At the pre-implantation blastocyst stage of development, the mammalian embryo is composed of a unique collection of cells of which three major populations predominate. The outermost layer the trophectoderm (TE) gives rise to the placenta, which acts to sustain the developing fetus connecting it to the mother host. The next is a cluster of cells known as the inner cell mass (ICM) these cells are said to be pluripotent (Fig. 1). A third group of cells known as the primitive endoderm, surrounds the ICM cells at the epiblast stage. As development proceeds the ICM cells rapidly divide and eventually begin to differentiate forming the three embryonic germ layers (ectoderm, mesoderm and endoderm). Effectively these pluripotent ICM cells are the precursors of all adult tissues. As these pluripotent cells commit to a specific cellular lineage, they lose their pluripotency. Embryonic stem (ES) cells are euploid pluripotent cell lines isolated directly from cultured preimplantation embryos. The first stable ES cell lines were isolated by immunosurgery from the ICM of implantation-delayed, mouse blastocysts (Martin, 1981; Evans and Kaufman, 1981). Mouse ES cells are very closely related to early ICM cells in terms of their developmental potential (Beddington and Robertson, 1989). This chapter will focus on mouse ES cells (mES) unless otherwise stated. Three features characterize mES cells;
They are isolated directly from the embryo (Robertson, 1987).
They can colonize the germ line when introduced to the embryo.
They possess unrestricted proliferative potential (Suda et al., 1987).
These features effectively mean that under appropriate conditions, a karyotype stable self-renewing, pluripotent population of cells can be propagated indefinitely
Early studies with mES cells showed that the use of mitotically inactivated STO cells (Ware and Axelrad, 1972; Hooper, 1997) was essential in the maintenance of self-renewal and the pluripotent state. Later it was found that the requirement for the feeder layer could be circumvented by the addition of the cytokine LIF in the presence of serum. In the absence of feeder layers or LIF, mES cells differentiate into a variety of cell types (Doetschman et al., 1985) depending on the developmental cue or signaling pathway activated. The process of differentiation can be seen as a loss of pluripotency and mES like their in vivo counterparts are capable of multi-lineage differentiation. mES cells undergo a controlled pattern of differentiation when injected and reintegrated into a pre-implantation blastocyst. Under these circumstance mES cells respond as ICM cells to
Differentiated mES cells from EBs can give rise to a wide variety of cell types including neuronal (Bain et al., 1995), hematopoietic (Suwabe et al., 1998), endothelial (Yamashita et al., 2000), cardiac (Maltsev et al., 1993), smooth muscle (Yamashita et al., 2000), chondrogenic (Kramer et al., 2000) and osteoblastic cells (Buttery et al., 2001). For stem cells there is a constant balancing act that must be maintained between self-renewal and the pluripotent phenotype versus cell lineage commitment and differentiation.
An understanding of the pathways and controlling factors involved in these fundamental cellular events is essential if we are to exploit the full potential of embryo derived stem cells for therapeutic uses in disease treatments and regenerative medicine in the future. This potential is real as it is clear that embryo derived stem cells are capable of unlimited self-renewal capacity and can differentiate into potentially any of over 200 cell types. The potency of these cells is maintained by a number of key regulatory factors, signaling pathways and extracellular signaling agents. The combination and interplay of these elements establishes a patterns of gene expression that sustains the pluripotent phenotype of ES cell. Some of the key regulators are transcription factors such as Oct4, Sox2 and Nanog, the signaling cascades involving phosphatidylinositol 3-kinase (PI3K) and the signaling molecules like LIF and Wnt proteins. Embryonic stem cells have great therapeutic potential however to fully realize the potential of these cells the signaling pathways that participate in controlling ES cell behavior must be better understood.
2. Pluripotency and Self-renewal
In the developing embryo, pluripotent cells first appear in the ICM, in the mouse blastocyst this is approximately at day 3.5 of embryonic development (E3.5). These cells persist as late as the pre-gastrulation stage. Thus
2.1. Transcriptional networks
Oct4 is a member of the POU (Pit-Oct-Unc) transcription factor family that regulates the expression of target genes by binding to a octameric sequence (Scholer et al., 1990). The key features of this family are the POU domain consisting of two sub-domains each of which bind to DNA. However the C-terminal is cell specific and may be essential for the expression of target gene in an orderly fashion as embryonic development proceeds. It is well established that the Oct4 gene (encoded by Pou5f1) is constitutively expressed in undifferentiated mES cells, in all pluripotent cells during mouse embryo development and is also an essential factor required in the generation of iPS cells (Niwa et al., 2000). Oct4 is also known as Oct3, Oct3/4, Otf3, and Otf4. In the mouse, Oct4 expression is up regulated beginning at the 4-cell stage and becomes localized to the pluripotent cell population (Yeom et al., 1996). The expression of Oct4 is common to human and mouse ES cells, and furthermore expression diminishes in both as cells differentiate.
