Approcci sperimentali e computazionali per l'investigazione dell'espressione di microRNA in relazione a malattie gastrointestinali

Gastrointestinal (GI) disorders have become one of the major health concerns since they show particularly high incidence in developed countries due to their correlation with Western dietary regimes and lifestyle. Among them CeD, inflammatory bowel diseases (IBD), diverticular diseases and polyps. Some of them are directly correlated to a high predisposition to the development of intestinal tumours, and a particularly high occurrence is observed in colorectal cancer (CRC) cases, associated with a relevant mortality rate. Currently, the diagnosis of GI pathologies is confirmed by means of invasive procedures after most of the symptomatology has already appeared. The design of new non-invasive screening strategies to monitor the onset of these conditions can anticipate diagnosis and improve prognosis. Biomarkers are molecules obtained in a relatively non-invasive manner that can be used for this purpose. Over the last few years, miRNAs have arisen as interesting biomarkers due to their high stability in surrogate tissue. Most strategies have assessed the expression of these short polynucleotides in blood circulation, in the attempt to correlate miRNA expression to specific pathologies. However, for GI diseases, stool samples can also provide reliable information due to their closer relationship with intestinal epithelium which continuously exfoliates and releases factors into the faecal stream. In this respect, the primary focus of this study was to assess stool miRNA profiles of diverse GI conditions to characterize pathology-linked molecules that can potentially be used as biomarkers. Firstly, small RNA-seq data analysis was performed on stool samples from CeD subjects with high transglutaminase antibody levels and compared to CeD patients with low antibody levels, revealing a peculiar miRNA signature associated with gluten-free diet adherence. Subsequently, with the aim to characterize all phases of CRC development, stool samples of GI diseases were analysed, trying to identify both common and conditions-specific molecular-printing. To support and to functionally understand this biomarker characterization, small RNA-seq and RNA-seq data from CRC tissues were integrated to computationally build a competing endogenous RNA (ceRNA) interaction network, providing a more complete model of gene expression regulation. Among all the differentially expressed miRNAs, 3 miRNAs revealed similar expression trends in CRC tumour samples and surrogate tissues obtained from subjects suffering from other GI disease: miR-148a-3p, miR-182-5p and miR-21-5p. The resulting miRNAs target relevant genes involved in cell cycle, apoptosis and inflammation, pathways that were found enriched when analysing stool differentially expressed miRNAs, suggesting a possible correlation between primary and surrogate tissue. Furthermore, the expression of these 3 miRNAs showed a certain degree of correlation to pathology severity. On the whole, the results presented in this thesis demonstrated a correlation between the expression of specific miRNAs measured in stool samples and the severity of GI disorders, suggesting that the miRNA profiling in surrogate tissue could become a powerful tool to monitor the progression of colorectal pathologies. Moreover, the construction of ceRNA interaction network added an extra layer of post-transcriptional regulation in CRC, adding insight to the critical players of tumour progression.