In vitro differentiation of sox1-GFP knock-in ES cells. Sox1-GFP ES cells were differentiated as described in Materials and methods, fixed at the NP stage (a–i) or ND stage (j–l) and analyzed by immunocytofluorescence. Confocal microscopy at 40× magnification of representative fields is shown. Scale bar = 50 μm.

Isolation of sox1-GFP-expressing NPs using FACS. (a) Sox1-GFP ES cells were differentiated to the NP stage, and GFP⁺ and GFP⁻ cell populations were isolated by FACS. J1 ES cells differentiated the same way were used as negative control. (b) FACS purification of Sox1-GFP precursors was re-analyzed after FACS purification to confirm purity. The black line represents FACS profile of J1 cells as a negative control and the green line represents the samples analyzed. Sox1-GFP⁺ cells were fixed 2 h after sorting and analyzed by bright field and fluorescent microscopy (c, d) for validation of sorting efficiency. Scale bar = 50 μm. (e) Sox1GFP⁺ cells were fixed at different time points during in vitro differentiation [4 days after FACS (NP d4), 5 days of neuronal differentiation stage (ND d5) and 10 days of neuronal differentiation stage (ND d10)]. Numbers of GFP⁺ cells and Hoechst⁺ cells were counted on multiple random fields for each sample, and the proportion of GFP⁺ cells was calculated by dividing the number of GFP⁺ cells by the number of Hoechst⁺ cells. Each group represents an average of three samples from each independent experiment.

NPs can be efficiently purified and pluripotent stem cells removed by FACS. GFP⁺ and GFP⁻ cells were fixed 4 days after FACS and analyzed by immunocytofluorescence using antibodies against nestin (a–p) or SSEA1 (q–v). Shown are representative fields using confocal microscopy at 40×. Scale bar = 50 μm. (w) Total RNAs were prepared from GFP⁻ and GFP⁺ cells at the NP stage after FACS and analyzed by RT-PCR analysis.

FACS-purified NPs generate neural cell-enriched cell populations. (a–h) FACS-isolated GFP⁻ and GFP⁺ cells were fully differentiated, fixed at the ND stage and analyzed by immunocytofluorescence using antibodies against βIII-tubulin and GFAP. Slides were counterstained using Hoechst. Shown are representative fields using confocal microscopy at 40× magnification. Scale bar = 50 μm. (i) The numbers of βIII-tubulin⁺, GFAP⁺ and Hoechst⁺ cells from sox1-GFP⁻ and sox1-GFP⁺ cells were counted on multiple random fields per each sample, and the proportions of neurons and astrocytes were calculated by dividing the numbers by the total number of Hoechst⁺ cells. Each group represents an average of three samples from each independent experiment. For neuronal phenotype, anova revealed F = 27.240, p < 0.05. Fisher’s PLSD post-hoc analysis was performed with a significance level of 0.05; *indicates significant difference from sox-1GFP⁻ cells. For astrocytic phenotype, anova revealed F = 99.536, p < 0.05. Fisher’s PLSD post hoc analysis was performed with a significance level of 0.05 (n = 3); *indicates significant difference from sox1-GFP⁻ cells.

FACS-purified NPs can generate dopaminergic neurons. FACS-purified GFP⁺ cells were analyzed at the NP stage by immunocytofluorescence using an antibody against the early midbrain marker En-1 (a–c). FACS-purified GFP⁺ cells were fully differentiated and analyzed by immunocytofluorescence using antibodies against dopaminergic markers such as TH (d–f), DDC (g–i) and DAT (j–l). Shown are representative fields using confocal microscopy at 40× magnification. Scale bar = 50 μm. (m) Fully differentiated sox1-GFP⁺ cells and sox1-GFP⁻ cells were stained, using antibodies against βIII-tubulin and TH, and counterstained with Hoechst. The numbers of TH⁺, βIII-tubulin⁺ and Hoechst⁺ cells from sox1-GFP⁻ and sox1-GFP⁺ cells were counted in multiple random fields for each sample, and the proportions of DA neurons among the total number of neurons and total cells were calculated by dividing the numbers of DA neurons by the number of Hoechst⁺ cells. Each group represents an average of three samples from each independent experiment. For proportion of TH⁺ cells among neurons, anova revealed F = 4.566, p < 0.05. Fisher’s PLSD post-hoc analysis was performed with a significance level of 0.05; *indicates significant difference from sox1-GFP⁻ cells. For the proportion of TH⁺ cells among total cells, anova revealed F = 17.364, p < 0.05. Fisher’s PLSD post-hoc analysis was performed with a significance level of 0.05; *indicates significant difference from sox1-GFP⁻ cells. (n) sox1-GFP⁺ cells and sox1-GFP⁻ cells, after full differentiation, released dopamine in response to membrane depolarization. GFP⁺ cells, GFP⁻ cells, a mixture of GFP⁺/GFP⁻ cells and mouse E12.5-derived primary DA cells (VM) were differentiated (stage 5 day 10) and treated with 50 mm KCl for 30 min, then the medium was collected and analyzed for dopamine content by HPLC. anova revealed F = 5.459, p < 0.05. Fisher’s PLSD post-hoc analysis was performed with a significance level of 0.05; *indicates significant difference from GFP⁻ and VM cells.

FACS-purified NPs can generate other neural subtypes. FACS-purified GFP⁺ cells were fully differentiated and analyzed by immunocytofluorescence using antibodies against serotonin (a–c), GABA (d–f) and glutamate (g–i). Shown are representative fields using confocal microscopy at 40× magnification. Scale bar = 50 μm. (j) sox1-GFP⁺ cells were fixed after neuronal differentiation stage (ND d10) and numbers of 5HT⁺, GABA⁺, glutamate⁺ and Hoechst⁺ cells were counted in multiple random fields per each sample. The proportion of each phenotype was calculated by dividing the cell numbers by the number of Hoechst⁺ cells. Each group represents an average of three samples from each independent experiment.

Transplantation analysis of FACS-purified cells. GFP⁺ and GFP⁻ cells were transplanted into mice striatum 4 days after FACS (NP stage), and the animals were killed 8 weeks after transplantation for histological analysis. Representative low power images of NeuN immunohistochemistry of grafts from GFP⁺ (a) and GFP⁻ (b) cells in naïve mouse striatum. TH immunohistochemistry on grafts from GFP⁺ cells using 20× objective (c; scale bar = 50 μm) and 63× objective (d; scale bar = 20 μm). (e) TH immunohistochemistry on contralateral side of the graft. Scale bar = 20 μm. (f) Total graft volume in GFP⁺ versus GFP⁻ grafts. anova revealed F = 18.457, p < 0.005. Fisher’s PLSD post-hoc analysis was performed with a significance level of 0.005; *indicates significant difference from GFP⁻ cells. (g) DA neuronal density expressed by DA neuronal number per mm³ graft volume in GFP⁺ versus GFP⁻ grafts. anova revealed F = 4.914, p < 0.05. Fisher’s PLSD post-hoc analysis was performed with a significance level of 0.05; *indicates significant difference from sox1-GFP⁻ cells. (h, i) Representative confocal images of SSEA1 immunohistochem-istry in graft from GFP⁻ cells. Scale bar = 50 μm. (j, k) Representative confocal images of Ki67 immunohistochemistry in graft from GFP⁻ cells. Scale bar = 50 μm. (l-o) Representative confocal images of TH/DDC/DAT immunohistochemistry in grafts of GFP⁺ cells. Scale bar = 50 μm.