Exploring transcriptomic regulation of the tissue-associated microbiome in oncogenic progression of colorectal cancer.

Abstract

3609 Background: The association between the tissue microbiome and colorectal cancer (CRC) etiology continues to be explored, and it has been proposed that certain bacteria drive oncogenesis. However, this field is hindered by a lack of mechanistic links. Considering the relationship between bacterial abundance and human gene expression allows us to traverse this functional gap. To date, work in this field has been largely restricted to considering the relationship between bacteria and Consensus Molecular Subtypes (CMS) of CRC. Here, we hypothesise that CRC-associated bacteria correlate with distinct expression profiles that drive CRC progression. Methods: Analysis was performed using BioCorteX’s knowledge graph and proprietary engines, v20250128_100953. 16S rRNA sequencing and whole RNA data were collated from tissue samples across 3 independent cohorts from CRC patients (n = 59) and healthy volunteers (n = 23). In CRC patients, paired tumour and normal adjacent samples were analysed yielding 108 samples in total. The Pearson correlation between differentially abundant (DA) OTUs and differentially expressed (DE) genes was calculated, and false discovery rates (FDR) were applied. Results: 14,460 DE genes (FDR < 0.05) were identified between tumour and normal adjacent tissues, 5,097 of which had an absolute log-fold change greater than 1. Seven OTUs were increased in CRC compared to normal adjacent (FDR < 0.05) and were present in at least 20% of samples; Fusobacterium nucleatum, Akkermansia muciniphila, and five Streptococcus species. Of the 61,164 OTU-gene pairs analysed, 108 were significant (FRD < 0.05) with a Pearson correlation greater than 0.5. The abundance of all five Streptococcus species was correlated with the expression of nine genes including: SPSB4 ( r = 0.51, p = 1.86x10-8), which has previously been associated with CRC metastasis, and REN ( r = 0.71, p = 1.21x10-17), which is associated with increased Wnt signaling. Moreover, the expression of LY6G6D, a recently discovered CRC-specific antigen, was positively associated with F. nucleatum abundance ( r = 0.53, p = 4.94x10-9). Conclusions: This study identifies novel interactions between the CRC-associated microbiome and expression profiles. The consistent relationships between five Streptococcus species and specific genes is highly indicative of a functional co-evolution between host transcriptomics and microbiome, and implies relationships have less to do with the specific bacteria and are more related to function. This analysis adds a mechanistic explanation for some of the associations previously identified between the microbiome and CRC. Further analysis is required to identify if the host transcriptome creates a niche for these species or whether the presence of these species is altering the microbiome. This study demonstrates the value of integrated multi-omic analysis in understanding CRC pathogenesis.