Mapping the gut bacteria linked to colorectal cancer
Colorectal cancer remains a leading cause of death worldwide, yet early detection methods often lack the sensitivity required to identify pre-cancerous lesions effectively. In a groundbreaking study(opens in new window) published in ‘Cell Host & Microbe’, researchers have identified a robust, generalisable microbial signature associated with the disease. This meta-analysis, supported by the EU-funded CartoHostBug project, re-examined nearly 6 800 gut microbiome profiles, offering unprecedented clarity on the link between gut bacteria and cancer. The study’s primary strength lies in its scale and innovative methodology. By aggregating data from 27 independent studies across Germany, the Netherlands and Switzerland, the research team overcame the limitations of small, inconsistent datasets that have plagued previous research. They developed advanced computational approaches and a machine learning classifier capable of integrating microbiome data generated via different sequencing methods. “The key tool is a machine learning algorithm that is trained to distinguish cancer from non-cancer microbiomes,” explains Georg Zeller, Visiting Team Leader at CartoHostBug project partner European Molecular Biology Laboratory (EMBL) in Heidelberg, Germany, in a recent news item(opens in new window). “It outputs a score of how ‘cancer-like’ a microbiome is.” By assigning a score to any gut microbiome sample, it reveals a consistent microbial signature that persists across diverse populations, geographies and age groups, including both early- and late-onset cases. “We can apply this to any existing human gut microbiome dataset, including from dietary intervention studies,” adds Zeller, who is also the study’s senior author and a professor at project partner Leiden University Medical Center in the Netherlands.
From stool to tissue
A critical finding was the correlation between stool-based signatures and microbes found directly within colorectal tumour tissue. The researchers analysed 906 intestinal tissue samples, discovering that microbes found in higher numbers in tumour tissue mirrored those detected in faecal samples. Notably, these cancer-associated microbes were present even in early-stage tumours, suggesting that microbiome changes occur early in disease development. However, the study also highlighted a current limitation. While the signature is strong for established cancer, detecting precancerous adenomas via stool samples remains challenging due to weaker microbial signals in these early lesions. “This limitation is important for future clinical translation,” comments co-author Michael Zimmermann, who is Group Leader at EMBL. “It suggests that more sensitive approaches, larger datasets, or combinations with other measurements may be needed before microbiome-based tools could contribute to the reliable detection of early precancerous lesions.”
Dietary fibre and bacterial insights
Diet emerged as a pivotal factor in modulating this microbial signature. The analysis revealed a clear inverse relationship between dietary fibre and the colorectal cancer microbiome score. Individuals with lower fibre consumption exhibited stronger cancer-like microbial patterns, while dietary intervention studies showed that increasing fibre intake could reduce these scores. The research also provided valuable insight into Fusobacterium, a bacterial group often linked to colorectal cancer. By analysing hundreds of bacterial genomes from this group, the team found that not all subspecies behave the same. Fusobacterium nucleatum subsp. Animalis was consistently more prevalent in cancer samples globally, whereas other subspecies showed geographically specific patterns. While the current classifier cannot yet be considered a diagnostic test, it established a vital reference framework for future clinical tools. The CartoHostBug (Functional cartography of intestinal host-microbiome interactions) project’s mission to map host-microbiome interactions at the cellular and molecular levels is proving essential for identifying reliable disease biomarkers. For more information, please see: CartoHostBug project
