br Materials and methods br

Materials and methods
Results
Discussion Others and we have demonstrated that ESC-based developmental modeling mimics a variety of neurodevelopmental events occurring in vivo using either 2-D or 3-D systems. In those studies, the hESC-based developmental model has been proposed to imitate either the neural plate (Elkabetz et al., 2008; Curchoe et al., 2012), the neural tube (Colleoni et al., 2010; Banda et al., 2015), or the cerebral cortex (Shi et al., 2012; Lancaster et al., 2013; Paşca et al., 2015; Ziv et al., 2015). In this study, we showed that the ESC-based system recapitulated the progression from the neural plate, to the neural tube, and to the cerebral cortex. We also found that the 2-D rosette-based developmental system exhibits a comparable capability in modeling neural development to that shown by the 3-D human cortical sphere (hCS) (Paşca et al., 2015).Both systems were able to capture the temporal appearance of neural cells, for which the sequence from early to late stages is NSCs, radial glial cells, intermediate progenitors, neurons, and astrocytes. This was corroborated with studies in vivo (Freeman, 2010).Additionally, the spatial locations of neural buy GDC0068 showed a resemblance in both systems, including NSCs and radial glial cells in the rosette, intermediate progenitors at the outer-layer of rosettes, and neurons at the superficial level in the 3-D system. In the 3-D system, neurons were trapped within the sphere, while neurons ran freely in the dish in our 2-D system. We noticed that cut-out rosettes were better than uncut rosettes while modeling neurodevelopment. First, stage-specific markers exhibited a stage-relative enrichment in gene expression in the former, but not in the latter. Second, the former enriched the population of neural cells so that neural gene expression was enhanced. Third, gliogenesis occurred earlier in the latter, which suggested that the environment in the latter may provide regulatory elements needed in glial cell production. Fourth, although we were able to detect GAD1 expression, a marker for interneurons, we only found very few GAD1+ cells, suggesting that GAD1 gene may be activated without producing proteins. Interestingly, our paradigm generated primarily glutamatergic neurons, which corroborated with our finding that our modeling system recapitulated neurodevelopment in the cerebral cortex. In this study, we further established the neural niche residing in our ESC-based developmental system also recapitulated that in in vivo. We explored one of the crucial elements of neural niches, FGF2 signaling, by defining its stage-specific roles during neural development, which showed characteristics reminiscent of its role in vivo, suggesting that the NSC niche in neural rosettes may resemble that in the VZ of the embryonic cortex. We found that the predominate apical localization of FGF2 and FGFR1 in human neural rosettes that resembled those in the VZ of the embryonic cortex in mice by gradually decreasing and restricting to the end-feet or the first layer of apical regions, which echoed other observations made in vivo (Raballo et al., 2000; Frinchi et al., 2008; Guillemot and Zimmer, 2011). In the rosette system, the inhibition of FGF2 signaling with various inhibitors accelerated NSCs exiting the cell cycle and leading to neurogenesis. Moreover, the inhibition decreased the size of neural rosettes, which corroborated in-vivo studies that the loss of FGF signaling by various FGFR knockouts leads to a smaller cortex, a thinner proliferative area, decreased self-renewal of VZ precursors, and premature neuronal differentiation (Jukkola et al., 2006; Kang et al., 2009; Rash et al., 2011). We further identified the potential role of NUMB in FGF signaling inhibition-mediated neurogenesis, in which we found that FGF2 signaling inhibition enriched the apical location of NUMB. Moreover, the enrichment of NUMB resulted in the disappearance of PAX6+ NSCs and the emergence of MAP2+ neurons, suggesting NUMB is involved in neurogenesis promoted by FGF signaling inhibition. It is known that NUMB is a negative regulator of Notch, which controls numerous cell fate decisions during development by maintaining a proliferating and undifferentiated fate (Cayouette and Raff, 2002; Chapman et al., 2006). The NUMB action in promoting neurogenesis in our study may occur via inhibiting the Notch pathway, because overexpressing NUMB in cells promoted cell cycle exit (reducing Ki67+ cells) and resonated the finding that the interplay between FGFs and Notch signaling regulates neuronal differentiation (Faux et al., 2001). Strikingly, FGF2 treatments neither enlarged the size of neural rosettes nor sustained PAX6+ NSCs, rather converting cells into PAX6‐ cells that remained highly proliferating. They possibly shifted to the glial fate, because we noticed that FGF2 treatments promoted astrogenesis that coincided with the role of FGF2 in the induction of cortical progenitors to adopt an astroglial fate at the expense of neuronal fates (Morrow et al., 2001; Qian et al., 2000), and FGF2 knockout reduced the number of GFAP+ cells in cortex and striatum (Reuss et al., 2003; Chen et al., 2008).