SEARCH

SEARCH BY CITATION

Keywords:

  • Embryonic stem cells;
  • Signal transduction;
  • Self-renewal;
  • Pluripotent stem cells;
  • Developmental biology;
  • Differentiation

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Stabilization of β-catenin, through inhibition of glycogen synthase kinase 3 (GSK3) activity, in conjunction with inhibition of mitogen-activated protein kinase kinase 1/2 (MEK) promotes self-renewal of naïve-type mouse embryonic stem cells (ESC). In developmentally more advanced, primed-type, epiblast stem cells, however, β-catenin activity induces differentiation. We investigated the response of rat ESCs to β-catenin signaling and found that when maintained on feeder-support cells in the presence of a MEK inhibitor alone (1i culture), the derivation efficiency, growth, karyotypic stability, transcriptional profile, and differentiation potential of rat ESC cultures was similar to that of cell lines established using both MEK and GSK3 inhibitors (2i culture). Equivalent mouse ESCs, by comparison, differentiated in identical 1i conditions, consistent with insufficient β-catenin activity. This interspecies difference in reliance on GSK3 inhibition corresponded with higher overall levels of β-catenin activity in rat ESCs. Indeed, rat ESCs displayed widespread expression of the mesendoderm-associated β-catenin targets, Brachyury and Cdx2 in 2i medium, and overt differentiation upon further increases in β-catenin activity. In contrast, mouse ESCs were resistant to differentiation at similarly elevated doses of GSK3 inhibitor. Interestingly, without feeder support, moderate levels of GSK3 inhibition were necessary to support effective growth of rat ESC, confirming the conserved role for β-catenin in ESC self-renewal. This work identifies β-catenin signaling as a molecular rheostat in rat ESC, regulating self-renewal in a dose-dependent manner, and highlights the potential importance of controlling flux in this signaling pathway to achieve effective stabilization of naïve pluripotency. Stem Cells 2013;31:2104–2115


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Recent success in the derivation of germ-line competent rat embryonic stem cell (ESCs) paves the way for applying contemporary genetic engineering techniques to the laboratory rat, one of the most important and widely used experimental models for studying vertebrate biology and human disease [1-5]. This breakthrough was achieved using a novel culture medium in which chemical inhibitors of the Ras/extracellular-signal regulated kinase (ERK) and GSK3 signaling pathways are used to suppress stem cell differentiation and support the continuous expansion of pluripotent ESCs [6]. This dual inhibitor system can also be used to derive ESC lines from strains of mice previously regarded as nonpermissive for ESC isolation and to stabilize expansion of naïve-like nonrodent induced pluripotent stem cells (iPSC)/ESCs when they are engineered to overexpress exogenous stem cell factors [7-9]. However, it remains to be established whether use of the two inhibitor (2i)-based culture medium provides a general approach for deriving normal naïve-type ESCs from other mammalian embryos. Improved understanding of how 2i culture conditions influence the ESC state will aid further development and refinement of strategies for the isolation of genuine naïve mammalian ESCs.

The importance of Ras/ERK signaling in mouse ECS (mESC) differentiation was uncovered through genetic and pharmacological disruption of the pathway [10-12]. Blocking the activity of MEK, the upstream activator of ERK, in a chemically defined serum-free medium eliminated unscheduled differentiation and permitted self-renewal of mESCs without additional input from exogenous factors such as leukaemia inhibitory factor (LIF) or bone morphogenic protein 4 (BMP4) [6, 13]. However, efficient blockade of MEK activity alone, particularly in low-density cultures, was insufficient to ensure effective cell survival. This limitation was overcome by combining MEK inhibition with blocking the activity of GSK3, a critical regulator of Wnt signaling, thus facilitating effective expansion of mESCs even at clonal densities [6].

GSK3 inhibition was initially reported to stimulate proliferation of both mESC and human ESC [14]. However, this role in regulating ESC growth and differentiation was controversial, partly due to off-target effects associated with BIO, the small molecule GSK inhibitor used in initial studies, and the general roles played by GSK3 in regulating cell metabolism [6, 15-17]. Subsequent re-evaluation of GSK3 function in recent studies, using mESC lines deficient in key components of the Wnt/β-catenin pathway, identified a direct role for GSK3 and β-catenin signaling in regulating the naïve ESC state [18-21]. In its simplest interpretation, inhibition of GSK3 by Wnt or small molecule inhibitors such as CHIR99021 (CH), stabilizes β-catenin which in turn binds to T-cell factor 3 (TCF3), nullifying the inhibitory effects of this transcription repressor on key targets of the pluripotency regulatory network, such as Nanog, Tbx3, and Esrrb [19]. Restricting TCF3 activity then blocks transition of naïve-type mESC to a “primed” epiblast stem cell (EpiSC) state, thus preventing differentiation. Significantly, neither β-catenin activation nor MEK inhibition alone are sufficient to avert mESC differentiation, but either in combination or individually in conjunction with LIF signaling will support mESC self-renewal. Indeed, the fact that β-catenin-deficient mESCs are readily propagated in culture further demonstrates that the requirement for β-catenin signaling is context-dependent [18].

