MiR-34a is dispensable for normal development but its loss accelerates medulloblastoma formation

Previous studies have evaluated the role of miRNAs in the initiation and progression of cancer. MiR-34a was found to be downregulated in several tumors, including medulloblastoma. We here analysed the function of miR-34a in vivo by targeted transgenesis to generate mice with constitutive deletion of the miR-34a gene, which resulted in the absence of mir-34a in all analysed tissues. Nevertheless, these mice were viable and fertile. A comprehensive standardized phenotypic analysis including more than 300 single parameters performed by the German Mouse Clinic revealed no apparent phenotype. Analysis of miR-34a expression in human medulloblastomas and medulloblastoma cell lines revealed significant downregulation as compared to human cerebellum. Re-expression of mir-34a in human medulloblastoma cells in vitro reduced cell viability, cell proliferation and induced apoptosis. Among the targets downregulated by miR-34a in human medulloblastoma cells were NMYC and SIRT1. Activation of the Shh pathway by targeted overexpression of SmoA1 causes medulloblastoma in mice, which is dependent on the presence and upregulation of NMYC. Analysis of miR-34a in ND2:SmoA1-derived medulloblastomas revealed significant suppression of miR-34a compared to normal cerebellum. Crossbreeding these mice with miR-34a knockout mice significantly accelerated medulloblastoma growth in mice deficient for miR-34a. Interestingly, NMYC and SIRT1 were highly expressed in medulloblastomas derived from these mice. We here demonstrate that miR-34a is dispensable for normal development, but that its loss accelerates medulloblastoma. Strategies aiming to re-express miR-34a in tumors could therefore represent an efficient therapy option. For genome-wide expression analysis total RNA from brain and thymus of three or four male miR34a and four control mice was isolated using RNeasy Midi kit (Qiagen, Hilden, Germany). The cDNA microarrays were generated, hybridized and analysed as described (Horsch et al 2009). Two chip hybridizations were performed with total RNA for each individual mutant mouse against a reference RNA pool of the same organ.