Unlike certain cancers, there is currently no way to determine a personalized approach for therapy in CD. MicroRNAs (miRNAs), a class of small non-coding RNAs that regulate gene expression at the post-transcriptional level, are stably present in tissue and blood samples. Therefore, miRNAs have the potential to serve as non-invasive biomarkers. Our group recently showed that protein-coding gene expression patterns define 2 clinically distinct molecular signatures in non-inflamed colon tissue from a cohort of adult CD patients. One signature largely resembled the expression profile of a normal colon (colon-like), while the other displayed expression patterns characteristic of the ileum (ileum-like). We found that these 2 CD subtypes associated with clinical phenotypes, such as presence of rectal disease and need for post-surgical biological therapy. We hypothesized that miRNA expression profiles would stratify these same molecular phenotypes, and that specific miRNAs are fundamental in the distinction of these major CD subtypes. To test this hypothesis, we analyzed colonic miRNA expression data from the same adult cohort as our previous study.Methods:
We collected samples from macroscopically non-inflamed regions of resected colon tissue from 21 CD patients and 13 non-CD patients. CD patients had been previously stratified into 11 colon-like and 10 ileum-like cases based on mRNA expression. High-throughput sequencing of small RNA and total RNA isolated from these samples was performed. Hierarchical clustering and principal component analysis were used to cluster samples with similar miRNA expression profiles. Specific differentially expressed miRNAs in disease versus control or between colon-like and ileum-like samples were identified using the DESeq2 package. Potential master miRNA regulators disrupted in CD pathogenesis that contribute the most to shaping changes in gene expression were determined using miRhub, a bioinformatic strategy that integrates miRNA and mRNA information to identify miRNAs that contribute the most to shaping changes in gene expression.Results:
We found that CD patients formed identical clusters based on miRNA expression patterns as those previously observed using mRNA expression profiles. Several miRNAs, most notably miR-31-5p, miR-196b-5p, miR-194-1-5p and miR-215, contributed significantly to the segregation of patients with CD into 2 main molecular phenotypes. In particular, miR-31-5p exhibited the greatest difference in expression, with significantly increased expression in ileum-like samples with respect to colon-like samples (P = 9.8e−09), and was a major driver in differentiating ileum-like samples from colon-like samples. Finally, using RNA-seq data from the same patient samples, we identified miR-31-5p as a candidate master regulator of gene expression pathways associated with ileum-like CD, a CD phenotype that is associated with the use of anti-TNF in the post-operative setting.Conclusions:
Our results show for the first time that miRNA levels in colon tissue segregate patients into 2 clinically distinct forms of CD. Specifically, our findings suggest that miR-31-5p contributes most to the discrimination between ileum-like and colon-like phenotypes. miR-31-5p is secreted into the circulatory system, therefore it is possible that plasma miR-31-5p levels could serve as an effective biomarker of CD subtypes, potentially providing an important clinical diagnostic and prognostic indicator of disease phenotypes.