Clinical utility of a STAT3-regulated microRNA-200 family signature with prognostic potential in early gastric cancer

Purpose: The majority of gastric cancer (GC) patients are diagnosed with late stage disease, for which distinct molecular subtypes have been identified that are potentially amenable to targeted therapies. However, there exists no molecular classification system with prognostic power for early stage GC (EGC) because the molecular events promoting GC initiation remain ill-defined. Experimental Design: miRNA microarrays were performed on gastric tissue from the gp130^F/F preclinical EGC mouse model, prior to tumor initiation. Computation prediction algorithms were performed on multiple datasets and independent GC patient cohorts. Quantitative real-time PCR expression profiling was undertaken in gp130^F/F-based mouse strains and human GC cells genetically-engineered for suppressed activation of the oncogenic latent transcription factor, STAT3. Human GC cells with modulated expression of the miR-200 family member, miR-429, were also assessed for their proliferative response. Results: Increased expression of miR-200 family members is associated with both tumor initiation in a STAT3-dependent manner in gp130^F/F mice, and EGC (i.e. Stage IA) in patient cohorts. Over-expression of miR-429 also elicited contrasting pro- and anti-proliferative responses in human GC cells dependent upon their cellular histological subtype. We also identified a miR-200 family-regulated 15-gene signature which integrates multiple key current indicators of EGC, namely tumor invasion depth, differentiation, histology and stage, and provides superior predictive power for overall survival compared to each EGC indicator alone.  Conclusions: Collectively, our discovery of a STAT3-regulated, miR-200 family-associated gene signature specific for EGC, with predictive power, provides a molecular rationale to classify and stratify EGC patients for endoscopic treatment.

http://ift.tt/2EDRijV