Cage environment beat genetics in shaping the gut microbiome, mouse study finds

If you or someone close to you is living with Crohn's disease or ulcerative colitis, you have probably been handed some version of this story: your genes loaded the gun, and something in the environment pulled the trigger. A mouse experiment published in Frontiers in Immunology puts unexpected weight on the second half of that sentence. When researchers deleted a gut gene tied to inflammatory bowel disease and then reshuffled the animals into new cages, the microbiomes sorted themselves by cage, not by genotype. Genetically identical mice ended up with different bacteria, different inflammation, and different disease severity based on little more than who they lived with.

Key Takeaways

• In these mice, the cage an animal lived in was significantly associated with gut microbiome composition (p = 0.001), while the disease-gene knockout itself was not (p = 0.162).
• After eight weeks of co-housing, the shared cage still drove microbiome differences (p = 0.003) and genotype still did not (p = 0.205) — and the two cages produced measurably different colitis severity in animals carrying the same mutation.
• The practical read: environment and household can outrank a single disease-risk gene in shaping a gut microbiome, which is part of why mouse microbiome studies so often disagree with one another.
• Headline caveat: this is a small mouse study (20 animals in the co-housing arm) at a single facility, and the paper reports statistical significance rather than a percentage of variance explained.

Adamdec1 knockout mice sorted by cage, not by genotype

The central result is that housing, not genetics, predicted which bacteria a mouse carried. Tomoko Kumagai, Silke Rath, Shuangqi Fan, Farooq Z. Rahman and Andrew M. Smith studied mice lacking Adamdec1, a gene encoding a metalloprotease (a protein-cutting enzyme) made almost exclusively in the gut and implicated in inflammatory bowel disease. In a first cohort of naive young animals, PERMANOVA — a statistical test that asks whether groups differ in overall community composition — found no genotype effect on the microbiome (p = 0.162), while cage (p = 0.001) and sibling status (p = 0.001) both came back significant.

That pattern held when the team looked harder. In a second group of 20 animals housed across six cages before any intervention, cage was again significant (p = 0.001) and genotype only borderline (p = 0.057) — and once the analysis accounted for cage, the genotype signal disappeared entirely (p = 0.315). The same fate met sibling status, which fell from p = 0.013 to p = 0.215 after cage adjustment. In other words, what looked like family resemblance was largely the cage that family shared. This mirrors an earlier finding that cage environment regulates gut microbiota independently of immune genotype in mice lacking Toll-like receptors, and a separate analysis in which cage identity and maternal lineage outweighed knockout genotype in predicting murine gut communities.

Why this reframes the genetics story for IBD patients

For anyone researching their own inflammatory bowel disease, the useful takeaway is that a known risk gene did not determine the microbiome — the surroundings did. More than 300 genetic risk loci have been mapped for IBD, and as the authors note, they still explain only a fraction of who gets the disease. This experiment offers one concrete reason why: the gene may set sensitivity to the environment rather than dictate an outcome on its own.

The disease data make that vivid. When the researchers induced colitis, mice missing Adamdec1 did not simply get sicker; they got sicker inconsistently. Weight loss among knockouts scattered far more widely than in wild-type animals, with a standard deviation of 4.69 g versus 2.20 g on day 9 (p = 0.008). After co-housing, knockout mice in one cage lost significantly more weight than knockout mice in the other (p = 0.047) despite being genetically identical. Cage I animals also showed more neutrophil (p = 0.003) and dendritic cell (p = 0.033) infiltration into the gut wall.

Co-housing 20 female mice to separate gene from environment

The methodological move that makes this credible is deliberate cage-shuffling. The team took 20 female mice — 8 wild-type and 12 Adamdec1-deficient — that had lived in six separate cages by litter and genotype for 10 weeks, then redistributed them into two larger cages that deliberately mixed genotypes and litters. Because mice eat each other's droppings, co-housing lets microbiomes cross-seed; if genotype truly shaped the community, the genotypes should have re-separated.

They did not. After eight weeks of co-housing, the animals clustered by their new cage (p = 0.003), not by genotype (p = 0.205) or sibling status (p = 0.426). The influence of the original cage, significant on its own (p = 0.004), vanished once the new cage was accounted for (p = 0.139). Microbial diversity shifted too: in Cage I, the Shannon index (a measure combining how many species are present and how evenly they are distributed) fell in both wild-type (p = 0.022) and knockout mice (p = 0.002). Communities were profiled by 16S rRNA gene sequencing — a method that identifies bacteria from a shared marker gene rather than growing them — on an Illumina MiSeq.

