br Results br Discussion FMRP plays a major

Results
Discussion FMRP plays a major role in the regulation of translation at the synapses. Several works have tried to scan for possible targets of FMRP that may explain the broad neurocognitive phenotype of FXS. Although many key neuronal proteins were found to be regulated by FMRP, there are still much data missing to explain the molecular pathology. In this work, we have chosen a different approach; rather than searching for FMRP targets, we have analyzed the global transcriptomic changes that occur in the absence of FMRP and linked these changes to the molecular phenotype of the disease. The strength of our model system enabled us to look at these changes at different developmental stages. FMRP may be expressed in human ESCs that carry a full mutation (Colak et al., 2014; Eiges et al., 2007). In this sense, the FXS-derived iPSCs are different, as the FMRP locus was found by us and others to be epigenetically resistant to the process of reprogramming (Alisch et al., 2013; Urbach et al., 2010). With this in mind, we have set to analyze the differences between FXS-derived and control iPSCs. We demonstrated that FXS and control derived iPSCs are highly similar, at the undifferentiated state (Figures 1A, 1B, and S1A). As FXS patients suffer from a neural pathology, we speculated that neurons derived from FXS-iPSCs would exhibit significant differences when compared with neurons derived from control iPSCs. We indeed found a large number of Z-IETD-FMK to be differentially expressed between the two groups (Figures 1C and S1A). The downregulated genes in FXS are mostly those related to neuron differentiation, axonogenesis, and the axon guidance pathway (Figures 1C and 1D). This result is in line with several other studies showing aberrations in neural development, abnormalities of dendritic spine morphologies, and deformities of growth cone development affecting axon guidance in the formation of FXS neurons (Antar et al., 2006; Bassell and Warren, 2008; Callan et al., 2010; Castrén et al., 2005; Comery et al., 1997; Egger et al., 2008; Irwin et al., 2001; Li et al., 2009; Luo et al., 2010; Tessier and Broadie, 2008). Many of the genes that take part in these cellular and molecular processes were found to be regulated by REST. In fact, different studies aimed to identify REST target genes indicated that REST is involved in processes such as nervous system development, neuron projection, and axonal guidance signaling. Some of these studies have also suggested REST to play a part in glutamate receptor signaling (Bruce et al., 2004; Johnson et al., 2007; Satoh et al., 2013). This finding may in fact connect aberrant regulation of REST to the abnormal activity of the glutamate receptor signaling seen in FXS neurons (Dölen et al., 2007) or may cause an additive effect. It is becoming clear from recent studies that REST is a key regulator of proper neural differentiation and development, and any perturbation in the regulation of REST will eventually lead to abnormalities in creating the neural network (Ballas et al., 2005; Covey et al., 2012; Paquette et al., 2000). Constitutive expression of REST in differentiating neurons was found to disrupt neural gene expression and caused significantly higher frequencies of axon guidance errors but did not prevent neurogenesis (Paquette et al., 2000). As we could not detect differences in REST levels in NPCs, we believe that the abnormal regulation seen in FXS occurs at the stage of mature neuronal development and network formation (Figures S2A and S2B). In this sense, the molecular phenotypes seen in FXS-derived neurons mimic the molecular phenotypes seen in neurons expressing higher levels of REST and reinforce our finding of aberrant REST regulation in FXS cells. As FMRP is involved in the miRNA machinery, we sought to explore the possibility that the regulation of REST by FMRP is mediated by miRNA levels in our iPSCs derived neurons. Global analysis of miRNAs expression in FXS cells and control cells revealed that at 4-fold cutoff we identify only miRNAs that are downregulated (and not upregulated) in FXS-derived neurons (Figure 3B). This finding may point to an important role for FMRP in the regulation of specific neural miRNAs. Of the candidate miRNAs identified, hsa-mir-382 was found to be enriched in the brain (Mor et al., 2013) and harbors two binding sites on the REST transcript (Figure 3C). Indeed, genetic manipulation of hsa-mir-382 affected REST expression in both iPSCs and FXS-derived neurons (Figures 3D and 4A–4C). Most importantly, overexpression of hsa-mir-382 was able to significantly upregulate the levels of the REST target axon guidance genes in FXS-derived neurons (Figure 4D). The specific role of FMRP in the maturation and function of neural miRNAs should be further studied in the future, as miRNAs are major posttranscriptional regulators affecting the levels of proteins, which are critical for proper neural differentiation and synaptic function.