Genome and piliome analysis of E. coli isolates from human colorectal cancer biopsies

Colorectal cancer (CRC) has been linked to specific members of the gut microbiota, including adherent-invasive Escherichia coli (AIEC), yet the strain-level features that underpin pathogenicity and clinical relevance remain incompletely defined. Here, we isolated E. coli directly from tumour and adjacent healthy colorectal biopsies of six CRC patients and performed whole-genome sequencing and functional phenotyping. Eight genetically distinct E. coli strains were identified, spanning phylogroups A and B2 and multiple pandemic sequence types, revealing substantial intra- and inter-patient heterogeneity. Only a subset of isolates displayed AIEC-like adhesion and invasion phenotypes. One isolate with increased recovery from tumour tissue exhibited pronounced cytotoxicity and harboured strain-specific oncogenic risk factors, including the pks colibactin genomic island and a high-virulence allele of the type 1 pilus adhesin FimH. In-depth analysis of fimH revealed intra-patient allelic variation, indicating that single representative isolates may not fully capture virulence potential within individual hosts. More broadly, comparative genomic analyses showed that chaperone-usher pilus repertoires and previously proposed AIEC and CRC biomarkers ( chuA, eefC, fitA ) were primarily associated with phylogroup rather than tumour-tropic colonisation. Together, these findings demonstrate that CRC-associated E. coli constitute a diverse population lacking uniform genetic signatures. Our results highlight the importance of strain-resolved genomic and functional profiling to better understand bacterial contributions to colorectal cancer.