The active contribution of Wnt/β-catenin signaling to mESC self-renewal highlights an important distinction between the actions of the two inhibitors used in 2i culture. While inactivation of MEK directly blocks activity within the fibroblast growth factor (FGF)/ERK differentiation pathway, inhibition of GSK3 switches off a constitutive repressive activity, leading to chronic activation of downstream signals such as β-catenin. Identifying an appropriate level of GSK3 inhibition is therefore potentially more challenging and may require adjustment depending on the context. In keeping with this possibility, chronic inhibition of GSK3 activity has been reported to interfere with chromosome alignment and stability [22, 23], and constitutively high β-catenin levels can disturb normal physiological patterns of interaction with downstream coeffectors, such as TCFs [24].

In this report, we have investigated the requirement of rESCs for the 2i inhibitors, focusing both on GSK3/β-catenin signaling and on how the response in naïve rat ESCs compares with that of similarly derived stem cells derived from mouse. Although rESCs are routinely propagated in the standard 2i medium [1-5, 25], we have found that their requirement for GSK3 inhibition differs from that of mESCs. It appears that rat ESCs (rESCs) exhibit generally higher levels of β-catenin signaling than mESCs. This influences their response to GSK3 inhibition and may enable their self-renewal in conditions that cannot support growth of the equivalent, naïve ESCs derived from mouse.

Materials and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Ethics Statement

Animal work conformed to guidelines for animal husbandry according to the UK Home Office and approval by the Roslin Institute Animal Ethics Committee. Animals were naturally mated and sacrificed under schedule 1; procedures that do not require specific Home Office approval.

Derivation and Culture of rESCs and mESCs

ESC derivation, culture, and differentiations were as previously described in Meek et al. [26], with minor modifications noted below. Cells were cultured on γ-irradiated (5 Gy) DIA-M mouse fibroblasts in 1i (N2B27 + 1 µM PD0325901 [PD]) or 2i (N2B27 + 1 µM + 3 µM CHIR99021 [CH]). Colonies were passaged every 2–3 days using TVP (0.025% trypsin, 1% chicken serum, and 1 mM EDTA) and plated at a density of 0.5–1 × 105 per centimeter square. Feeder-free cultures were plated on wells coated with 10 µg/mL laminin (Sigma http://www.sigmaaldrich.com). MEK and GSK3 inhibitors were supplied by Axon Medchem (http://www.axonmedchem.com), Stemgent (https://www.stemgent.com), and Sigma. FH535 and Quercetin were supplied by Sigma.

Vector Construction

The mutant β-catenin human cDNA, S33, was polymerase chain reaction (PCR) amplified [27] (NEB, phusion), TOPO cloned into the Gateway Entry vector pCR8/GW/TOPO (Life Technologies, Rockville, MD, http://www.lifetech.com), and sequence verified before cloning into a Gateway-modified pCAGGs-IP expression vector using LR recombinase (Life Technologies).

Stable Transfection

rESCs (5 × 105) were transfected with 500 ng of the S33 β-catenin mutant or an eGFP expression vector using Lipofectamine LTX (Life Technologies). The cells were then replated for 36 hours after transfection and puromycin (0.5 µg/mL) selection was applied for a further 4 days to obtain stable transfectants.

Cell Proliferation Assay

Cells were plated in replicates at a density of 1.5 × 104 per well into wells of a 96-well plate. Cell proliferation was assayed using the CyQuant kit (Invitrogen) at 24, 48, and 72 hours according to manufacturer's instructions.

Colony Assay

2 × 103 cells were plated on feeders in a 4 cm2 well in 1 µM PD containing 0, 3, or 6 µM CH for 6 days. All colonies were stained for alkaline phosphatase and scored as uniformly low, mixed, or uniformly high according to the degree of staining.

Real-Time Reverse Transcriptase PCR

RNA (1 µg) purified using RNeasy Mini Kit (Qiagen, Hilden, Germany, http://www1.qiagen.com) was used to synthesize cDNA using SuperScript First-Strand Synthesis System (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Approximately 1/60 of the cDNA was amplified using Platinum SYBR Green QPCR kit (Invitrogen) using the conditions; 50°C for 2 minutes, then 95°C for 2 minutes followed by 40 cycles of 95°C for 15 seconds, then 60°C for 30 seconds, with a final cycle consisting of 95°C for 1 minute, 60°C for 30 seconds and 95°C for 15 seconds.

Karyotype Analysis

Cells were cultured for 2 hours in 0.1 mg/mL colcemid, swollen in 0.56% KCl hypotonic solution, and fixed on ice in 3:1, methanol:acetic acid. Metaphase spreads were prepared on glass slides and stained with Giemsa.

Microarray

RNA was harvested from three independent 1i and 2i lines. Prior to harvesting, each line was split into nine wells (three wells each of three six-well plates). Each plate consisted of one 1i and one 2i line. One replicate from each plate was pooled to give a total of three replicates for each cell line. Duplicates for each line were hybridized to the Affymetrix Rat Gene 1.1 ST array. Affymetrix data were normalized and summarized using robust multi-array average method [28], and then GeneSpring GX version 7 (Agilent, Palo Alto, CA, http://www.agilent.com) was used to average the replicate grouped together before final visualization of the expression results. Candidate genes were selected using average fold change filtering with a threshold of twofold.