Nanog is an homeobox containing transcription factor of approximately 280 amino acids. In the developing mouse embryo Nanog plays a key role in determining the fate of the ICM cells, acting to sustain pluripotency and preventing differentiation (Chambers et al., 2003). Nanog was identified as a factor, which when over expressed, supported pluripotency even in the absence of a LIF based signal. In the embryo Nanog expression is first seen at the compacted morulae stage before becoming restricted to the ICM, post-implantation stage Nanog expression is drastically reduced.
Sox2 is a DNA-binding protein of the HMG family. In the mouse Sox2 is expressed predominatley at the blastocyst stage (Avilion et al., 2003). However unlike Oct4, Sox2 has a major role to play also later in development and in adult stem cells (Wood and Episkopou, 1999; Zappone et al., 2000). At early stages of development and in mES cells, Sox2 activates target genes through interaction with Oct4. Sox2 knockouts are lethal to mouse embryos and they fail to fully develop, furthermore ES cells derived from these embryos are unable to proliferate or self-renew (Avilion et al., 2003). As outlined for Oct4, a precise level of Sox2 appears to be key for pluripotency and to sustain self-renewal. Many studies have highlighted how Oct4 and Sox2 can in a direct way drive the expression of genes required for pluripotency including positive feedback on their own expression and that of Nanog (Chew et al., 2005; Tomioka et al., 2002). Together with the transcription factor Klf4, they activate the expression of Lefty1 (Nakatake et al., 2006). In mES cells a wide range of studies have focused on and delineated the functional role of Oct4, less is known about Sox2. Recent studies are beginning clear up the role of Sox2. As might have been anticipated mES cells deficient in Sox2 lose pluripotency and quickly differentiate supporting the perceived role of Sox2 in maintaining self-renewal. What is interesting is that in the Sox2 protein deficient system Oct-Sox enhancers are still active and up-regulation of Oct4 alone is sufficient to rescue these cells from differentiation. Thus it has been suggested that potentially other members of the Sox family may substitute for Sox2 in the co-activation process mediated in partnership with Oct4. What has become clear is that Sox2 plays a role in regulating many transcription factors that can affect Oct4 levels including Nanog. Furthermore in cellular reprogramming studies up-regulation of Oct4 in combination with Sox2 is sufficient to generate pluripotent cells (Takahashi and Yamanaka, 2006; Okita et al., 2007). Some fascinating studies looking at global protein phophorylation patterns in hES cells have revealed some interesting dynamics in the Oct4 and Sox2 pattern of activation (Burdon et al., 2002). Thus exploring phophorylation pathways from extracellular signals to gene transcription effects will be key to furthering our understanding of self-renewal, in this context, pathways like those involving LIF and PI3K will be key to disentangling the signaling and transcription circuits involved.
3. Signal transduction pathways
3.1. Leukemia inhibitory factor (LIF)
LIF is expressed in mouse preimplantation embryos from fertilization to the blastocyst stage but not in TE cells (Nichols et al., 1996). LIF transcripts are also found in mES cells (Rathjen et al., 1990) and endometrial glands of the mouse uterus which stops once implantation has occurred. Mouse LIF gene knockouts result in growth retardation and fertilized blastocysts fail to implant (Stewart et al., 1992). Historically mES cells were derived and maintained on a feeder layer of embryonic fibroblast. Subsequently it was found that the use of conditioned media from these fibroblast cultures was sufficient to maintain mES cell self-renewal. It was then shown that a the active agent produced by the feeder layer capable of blocking mES cell differentiation was in fact a cytokine later identified as leukemia inhibitory factor (LIF) (Smith et al., 1988). LIF is the best-characterized effector of self-renewal in mES cells. It is a multifunctional cytokine, which has a wide variety of effects on various cell types (Hilton and Gough, 1991). The name LIF is based on initial observations that
The cellular actions of LIF are effected via a specific cell membrane receptor. The LIF receptor is a heterodimeric complex composed of a glycoprotein subunit gp130 and the receptor subunit LIFR (also called LIFRβ) (Ernst and Jenkins, 2004). Studies in mES have shown that the gp130 subunit is the essential component in transmitting self-renewal signals (Nakamura et al., 1998). Binding of LIF to the LIFR subunit induces dimerization with gp130, resulting in the formation of a high affinity receptor complex. The activated receptor switches on the constitutively bound tyrosine kinase Janus kinase (JAK). Activated JAK, phosphorylates both receptor subunits forming SH2 domain bind sites, which are capable of recruiting other signal transduction partners.
