Microbiota-host interaction in colorectal cancer: emerging computational technology, multi-omics integration, and mechanisms

Yinghong Lu; Jun Yu

doi:10.20892/j.issn.2095-3941.2025.0762

Article Figures & Data

Figures

Tables

Download figure
Open in new tab
Download powerpoint
Figure 1
Microbiome and host interaction in CRC. pks⁺ Escherichia coli: Produces the genotoxin colibactin, which induces DNA double-strand breaks and genomic instability. Enterotoxigenic Bacteroides fragilis (ETBF): Secretes BFT, which contributes to ROS generation and induces DNA damage; BFT also disrupts E-cadherin-mediated cell adhesion, activating β-catenin signaling. ETBF downregulates the tumor-suppressive microRNA, miR-149-3p. Fusobacterium nucleatum: Utilizes virulence factors (FadA and Fap2) to bind host E-cadherin and Gal-GalNAc, respectively. F. nucleatum promotes TNFSF9 gene expression, upregulates the lncRNA ENO1-IT1, and suppresses the m6A “writer” enzyme (METTL3) via the YAP signaling pathway. Other bacteria promote cholesterol biosynthesis. Commensal-derived metabolites, such as butyrate, function as HDAC inhibitors, while microbial LPS activates TLR4 and NF-κB signaling, collectively fostering an inflammatory and pro-tumorigenic microenvironment. BTF, Bacteroides fragilis toxin; CRC, colorectal cancer; Gal-GalNAc, galactose-N-acetyl-d-galactosamine; HDAC, histone deacetylase; LPS, lipopolysaccharide; ROS, reactive oxygen species; TLR4, Toll-like receptor 4.
Download figure
Open in new tab
Download powerpoint
Figure 2
Overview of the multi-omics integration method to reveal host-microbiota interaction. Principal coordinates analysis is a fundamental linear method that transforms the original high-dimensional data into a set of orthogonal axes capturing major sources of variation. Hierarchical clustering organizes microbiota samples into dendrograms through pairwise distance metrics, such as Bray-Curtis or Jensen-Shannon divergence. The Spearman rank correlation is widely used due to robustness of the non-normal distributions typical of microbial data. Random forest models can handle high-dimensional data and provide measures of feature importance, ranking microbes by the predictive power. Network analysis provides a systems-level framework for integrating multi-omics data by representing complex microbiota-host interactions as unified graphs.

Tables

Figures

View popup

Table 1

Overview of the application of multi-omics technologies and the main findings in the CRC-related microbiome research