Luciferase Assay

Cells were plated at a density of 0.75 × 105 per cm2 and transfected with 360 ng reporter plasmid (SUPERTOPFLASH or FOPFLASH) and 6.25 ng pEF1α-renilla [29] using Lipofectamine LTX and PLUS reagents (Invitrogen). Cultures were lysed after 48 hours and analyzed using the Dual-Luciferase Reporter System (Promega, Madison, WI, http://www.promega.com). Luciferase levels were normalized to renilla activity, and data are presented as the ratio of SUPERTOPFLASH to FOPFLASH activity.

Chimaera Generation

Rat blastocysts at E4.5 days postcoitum were collected by noon on the day of injection and cultured for 2–3 hours in KSOM embryo culture medium to ensure cavitation. Cells were disaggregated in TVP, pelleted in N2B27, then resuspended in N2B27 containing 1 M HEPES buffer, and kept on ice prior to injection. Blastocycts were injected with 10–12 cells, then transferred into the uteri of pseudopregnant Sprague Dawley rats.

siRNA Knockdown

rESCs were transfected at a density of 0.5–5 × 104 per cm2 with β-catenin (Ambion, Austin, TX, http://www.ambion.com, s136458, s136459, and s136460), adenomatous polyposis coli (APC) (s127456) or negative control (Ambion, 4390846) small interfering RNA (siRNA) at a final concentration of 20–160 nM using Lipofectamine LTX (2.25 mL LTX + 0.75 mL PLUS reagent, Invitrogen) according to manufacturer's instructions. A custom-made negative control siRNA, in which four G to C base changes were introduced to the sequence of β-catenin siRNA s136459 (GCAUGGAGGAGAUAGUUGATT), was also used. Transfection reagents were removed 16 hours post-transfection, and the cells were cultured for a further 2 days prior to harvesting for RNA or for a further 5 days prior to staining for alkaline phosphatase (Sigma, 86R-1KT).

Immunostaining

Cells were fixed with 4% paraformaldehyde, permeabilized with methanol at −20°C, and blocked with 10% (v/v) fetal calf serum (FCS) for 1 hour. Primary antibodies were applied in 2% FCS (v/v in phosphate buffered saline: PBS) for ≥2 hours at room temperature. Secondary antibodies were applied for 1 hour in dark followed by three ashes with PBS. Stained cells were then mounted in Vectashield (Burlingame, CA, http://www.vectorlabs.com), and images were taken using a NIKON EC-1 confocal microscope. The antibodies were: anti-Brachyury(C-19) (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com, sc-17745), anti-Cdx2 (Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com, 3977), anti-Oct3/4 (C-10) (Santa Cruz, sc-5279), anti-Nanog (Abcam, Cambridge, U.K., http://www.abcam.com, ab80892).

Time-Lapse Imaging

Cells were cultured on feeders in 2i+LIF for 24 hours. The medium was replaced with 2i containing 9 µM CH (Stemgent) and time-lapse recordings were taken every 30 minutes for 4 days using a Zeiss Live Cell Observer System.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Stable expansion of authentic rESC lines can be achieved using the 2i medium in conjunction with feeder-support cells [1, 2]. To assess the extent to which rESCs rely on chemical blockade of MEK and GSK3 regulated pathways, we assessed the growth of an established rESC line [1] on feeders in media where we varied the concentration of MEK or GSK3 inhibitors. Reducing the concentration of MEK inhibitor PD decreased colony size and cell numbers, while increasing signs of differentiation and cell debris, indicating that rESCs do not thrive when MEK activity is unimpeded (Supporting Information Fig. S1A, S1B). In contrast, withdrawal of CH did not appreciably affect cell growth or stem cell marker expression indicating that an established rESC line readily tolerated removal of the GSK3 inhibitor (Supporting Information Fig. S1B–S1D).

To investigate whether efficient growth in MEK inhibitor-only (1i) medium could be extended to derivation of lines de novo, we compared the efficiency of ESC isolation from embryos of three strains of rats in 1i and 2i medium. For DA and Lewis strains, there was little difference in the efficiency with which ESC lines were established in 2i and 1i medium (Supporting Information Table S1). A slight reduction in success rate with SD lines notwithstanding the overall high efficiency of deriving rESC lines in 1i medium on feeders indicates that establishment and expansion of rESCs does not require chemical inhibition of GSK3 activity.

To evaluate the phenotype of 1i rESCs, we compared stem cell identity and growth of three individual DA rESC lines derived in 1i or 2i medium. The 1i rESCs formed compact stem cell colonies, similar to those in 2i medium (Fig. 1A). The undifferentiated status of these 1i cell lines was supported by the transcription of ESC markers Oct4, Nanog, Sox2, Rex1/ZFP42, Klf2, and Klf4, at similar levels to matched 2i cultures (Fig. 1B). Oct4 and Nanog protein was detected in the nuclei of 1i rESCs (Fig. 1C). We also compared the growth rates of the matched 1i and 2i lines over a 5-day culture period. With the exception of the slightly more rapidly growing DA27 1i line, the 1i and 2i cell lines showed equivalent rates of growth, further confirming that rESCs cultured on feeders readily tolerated omission of the GSK3 inhibitor (Fig. 1D).

image

Figure 1. Derivation of germline-competent rat embryonic stem cell (rESC) in the absence of CHIR99021. (A): Bright-field images of DA rESCs derived in 1i (1 µM PD0325901) or 2i (1 µM PD0325901 + 3 µM CHIR99021). (B): Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) analysis for ESC markers in 1i and 2i rESCs. Mean and SD of three biological replicates. (C): Immunostaining for Oct4 (middle panels) and Nanog (right panels) in 1i and 2i rESCs (magnification ×100). (D): Proliferation assay of three independent 1i and 2i lines. Mean and SD of three experimental replicates. (E): Immunostaining for Nestin and Tuj1 (left panel) and skeletal myosin (right panel) following an 11-day neuronal or myogenic monolayer differentiation protocol respectively (magnification ×100). (F): Photograph of DA27 chimeric mother with pigmented germline offspring and albino littermate. Abbreviation: DAPI, diamidino-2-phenylindole.

