Meta-analysis of microRNAs expression in head and neck cancer: uncovering association with outcome and mechanisms

Head and neck squamous cell carcinoma (HNSCC) is often diagnosed at advanced stages, incurring significant high mortality and morbidity. This review explored the risk stratification of miRNAs, and investigated the impact of miRNA networking in HNSCC prognostication. We performed a meta-analysis and a systematic literature search on online databases for papers published prior to December 1, 2016. The list of miRNAs was uploaded to MetacoreTM to construct a protein-protein interaction network, which was used to identify targets of the miRNAs and potential drugs. In addition, a representative network was further validated by immunohistochemistry in a cohort of 100 patients. We found 116 studies that included 8,194 subjects, in which the relationship between miRNA expression and prognosis of HNSCC were analyzed. Significant elevated expressions of 27 miRNAs and decreased expression of 26 miRNAs were associated with poor outcome. After excluding the studies causing heterogeneity, a fixed model was applied, which showed a statistically significant association between increased expression of miR-21 and poor survival (Pooled HR = 1.81,95% CI = 0.66-2.95, P < 0.005). We identified four networks affected by the miRNAs expression and enriched in genes related to metabolic processes and regulation of cell mitogenesis in response to extracellular stimuli. One network point out to 16 miRNAs directly or indirectly involved in the regulation of androgen-receptor (AR). Evaluation of AR protein expression in our cohort revealed that patients with upregulation of AR had poor survival rates (log-rank test, P < 0.005). This study showed that miRNAs have potential prognostic value to serve as screening tool for HNSCC during the follow-up. In addition, the implementation of a network-based analysis may reveal proteins with potential to be used as a biomarker.