The SH2 domains facilitate the binding of signal transducers and activators of transcription (STAT) 1 and STAT3, which are phosphorylated by JAKs (Stahl et al., 1995). The activated STAT proteins form homodimers or heterodimers which then move to the nucleus, where they act as transcription factors (Auernhammer and Melmed, 2000). STAT3 is the principal STAT protein activated in mES cells stimulated with LIF (Niwa et al., 1998). Activation of STAT3, has been shown to be critical for LIF/gp130 dependent self-renewal in mES cells (Niwa et al., 1998). Using a tamoxifen inducible form of STAT3 (fusion of STAT3 to estrogen receptor) it has been shown that activation of STAT3 is capable of sustaining self-renewal of mES in the presence of serum (Mastuda
In the absence of fetal calf serum, in the presence of activated STAT3, BMP4 signaling maintains pluripotency. However, for hES cells LIF-STAT3 signaling cannot maintain pluripotency (Reubinoff et al., 2000) additional factors independent of LIF-STAT3 are required including basic fibroblast growth factor (bFGF) in the presence of Noggin which acts as a BMP pathway inhibitor. The exact mechanism of LIF-STAT3-dependent mES cell self-renewal is still not fully elucidated although models are arising (Fig. 2). A notable target for STAT3 is the transcription factor Myc (Cartwright et al., 2005) which along with others (Klf4, Oct4 and Sox2) has a role in cellular reprograming of somatic cells to a pluripotent state (Takahashi and Yamanaka, 2006). The forced up regulation of Myc supports self-renewal in the absence of LIF. Whereas cessation of LIF signaling results in a decrease in Myc expression presumably through a down-regulation of STAT3. Apart from the above-mentioned STATs a wide range of other downstream effector molecules can be activated through LIF receptor activation including extracellular regulated kinases (ERK), mitogen-activated protein kinases (MAPK) and phosphatydilinositol-3 kinase (PI3K). The network of interactions between intracellular pathways and extracellular ligands continues to develop a pace, with numerous overlaps being identified. In this context another kinase, glycogen synthase kinase 3 (GSK3) a key enzyme in the Wnt pathways is quickly activated resulting in Myc phosphorylation and its degradation. The activity of GSK3 may be controlled by PI3K either directly or indirectly due to LIF signaling. Another possible network connection is that between LIF, PI3K and the Wnt pathway in self-renewal comes from the data that shows improved results in the derivation of mES cells in the presence of the GSK3 inhibitor BIO. Thus from a signaling perspective multiple pathways may be involved in the maintenance of low levels of GSK3 activity to promote pluripotency and mES cell self-renewal. The array of signaling pathways and the level of crosstalk that exist between them and the LIF-STAT3 pathway in mES is slowly being deciphered giving us a clearer picture of the connections between LIF signaling and the transcriptional machinery controlling self-renewal.
3.2. PI3K Pathway
Phosphatidylinositol 3 kinases (PI3Ks) are recognized to modulate a wide range of cellular functions from growth, proliferation and self-renewal to simple metabolic control. They are a family of enzymes, which phosphorylate the 3′-OH position of the inositol ring of phosphoinositides. In 1987 (Whitman et al., 1987) identified two distinct phosphatidylinositol kinases (PIKs) isolated from fibroblasts. They further demonstrated that one of these enzymes associated with activated tyrosine kinase receptors. They called this kinase type I PIK. Subsequently the same group showed that the most abundant form of the previously identified enzymes, type II PIK, phosphorylates the D-4 position on the inositol ring and that type I PIK phosphorylated the inositol ring at the D-3 position.
