miRNAs have been shown to specify cell fates in the nervous system in worms [13
] and brain morphogenesis in zebrafish [15
], and these factors, as well as their distinct expression patterns during mammalian brain development, suggest a role in neural differentiation in mammals [1
]. However, a functional role of miRNAs in mammalian neurogenesis has not been described yet. Here, we demonstrate that in vitro transient overexpression or inhibition of brain-specific miRNAs in ES cell–derived neural precursor cells significantly reduced differentiation along either glial or neuronal cell lineages and could alter the balance between neurogenesis and gliogenesis.
For in vitro neurogenesis, we used a previously published five-step mouse ES cell differentiation protocol, which resembles several steps of embryogenesis, such as EB formation and differentiation of NPs into neurons and astroglia [19
]. To determine miRNA expression profiles during ES cell differentiation, we performed miRNA array analysis and observed temporal expression of groups of miRNAs at specific stages of cell development. Since multiple miRNAs are co-induced during the NP to ND transition, they likely have multiple targets and pleurotropic combinatorial effects on differentiation. In addition, several miRNAs may cooperatively target the same mRNA, modulating expression of the encoded protein in a combinatorial way. These scenarios imply unlimited numbers of different regulatory combinations created by a network of co-expressed miRNAs that could contribute to a highly complex cell response. In this study, we have focused on five highly abundant miRNAs from those simultaneously induced in the transition from NP to ND cell stage. Among these miRNAs, miR-124a, and miR-9, which are expressed almost exclusively in the brain [12
], also show prominent effects on ES cell–derived neurogenesis. Early overexpression of these miRNAs in neural precursors reduced the number of GFAP+
cells (astrocytes) differentiated in culture. Inhibition of miR-9 expression alone or in combination with miR-124a caused a reduction of Tuj1+
cells (neurons). Since miR-124a is expressed preferentially in embryonic neurons, whereas miR-9 is expressed in both neurons and glia ([37
] and our unpublished data), our results suggest that early overexpression of miR-124a in NPs prevents gliogenesis, whereas miR-9 expression contributes to neurogenesis. It should be noted that overexpression or inhibition of these miRNAs could have also affected other cells or factors in the differentiation cultures. Although flow cytometry with specific antibodies against Tuj1 and GFAP can be used for quantitative studies to discriminate neurons from astrocytes [25
], GFAP-positive early neuronal (granule cell) and astrocytic neural precursors from the subventricular zone have been described [38
]. Therefore, the observed effects of the miRNAs in our studies might, in part, be on a common neuronal and glial precursor, which has not been committed to either lineage. In addition, it is interesting to note that another class of small RNA molecules, dsRNA, can specify fates of adult neural stem cells by acting at the level of transcription [41
]. It seems, therefore, that determination of the neural fate can be influenced by different types of small noncoding RNA modulators, which are involved in both transcriptional and post-transcriptional regulation.
The brain-specific miRNAs miR-9 and miR-124a, which we identified as mediators of neurogenesis, are conserved in mammals and associated with polysomes in the brain [10
], suggesting that they act, at least partially, at the level of translation. Hundreds of mRNA targets have been predicted computationally for these two miRNAs [42
]. Some of them seem to be related to neurogenesis and/or to gliagenesis; in particular, our results imply that miR-9 and miR-124a act in the STAT3 signaling pathway, targeting some STAT3 upstream factor or factors but not the STAT3 mRNA itself. Though activation of STAT3 promotes differentiation of astrocytes in the developing central nervous system [33
] and suppresses neurogenesis of neural stem cells [32
], there is no clear understanding of STAT3 signaling in ES cell neuro- and gliogenesis. Multiple upstream factors, including Janus kinases, cytokines (LIF and CNTF) and their receptors (GP130, LIFR, CNTFR) and functional partners, could affect STAT3 phosphorylation. To our knowledge, none of these molecules have been consistently predicted as a very potent target for miR-9 and/or miR-124a. Therefore, further experiments in combination with bioinformatics approaches will be required to determine the molecular mechanism of miR-9 and miR-124a function in ES cell– derived neural cell differentiation.