Download figure to PowerPoint

It has been speculated that aneuploidy observed in rESC cultures might be associated with chronic inhibition of GSK3 activity [2]. We therefore compared chromosome counts in 1i and 2i rESC cultures at early and late passages, and at low plating densities, and found similar overall percentages of euploid cells in the different culture media, suggesting that addition of CH did not have a significant effect on karyotypic stability in rESCs (Supporting Information Fig. S2A, S2B). The reduction in karyotypic stability observed in low-density cultures, however, points to autocrine factors or intercellular contact being important for supporting growth of rESCs, even in the presence of feeder cells.

The developmental potential of 1i rESCs was examined by assessing their differentiation capacity in vitro and in vivo. Differentiation marker analysis in embryoid body outgrowths showed that 1i cultures generated derivatives of all three germ layers, including endoderm (Sox17, Afp), mesoderm (Brachyury, Kdr1), and ectoderm (Nestin, Pax6) (Supporting Information Fig. S2C). 1i monolayer cultures also generated nestin-positive neural progenitors and neurons, as well as mesodermal derivatives such as skeletal muscle (Fig. 1E). The in vivo developmental potential of the 1i cells was tested by blastocyst injection and chimaera formation. Using “hooded” albino strain blastocysts as host embryos we obtained coat color chimaeras with two 1i DA cell lines, that exhibited the typical “reverse-hooded” coat-color phenotype characteristic of chimaeras between hooded albino and nonhooded pigmented strains (Table 1, Fig. 1F) [30]. When mated with albino rats, two RIDA27 1i chimaeras produced agouti pups demonstrating transmission of the 1i ES cells through the germ line (Fig. 1F, Table 1). This was confirmed by microsatellite analysis of DNA prepared from tissue samples (Supporting Information Fig. S2D). The contribution of 1i rESCs to chimaeras, including the germ line, approached efficiencies we obtained with an established 2i cell line [31] (Table 1), confirming that inhibition of MEK signaling alone in the presence of feeders allows efficient derivation and propagation of authentic rESCs.

Table 1. Chimaera and germline transmission efficiency
Cell lineSexPassageTransferredBornChimaerasSexGLT
  1. Summary of chimaera generation by rat embryonic stem cell (ESC) injection into recipient blastocysts (SD or Fischer 344), outlining the number of embryos transferred, the total number of live pups, the number of chimaeras, and the proportion of those that transmitted ESCs through the germ line.

  2. One chimaera was sacrificed due to jaw deformity.

  3. Chimaeras were generated from passage 11 cells.

  4. No pups were born.

  5. Abbreviations: GLT, germline transmission; LIF, leukaemia inhibitory factor.

DA27 (1i)F9–18633186 (29%)25 (13%)11F2/10
DA38 (1i)F7, 112517 (68%)2 (12%)1F0/1
DAK31 (2i+LIF)M18–2017862 (35%)21 (34%)10M4/10

In light of recent reports describing a critical role for the Wnt/GSK3/β-catenin signaling pathway in maintaining the naïve status of mESCs [18-21, 32], it was important to more comprehensively compare the molecular profile of 1i cultures. We therefore compared the mRNA expression profiles of three cell lines established in each growth condition by microarray (Fig. 2A and Supporting Information Tables S2 and S3, URL link to complete data set on acceptance). This showed that 1i and 2i transcription patterns were similar, with >99% genes varying by less than 2-fold and 255 genes differing between the two conditions. In line with quantitative reverse transcriptase PCR (qRT-PCR) results, most ESC-associated markers were expressed at equivalent levels. Two exceptions were the transcription factors Gbx2 and Otx2, which were expressed 2.3-fold lower and 3.7-fold higher in 1i cultures, respectively (Fig. 2A, 2B). While this difference might hint at a bias of 1i cells toward a primed state [33] evidence from the studies in Xenopus suggests that Gbx2 and Otx2 are directly activated and repressed targets of Wnt signaling, respectively [34]. Indeed, the 166 genes that exhibited higher expression in 2i include standard Wnt signaling response genes Axin2 (2.9-fold) and Cdx1 (3-fold), as well as the transcription factors Brachyury and Cdx2, two known regulators of mesendoderm differentiation and markers of the primitive streak during early gastrulation [35, 36]. qRT-PCR confirmed the upregulation of Brachyury and Cdx2 expression in 2i cultures (Fig. 2B). This response to CH was also observed in cell lines irrespective of sex, strain, or passage, confirming that Cdx2 and Axin2 induction was a general feature of rESCs (Supporting Information Fig. S2E, S2F). Increased RNA expression was accompanied by upregulation of Cdx2 and Brachyury nuclear protein throughout 2i rESC colonies (Fig. 2C, 2D). Overall, these molecular analyses confirmed that rESCs accommodate the absence of CH. However, they also identified expression of differentiation markers induced by the presence of the GSK inhibitor in 2i rESC cultures.

