Identification of low miR-105 expression as a novel poor prognostic predictor for human glioma.

Glioma is the most common and aggressive brain tumor with poor clinical outcome. Identification and development of new biomarkers could be beneficial for diagnosis and prognosis of glioma patients. Recent studies have showed evidences that dysregulation of microRNAs (miRNAs) is involved in glioma tumorigenesis. Therefore, we attempted to identify specific miRNAs as prognostic and predictive markers for glioma. We statistically compared expression profile of 365 miRNAs between WHO grade IV and grade III gliomas, by qRT-PCR. MiR-105 was identified as a remarkably decreased miRNA in grade IV gliomas compared with grade III gliomas (P=0.012, fold change =0.04). We subsequently examined its expression levels in an independent series of gliomas, and statistically analyzed the associations between miR-105 expression and clinicopathological characteristics and survivals of these glioma patients. MiR-105 showed remarkably decreased expression in gliomas as compared to non-neoplastic brains. And grade IV gliomas had significantly lower miR-105 expression compared with grade III and II gliomas (both P<0.001). Additionally, low miR-105 expression was statistically associated with advanced tumor grade, advanced patient's age and low pre-operative Karnofsky performance score (all P<0.001). Furthermore, patients with low miR-105 expression had significantly poorer survival by Kaplan-Meier method (P<0.001). Multivariate analysis indicated miR-105 as an independent prognostic indicator for glioma patients (P=0.018, risk ratio =4.2). Our results suggested that low expression of miR-105 may correlate with unfavorable clinical outcome and be involved in tumorigenesis and aggressive progression of glioma. And miR-105 may be a novel biomarker in prognostic prediction for glioma.