Colorectal cancer is often described as a disease written into the genome, but the authors of a new study argue that some of its most revealing lines may have been penned by microbes living in the intestine. In research published in Nature Genetics, Shiba, Yachida, Mizutani and colleagues investigated how frequently colibactin-associated mutations occur in Japanese colorectal cancers, when those mutations arise during tumor development, and how they relate to the microbial communities found in affected tissues. Their analysis focuses on a chemical weapon produced by certain strains of Escherichia coli—a molecule that can damage DNA and leave behind a recognizable mutational fingerprint.
Colibactin is not a conventional toxin that simply kills cells. It is a genotoxic compound produced by bacteria carrying a cluster of genes known as the pks, or polyketide synthase, island. After colibactin reaches host cells, it can form covalent links between the two strands of DNA. These interstrand crosslinks obstruct replication and transcription, forcing cells to rely on repair pathways that can introduce permanent genetic errors. When a cell survives this damage while retaining the resulting mutations, the scars may become part of the evolutionary history of a tumor. Because the pattern is distinctive, researchers can search cancer genomes for evidence that colibactin was present even when the original bacteria are no longer detectable.
The study examines these scars through mutational signatures, statistical patterns created by classifying DNA changes according to the bases affected and their surrounding sequence context. Colibactin exposure is associated particularly with a single-base substitution signature commonly called SBS88, as well as an insertion–deletion pattern known as ID18. Rather than identifying one mutation in isolation, signature analysis measures a broader distribution of substitutions and small deletions across the genome. This approach allows scientists to distinguish damage caused by colibactin from alterations associated with other processes, such as defective mismatch repair, tobacco exposure or spontaneous DNA replication errors.
A central question for the Japanese research team was prevalence. The presence of a colibactin signature in a tumor does not necessarily mean that the bacterium is still living in the tumor or that exposure occurred immediately before diagnosis. Mutations are durable records, while microbial populations can change rapidly in response to diet, inflammation, medication, age and cancer treatment. By examining colorectal cancer genomes from Japanese patients and comparing their mutational profiles with microbiome information, the researchers sought to determine how broadly colibactin-associated processes are represented in this population. The work places Japanese colorectal cancer within a growing international effort to understand why microbial mutagens appear in some tumors but not others.
The chronology of the mutations is especially important. Tumors do not emerge in a single step; they evolve as abnormal cells acquire successive genetic and epigenetic changes. If colibactin-associated mutations are concentrated in the earliest branches of a tumor’s evolutionary tree, the findings would suggest that exposure can occur before malignant transformation or during the initial expansion of precancerous cells. If the signature appears mainly in later, genetically distinct subclones, it would point to continuing or renewed exposure after the tumor has already formed. By comparing mutations shared across tumor regions with those confined to individual branches, the investigators reconstruct when colibactin-related damage entered the cancer’s history.
This evolutionary perspective may help explain why a temporary microbial encounter can have a permanent effect. A person does not need to carry the same colibactin-producing bacteria throughout life for the exposure to matter. A short period of colonization could damage DNA in an intestinal stem cell, and the mutation-bearing cell could then persist as the tissue renews itself. If that clone gains additional advantages—such as altered control of cell division, resistance to cell death or improved ability to evade the immune system—it may become the seed of a tumor. The microbial exposure may disappear, but the genomic record remains embedded in every descendant of the affected cell.
The researchers also examined what the study describes as the microbiome spectra associated with colibactin-related mutational processes. The word “spectra” refers to the characteristic microbial patterns that accompany different mutational states, rather than to a single bacterium acting alone. Tumors with evidence of colibactin damage may be associated with distinct combinations of bacterial taxa, ecological relationships or functional gene pathways. Such associations do not automatically prove that every detected microbe caused DNA damage. The microbiome is an interacting community, and one organism may create conditions that allow another to expand, compete with it or alter its access to the intestinal lining. Nevertheless, linking microbial composition with tumor-genome signatures can reveal ecological contexts in which genotoxic bacteria are more likely to matter.
The findings carry implications beyond cataloguing cancer mutations. If colibactin exposure contributes to the earliest stages of colorectal tumor formation, screening strategies might eventually combine genomic signature detection with microbial or metabolite measurements. A stool test could, in principle, look for pks-positive bacteria, while tumor or precancerous tissue could be examined for SBS88 and ID18. These approaches would answer different questions: microbial tests would indicate current or recent potential exposure, whereas mutational signatures would reveal whether DNA damage had already occurred. Neither would by itself establish an individual’s future cancer risk, but together they could provide a more complete picture of the interaction between microbial ecology and tumor evolution.
The study also underscores the limits of treating the gut microbiome as simply “good” or “bad.” Most intestinal bacteria are harmless or beneficial under ordinary conditions, and even the presence of a pks-containing strain does not guarantee that colibactin will reach vulnerable cells in sufficient quantities to cause detectable mutations. Barrier integrity, inflammation, bacterial localization, immune responses and the duration of colonization may all influence risk. The Japanese data therefore add to a more precise model of microbiome-associated cancer: disease may depend not only on which microbes are present, but also on where they are, what genes they carry, how long they remain and when their chemical activity intersects with the changing genome of an intestinal cell.
By combining cancer genomics, evolutionary timing and microbiome analysis, Shiba and colleagues present colorectal tumors as records of both human biology and microbial chemistry. Their work suggests that some cancers carry evidence of an encounter between a bacterial metabolite and the DNA-repair machinery of a host cell—a molecular event that may have occurred years before diagnosis. The broader message is that cancer prevention may eventually require attention to the ecological communities living alongside human tissues, not just to inherited risk and lifestyle. Colibactin is one example of how a microscopic resident can leave a macroscopic mark, transforming a transient microbial exposure into a permanent signature in the genome.
Subject of Research: Colibactin-associated mutational processes, their timing, prevalence and relationship with the gut microbiome in Japanese colorectal cancer.
Article Title: Prevalence and chronology of colibactin-associated mutational processes and their microbiome spectra in Japanese colorectal cancer.
Article References: Shiba, S., Yachida, S., Mizutani, S. et al. Prevalence and chronology of colibactin-associated mutational processes and their microbiome spectra in Japanese colorectal cancer. Nat Genet (2026). https://doi.org/10.1038/s41588-026-02692-x
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s41588-026-02692-x
Keywords: colorectal cancer, colibactin, gut microbiome, bacterial genotoxins, mutational signatures, SBS88, ID18, Escherichia coli, cancer genomics, tumor evolution, DNA damage, Japanese cancer research
Tags: colibactin-induced DNA mutationscolorectal cancer microbiomeDNA damage from bacterial toxinsEscherichia coli genotoxinintestinal microbiota and cancerJapanese colorectal cancer genetic analysismicrobial influence on tumor geneticsmicrobiome signatures in colorectal cancermutational fingerprint of colibactinpks island and tumor developmenttiming of mutational events in cancertumor evolution and microbial interactions