image

Figure 2. Upregulation of Brachyury and Cdx2 expression by CHIR99021 in rat embryonic stem cell (rESC). (A): Scatterplot of microarray gene expression data comparing three 1i and three 2i rESC lines. (B): Quantitative reverse transcriptase polymerase chain reaction validation of microarray data. Mean and SD of three biological replicates. (C): Immunostaining of two 1i (DA27 and DA38) and two 2i (DA18 and DA22) rESC lines for Oct4 (red) and Cdx2 (green) (magnification ×400). (D): Immunostaining of the same rESC lines as panel (C) for Oct4 (red) and Brachyury (green) (magnification ×400). Abbreviation: DAPI, diamidino-2-phenylindole.

Download figure to PowerPoint

To assess whether growth in MEK inhibitor-only 1i medium is a general property of naïve-type ESCs, we directly compared the CH-dependence of mESC and rESC lines derived in identical 2i conditions. After four passages in 1i medium, mESCs, in contrast to rESCs, exhibited signs of differentiation (Fig. 3A). The colonies in 1i medium were either small, irregular-shaped, heterogeneous aggregates, or groups of flattened epithelial cells, distinct from the compact and uniform colonies formed in 2i. These overt signs of differentiation were accompanied by downregulation of stem cell markers, including the naïve ESC-specific transcription factor Rex1 (particularly evident in C57BL/6 lines) as well as dramatic upregulation of the “primed” postimplantation epiblast/EpiSC marker Fgf5 in all lines (Fig. 3B). In striking contrast to rESCs, mESC cultures also exhibited upregulation of Brachyury and Cdx2 in 1i culture. This strongly suggests that even when cocultured with feeder cells, relaxation of GSK3 inhibition in 1i medium caused mESC to differentiate toward a primed, late-epiblast state. These results suggest that, in contrast to rESCs, MEK inhibition together with feeder-derived factors cannot block the exit of mESCs from the naïve ESC state, a reversal of the situation that occurs in standard serum/LIF culture conditions [1, 37].

image

Figure 3. Differential responses of mouse (m) and rat embryonic stem cell (rESC) to withdrawal of CHIR99021. (A): Bright-field images of rESC and mESC after culture in 1i or 2i medium for four passages (magnification ×100). (B): Quantitative reverse transcriptase polymerase chain reaction analysis of two rESC lines and four mESC lines (two CBA/C57BL/6 and two OLA/129 lines) after culture for four passages in 1i or 2i medium. Data are presented as expression of 1i relative to 2i.

Download figure to PowerPoint

The lower level of Axin2 induction in 2i mESC cultures suggested that they might be generally less responsive to the GSK3 inhibitor than rESCs (Fig. 3B). To compare the CH dose-response of mESCs and rESCs, we assessed colony morphology and stem cell/differentiation marker expression in ESCs cultured for 3 days on feeders in 0.37–12 µM CH (Fig. 4; Supporting Information Fig. S3). As expected, rESCs showed little variation in colony morphology up to 3 µM CH (Fig. 4A; Supporting Information Fig. S3A). However, at the higher concentrations of CH, many of the colonies flattened, signifying the onset of differentiation. This morphological change was accompanied by marked downregulation of stem cell markers and dramatic increase in Cdx2 and Brachyury expression (Fig. 4B, 4C; Supporting Information Fig. S3B). mESCs formed flattened colonies in the absence of CH but grew as progressively more compact colonies as the concentration of CH increased (Fig. 4A; Supporting Information Fig. S3A). Interestingly, expression of both Cdx2 and Brachyury was also induced in mESCs in response to CH (Fig. 4B; Supporting Information Fig. S3B), but required higher concentrations than in rESCs, and the degree of induction was almost an order of magnitude lower than in rESCs at 6 and 12 µM CH. In addition, expression of Cdx2 and Brachyury in mESCs was not accompanied by downregulation of stem cell markers, as observed in rESCs (Fig. 4B, 4C; Supporting Information Fig. S3B).

image

Figure 4. Mouse (m) and rat embryonic stem cell (rESC) differ in their dose-response to CHIR99021. (A): Bright-field images of rESC line DA22 (upper panels) and mESC line m1.1 (lower panels) cultured in 0 µM, 3 µM, or 6 µM CH for 3 days (magnification ×100). (B): qRT-PCR analysis of the same rESC (upper panel) and mESC (lower panel) lines as used in panel (A), cultured in various concentrations of CH for 3 days. Expression levels are normalized relative to those in 1i (0 µM CH). (C): Confocal images of immunostained rESC (left panels) and mESC (right panels) for Oct4 and Cdx2 following culture in 0 µM, 3 µM, and 6 µM CH for 3 days. (D): Quantification of alkaline phosphatase stained rat and mouse colonies formed at clonal density in 0, 3, and 6 µM CH. Abbreviation: DAPI, diamidino-2-phenylindole.

