Colorectal cancers (CRC) have an important impact on both incidence and mortality worldwide, with varying rates in different parts of the world. Despite an overall decrease in incidence in the past years, an alarming increase in early onset colorectal cancer (CRC in patients under 50 years of age) rates has been observed. As part of the Cancer Research UK Grand Challenge Mutographs project, we aim to better understand the underlying mutagenic causes contributing to the differences in CRC incidence rates through mutational signatures.

We collected epidemiological data and performed whole genome sequencing and mutational signature analysis on 981 CRC tumor samples from 11 countries with varying incidence rates, including intermediate incidence regions (Iran, Colombia, Thailand, and Brazil) and higher incidence regions (Argentina, Russia, Canada, Poland, Czechia, Serbia, and Japan).

The average mutational profiles were similar across the countries. Mutational signatures associated with DNA repair deficiencies, including POLE and POLD1 associated signatures (SBS10a/b/c/d and SBS28), MUTYH (SBS36), NTHL1 (SBS30), homologous recombination deficiency (SBS3), and a plethora of microsatellite instability (MSI) signatures, were found in 177 cancers (18% of all samples) and at comparable levels across all countries. Multiple signatures with known etiologies were found in 802 DNA repair proficient colorectal cancers including clock-like (SBS1 and SBS5), reactive oxygen species (SBS18), APOBEC (SBS2 and SBS13), and the microbiome-product colibactin (SBS88) signatures. The colibactin signature SBS88 was observed in 14% of samples overall, being more prevalent in distal and rectum tumors. SBS88 was significantly enriched in younger patients (33% <40y, 23% 40-49y, 20% 50-59y,10% 60-69y, 10% ≥70y; p-value = 0.0001). Tumors from early onset patients also showed higher prevalence of another mutational signature with unknown etiology (19% <40y, 16% 40-49y, 13% 50-59y, 6% 60-69y, 7% ≥70y, p-value = 0.005).

These results indicate a potential association between SBS88 with early onset CRC, suggesting that the recent increase in incidence may be at least partially explained by the genotoxic compound colibactin, and associated bacteria.