Global characterisation of mutational signatures reveals bacterial toxin as a potential driver of early-onset colorectal cancer

In many countries, colorectal cancer incidence in people under the age of 50 has doubled in the last two decades. Despite this dramatic increase, the causes driving it have remained mostly uncovered.

One way to investigate the causes of cancer is to look for the genetic 'fingerprints' that DNA-damaging agents leave in tumour genomes. We know that environmental exposures such as UV light and lifestyle choices like tobacco smoke damage our DNA. However, they actually leave distinct patterns of damage in the genome, called mutational signatures, that we can use to track them using advanced computational methods.

By studying which signatures are more common in areas of the world with higher incidence rates, we can uncover the factors potentially responsible for the observed rise in the number of colorectal cancer cases among young adults worldwide. 

We collected and sequenced the genomes of 981 colorectal tumours from patients across eleven countries on four continents, representing a wide range of incidence rates.

We then performed differential signature analysis and found that multiple mutational signatures showed country-specific differences, suggesting that local environmental exposures may contribute to colorectal cancer risk. In most cases, however, the causes behind these signatures remain unknown.

One particular set of signatures stood out due to their striking correlation with the incidence rate of colorectal cancer. Furthermore, we discovered that these signatures were significantly enriched in early-onset cases, being over three times more common in patients under 40 compared to those over 70.

The signatures in question, termed SBS88 and ID18, are known to be caused by colibactin, a bacterial toxin produced by certain strains of E. coli which carry the pks pathogenicity island. Previous research has shown that mutations derived from colibactin occur mainly in the first decade of life. Our study now provides evidence that these early-life alterations may be contributing to an increased risk of colorectal cancer many years later.

As to how colibactin influences the risk of developing cancer, we have shown that colibactin-related signatures are associated with mutations in specific cancer driver genes, particularly APC, which is regarded as the ‘gatekeeper’ tumour suppressor gene in the colon. Inactivation of APC is a crucial early event in colorectal tumourigenesis, and our data show that colibactin-associated signatures are significantly more likely to generate certain hotspot mutations and small insertions and deletions (indels) in APC compared to other mutational processes. This provides mechanistic insight into how a microbial exposure early in life may initiate cancer decades later.

Overall, our work highlights the potential role of the microbiome in the development of colorectal cancer, specifically in young patients. While we have shown that there is an association between colibactin-related mutational signatures and early-onset cases, further research is necessary to establish causation and clarify the molecular mechanisms involved. Moving forward, the international collaboration established through team Mutographs to address the Cancer Grand Challenges unusual mutation patterns challenge will continue to explore these critical questions, aiming to uncover actionable insights to address the rising incidence of colorectal cancer among young adults globally.

Originally trained in engineering, I have focused my career on computational biology and bioinformatics applied to cancer research. I have a strong interest in genomics data analysis and AI health applications, which I developed during my PhD at the IDIBAPS Research Institute (Barcelona, Spain) and my postdoctoral training at the University of California San Diego (USA).

Currently, I lead the Digital Genomics group at the Spanish National Cancer Research Center (CNIO, Madrid, Spain). Our research focuses on developing new computational methodologies to identify the underlying mechanisms of cancer predisposition and progression based on specific signatures of somatic mutations. We are interested in applying these innovative methods to diverse cohorts of cancer patients to better understand the dynamics and evolution of tumours in different worldwide populations, as well as to develop next-generation genomics biomarkers to characterise tumour prognosis and treatment response.

I completed my PhD at the University of Cambridge working on the identification of driver mutations in blood cancers. After this I transitioned from wet lab to computational biology when I joined the Stratton group at the Wellcome Sanger Institute in 2018 to work on the Cancer Grand Challenge Mutographs project. I am currently a senior staff scientist, specialising in cancer genomics with a focus on mutational signature analysis. By leveraging large-scale sequencing data and computational methods, my work explores the patterns of DNA mutations across cancer genomes to uncover the underlying biological processes driving tumour development and evolution.