As a first step to assess
As a first step to assess the cellular composition in EZ spheres, we performed PCR analysis for a variety of pluripotent and region specific markers (Fig. 3A). At passage 12, 21.8 iPSC-derived EZ spheres showed expression of a number of early neural progenitor (nestin, SOX1, Dach1) and posterior markers (HOXB4, GBX2), but limited expression of more anterior markers (FOXG1, OTX2) suggesting EZ spheres may also undergo regional specification during extended culturing as has been shown by others (Elkabetz et al., 2008; Koch et al., 2009). To more thoroughly assess this question, we examined gene expression by PCR at every passage out to 32 using a set of regional identity markers: neural ectoderm (SOX2, nestin, PAX6), anterior (FOXG1, OTX2), posterior (GBX2), dorsal (PAX7), neural crest (HNK1), and undifferentiated stem purchase propyl (OCT4). Representative PCR gels for undifferentiated colonies and EZ spheres at passages 0, 9, 18, and 27 (Fig. 3B) show basal expression levels among the lines and passages (a compiled table indicating relative expression at all passages is presented in Table S2). OCT4 expression was detected only in early passages and was silenced in all lines consistent with the immunocytochemistry results (Figs. 2A, 3B). Nestin, SOX2, and HNK1 were consistently expressed in all three lines through late passages, indicating maintenance of a neural progenitor cell population amenable to both central and peripheral nervous system subtypes (Fig. 3B, Table S2). Nestin was detected in undifferentiated colonies in all lines, which is consistent with previously reported data (Keil et al., 2012). PCR analysis for region specific markers showed fluctuations in gene expression levels over successive passages while maintained in EGF/FGF-2 growth conditions. For example, OTX2 was expressed at high levels in nearly every passage of H9 hESC- and 4.2 iPSC-derived EZ spheres, but not in 21.8 iPSC-derived EZ spheres (Fig. 3B, Table S2). On the other hand, PAX6, PAX7, FOXG1, and GBX2 were detected only sporadically at early and late passages in all lines (Fig. 3B, Table S2). Taken together, these data highlight that EZ spheres maintained long-term in EGF and FGF-2 are dynamic across passages, but that there is not a selective growth advantage for one regional progenitor cell over another. Additionally, gene expression patterns appear to be influenced by inherent properties of each cell line, similar to previously reported results (Boulting et al., 2011), rather than the culture conditions. Columnar neuroepithelial-like cells identified within rosette structures are a hallmark of neural progenitor cells and are found transiently in EBs or monolayer differentiation protocols during lineage restriction of hESCs toward neural subtypes (Chambers et al., 2009; Zhang et al., 2001). We therefore tested whether cells in EZ spheres were representative of a pre-rosette neural stem cell population that could be extensively expanded. We examined rosette markers by immunocytochemistry using intact spheres at multiple EZ sphere passages (13, 16, 21, and 28). Sections through 21.8 iPSC EZ spheres grown in EGF and FGF-2 did not reveal immunolabeling for PLZF, ZO-1, or N-cadherin (Figs. 4A, B), which are all standard markers used to identify the apical orientation of rosettes (Abranches et al., 2009). However, following 7days of neural induction initiated by removing EGF and FGF-2 and supplementing with N2 and BDNF, rosettes characterized by a central lumen and radially disposed neuroepithelial cells appeared and robustly expressed PLZF in conjunction with apical rosette markers ZO-1, PAR3, β-catenin, and N-cadherin (Figs. 4C–J). Importantly, rosettes were also detected in whole sections through 4.2 iPSC EZ spheres as well as upon plate-down of H9 hESC EZ spheres following 1week of neural differentiation (Fig. S2). These data indicate that neural stem cells within EZ spheres are captured in a self-renewing, pre-rosette stage, which precedes the long-term neural stem cell (LT-hESNSC) stage described by Koch and colleagues (Koch et al., 2009).