The signals of FGFs on the neurogenesis of embryonic stem cells
© Chen et al; licensee BioMed Central Ltd. 2010
Received: 28 December 2009
Accepted: 29 April 2010
Published: 29 April 2010
Skip to main content
© Chen et al; licensee BioMed Central Ltd. 2010
Received: 28 December 2009
Accepted: 29 April 2010
Published: 29 April 2010
Neural induction is a complex process and the detailed mechanism of FGF-induced neurogenesis remains unclear.
By using a serum-free neural induction method, we showed that FGF1 dose-dependently promoted the induction of Sox1/N-cadherin/nestin triple positive cells, which represent primitive neuroblasts, from mouse embryonic stem (ES) cells.
We demonstrated that FGF1, FGF2, and FGF4, but not FGF8b, enhanced this neurogenesis. Especially, FGF-enhanced neurogenesis is not mediated through the rescue of the apoptosis or the enhancement of the proliferation of Sox1+ cells. We further indicated that the inactivation of c-Jun N-terminal kinase-1 (JNK-1) and extracellular signal-related kinase-2 (ERK-2), but not p38 mitogen-activated protein kinase (MAPK), inhibited the neural formation through the inhibition of ES differentiation, but not through the formation of endomesodermal cells.
These lines of evidence delineated the roles of FGF downstream signals in the early neural differentiation of ES cells.
In the early gastrula of the chicken, temporary treatment of the primitive ectoderm with Hensen's node for 5 hours steers the ectoderm to become the neural fate [1, 2]. FGF was shown to be responsible for this instructive ability of node and for the maintenance of later neural instructive signals [3, 4]. FGF first activates ERNI during early gastrulation and consequently triggers the zinc-finger transcriptional activator, Churchill, and its downstream target Sip1 in late gastrulation . In Xenopus, the study of neural induction has revealed the essential role of Ras/MAPK activation for neurogenesis in uncommitted ectoderm and in dissociated animal cap cells, suggesting that the requirement of FGF signals in neural induction is conserved in chordates .
ES cells, which resemble epiblast cells in the blastocyst, provide an alternative approach to the study of early development in mammals [6, 7]. Several one-step neural induction models have been established. Trans-retinoic acid (RA), a pro-neural inducer, enriches the neural population in a serum-containing embryoid bodies (EBs) system [8, 9]. However, RA treatment has several drawbacks, including the caudalization of the neural fate, blockage of forebrain induction, and the disruption of normal embryogenesis [9–11]. Co-culture of ES cells with mouse skull-derived stromal cells, such as PA6 cells, or bone marrow-derived cells, such as MS5 cells, efficiently induces the ES cells to become neuron lineages [8, 12]. However, the factors contributing to this stromal-derived inducing activity are still uncharacterized. ES cells cultured in serum-free Neurobasal medium with N2B27 supplement efficiently differentiate into Sox1+ neural precursors, which represent the earliest committed neuroblast cells in the developing embryo [13, 14]. Specific neuronal subtypes, such as dopaminergic and serotoninergic neurons, are derived from the Sox1 neuroblasts by the addition of defined patterning factors. Although the Neurobasal/N2B27 model provides a simple monoculture differentiation system for ES cells, these cells often undergo apoptosis on days 3 to 5. Recently, an efficient neural-induction monoculture system with a high survival rate for differentiating ES cells was developed and termed as serum-free embryoid bodies formation (SFEB) method . This simple and reproducible system consists of defined components and is suitable for the exploration of downstream FGF signals in the early neurogenesis of mammals.
Sox1-GFP knock-in ES cells (46C), from Dr. Austin Smith (University of Cambridge, UK), and ESC 26 cells, were both well-characterized and germline transmissible [14, 16]. The culture condition of both cells [14, 16] and the SFEB method  has been described previously in detail.
Human recombinant FGF2, FGF4 and FGF8b were all from R&D Systems. Recombinant human FGF1 was prepared from Prof. Chiu in Institute of Cell and Systems Medicine, the National Health Research Institutes, Taiwan . Synthetic inhibitors of FGF signaling, including SU5402, LY294002, SB203580, and SP600125, were from Calbiochem; U0126 was purchased from Tocris.
The plasmid Flag-DsRedT4-NLS was a gift from Tim Shroeder at Helmholtz Center Munich, Institute of Stem Cell Research, Germany. The genes of JNK dominant negative mutants, Flag-JNK1a1apf and Flag-JNK2a2apf [18, 19], were obtained from Addgene http://www.addgene.org and fused with a IRES-DsRed as a reporter. The plasmids were transfected into ES cells with lipofectamine 2000 (Invitrogen). After selection with 0.4 mg/ml G418 for two weeks, stable clones with red fluorescence were picked up and maintained with 0.2 mg/ml G418. The selected ES cells showed normal ES cell morphology and pluripotent gene expression (data not shown).
