Isolation and propagation of enteric neural crest progenitor cells from mouse embryonic stem cells and embryos

Abstract

Neural crest is a source of diverse cell types, including the peripheral nervous system. The transcription factor Sox10 is expressed throughout early neural crest. We exploited Sox10 reporter and selection markers created by homologous recombination to investigate the generation, maintenance and expansion of neural crest progenitors. Sox10-GFP-positive cells are produced transiently from mouse embryonic stem (ES) cells by treatment with retinoic acid in combination with Fgf8b and the cytokine leukaemia inhibitory factor (Lif). We found that expression of Sox10 can be maintained using noggin, Wnt3a, Lif and endothelin (NWLE). ES cell-derived Sox10-GFP-positive cells cultured in NWLE exhibit molecular markers of neural crest progenitors. They differentiate into peripheral neurons in vitro and are able to colonise the enteric network in organotypic gut cultures. Neural crest cells purified from embryos using the Sox10 reporter also survive in NWLE, but progressively succumb to differentiation. We therefore applied selection to eliminate differentiating cells. Sox10-selected cells could be clonally expanded, cryopreserved, and multiplied for over 50 days in adherent culture. They remained neurogenic in vitro and in foetal gut grafts. Generation of neural crest from mouse ES cells opens a new route to the identification and validation of determination factors. Furthermore, the ability to propagate undifferentiated progenitors creates an opportunity for experimental dissection of the stimuli and molecular circu that govern neural crest lineage progression. Finally, the demonstration of robust enteric neurogenesis provides a system for investigating and modelling cell therapeutic approaches to neurocristopathies such as Hirschsprung's disease.

Fig. 1

Characterisation of Sox10 reporter (S10G) and Sox10 constitutive expression (S10G-S) mouse ES cells. (A) Fluorescence image of E11.5 chimera produced by morula aggregation with S10G cells. (B) Flow cytometry analysis of Sox10-GFP-positive cells from embryoid bodies made with S10G cells. (C) Effect of Lif and Fgf8b on generation of Sox10-GFP-positive cells from embryoid bodies. Bars represent mean value of three independent plates. (D) Flow cytometry scatter plot for sorting Sox10-GFP-positive cells and bright-field images of sorted cells cultured overnight. (E) RT-PCR analysis of pluripotency and neural crest markers in undifferentiated ES cells, Sox10-GFP-negative, and Sox10-GFP-positive populations. (F) Sox10-GFP-positive purified cells cultured for 10 days in the presence of serum and immunostained for smooth muscle actin. (G) GFP-positive cells quantified by flow cytometry during differentiation of S10G-S cells and cells transfected with empty vector (Mock). RA was applied on day 2 at 10^–8 M. (H) FACS purification of Sox10-GFP-positive S10G-S cells from day 3 embryoid bodies and subsequent neurogenesis in serum-free N2B27 medium supplemented with Bmp4. Images show immunostaining for HuC/D. (I) Sox10-GFP-positive cells purified as in H differentiate into neurons upon culture in Bmp4 and Gdnf for 4 days. Fixed cells were immunostained for the indicated markers. Scale bars: 500 μm in H; in I, 200 μm left, 50 μm centre, 100 μm right-hand columns.

Fig. 2.

Maintenance of Sox10-GFP-positive neural crest progenitor cells in vitro. (A) Sox10-GFP-positive S10G-S cells cultured in neural crest cell (NCC) medium in the absence or presence of chick embryo extract (CEE) for 3 days. (B) Effect of cytokines on the maintenance of Sox10-GFP-positive cells. Sorted cells were cultured in NCC medium supplemented with the indicated combinations of cytokines for 3 days and analysed by flow cytometry. Bars represent mean ± s.d. of triplicate assays on three different S10G-S clones. The frequency of GFP-positive cells in NCC medium without any additive is ∼5%, 3 days after plating. Student's t-test was performed against untreated samples; ^*, P<0.05, with P-values greater than 0.05 shown. (C) Neuronal differentiation of purified Sox10-GFP-positive cells from S10G-S cells cultured in NCC medium containing noggin, Wnt3a, Lif and Et-3 for 7 days and subsequent culture in serum-free N2B27 medium containing Bmp4 and Gdnf for 7 days. Cells were immunostained for HuC/D and TuJ1. (D) Representative fluorescence images of injected E11.5 gut immunostained for DsRed and TuJ1. The boxed region is shown at higher magnification beneath. White arrows indicate DsRed (red) and TuJ1 (green) double-positive cells. Nuclei are stained with DAPI (blue). (E) Serial confocal images through a region of E11.5 gut graft. (F,G) Confocal images of E11.3 distal hindguts immunostained for DsRed (F) and merged DsRed with HuC/D (G). Right-hand panels show control hindgut cultured without grafting. Scale bars: 200 μm in C; in D, 500 μm above, 100 μm below.

