Identification of miR-34a as a prognostic factor and possible therapeutic target for Ewing's sarcoma

Author(s)：

F. Nakatani¹, H. Chuman¹, A. Kawai¹, U. Yamaguchi¹, Y. Tanzawa¹, E. Kobayashi¹, N. Setsu¹, K. Ogura¹, T. Fujiwara¹, D. Kubota¹, P. Picci ² and K. Scotlandi ²

Affiliation(s)：

¹⁾ Department of Musculoskeletal Surgery, National Cancer Center Hospital, Tokyo, Japan
²⁾ Laboratory of Experimental Oncology, IOR, Bologna, Italy

Clinical outcomes of Ewing's sarcoma were improved after the introduction of multimodal therapies including intensive chemotherapy. However current therapeutic approaches still leave us the problems such as short-term toxicities of chemotherapy, severe morbidity resulted from repetitive major surgeries as well as serious late adverse effects like second malignancy and fatal cardiac failure even for long survivors. Therefore, we are in the urgent need to develop the novel biomarkers for an appropriate stratification of patients at diagnosis to establish more effective and less invasive therapy for all Ewing's sarcoma patients. MicroRNAs (miRNAs) are newly recognized small non gene-coding RNA family which regulate gene expression through translational suppression. More recently, they have been increasingly accepted as playing crucial roles in pathogenesis of most human malignancy as well as normal cell differentiation. However, little is known regarding to the possible role of miRNAs on sarcoma including Ewing's sarcoma. Here we focus on the whole profiling of the miRNA expressions in Ewing's sarcoma for the aim to elucidate the novel miRNA biomarkers in the treatment of Ewing's sarcoma.
Randomly selected 34 primary biopsy samples of Ewing's sarcoma have been analyzed for complete miRNA expression profiles. All of the patients were treated at a single institute from 1991 to 2008 with multimodal therapies. Microarray and statistical analyses were carried out using Agilent's miRNA microarray kit (v2) and GeneSpring GX analysis software, respectively. To identify miRNAs associated with the prognosis of Ewing's sarcoma patients we first compared the expression profiles of 21 patients who relapsed within 5 years with 13 patients who remained free of disease throughout the follow-up. As a result, we identified 16 miRNAs differentially expressed between the two groups (fold change > 1.5, p<0.1). We have also performed univariate Cox regression analysis to elucidate the miRNAs, whose expression level are related to the length of event-free survival (EFS) and/or overall survival (OS) and found out 15 miRNAs and 32 miRNAs that significantly correlated with EFS and OS, respectively. We then validated these data with quantitative real time PCR using extended series of 50 primary biopsy samples. After the validation, we have confirmed that low expressions of miR-34a were strongly associated with worse prognosis (P<0.001, Logrank test) and remained most significant biomarker after multivariate analyses. Moreover expression levels of miR-34a using real time PCR were highly correlated with the level of corresponding microarray data (Pearson correlation test: p=0.00000716). Over-expression of miR-34a inhibited the colony formation in the soft agar and enhanced chemosensitivities of Ewing's sarcoma cell lines. In conclusion, we have identified miR-34a as a powerful biomarker to distinguish the prognosis of Ewing's sarcoma patients through the whole miRNA profiling analyses.