Cells were fixed in 4% cold paraformaldehyde and permeabilized with 0.3% Triton-X 100. Immunocytochemistry was performed with the following primary antibodies: OCT3/4 (1:500, Santa Cruz), Nanog (1:100, Cosmo Bio, Japan), Sox2 (1:4000, Chemicon), N-cadherin (1:100, DSHB, Iowa), FGF receptor 1 (FGFR1) and FGFR3 (both 1:100, Santa Cruz), FGFR2 (1:500, Abcam) and GFP (1:1000, Aves Labs). Images of immunostaining were captured usinga fluorescent microscope (Nikon ECLIPSE 80I) or confocal microscope (LSM510 Meta, Zeiss).
Sox1-GFP ES cells were fully dissociated and analyzed with flow cytometry (FC500, Beckman Coulter). Apoptosis was measured by staining for Annexin V (AbD Serotec) at room temperature for 10 min in the dark.
Total RNA was isolated from ES cells using REzol™ C&T reagent (Protech technology, Taiwan). Primers were applied to detect the expression of FGFR1 (5'-CAC ACT GCC TTC TCC TCC TC-3', 5'-CTC TGC CTC CCT GTC TTC TG-3'), FGFR2 (5'-GGG GAT GTG GAG TTT GTC TG-3', 5'-GCT TCT TGG TCG TGG TCT TC-3'), FGFR3 (5'-CGG CTA CCT GTG AAG TGG AT-3', 5'-GCT TGG TCT GTG GGA CTG TT-3'), FGFR4 (5'-AGG AAA TGT GGC TGC TCT TG-3', 5'-GGT GTG TCC AGT AGG GTG CT-3'), Sox1 (5'-CCT CGG ATC TCT GGT CAA GT-3', 5'-TAC AGA GCC GGC AGT CAT AC-3'), and G3PDH (5'-GTG AAG GTC GGT GTG AAC G-3', 5'-GGT GAA GAC ACC AGT AGA CAC TC-3').
ES cells were lysed in RIPA buffer (50 mM Tris pH7.5, 150 mM NaCl, 10 mM EDTA, 1% NP-40, 0.1% SDS) plus a cocktail of proteinase inhibitors (Sigma-Aldrich). Denatured proteins were separated by 10% SDS-PAGE and then transferred to PVDF membranes. Samples were detected with antibodies to ERK1/2, phosphoERK1/2 (pERK1/2), p38 and pp38, JNKs and pJNKs, AKT and pAKT. All MAPK-related antibodies were from Cell Signals and diluted 1:1000 for immunoblotting. Chemiluminescence of immunoreactive bands was detected using secondary horseradish peroxidase-conjugated antibodies (Jackson ImmunoResearch) and ECL reagents (Amersham).
We next tested the effects of different FGFs on neural formation of ES cells. FGF1, FGF2, and FGF4 all showed significantly elevated neural induction in 46C cells (Fig. 2A). However, FGF8b, even at the high concentration of 80 ng/ml, failed to enhance the neural induction of ES cells (Fig. 2A). We further investigated the expression of FGFRs in ES cells during neural induction and found that the expression of FGFR4 gradually declined (Fig. 2B), which is in agreement with the finding that FGFR4 is excluded from the neuroectoderm of mouse embryos . In contrast, FGFR1, FGFR2, and FGFR3 expressions were significantly increased during the conversion of ES into neuroblast cells. Immunocytostaining revealed that both FGFR1 and FGFR3 were detected in cytosol and nuclei in neural derivatives (Fig. 2C). On day 6, GFP+ signals were colocalized with FGFR1- and FGFR3-expressing cells, suggesting that both signals may be involved in neurogenesis (Fig. 2C). RT-PCR and immunostaining, shown in Figs. 2B and 2C, indicated that the expression of FGFR2 in differentiating ES cells was robustly induced and was localized on the cell membrane and cytosol, rather than in the nucleus. We also found that FGFR2 was not completely coexpressed with the GFP in 46C cells on day 6 (Fig. 2C), suggesting that FGFR2 is involved in the formation of subtypes of neurons. Taken together, these results suggest that FGFR1 and FGFR3 are generally required for neural induction and FGF8b is incompetent on the enhancement of neurogenesis of ES cells.
The increase of Sox1+ cells in the FGF1-treated condition may result from enhanced proliferation and/or reduced apoptosis of neuroblast cells. To test these possibilities, FGF1 was incubated with the 46C cells, and the apoptosis and proliferation of Sox1+ cells were analyzed by staining of activated caspase-3 and Ki67, respectively. Double staining of cleaved caspase-3 and GFP revealed that less than 5% double positive cells were detected (Fig. 3B). Similar results were obtained in FGF1-treated Sox1+ cells (data not shown). The percentages of Ki67+ cells in Sox1+ population were 24.75% (196/792) and 25.48% (362/1421) in SFEB- and SFEB/FGF1-treated cells respectively (Fig. 3C and 3D), demonstrating that FGF-triggered neurogenesis may not mediated through the enhancement of Sox1 cell proliferation.
