Lack of microRNA-101 causes E-cadherin expression and function deregulation through EZH2 up-regulation in gastric cancer. Homo sapiens

Background & Aims: E-cadherin expression disruption is commonly observed in epithelial cancers and metastasis. Such event is also recognized as a crucial step in gastric cancer (GC) initiation and progression. As aberrant expression of microRNAs often perturb the normal expression and function of pivotal cancer-related genes, we characterized and dissected a pathway initiated by loss of microRNA-101 that causes E-cadherin dysfunction through upregulation of EZH2 expression in GC. Methods: Microarrays were used to profile the expression of microRNAs in human GC. Array-CGH revealed DNA copy number changes that were validated by genomic quantitative PCR and Snapshot. Expression levels of microRNAs, mRNA and protein were determined by quantitative-real time PCR and western-blot/co-immunofluorescence. CDH1 inactivating mechanisms were analyzed. Gain and loss of function experiments were done in KatoIII cells. E-cadherin functionality was assessed by immunofluorescence and flow cytometry. Results: MiR-101 expression was significantly decreased in tumors in comparison with normal gastric mucosas (P<.0001). In 65% of the analyzed GC cases, miR-101 downregulation was caused by deletions and/or microdeletions at miR-101-2 locus. EZH2 overexpression and consequent loss/aberrant E-cadherin expression was found in concomitance with miR-101 downregulation in 41% of the analyzed GC cases. This occurred preferentially in cases retaining allele(s) untargeted by classical CDH1 inactivating mechanisms. MiR-101 gain of function experiments or direct inhibition of EZH2, led to a strong depletion of endogenous EZH2 protein and consequent rescue of functional E-cadherin at the cell membrane, mimicking results obtained with clinical GC samples. Conclusions: Deletions and/or microdeletions at miR-101-2 locus underlying mature miR-101 downregulation and consequent EZH2 overexpression represented a novel cascade of genetic events leading to E-cadherin disruption, preferentially affecting the intestinal-type of GC. Overall design: In the present study, 37 primary GCs and a pool of 10 normal gastric mucosas were used to acquire the expression of 703 known human mature miRNAs deposited in the Sanger miRBase Sequence Database, Release 10.0 (http://microrna.sanger.ac.uk), as well as 393 novel human miRNAs discovered through deep sequencing and validated by qRT-PCR and array profiling (NCode™ Human miRNA microarray probe set V3 from Invitrogen). The 10 normal gastric mucosa RNA samples were pooled to obtain the normal gastric reference sample, which was labelled and hybridised four times to check arrays' reproducibility.