Download figure to PowerPoint

To further confirm the species differences in response to CH, we compared mESCs and rESCs in a colony assay at 0, 3, and 6 µM CH (Fig. 4D). In line with stem cell marker expression, rESCs self-renewed efficiently in the absence of CH but differentiated at 6 µM CH, while mESCs showed the opposite response. Since colony viability was reduced in rESC treated with high concentrations of CH, we also examined whether the effect on rESC might be due to selective survival of pre-existing differentiated cells from the starting cultures by performing time-lapse microscopy (Supporting Information Video 1). By following rESC growth, it is clear that the majority of stem cell colonies differentiate in high CH concentrations, and it is the poor survival of cells differentiating during the later stages of the culture that accounts for colony loss.

The species-specific variation in response to CH pointed to differences in signaling induced by the GSK3 inhibitor in mESC and rESCs. To verify that the prodifferentiative effects of CH on rESC were mediated through GSK3, we tested the effects of a different source of CH and another GSK3 inhibitor BIO. We also tested whether siRNA-mediated knockdown of APC, a scaffold protein that coordinates GSK3-mediated degradation of β-catenin, affects rESC self-renewal. In each case, interference in GSK3 inhibition, as measured by increased β-catenin activity, correlated with loss of stem cell colony morphology, alkaline phosphatase activity, and downregulation of the stem cell markers such as Nanog in rESC (Fig. 5A, 5B; Supporting Information Fig. S4A–S4D). To examine if variation in β-catenin activity accounts for differences between mESC and rESC, we measured transcriptional activity of β-catenin directly using a TCF-dependent (TOPflash) luciferase reporter assay. Assessment of β-catenin-dependent transcription in ESCs treated with different doses of CH, as measured by the ratio of TOPflash to FOPflash reporter activity, showed that β-catenin activity increased in a dose-dependent manner in both ESCs (Fig. 5C; Supporting Information Fig. S4E). However, β-catenin activity was at least 2-fold higher in rESCs compared to mESC in all conditions, including in the absence of CH. Indeed, in CH-free medium, comparison of TOP:FOP flash activity in the presence and absence of feeders demonstrated that rESCs had approximately 5-fold more basal activity than mESCs, on feeders (Fig. 5D). This feeder-dependent response could be suppressed by siRNA-mediated knockdown of β-catenin, confirming that it was due to β-catenin activity (Supporting Information Fig. S4F, S4G).

image

Figure 5. Rat embryonic stem cell (rESC) exhibit higher levels of tonic and induced β-catenin activity than mouse (m) ESC. (A): Alkaline phosphatase staining of two rESC lines (DA18 and DA22) cultured in 3 µM CHIR99021, 3 µM BIO, or 3 µM CHIR + 3 µM BIO. (B): Alkaline phosphatase staining of two rESC lines (DA18 and DA22) transfected with either a negative (−ve) control siRNA or a siRNA to APC. (C): Graph showing ratio of TOPFlash:FOPFlash activity of rESC and mESC cultured in 0, 1.5, 3, 6, or 12 µM CH for 3 days. (D): Histogram showing the ratio of TOPFlash:FOPFlash activity of rat and mouse 2i ESC lines cultured in 1i (PD only) on laminin or feeder cells. Mean and SD of three biological replicates. (E–H): siRNA knockdown of β-catenin. Alkaline phosphatase staining (E) and quantitative polymerase chain reaction (qRT-PCR) analysis (F) of rESC lines cultured in 1i for 6 days under selection for stable transfection with the β-catenin mutant (S33) or an eGFP negative control (C) vector. Low power (left panel) and high power (middle and right panels, ×100) magnification of stably transfected rESC. Residual S33 transfectants formed small, irregular and weakly staining colonies (highlighted by arrows). (G–J): Alkaline phosphatase staining of rESC line DA22 cultured for 6 days post-transfection in 2i, containing (G) 6 µM or (I) 3 µM CH. qRT-PCR analysis of rESC line DA22 cultured for 3 days post-transfection in 2i, containing (H) 6 µM or (J) 3 µM CH. Abbreviations: APC, adenomatous polyposis coli; siRNA, small interfering RNA.

Download figure to PowerPoint

To directly investigate the functional contribution of β-catenin in rESC, we monitored self-renewal and gene expression following β-catenin overexpression and siRNA knockdown. Expression of the S33 activated form of β-catenin [27] in 1i conditions markedly reduced formation of undifferentiated ESC colonies, and in residual cells induced Cdx2 and Brachyury and suppressed Nanog expression to levels achieved with ≥3 µM CH (Fig. 5E, 5F; Supporting Information Fig. S4K, S4L). In cells treated with 6 µM CH, β-catenin siRNA-mediated knockdown rescued both the growth of undifferentiated rESC colonies and stem cell marker expression, while reducing Cdx2 and Brachyury expression (Fig. 5G, 5H; Supporting Information Fig. S4I, S4M, S4N). Treatment with two different β-catenin/TCF inhibitors, FH535 and Quercetin, also improved colony growth in high levels of CH, further confirming that β-catenin plays an essential role in CH-mediated loss of rESC self-renewal (Supporting Information Fig. S4J). In medium containing 3 µM CH (2i), a similar decrease in β-catenin also reduced Cdx2 and Brachyury levels but did not significantly affect ESC marker expression or colony growth (Fig. 5I, 5J; Supporting Information Fig. S4O, S4P). In the absence of CH (1i), β-catenin knockdown had little effect on colony morphology or expression of ESC markers (Supporting Information Fig. S4Q–S4T). Treatment with the Wnt antagonist Dkk1 blocked exogenous Wnt induction of a Topflash luciferase reporter, but was unable to further suppress Topflash activity in 1i rESC cultures, implying that basal Wnt signaling in rESC is at low levels in feeder cocultures (Supporting Information Fig. S4H). These experiments suggest that rat ESC are sensitized to β-catenin-mediated signaling, and that excessive activity of this pathway by CH destabilizes self-renewal.