Fig. 3.

Culture of embryo-derived neural crest cells in the presence of noggin, Wnt3a, Lif and Et-3. (A) Flow cytometric purification of neural crest cells from E10.5 S10G chimaera. (B) qRT-PCR analysis of marker mRNA expression in: S10G-S ES cells; Sox10-GFP-positive cells purified from S10G-S embryoid bodies and cultured in NWLE for 5 days; Sox10-GFP-positive cells purified from E10.5 S10G chimaeras and cultured in NWLE for 7 days. The highest level of expression of each gene was defined as 100%. Bars represent means from duplicate PCR reactions. (C) Phase-contrast and fluorescence images of chimaera-derived neural crest cells in NCC medium with or without NWLE for 5 days. (D-K) Expansion of embryo-derived neural crest cells with blasticidin selection for Sox10 expression. (D,E) Proportion (D) and proliferative index (E) of Sox10-GFP-positive cells 6 days after purification from E10.5 S10G-B chimaeras. GFP-positive cells were detected by flow cytometry. Bars represent mean ± s.d. of six independent wells. Proliferation index was determined by dividing the number of GFP-positive cells by the initial cell number plated. (F) Long-term culture and growth of Sox10-GFP-positive neural crest cells purified from chimeric embryos and cultured in NWLE with blasticidin. Bars represent total number of cells at indicated days and passage number. The proportions of GFP-positive and -negative cells, as quantified by flow cytometry, are shown by white and black boxes, respectively. (G) Metaphase spread from neural crest progenitor culture in NWLE plus blasticidin. (H) Phase-contrast image of culture at passage 5 expanded for over one month. (I) Flow cytometry analysis of Sox10-GFP and c-Ret in cultured neural crest cells in NWLE and 5 days after transfer to N2B27 with Bmp and Gdnf. (J,K) Neuronal and glial differentiation of neural crest cells after 8 passages (over 50 days) in NWLE. Cells were transferred to medium containing bFgf/Bmp4/Gdnf for 5 days then fixed and immunostained for HuC/D, TuJ1 and Gfap. Scale bars: 200 μm in H,J; 100 μm in K.

Fig. 4.

Clonal expansion and differentiation of Sox10-positive neural crest progenitors. (A) Colony derived from expansion of a single Sox10-GFP-positive cell purified by flow cytometry from E10.5 S10G-B chimaera. Cells were cultured in NWLE medium for 30 days with periodic blasticidin selection. (B) Phase-contrast and immunofluorescent images of clonal culture after transfer to adherent culture in (left to right) NWLE medium, bFgf/Bmp4/Gdnf, or TGFβ3. Scale bars: 200 μm in left-hand panels; 500 μm in centre and right-hand panels.

Fig. 5.

Grafting of culture-expanded embryo-derived neural crest cells into foetal gut ex vivo. (A,B) Immunostaining for DsRed and TuJ1 of guts grafted with Sox10-positive cells after culture in NWLE medium for 7 (A) or 30 (B) days. Right-hand images are confocal sections. (C) Aggregate formation from neural crest cells cultured in the presence of Bmp4/Gdnf. Merge of phase-contrast and fluorescence (DsRed) images. (D) Immunostaining for DsRed, HuC/D, c-Ret, Npy, Vip and Th in TuJ1-positive neurons differentiated in vitro from aggregates shown in C. (E) Migration of neural crest cells and extension of axon-like structure from grafted aggregate after 24 hours. Merged phase-contrast and fluorescence images. (F) Immunostaining for TuJ1 and DsRed showing contribution of grafted neural crest aggregate after 9 days. The boxed region is shown at higher magnification beneath. Arrows indicate DsRed (red) and TuJ1 (green) double-positive cells. (G-I) Immunostaining of neuronal network formed in the distal hindgut by grafted cells. Embryo-derived Sox10-GFP cells were cultured for 12 days in NWLE medium prior to grafting. Shown are a confocal image of DsRed immunostaining (G), a merged image of DsRed and TuJ1 immunostaining (H), and a high-magnification image of TuJ1 and HuC/D immunostaining (I). Scale bars: 100 μm in A,B; 200 μm in C,E; in D, 100 μm left and middle, 200 μm right; in F, 200 μm above, 100 μm below.