The taxonomic differences between the two cages were specific. Among knockout mice, Muribaculaceae made up 64.90% of the community in Cage I versus 56.86% in Cage II (p = 0.017), while Sutterellaceae ran 0.05% versus 2.04% (p = 0.004). The pathobiont Duncaniella muricolitica was enriched in the sicker Cage I knockouts at 3.12% versus 0.91% (p = 0.002) — but it was also elevated in Cage I wild-type mice, who stayed well. That detail matters: the same bacterium was harmful only against the knockout background.

What a cage effect means for your own microbiome

The reader-facing lesson is that a meaningful share of "your" microbiome is really your household's. Human work points the same direction: a landmark Nature analysis found that environment dominates over host genetics in shaping the human gut microbiota, with genetically unrelated people sharing a home showing notable microbiome similarity. Practically, that reframes what is worth attention — shared food, water, pets, home environment and cohabitants are not background noise.

What this study does not license is any specific intervention. No probiotic, diet or supplement was tested here, and nothing in this paper establishes that changing a household changes disease risk in people. If you have IBD, the honest question to bring to a gastroenterologist is not "which bacterium do I need" but whether environmental exposures are being considered alongside genetics at all.

Twenty mice at one facility: the real limits

The biggest limitation is scale and setting. The co-housing arm involved 20 female mice at a single facility on a single genetic background, so the particular bacteria involved may be specific to that animal room rather than generalisable. Notably, the paper reports significance tests only — it does not publish an R-squared value quantifying what share of microbiome variance cage explained versus genotype, so the honest claim is that cage was significant and genotype was not, not that cage explained some specific percentage.

Colitis was also chemically induced using dextran sodium sulfate, which models intestinal injury rather than reproducing spontaneous human IBD. And the taxa linked to severity are correlational: no bacterium was transferred into germ-free animals to test causation. The work was funded by the Medical Research Council (grant MR/R001901/1) and the NIHR University College London Hospitals Biomedical Research Centre (BRC784), with the authors declaring no commercial or financial conflicts.

Why mouse microbiome studies keep contradicting each other

The methodological moral is that cage is a confounder, and studies that ignore it can mistake housing for biology. If two labs run the same knockout and get opposite results, differing cage microbiota is a live explanation. Established guidelines for designing mouse microbiome experiments already recommend co-housing animals before an experiment, using littermate controls, mixing treatment groups within cages, and maximising the number of cages — advice this study supplies fresh evidence for. Expect the near-term follow-up to be causal: transferring the Cage I community, including Duncaniella muricolitica, into germ-free knockout mice to test whether those bacteria actually drive the worse colitis.

Common Questions

Does this mean genetics don't matter for IBD? No. The knockout mice were still more susceptible to colitis overall than wild-type mice, with far more variable weight loss (standard deviation 4.69 g versus 2.20 g, p = 0.008). The finding is that the gene set sensitivity while the environment determined how badly the disease played out.

How much of my microbiome comes from my household? This study did not measure that in humans and reported no percentage. Human research does show that people sharing a household have more similar gut microbiomes than genetics alone would predict, but no reliable single number applies to an individual.

Can I change my microbiome by changing my environment? The mice changed substantially after eight weeks of shared housing, and diversity measurably shifted. Whether deliberate environmental changes improve symptoms in people with IBD or IBS is untested here, and no specific intervention is supported by this paper.

Was any bacterium proven to cause the worse colitis? No. Duncaniella muricolitica was enriched in the sicker cage (3.12% versus 0.91%, p = 0.002), but it was also elevated in healthy wild-type mice in that same cage, so it is associated with severity in a susceptible host rather than proven to cause it.

References

Tomoko Kumagai, Silke Rath, Shuangqi Fan, Farooq Z. Rahman, Andrew M. Smith (2026). Environment shapes gut microbiome and determines susceptibility to DSS-colitis in Adamdec1-/- mice. Frontiers in Immunology. https://doi.org/10.3389/fimmu.2026.1824007