Microbiome spectra and prevalent colibactin-associated mutational process in Japanese colorectal cancer

Abstract The recent rapid increase in colorectal cancer (CRC) cases, particularly among young patients, poses a significant public health concern in Japan and other countries. The association between the gut microbiota and CRC has been well established. Here, we present a detailed analysis of this association by combining whole-genome sequencing (WGS) and whole-transcriptome data from CRC tissues with whole-genome metagenomic sequencing (WGMS) of fecal samples. The utilization of explanatory artificial intelligence analysis enabled classification of the disease into four distinct subtypes. This microbiome-based subclassification was found to be correlated with specific clinical and molecular characteristics, and included a subgroup having a higher proportion of younger patients with an unfavorable prognosis. The WGS analysis revealed mutational signatures (SBS88 and ID18) associated with colibactin exposure; these were considered early clonal events and found in 70.3% of patients (97 out of 138), with higher prevalence in the younger subtype. APC hotspot mutation and frameshift deletion, which matched the colibactin signature and the colibactin-specific motif, were observed in 12 patients (8.7%). CRCs exhibiting colibactin-associated mutational signatures were more common in younger patients and had an immunosuppressive microenvironment. Colibactin exposure is a highly prevalent but modifiable risk factor for CRC in the Japanese population.