To assess the requirement of rESCs for GSK3 inhibition in the absence of feeders, we plated rESCs on laminin and monitored their response to CH. We used laminin because it allows reliable attachment of ESCs, supports short-term rESC self-renewal [1], and contributes negligible β-catenin transcriptional activity in the absence of CH (Fig. 5D). In contrast to the situation on feeders, expression of stem cell markers was dramatically reduced in both rESC and mESCs plated on laminin without CH and was accompanied by poor cell growth and morphological differentiation (Fig. 6A; Supporting Information Fig. S5A). Increasing the dose of CH from 0.37 to 3 µM produced a graded recovery of ESC marker expression and growth for both rESCs and mESCs (Fig. 6B; Supporting Information Fig. S5B). However, above 3 µM CH, the response of ESCs diverged. rESCs showed a sharp decline in expression of ESC markers, impaired growth and signs of differentiation. A similar response was observed with rESC cultures prepassaged twice on laminin to eliminate residual conditioning associated with feeders, before setting up the experiment, confirming that CH-induced differentiation was intrinsic to rESCs (Supporting Information Fig. S5C, S5D). Interestingly, feeder-conditioned medium increased rESC colony growth on laminin in the presence of 2i, supporting the notion that the feeders may provide GSK3/β-catenin independent support for rESC growth (Supporting Information Fig. S5E). In contrast to rESC, mESCs plated on laminin formed increasingly compact colonies at higher doses of CH, and no significant reduction in ESC marker expression (Fig. 6A, 6B; Supporting Information Fig. S5A, S5B). These results show that moderate levels of GSK3 inhibition promotes rESC self-renewal in a pattern broadly similar to that of mouse ESC, but that higher concentrations of CH induced differentiation, as observed in the feeder cocultures. Collectively, results from these feeder-free cultures suggest that the higher tonic level of β-catenin activity in rESCs requires that GSK3 activity is more finely tuned to limit differentiation and effectively promote self-renewal and maintenance of pluripotency.

image

Figure 6. Tuning GSK3 inhibition is required to support rat embryonic stem cell (rESC) growth on laminin. (A): Bright-field images of rESC line DA22 (upper panels) and mouse (m) ESC line m1.1 (lower panels) cultured on laminin in 0 µM, 3 µM, or 6 µM CH for 3 days (magnification ×100). (B): Quantitative reverse transcriptase polymerase chain reaction analysis of the same rESC (upper panel) and mESC (lower panel) lines as used in panel (A), cultured on laminin in various concentrations of CH for 3 days. Expression levels are normalized relative to those in 1i (0 µM CH).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Investigation of how signaling pathways regulate mESCs led to the suggestion that authentic naïve ESCs have an intrinsic capacity for self-renewal, when shielded appropriately from prodifferentiative signals [6, 17]. An important component of this shield is the suppression of autocrine FGF/ERK signaling through inhibition of MEK [6, 12, 13]. More recent reports, however, clearly show that this barrier alone is insufficient, and that mESCs require additional active inputs from either Wnt/β-catenin or LIF/STAT3 signaling pathways to support long-term self-renewal [18, 19, 21, 32]. Our investigation into the requirement for MEK and GSK inhibitors in rESC cultures has demonstrated that on feeders, MEK inhibition alone (1i conditions) allowed efficient derivation and expansion of rESC cultures. Crucially, 1i rESCs contributed to chimeric animals at efficiencies comparable to 2i cell lines and were transmitted through the germ line, demonstrating that these 1i cultures represented authentic naive ESC. These findings also drew our attention to novel aspects of self-renewal signaling in ESCs, in particular the dose-dependent response to GSK3 inhibition, and how this differs between naïve ESCs derived from rat and mouse. A key factor underlying these differences was the higher levels of β-catenin-mediated transcriptional activity seen in rESCs. While induction of β-catenin activity above basal levels initially promoted self-renewal of ESC from both species, further stimulation caused differentiation of rat cells, but not in mESC. This shows that β-catenin activity plays a key role in modulating the balance between self-renewal and differentiation in rESC and raises the general possibility that tuning the level of activity within this signaling pathway could be an important element in stabilizing the naïve-type pluripotent ESC state in species other than the mouse.