Currently the family is divided into 3 classes based on structure and substrate preference (Wymann and Pirola, 1998; Vanhaesebroeck et al., 2001). Class I PI3Ks form heterodimers, consisting of a ∼110 kDa catalytic subunit, and a regulatory subunit. The regulatory subunit comes in 4 main flavours (p85a, p55a, p50a, p85b, p55g) and a catalytic subunit in 3 major types (p110a, p110b, p110d) (Engelman et al., 2006). In vivo the primary substrate is phosphatidylinositol-4,5, bisphosphate (PtdIns(4,5)P2 or PIP2), which is converted to phosphatidylinositol-3,4,5, triphosphate (PtdIns(3,4,5)P3 or PIP3) (Cantley, 2002). This class of PI3Ks are activated by an array of plasma membrane receptors (for a review see Wymann et al., 2003). Class II PI3Ks produce PI(3)P and PI(3,4)P2 in vitro, but in vivo targets are less clear but the enzyme itself has been localized to the Golgi network. Class III PI3Ks produce only PI(3)P. Much of what we know about the functions of PI3K is because a potent and quite specific inhibitor is available. Wortmannin and LY294002 act as competitive ATP binders targeting the ATP-binding site of catalytic p110 subunit. The most interesting early finding was that wortmannin in the low nanomolar range blocked the respiratory burst of neutrophils (Baggiolini et al., 1987). Studies on purified enzymes have shown that the mammalian PI3K is the most sensitive to wortmannin (Yano et al., 1993). The use of these inhibitors has proved invaluable in the study of PI3K and its cellular effects (reviewed by Nakanishi et al., 1995). The best known product of PI3K action is PIP3 which has been shown to be an important second messenger capable of recruiting AKT and involved in numerous cellular pathways associated with growth, proliferation and survival (Cantley, 2002). The production of PIP3 facilitates the recruitment of pleckstrin homology (PH) domain containing proteins an important example of which is the protein kinase Akt which itself has multiple intracellular targets (Toker, 2002). Commonly in transformed cells the PI3K/Akt pathway is directly activated by the loss of PTEN, a negative regulator of PIP3 formation and an identified tumor suppressor. Maybe unsurprisingly in mES cells the role of PI3Ks was highlighted by the fact that in PTEN null mES cells, accelerated cell cycle progression was observed (Sun et al., 1999) which can be blocked by the PI3K inhibitor LY294002. However a role for PI3K signaling events has also been identified in the maintenance of pluripotency in mES cell derived for a number of species (Fig. 3) (Armstrong et al., 2006). Blocking PI3K signaling events results in elevated ERK/MAPK signaling (Paling et al., 2004) and there is evidence to suggest that ERK (Hamazaki et al., 2006) and Wnt (Sato et al., 2003) signaling are required to sustain pluripotency in both mouse and human ES cell lines. In the
case of mES cells inhibition of PI3K pathways can induce differentiation in the presence or absence of LIF (Paling et al., 2004; Armstrong et al., 2006). However interestingly up regulation of Akt signaling is sufficient to maintain pluripotency of m ES cells (Watanabe et al., 2006). Another linkage for PI3K signaling and self-renewal comes from evidence that Nanog expression as well a number of Nanog target genes are modulated by PI3K signaling. Results have shown that the loss of pluripotent phenotype associated with PI3K blockage by LY294002 can be rescued by exogenous Nanog expression. Also regulation of GSK3 activity acting downstream of PI3Ks plays a role in Nanog expression. The evidence is clearly points out that PI3Ks play an important role in the signaling and maintenance of Nanog expression. PI3K effects are not limited directly to Nanog alone, inhibition of PI3K pathways results in the repression of rfx4, an identified Nanog target (Storm et al., 2007). However, interestingly, of the triad of master factors Oct4, Sox2, and Nanog, it appears that Nanog alone is sensitive to PI3K signaling pathways. However recently it has been shown that suppression of PI3K leads to a reduction in other self-renewal transcription factors including Klf4 (Storm et al., 2009), one of the targets in iPS generation. The role of PI3K in ES cells is complicated by the fact that self-renewal and cell proliferation are linked, and PI3Ks have been cast in major roles for both cellular processes.
