Large-scale integrative analysis of fecal samples sequencing data from healthy and CRC patients supports novel miRNA-mediated host-microbiota interactions

Introduction Dysregulated microRNA (miRNA) profiles were repeatedly reported in studies on Colorectal Cancer (CRC) tissues. More recently, analyses of stool from CRC patients have clearly shown that altered miRNA levels reflect the presence of neoplastic lesions. Moreover, studies in model organisms reported that miRNAs released into the gut lumen, and detectable in stool, can modulate microbial gene expression. However, a comprehensive investigation of miRNA-microbiota interactions in humans is still lacking. Combined profiling of fecal miRNAs and microbial species in a large population of CRC patients and healthy individuals could provide novel insights into host-microbiota interactions dysregulated in this tumor. Materials and methods An integrative analysis of fecal shotgun metagenomic sequencing and small RNA-Seq (sRNA-Seq) data of 1,052 subjects from eight European cohorts was performed. The study included 332 CRC patients, 87 subjects with precancerous lesions, 165 with non-tumoral gastrointestinal diseases, and 468 healthy controls. Microbial profiling was performed using MetaPhlAn 4.1 while a Docker4Seq pipeline was applied for fecal miRNA quantification. Diversity analysis of miRNA and microbial profiles (by Richness, Shannon, Inverse Simpson, Evenness indices) across study groups was performed with an investigation of significant and consistent miRNA-microbial associations across cohorts. Results From the fecal miRNome/microbiome characterization, a significant decrease (p <0.05) of miRNA heterogeneity was observed from healthy to CRC, while the microbial diversity increased in CRC was driven by the oral-to-gut introgression of oral species. 216 detectable miRNAs across cohorts were confirmed to be transcriptionally active in the intestinal tissue and some of them were dysregulated in CRC tissue, as also reported in public datasets. Analysing miRNA levels with respect to abundances of 461 commonly detected microbial species, 2,223 significant and coherent miRNA-microbial associations across cohorts were identified. Despite the observed associations were predominantly positive and detected in CRC (n=916, 47%), 780 associations (35%) were uniquely detected in controls while only 53 (3%) showed opposite direction among controls and CRC. The associations involved oral species including the previously reported miR-1246 and Fusobacterium nucleatum association. The associations changed from healthy to late-stage CRC, with only 3% of them shared among main disease categories. Conclusions Our findings indicate that gut microbial composition is associated with fecal miRNA profiles. Most of the associations are specific to CRC and may reflect alterations in both the fecal miRNome and microbiome. To conclude, these data support the role of miRNAs as potential mediators of host-gut microbiota interactions.