We chose to use the human GFAABC D promoter to

We chose to use the human GFAABC1D promoter to drive TagRFP in human astrocytes because of its accuracy to specifically direct reporter expression in astrocytes in several CNS regions, at embryonic and adult stage, in rodent (Lee et al., 2008). Interestingly, during the pre-screening stage, at 70 DIV, not all GFAP-expressing astrocytes co-expressed GFAP::TagRFP (Fig. 2C). This demonstrated that increasing specificity of reporter expression using the GFAABC1D promoter that lacks several regulatory elements resulted in the loss of large quantities of astrocytes, perhaps also revealing astrocyte heterogeneity within the culture. Interestingly, at day 45 of differentiation, no positive ambroxol hydrochloride were observed in cultures of any clones differentiated using the midbrain protocol. In marked contrast, positive cells were observed in cultures aged 45 DIV for several clones differentiated using the spinal cord protocol. Moreover, TagRFP expression was stronger in spinal cord astrocyte cultures than midbrain astrocytes, for the majority of the clones tested (data not shown). These findings were common to both hESC and hiPSC, and may reflect the dynamic expression of the promoter, in different part of the central nervous system. Similar to our observation was that of Gu and coworkers who showed that in wildtype mice, GFAP expression is much stronger in the spinal cord than in the brainstem (Gu et al., 2010). Nonetheless, even if TagRFP expression appeared to be more prominent and more robust in and between spinal cord clones compared to midbrain clones, purification of TagRFP-expressing astrocytes generated using either of the protocols resulted in homogeneous populations of astrocytes expressing canonical markers NF1A, CX43, GLAST, GS, S100β and CD44, We also observed morphological differences and varying staining intensity in Reelin between spinal and midbrain astrocyte cultures (data not shown). In several instances (especially observed for midbrain clones), the percentages of GFAP-expressing astrocytes and TagRFP-expressing cells in purified cultures aged 130 DIV were lower than the initial FACS-purified + astrocytes (Fig. 4A). This could be attributed to GFAP down-regulation in maturing astrocytes (Roybon et al., 2013). Moreover, the difference between reporter expression and endogenous GFAP can be attributed to half-lives of intermediate filament proteins in astrocytes which are most likely relatively short (Chiu and Goldman, 1984) compared to stable fluorescent proteins such as TagRFP. Nevertheless, the clones we selected (C32 for hESC and C70 for hiPSC) showed high overlap between GFAP and TagRFP, and their purification by FACS resulted in homogeneous cultures of astrocytes expressing canonical astroglial markers. Notably, astrocytes generated from the human ESC reporter line could be used in several assays. Interestingly, almost all cytokines and chemokines detected in the protein array study were also detected in a recent similar study employing human fetal astrocytes (Choi et al., 2014). In concordance with our data, Choi and coworkers found MCP-1 secretion reduced. In contrast, while our data indicated no change in MIF secretion between FBS and IL-1β treated cultures, they reported that the secretion of this anti-inflammatory cytokine decreased following a one-day treatment period with two cytokines: IL-1β and TNFα. This decrease in MIF secretion may only be detectable after short treatment, or attributed to the presence of TNFα. Nevertheless, all other cytokines and chemokines were commonly secreted in both studies. It is worth mentioning that our new models have additional uses, such as the tracking of pathogenic protein labeled with ATTO488 in astrocytes in unpurified cultures, where astrocytes generated from regionalized progenitors are labeled using GFAP::TagRFP. Finally, as a proof-of-principle, we showed that the reporter lines could also be generated using human iPSCs. The hiPSC reporter lines differentiated with the spinal cord and midbrain protocols generated astrocytes strongly resembled those generated from hESC, as previously reported for other hESC and hiPSC lines (Roybon et al., 2013). The hiPSC clone C70 generated astrocytes responding to IL-1β treatment, assessed by ELISA for IL-6 secretion, and able to take up α-synuclein protein (Fig. 7E).