Our investigation of the effects of CH on ESCs revealed that generally rESCs exhibit significantly higher levels of β-catenin signaling than mESCs. This was strikingly evident from the marked upregulation of the differentiation markers Brachyury and Cdx2 in 2i rESC cultures. Brachyury, a key regulator of mesoderm development during gastrulation, and direct target of β-catenin, is normally expressed at very low levels in naïve mESCs but can be induced in primed-type EpiSCs and therefore represents a useful indicator of early-stage mESC differentiation [32, 38]. Cdx2 expression, in contrast, was until recently frequently cited as a definitive marker of trophoblast in differentiated ESCs, but expression within the primitive streak and induction by Brachyury and FGF in human ESCs, have highlighted association of this transcription factor with mesoderm differentiation [39]. Given that rESCs maintained in 2i are demonstrably pluripotent, expression of Brachyury and Cdx2 nuclear protein in rESCs is clearly tolerated and does not appear to compromise self-renewal in these particular conditions. However, higher concentrations of CH, where the level of Brachyury and Cdx2 mRNA increased by more than 10-fold, did induce downregulation of stem cell markers and rESC differentiation. This response is reminiscent of a recent report describing how Wnt signaling and GSK3 inhibitors induce mesoderm differentiation in human ESCs [40], and points to the possibility that β-catenin plays distinct roles in rESCs depending on its level of activity. It is noteworthy in this context that studies on mESCs suggest that β-catenin protein targets the TCF factors, make distinct contributions to the regulation of ESCs [19]. This is illustrated by the contrasting activities of TCF3 and TCF1, respectively, repressing and activating generic TCF reporter constructs in ESCs [18, 19]. Perhaps distinct species-specific patterns of TCF engagement defined by their abundance and the different levels of β-catenin activity [41, 42] underlie the β-catenin-dependent switch between self-renewal and differentiation in rESCs.

The lower levels of β-catenin-dependent activity seen in mESC, either in response to coculture with feeders, or through chemical inhibition of GSK3, indicates that flux within this pathway in mESCs is limited compared with rESC. This restriction, which evidence suggests is downstream of GSK3, would be responsible for the relative insensitivity of mESC to feeder-dependent activation of the pathway and is also consistent with a recent report that coculture with MEFs is incapable of supporting robust growth of mESC in serum free conditions, even in the presence of the MEK inhibitor [21]. Restricted flux within the Wnt/β-catenin pathway may also account for the resistance of mESCs to the prodifferentiative effects of high levels of CH that are observed in rESC. Indeed, although mESC cultures exposed to higher concentrations of CH expressed Brachyury and Cdx2, expression of stem cell markers and typical naïve-type ESC colony morphology was maintained. The reduction in mESC colony size we observed at high concentrations of CH has also been noted previously by others using a range of chemically distinct GSK3 inhibitors, including some with greater potency than CH [18]. This implies that even when inhibition of GSK3 is near complete, mESCs are relatively resistant to differentiation: a conclusion supported by the continued suppression of the primitive ectoderm/EpiSC marker Fgf5 in these cultures. While the restricted dynamic range of β-catenin activity in mESCs may dictate the level of GSK3 inhibition compatible with self-renewal, it might also provide resistance to fluctuations in other cooperating regulatory signals or growth conditions that would otherwise drive ESCs beyond the bounds of the stem cell regulatory network.

Conclusion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Our analysis of rESCs provides insights into how GSK3/β-catenin signaling may affect self-renewal of naïve-type ESC in species other than the mouse. These findings suggest that the level of β-catenin signaling plays an important role in controlling ESC self-renewal. The stability of mESCs in established culture conditions may have until now masked more general consideration of how flux within signaling pathways controls ESC self-renewal. This issue is especially relevant where resistance to ESC differentiation relies upon continuous active input from extrinsic self-renewal signals, for example, Wnt/β-catenin or LIF/STAT3 [41]. Indeed, comparison between the responses of rESCs and mESCs indicate that while it may be relatively straightforward to neutralize the prodifferentiative activity of ERK, it may be more challenging to achieve an appropriate level of GSK3 inhibition, or more specifically β-catenin activity, in different contexts. This concept may apply more broadly to other self-renewal signals and account for the significant difficulties in isolating and propagating unmodified naïve ESCs from other mammals [43]. Finally, the unexpected finding that rat ESCs self-renew more effectively in 1i conditions than the archetypal ESCs derived from mouse highlights the value of interspecies studies for constructing robust general models of how signals control the core regulatory network of naïve pluripotent ESCs.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

We thank the members of the transgenic animal facilities at the Institute of Stem Cell Research and Biomedical Research Resources at the University of Edinburgh, for their support; Tilo Kunath, Denis Headon, Ian Chambers, and Helen Sang for their comments on the manuscript; Michael Clinton for his expert editorial input; Bob Fleming for his help with time-lapse microscopy; Hans Clevers and Marc van de Wetering for providing β-catenin expression vectors; Peter Hohenstein, Karamjit Singh-Dolt, Katherine Norrby, and Melany Jackson for assistance with Gateway cloning; Austin Smith and Kathryn Blair for discussing data and providing the DAK31 rESC line. This work was supported by funding from the Biotechnology and Biological Sciences Research Council, the Roslin Foundation and from the European Community's Seventh Framework Programme (FP7/2007–2013) under grant agreement No. HEALTH-F4-2010-241504 (EURATRANS).

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Additional Supporting information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1466-sup-0001-suppfig1.tif5411KSupporting Information Figure S1
stem1466-sup-0002-suppfig2.tif442KSupporting Information Figure S2
stem1466-sup-0003-suppfig3.tif775KSupporting Information Figure S3
stem1466-sup-0003-suppfig4.tif4273KSupporting Information Figure S4
stem1466-sup-0004-supptab1.docx34KSupporting Information Table S1
stem1466-sup-0005-supptab2.docx18KSupporting Information Table S1
stem1466-sup-0006-supptab3.docx76KSupporting Information Table S1
stem1466-sup-0007-supptab4.m4v3880KSupporting Information Table S1
stem1466-sup-0008-supp.doc32KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.