mice.

Acetate rather than butyrate carried the protective signal

The metabolite doing the work appears to be acetate, which is a mild surprise in a field that has spent a decade focused on butyrate. Targeted metabolomics showed putrescine raised total short-chain fatty acids, and the increase was driven by acetate specifically, not propionate or butyrate. Acetate levels correlated with the abundance of bifidobacteria and with B. pseudolongum measured by qPCR. Putrescine also increased three bile acids: omega-muricholic acid, lithocholic acid, and hyodeoxycholic acid.

Both interventions activated PPAR-γ (peroxisome proliferator-activated receptor gamma), a transcription factor in intestinal epithelial cells whose microbiota-driven activation is known to restrain inflammation and limit the oxygen availability that lets Enterobacteriaceae bloom. When the researchers grew B. pseudolongum in culture, its supernatant was rich in acetate but not butyrate or propionate, and that supernatant slowed NC101 growth at every dilution they tested. That last experiment is the cleanest evidence that the bacterium is not merely correlated with a calmer gut but is producing something that directly disadvantages the pathogen.

The finding also converges with independent work: a separate group has reported that B. pseudolongum protects the intestinal barrier in colitic mice through a PPAR-γ pathway, which is a reassuring sign that this is not a one-lab artifact.

What this means if you are managing IBD right now

The practical answer, for now, is that this changes nothing you should buy. B. pseudolongum is not the species in the Bifidobacterium products on pharmacy shelves, which are typically B. longum, B. breve, B. lactis, or B. bifidum, and no commercial product has been tested against pks+ E. coli in a person. Taking a bifidobacteria supplement is unlikely to hurt, but presenting it as a way to suppress a colibactin-producing strain would be running well ahead of the evidence.

What is reasonable is to know the concept exists. If you have IBD, particularly with a family or personal history of colorectal polyps, colibactin-producing E. coli is a legitimate thing to ask a gastroenterologist about, along with whether your surveillance schedule accounts for your risk profile. Research testing whether specific commensals can be used deliberately against specific pathobionts is now active enough that it is worth tracking.

The honest limits: mouse colons, short experiments, one strain

The constraints are real and the authors are direct about them. This is an acute chemical colitis model in mice, which reproduces mucosal injury but not the chronic relapsing course of human ulcerative colitis. The animals were pre-treated with antibiotics to make colonization possible, so the starting community was already artificial. Group sizes of 9 to 13 are normal for this kind of work but small, and the evidence that acetate is the operative molecule is described by the authors themselves as preliminary in vitro support rather than a settled mechanism, since they did not knock out acetate production and repeat the experiment.

The study was funded by the Canadian Institutes of Health Research (grant PJT-196003), with the authors declaring no commercial or financial relationships, which matters in a probiotics literature where manufacturer involvement is common.

What comes next for pathobiont-targeted probiotics

The obvious follow-ups are a chronic colitis model, an acetate-deficient mutant to test the mechanism directly, and eventually human strains and human guts. None of that happens quickly. A first-in-human trial of a defined commensal against a defined pathobiont is realistically years out, and would have to answer whether a strain that works in an antibiotic-emptied mouse gut can establish itself in a full human one.

Common Questions

Is Bifidobacterium pseudolongum available as a supplement? Not in any meaningful sense. It is a common commensal in animals and is present in some human guts, but it is not a standard ingredient in commercial probiotics, and the study used a specific reference strain, DSM 20099, under controlled dosing.

Does this mean butyrate is overrated? No. It means that in this particular model, acetate was the short-chain fatty acid that rose and tracked with protection. Acetate is also a substrate that butyrate-producing bacteria feed on, so the two are linked rather than in competition.

How would I know if I carry pks+ E. coli? Standard stool testing does not report it, and there is no routine clinical assay for the pks island. It is currently a research measurement, not something you can order and act on.

Is putrescine something I should take? No. It was used here as a research tool delivered in drinking water at controlled doses in mice, and polyamine supplementation in humans has not been established as safe or effective for colitis.

References

1. Ayodeji Samuel Ajayi, Gabriela Fragoso, Annie Calvé, et al. (2026). Expansion resistance against genotoxic E. coli: protective effects of Bifidobacterium pseudolongum in DSS-induced colitis in mice. Frontiers in Microbiology. https://doi.org/10.3389/fmicb.2026.1877186
2. Nilmara de Oliveira Alves Brito, Guillaume Dalmasso, Darja Nikitina, et al. (2023). The Colibactin-Producing Escherichia coli alters the tumor microenvironment to immunosuppressive lipid overload facilitating colorectal cancer progression and chemoresistance. https://doi.org/10.1101/2023.03.13.523827
3. Mariana X. Byndloss, Erin E. Olsan, Fabian Rivera-Chávez, et al. (2017). Microbiota-activated PPAR-γ signaling inhibits dysbiotic Enterobacteriaceae expansion. Science. https://doi.org/10.1126/science.aam9949
4. Weiling Guo, Bingyong Mao, Shumao Cui, et al. (2022). Protective Effects of a Novel Probiotic Bifidobacterium pseudolongum on the Intestinal Barrier of Colitis Mice via Modulating the Pparγ/STAT3 Pathway and Intestinal Microbiota. Foods. https://doi.org/10.3390/foods11111551