For a few years now, the idea that tumors carry their own private microbiomes has been one of the more provocative claims in cancer biology. Some high-profile papers reported bacterial DNA lurking inside all sorts of cancers, from breast to bone, and suggested those microbes might shape how tumors grow or respond to treatment. Other researchers pushed back hard, arguing that much of the signal was contamination. A large analysis published this week in Cell lands closer to the skeptics, with one striking exception.
The team, led by Anders Dohlman and colleagues, went back to the raw data with a deliberately paranoid method. They built a pipeline to strip out human DNA and carefully classify whatever microbial sequences remained, then benchmarked it against known sources of contamination. Then they ran it across 16,369 high-depth tumor whole genomes from the UK 100,000 Genomes Project. That is one of the largest and cleanest datasets of its kind, which matters when the whole debate hinges on telling real biology from lab noise.
When the signal mostly vanishes
After decontamination, the result was blunt. In most cancer types, the microbial signatures were indistinguishable from background. In other words, once you account for the contamination that creeps into sequencing, the supposed microbiomes of many tumors dissolve. That is a useful, deflating finding. It suggests that some earlier reports of near-universal tumor microbiomes were reading tea leaves in the noise.
But one group of tumors held up. In orodigestive cancers, meaning cancers of the mouth, throat and digestive tract, the microbes were real and abundant. These were not lone bacteria either. The authors found multi-kingdom polymicrobial communities: bacteria, fungi, viruses, archaea, and in some cases Trichomonas, a single-celled parasite you would more often associate with other parts of the body. Whole ecosystems, sitting inside the tumor.
The tumors that let microbes in
The communities were not random. They varied by tumor site and subtype, which is what you would expect if they reflect the local environment rather than stray contamination. More intriguing, colonization was higher in tumors with specific genetic features. Cancers that were microsatellite-instable, or that carried mutations in the DNA-copying enzymes POLE and POLD1, tended to harbor more microbes. Across orodigestive cancers, microbial load tracked with tumor mutation burden, the overall number of mutations a tumor has accumulated.
That link is the part worth sitting with. Tumors with broken DNA-repair machinery pile up mutations, and those same tumors appear more welcoming to microbial colonists. The paper does not claim the microbes are causing the mutations, or the other way around. It reports a correlation and connects the tumor microbiome to the host's genomic context. Which came first, and whether the microbes are passengers or players, remains open.
What this does and does not settle
A few caveats are worth stating plainly. This is a snapshot from sequencing data, not an experiment. It can show that microbes are present and that their abundance correlates with mutation load, but it cannot tell you whether they help a tumor grow, hinder it, or simply move in once the tissue is already damaged. The findings come from one national dataset, drawn largely from a UK population, so the exact communities may look different elsewhere. And the strength of the work, that aggressive decontamination, is also a limit. A conservative pipeline that errs toward calling things contamination will miss genuinely low-abundance microbes. Absence of a signal in breast or bone tumors here is not proof that nothing lives there.
Still, the shape of the answer is clarifying. The tumor microbiome, as a universal feature of cancer, does not survive careful scrutiny. But in the mouth and gut, where the body's tissues are already steeped in a dense resident flora, tumors genuinely do carry microbial communities, and those communities line up with the genetic damage inside the cells. For a field that has swung between hype and dismissal, that is a more honest map to work from. The next question is the hard one. In those orodigestive cancers, are the microbes doing anything at all, or just along for the ride.
Comments