Neural differentiation of embryonic stem cells in vitro: a road map to neurogenesis in the embryo

Abstract

Background: The in vitro generation of neurons from embryonic stem (ES) cells is a promising approach to produce cells suitable for neural tissue repair and cell-based replacement therapies of the nervous system. Available methods to promote ES cell differentiation towards neural lineages attempt to replicate, in different ways, the multistep process of embryonic neural development. However, to achieve this aim in an efficient and reproducible way, a better knowledge of the cellular and molecular events that are involved in the process, from the initial specification of neuroepithelial progenitors to their terminal differentiation into neurons and glial cells, is required.

Methodology/principal findings: In this work, we characterize the main stages and transitions that occur when ES cells are driven into a neural fate, using an adherent monolayer culture system. We established improved conditions to routinely produce highly homogeneous cultures of neuroepithelial progenitors, which organize into neural tube-like rosettes when they acquire competence for neuronal production. Within rosettes, neuroepithelial progenitors display morphological and functional characteristics of their embryonic counterparts, namely, apico-basal polarity, active Notch signalling, and proper timing of production of neurons and glia. In order to characterize the global gene activity correlated with each particular stage of neural development, the full transcriptome of different cell populations that arise during the in vitro differentiation protocol was determined by microarray analysis. By using embryo-oriented criteria to cluster the differentially expressed genes, we define five gene expression signatures that correlate with successive stages in the path from ES cells to neurons. These include a gene signature for a primitive ectoderm-like stage that appears after ES cells enter differentiation, and three gene signatures for subsequent stages of neural progenitor development, from an early stage that follows neural induction to a final stage preceding terminal differentiation.

Conclusions/significance: Overall, our work confirms and extends the cellular and molecular parallels between monolayer ES cell neural differentiation and embryonic neural development, revealing in addition novel aspects of the genetic network underlying the multistep process that leads from uncommitted cells to differentiated neurons.

Figure 1. ES-cell derived NPs culture analysis.

A) Percentage of GFP⁺ cells in monolayer cultures grown for 6 days without replating in RHB-A and N2B27 media (* p-value = 0.005; ** p-value = 0.052). B) Fold increase (FI) for monolayer cultures grown for 6 days in RHB-A and N2B27 media. C) Semi-quantitative RT-PCR analysis for selected markers of pluripotency and lineage commitment in day 0–6 RHB-A cultures; mRNA from E10.5 mouse embryos was used as positive control. D) FI (filled squares) and viability (open squares) for RHB-A cultures maintained for 20 days in culture and replated every 4 days at the same initial cell density. E) Percentage of Sox1-GFP⁺ cells along 20 days in culture in RHB-A, with replating every 4 days. In all graphs data are means±SEM from at least three independent experiments. F) After replating in laminin (day 5), Sox1-GFP⁺ cells organize in rosettes, with N-Cadherin (in red) present at the centre of these cell clusters. G) ZO-1 accumulates in the cell processes that coalesce at the centre of rosettes, like it does in the apical domain of NPs in the embryonic neural tube (H). I) Anti-PAR3 immunostaining reveals well-defined “apical” domains at the centre of rosettes, where it co-localizes with ZO-1 (J). K) aPKC, another known apical marker is also present at the centre of rosettes and co-localizes with N-Cadherin. L) Adherent junctions' components, ß-catenin and N-Cadherin, co-localize at the central, apical region of rosettes. M) Anti-γ-tubulin staining (in green) shows “apically” localized centrosomes. N) Mitotic figures (ppH3) are localized centrally in rosettes while S-phase nuclei (BrdU) are located at the periphery. O) Differentiating Tuj1⁺ neurons accumulate at the periphery of rosettes. Nuclei counterstained with DAPI (blue). Scale bar: 50 µm.

Figure 2. Chemical inhibition of the Notch activity by γ-secretase inhibitor LY411575.

A) Expression of Notch pathway genes during monolayer ES cell differentiation, from day 0 to day 20, by RT-PCR analysis. mRNA from E10.5 mouse embryos was used as control. B) Detection by ISH of Dll1, hes5 and hes6 transcripts in control (DMSO-treated) and LY411575-treated rosette cultures. Treatment was done in day 6 cultures for 24 hours. Nuclei counterstained with DAPI. C) After 48 h of LY411575 treatment, starting at day 6, massive neuronal differentiation is observed by Tuj1 immunostaining. D) Notch inhibition with LY411575 at day 8 or 12 of the monolayer protocol results in increased neuronal production, detected by HuC/D imunostaining. No change was detected when inhibition was done in day 16 rosettes. Bars in D represent SEM for the minimum of three independent experiments. * p-value = 0.025; ** p-value = 0.002. Scale bars in B,C: 50 µm.

Figure 3. Timing of production of neurons and glia in rosette cultures.

A) Percentage of HuC/D⁺ and GFAP⁺ cells in rosette cultures, relative to the total number of cells in culture. A decrease in neuronal production is observed at day 16, concomitant with an increase of glial cells. B) Semi-quantitative RT-PCR data showing fold change of expression (relative to day 0) for neuronal (tau) and glial (gfap) markers at successive timepoints of rosette cultures. Data normalized to gapdh. C–H) Rosette cultures at day 8+3, 12+3 and 16+3, labelled with anti-HuC/D and anti-GFAP antibodies to visualize neurons and glial cells, respectively. Few GFAP⁺ cells appear in day 8+3 cultures (C), with the number increasing at day 12+3 (E) and 16+3 (G). In contrast, a decrease in the number of HuC/D⁺ neurons is detected at day 16+3 (H). Nuclei counterstained with DAPI (blue). Scale bars: 50 µm.

Figure 4. NS cell potential of the in vitro neuroepithelial rosette cultures.

A) Floating aggregates of NS cells derived from day 4 monolayer cultures (Sox1-GFP 46C cells; phase contrast and GFP fluorescence images). B) Efficiency of derivation of NS cells-derived floating aggregates from several rosette cultures time points (days 4, 8, 12, 16 and 20), expressed as number of aggregates formed per 1000 cells. Bars represent SEM for 3 independent experiments. C) Floating aggregates of NS cells (derived from day 4 monolayer cultures of 46C) were left to attach for 4 days onto laminin substrate and stained for Sox2 (NP marker), Doublecortin (DCX, neuronal marker), GFAP (glia) and O4 (oligodendrocytes). Scale bars: 50 µm.

Figure 5. Validation of microarrays results.

A) Histogram of sorted Sox1-GFP populations from day 3 monolayers. GFP negative (GFP-) cells were discarded, while two GFP positive populations were collected individually, according to their levels of GFP expression (GFP+ and GFP++). B) RT-PCR analysis of RNA samples collected for microarray analysis for the genes Oct4, nanog, hes5, and blbp. C) Fold changes, relative to day 0, obtained from Affymetrix profiling for the genes Oct4, nanog, hes5, and blbp.

Figure 6. Clustering analysis of differentially expressed genes.

A) Frequency distribution of the expression levels of the genes belonging to the five defined groups. B) Dendogram of the relationship of expression of genes belonging to each group (with biological replicates being represented by the letters A, B, C and D) and examples of genes that are present in the five defined groups. C) Schematic representation of the successive cellular states that occur along the path to neural differentiation (see text for definitions of stages).