Project	Sample type	Sample size	Sequencing	Integration strategy	Main findings
Dziubańska-Kusibab et al.¹⁹	Human CRC tissue	n = 1372	Whole exome seq, whole genome seq	Correlation, Mann–Whitney U-tests	Mutation rates in AAWWTT motifs were enriched compared to all other WWWWWW motifs in CRCs with pks⁺ E. coli infection.
Kadosh et al.²⁰	Mouse CRC tissue; stool	n = 41	RNA-seq; ChIP-seq; 16S rRNA seq	One-sided student’s t-test	The gut microbiome switches mutant p53 from tumor-suppressive-to-oncogenic.
Chen et al.²¹	Human CRC tissue; stool	n = 40	Metagenomic seq; RNA-seq	PCoA, Pearson correlation, network	KRAS mutations affect the intratumoral colonization of ETBF in CRC through the miR3655/SURF6/IRF7/IFNβ axis.
Zhu et al.²²	Human and mouse CRC tissue	n = 278	16S rRNA seq; targeted gene seq	Correlation, clustering	F. nucleatum promotes tumor progression in KRAS p.G12D-mutant CRC by binding to DHX15.
Li et al.²³	Mouse CRC tissue; stool	n = 426	16s rRNA seq; RNA-seq	Mediation analysis; consensus clustering, network	Mediation analysis revealed gut microbiome as mediators partially exerting the effect of SNP UNC3869242 within Duox2 on colorectal tumor susceptibility.
Zou et al.²⁴	Human CRC tissue; stool	n = 41	Metagenomic DNA seq; exome DNA seq; RNA seq	Spearman correlation, clustering	F. nucleatum modified the tumor immune environment by TNFSF9 gene expression.
Galeano Niño et al.²⁵	Human CRC tissue	n = 11	16S rRNA seq; 10x Visium spatial single-cell RNA seq	PCoA, clustering, Spearman correlation	Bacteria localize within specific intratumoral microniches characterized by the upregulation of immunosuppressive pathways, and a specific microorganism, including Fusobacterium and Treponema, are predominantly associated with epithelial and macrophage cell types, driving transcriptional changes linked to metastasis and inflammation.
Hong et al.²⁶	Human CRC tissue	n = 289	RNA-seq; ChIP-seq	Spearman correlation	F. nucleatum activated lncRNA ENO1-IT1 transcription via upregulating the binding efficiency of transcription factor, SP1, to the promoter region of lncRNA ENO1-IT1.
Yu et al.²⁷	Human CRC tissue	n = 296	RNA-seq	Cox regression, clustering	F. nucleatum targeted TLR4 and MYD88 innate immune signaling and specific microRNAs to activate the autophagy pathway and alter colorectal cancer chemotherapeutic response.
Ansari et al.²⁸	Mouse CRC tissue	n = 15	Whole-genome bisulfite seq; ATAC-seq; RNA-seq; ChIP-seq	Hidden Markov modelling	Exposure to commensal microbiota induced localized DNA methylation changes at regulatory elements, which are TET2/3-dependent.
Xia et al.²⁹	Human CRC tissue	n = 33	Methylated-DNA capture seq	Zero-inflated negative binomial regression, Spearman correlation	F. nucleatum and H. hathewayi upregulated DNA methyltransferase. H. hathewayi inoculation also promoted colonic epithelial cell proliferation in germ-free and conventional mice.
Liu et al.³⁰	Human CRC tissue; stool	n = 24	16s rRNA seq; shotgun metagenomic seq; whole-genome bisulfite seq	Lasso penalized regression, network	Gut microbiota and pathogenic bacteria in dynamically shaping DNA methylation patterns, impacting physiologic homeostasis, and contributing to CRC tumorigenesis.
Chen et al.³¹	CRC cell line	n = 2	m6A seq, RNA-seq	Clustering	F. nucleatum induces a dramatic decline of m6A modifications in CRC cells and PDX tissues by downregulation of an m6A methyltransferase METTL3, contributing to induction of CRC aggressiveness.
Gao et al.³²	Human CRC stool	n = 225	Metagenomic seq; mass spectrometry	PCoA, clustering, random forest, Spearman correlation, network	CRC-associated metabolites were linked to cross-cohort gut microbiome signatures of the disease.
Xu et al.³³	Human CRC tissue	n = 30	16S rRNA; RNA-seq	Spearman correlation, clustering	Intestinal microbiota can affect CRC progression through arginine catabolism.
Chen et al.³⁴	Human CRC tissue	n = 905	Single cell RNA seq; RNA seq; 16S rRNA seq; metagenomic seq	PCoA, clustering, random forest, Spearman correlation	Host urea cycle metabolism is significantly activated during colorectal tumorigenesis, accompanied by the absence of beneficial bacteria with ureolytic capacity, such as Bifidobacterium, and the overabundance of pathogenic bacteria lacking ureolytic function.
Liu et al.³⁵	CCSCs	n = 3	RNA-seq	Clustering	F. nucleatum directly manipulates colorectal cancer cell fate and reveals the mechanism of lipid droplet-mediated Numb degradation for activating Notch signaling.

CCSCs, colorectal cancer stem-like cells; CRC, colorectal cancer; ETBF, enterotoxigenic Bacteroides fragilis; m6A, mRNA N⁶-methyladenosine; PCoA, principal coordinates analysis; PDX, patient-derived xenograft.