FUS binds to DNA as well as to

FUS binds to DNA as well as to RNA and regulates the expression of many transcripts in multiple steps of the gene regulation process (Dormann and Haass, 2013). In this study, we used FALS MPCs efficiently oriented into neuronal lineages including HB9-positive motor neurons to observe altered gene regulation in FALS-derived cytotoxicity and to discover an early disease-related diagnostic marker. FUS has been shown to associate with RNA polymerase II and TFIID, thereby participating in the general transcriptional regulation process (Bertolotti et al., 1996). In addition, FUS binds to TBP and TFIID to repress transcription by RNAPIII, suggesting that FUS controls the cross-regulation between RNA polymerases (Tan and Manley, 2010). Recent studies, however, have shown that the recruitment of FUS proteins to promoter regions with lncRNA represses transcription (Wang et al., 2008). Furthermore, FUS binding to the antisense RNA transcribed by RNAPIII from promoter regions downregulates transcription (Ishigaki et al., 2012). These past studies suggest that FUS regulates transcriptional repression by various mechanisms in specific target genes. This finding could reflect our observations from our microarray assay that 78% (124 out of 159 genes) of the significantly changed genes in transcript levels are upregulated in FALS MPCs. Furthermore, with more specific analysis that focused on FUS direct targets that we defined by using the CLIP-seq dataset, we found that 95.6% of genes (of 23 genes) are upregulated in FALS. In one proteomics study, FUS/TLS proteins were also identified as general splicing factors, which may be an early stage of the splicing process (Hartmuth et al., 2002). Our exon array analysis also revealed aberrant gene splicing events in RSU1, RPH3AL, and EFCAB13 genes in the FALS patients. Of these, aberrant RNA processing of RPH3AL was also confirmed by our isogenic FUSH517D/H517D lines, suggesting the direct effects of FUS H517D mutation. However, in the other two alternative exons, we did not detect significant differences between control and isogenic FUSH517D/H517D lines, suggesting that FUS H517D mutation is not likely to have direct effects on their splicing regulation. RPH3AL is an associated protein of RAB3A (Haynes et al., 2001) and RAB27A (Fukuda, 2003) and regulates exocytosis in dense-core granules from endocrine cells (Haynes et al., 2001). In addition, mutant RPH3AL is mis-localized throughout the cytosol, whereas WT-RPH3AL is localized in the distal portion of the neurites (Fukuda et al., 2004). To date, there has been no reports of functional analyses of products from each alternative splicing event in these three genes. This may be involved in the pathological features of FUS-mediated pathologies and may also be useful as early diagnostic markers for ALS. Our iPSC model for FALS may thus represent a useful tool for observing gene expression levels in the motor neuron lineage, as gene expression analysis in motor progenitor cells oriented into HB9-positive motor neurons cannot be achieved using conventional disease models. We also addressed whether pathological features observed in other ALS models are observed in our in vitro FALS model. Motor neurons in patients with sporadic ALS express abundant unedited GLUA2 forms in the AMPA receptor subunit (Kawahara et al., 2004; Kwak et al., 2010), but the unedited GLUA2 form in FALS MPCs and neurons derived from iPSCs was not detected (Figure S6). ER stress is increasingly recognized as an important pathway leading to cell death in animal and cellular disease models based on mutant SOD1 (Atkin et al., 2006; Saxena et al., 2009), and mutant FUS proteins would also be predicted to induce ER stress and to interact with protein disulfide isomerase, similar to mutant SOD1 (Farg et al., 2012). In this study, however, the expression of ER stress-related genes (BIP, CHOP, spliced XBP1, CASP4, and ASK1) almost did not change in FALS and FUSH517D/H517D MPCs and neurons (Figure S7).