3.3. Wnt pathway
The name “Wnt” comes from the fusion of the two names, int (based on the proto-oncogene integration-1 (Tanaka et al., 2002) and wg (based on wingless the segment polarity gene in Drosophila). The Wnt proteins are defined by amino acid sequence rather than by noted functional activities, but all Wnts share a number of common properties like numerous glycosylation sites and target sequences for secretion (Nusse and Varmus, 1992). Upon Wnt binding to its specific receptor, a signaling cascade is activated ultimately upregulating Wnt target genes. The Wnt signaling system is a highly conserved network controlling numerous other signaling transduction pathways from embryonic development to adult tissue homeostasis. Approximately 19 different WNT proteins have been identified acting on at least three different signaling pathways (Nusse and Varmus, 1992). The three pathways are the canonical Wnt pathway, acting via β-catenin and Tcf/Lef factors; the planar cell polarity (PCP) pathway; and the Wnt-Ca2þ pathway (Staal et al., 2008). This section will focus only on the canonical pathway. β-catenin is a well-known cytoplasmic protein and has a role in cell-cell adhesion acting to link membrane bound cadherins to the actin elements in the cytoskeleton. However it is now known to also act as a signaling molecule inside cells as part of the canonical Wnt signaling pathway (Reya and Clevers, 2005). In the absence of Wnt, β-catenin exists in a phosphorylated state in a complex marked for degradation by the ubiquitin-associated proteases. The β-catenin degradation complex includes the tumor suppressor proteins adenomatous polyposis coli gene (APC), Axin, and GSK3. Wnt signaling involves the Wnt ligand binding to the membrane receptor named Frizzled (Fz). Frizzled is a seven transmembrane receptor and the first receptor identified to bind the Wnt ligand (Bhanot et al., 1996). Activation of signal transduction by Wnt binding the Fz receptor requires a co-receptor attachment with a member of the low-density lipoprotein (LDL) family called Lrp5 and −6, this interaction is required for activation of the canonical Wnt signaling pathway (Li and Bu, 2005). Activation of Fz by Wnt results in the protection of β-catenin from proteosomal degradation. Thus the action of Wnt is to maintain the intracellular levels of β-catenin which then translocates to the nucleus where it forms a transcription complex with one of a number of transcription factors including Tcf1, Tcf3, Tcf4, or Lef1 (Okamura et al., 1998). Tcf1 is found mainly in T lymphocytes, Tcf4 is widely expressed and found in stem cells of gut while Tcf3 is expressed in mES cells. In mES cells there is growing but often-conflicting evidence that Wnt signaling pathways are important components of mES cell self-renewal. Wnt pathways have been shown to sustain pluripotency but also are important for of adult progenitor cell proliferation. The focus on Wnt signaling and its role in pluripotency comes from studies using the GSK3 inhibitor 5-bromoindirubin-3-oxime (BIO) (Fig. 4) (Meijer et al., 2003; Sato et al., 2003).
Inhibition of GSK3 prolongs the existence of β-catenin, causing it to accumulate, increasing the pool, which can translocate to the nucleus and activate gene expression. BIO has been shown to be able to maintain pluripotency of mouse and human ES cells in the absence of LIF (Sato et al., 2003). In a similar vain activation of Wnt signaling indirectly by removing the inhibitory effect of APC sustains pluripotency, suggesting the Wnt signaling is required for self-renewal (Kielman et al., 2002). In addition, treatment with Wnt3a was found to stimulate hES cell proliferation (Singla et al., 2006). Oct4 over-expression increased β-catenin transcriptional activity in progenitor cells. The Wnt controlled transcription factor Tcf3 has been shown to repress Nanog and thus promote differentiation. More recent studies have shown that Lef1 acting along with β-catenin is able to up-regulate Oct4 expression and interact with Nanog and thus promote self-renewal. All these data suggest that Wnt/β-catenin signaling has some role in the mES cell self-renewal (Takao et al., 2007). In obvious contrast to LIF and BMP signaling in mouse and human ES cells there is no difference between the cell types with regard to Wnt/β-catenin signaling self-renewal (Hao et al., 2006). However, contrary to the above-portrayed role of Wnt in self-renewal, Wnt action has been shown to facilitate differentiation of mES cells into neural precursors and increases the expression of Brachyury a mesoderm marker (Yamaguchi et al., 1999). More work is required to elucidate the role of Wnt signaling in mES cell self-renewal and pluripotency and potential other effect for the non-canonical Wnt pathways.
4. Summary and conclusion
Even after prolonged periods and numerous expansions in culture ES cells retain the ability to respond to normal developmental signals and display no apparent bias for any one cell lineage when reintegrated to a developing embryo. Constructing a stable and coherent map of how ES cells achieve such a feat is a major challenge that must be met if the true potential of these cells is to be realized in a clinical setting. A fundamental breakthrough in this area came with the generation of a tetracycline-suppressible