We also found that on day 1 through day 4, the total number of apoptotic cells was not reduced after treatment with 40 ng/ml FGF1, or with 5 μM of a pan-caspase inhibitor, z-VAD-fmk. Even after the addition of both FGF1 and z-VAD-fmk, the rescue of apoptotic cells was not significant (Fig. 3E). The total ES cell population was also counted on differentiation days 1 to 4. No statistical significance in number was seen after treatment with FGF1 and/or z-VAD-fmk (Fig. 3F). In sum, these results suggest that the FGF-steering neurogenesis mainly depends on the enforcing differentiation of ES cells, rather than on anti-apoptosis or cell proliferation.
Specific pharmacological inhibitors of MAPKs, shown affecting their respective kinase targets in Fig. 4B, were administrated to delineate the kinases involved in neurogenesis. We found that a PI3K/AKT inhibitor, LY294002, significantly reduced the formation of Sox1-GFP+ cells under SFEB and SFEB/FGF1 conditions (Fig. 4C and 4D).
Neural induction requires sequential signals to direct uncommitted ectoderm into the definitive neural plate . Cumulative evidence supports the fact that FGF is an essential factor for neurogenesis [26, 27]. Interestingly, activation of the Ras/MAPK pathway, rather than the diluted BMP ligands, has been shown to be responsible for the neural cell fate of the fully dissociated animal cap cells, arguing against the simplistic neural default model . The primitive streak- or organizer-derived BMP inhibitors are not the only signals required for neurogenesis. FGF and the other developmental cues, such as Wnt and Notch, also participate in neural induction in a sophisticated manner .
It is noteworthy to emphasize that the activation of MAPK during ES differentiation may not solely depend on FGFR signals and other neural instructing factors could also contribute to the neural induction through JNK or ERK activation, such as insulin-like growth factor (IGF) . Treatment of JNK and ERK inhibitors should simultaneously abolish the endogenous receptor tyrosine kinase signals of differentiating ES cells. Here we showed that neural induction of ES cells was accompanied with the elevated expression of FGFRs and the activation of MAPK pathway (Figs. 2B, 4A and 4B). Pharmacological evidences (Fig. 4C) further supported that differentiation into primitive neuroepithelial cells relied on the activation of both JNK and ERK pathways, but not the p38 MAPK pathway (Fig. 4C). Exogenous FGF-triggered neurogenesis was completely reduced by the JNK and ERK inhibitors (Fig. 4D). Taken together, these data highlights the importance of FGFR activation and of individual MAPK signals in the ES-neuron conversion.
Both pharmacological and genetic evidences support the important role of JNK1 for the neural induction of ES cells (Fig. 4C, D and 5). These results are consistent with the previous finding that JNK1-/- ES cell has a significant reduction in RA-triggered neurogenesis and that JNK/Stress-associated activated protein 1 (JSAP1) is involved in early embryonic neurogenesis [29, 30]. While a neural tube defect is only observed in JNK1/JNK2 double-knockout mice and a JNK1 and JNK2 single-null embryo is normal . It is important to further explore the reason of discrepancy between in vitro and in vivo data and the JNK regulatory networks which participate in neural fate decision and the development of primitive neuroectoderm.
Genetic manipulation has shown that ERK1-null mice are healthy after birth, whereas disruption of the ERK2 gene results in abnormal trophectodermal and mesodermal development [32, 33]. In vitro ES differentiation has also revealed that inhibition of ERK2 completely blocks neural and mesodermal formation, suggesting that ERK2 is essential for the initiation of cell fate commitment of epiblast cells [21, 24]. In this study, we showed that inhibition of MAPK signals sustained the undifferentiated status and the expression of pluripotent markers under the SFEB condition. In future studies, it will be important to understand how the regulatory networks of MAPKs are affected after deprivation of LIF and how they initiate somatic cell induction in ES cells.
Based on a simple and efficient neural induction method, we demonstrate that FGF-triggered neurogenesis of ES cells is not involved in cell proliferation or inhibition of apoptosis. Activation of the ERK2 and JNK1 pathways, rather than p38 MAP kinase, is mainly responsible for the neural induction of ES cells. Release of pharmacological inhibition re-initiated the ES differentiation and neurogenesis, indicating that the FGF pathway participates in the initiation of ES commitment into embryonic cell lineages.
embryonic stem cell
fibroblast growth factor
mitogen-activated protein kinase
serum-free embryoid body-like formation.
This work was supported by the Changhua Christian Hospital (C.S.L.), National Health Research Institutes (H.L.S.) as well as the National Science Council (H.L.S.) of Taiwan. This work was also granted from the Taichung Veterans General Hospital and National Chung Hsing University (TCVGH-NCHU-9776614 and -977602; to H.L.S and H.C.P.), Taichung, Taiwan. We also thank for the support from the core laboratory of tissue engineering and stem cells center in NCHU.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.