To keep ESCs in an undifferentiated, pluripotent and self- renewal state, a complex network of signal transduction pathways in combination with transcription factors needs to be maintained. Key signaling pathways include the leukemia inhibitory factor LIF/ JAK/ Stat3, PI3K/ Akt, BMP2/ Smad and wnt signaling in combination with the dual inhibition of the FGF/
Erk and GSK3 signaling pathways (Niwa et al., 1998, 2009; Matsuda et al., 1999; Sato et al., 2003; Ying et al., 2003a, 2008; Paling et al., 2004; Watanabe et al., 2006; Berge et al., 2011;
Griffiths et al., 2011) (Figure 8). These conditions promote the expression of master transcriptional regulators of the pluripotency network, oct4 (also called pouf5I), nanog and sox2 (Okamoto et al., 1990; Nichols et al., 1998; Avilion et al., 2003; Chambers et al., 2003;
Mitsui et al., 2003). Together with epigenetic modifiers, non- coding RNA and the c-Myc transcriptional network, these crucial regulators form a core regulatory transcriptional network that promotes expression of pluripotency associated genes and represses genes for lineage commitment and differentiation (Boyer et al., 2005; Loh et al., 2006; Kim et al., 2008a) and reviewed in Ng and Surani, 2011; Orkin and Hochedlinger, 2011).
Introduction
24
Figure 8. Overview of signaling pathways that maintain pluripotency and self-renewal in mouse ESCs.
Several pathways synergistically act to keep ESCs in a pluripotent and self- renewal state. LIF can act via two different pathways: Binding to the gp130/ LIF receptor (LIFR) activates the Janus Kinase (Jak) which then phosphorylates signal transducer and activator of transcription 3 (Stat3). Activated Stat3 promotes the transcription of c-Myc and hence self- renewal and pluripotency. LIF can also activate the Phosphatidyl Inositol 3 Kinase (PI3K) which ultimately leads to activation of a serine/ threonine kinase (Akt), subsequent modulation of mTOR signaling and hence stimulation of proliferation and suppression of cell death. At the same time, Akt inhibits the glycogen synthase kinase 3ß (GSK3) normally involved in phosphorylation of ß-catenin followed by its degradation. By inhibiting GSK3, either by wnt signaling or Akt, ß-catenin can shuffle to the nucleus and stabilizes pluripotency by inhibiting Tcf3, a repressor of core pluripotency- associated transcription factors. In addition, the bone morphogenetic protein (BMP)/ Smad signaling contributes to pluripotency. Phosphorylated Smad1 triggers the expression of Inhibitor of differentiation (Id) proteins which block transcription factors involved in lineage commitment. Also ESCs themselves contribute to the metastable state of pluripotency by producing Fgf4 which drives ESCs to lineage commitment. In the pluripotent state, nanog is highly and - unlike oct4 and sox2 - heterogeneously expressed and counteracts Erk signaling. Based on the heterogeneous expression of nanog, cells with low nanog levels are more prone to Erk signaling and likely induce differentiation programs. (reviewed in (Okita and Yamanaka, 2006; Niwa, 2011; Welham et al., 2011).
A deeper understanding of how pluripotency is established and maintained by transcription factors and epigenetic modifications is of great interest not only for the facilitation of directed programming of ESCs to specific lineages but also for somatic cell reprogramming and hence holds great promises for the development of new therapies for diseases and regenerative medicine. Especially the master regulators Oct4, Sox2 and Nanog of the core transcriptional network have been extensively studied over the past years. Oct4, a POU domain-containing transcription factor encoded by Pouf5I, lies in the center of the network and is essential for the formation of a pluripotent founder cell population in the embryo (Nichols et al., 1998). The Oct4 levels within a cell need to be tightly controlled as already two fold changes in expression drastically affect the stem cell fate. Whereas an increase of oct4 results in differentiation towards primitive endoderm and mesoderm, a decrease of oct4
Introduction
25
leads to loss of pluripotency and differentiation to the trophoectodermal lineage (Niwa et al., 2000). Oct4 is one of the critical, so far irreplaceable factors of somatic cell reprogramming to iPS cells (Takahashi and Yamanaka, 2006; Nakagawa et al., 2007a).
Sox2 has been shown to function as a transcriptional partner of Oct4 (Avilion et al., 2003).
More specifically, an enhancer highly active in ESCs contains binding motifs for Oct4 and Sox2 and regulates the expression of pluripotency- associated genes, including nanog and fgf4. It has been shown that Oct4 and Sox2 collaborate to synergistically activate the enhancer, thereby promoting the expression of genes important to maintain pluripotency (Yuan et al., 1995; Rodda et al., 2005; Masui et al., 2007). The Oct4- Sox2 enhancers also regulate the expression of oct4 and sox2 themselves by a positive- feedback loop (Tomioka et al., 2002; Chew et al., 2005; Okumura-Nakanishi et al., 2005). Sox2 is part of the SRY-related HMG box protein family and its genetic ablation in mice results in early embryonic lethality (Avilion et al., 2003) and deletion of sox2 in ESCs leads to differentiation primarily into the trophoectodermal lineage (Masui et al., 2007).
The third member of the core transcriptional network, Nanog, is a homeodomain containing protein. Whereas lack of nanog in ESCs leads to loss of pluripotency and differentiation to the endodermal lineage, overexpression of nanog prevents the induction of differentiation and maintains pluripotency independently of the LIF/ Stat3 pathway (Chambers et al., 2003;
Mitsui et al., 2003). Consequently, Nanog is often referred to as the gate keeper of pluripotency, as it needs to be tightly controlled to allow reprogramming and differentiation (Silva et al., 2009); see also Figure 8). Nanog is known to show a very heterogeneous expression and latest data suggest that the heterogeneity of nanog expression is regulated on a chromosomal level. In early pre-implantation embryos, nanog is monoallelically expressed. However, its expression is gradually switched to biallelic expression as the inner cell mass (ICM) matures to the naïve epiblast and acquires ground- state pluripotency (Miyanari and Torres-Padilla, 2012).
Genome- wide comparative binding data of Oct4, Sox2 and Nanog revealed substantial overlapping binding sites on both, active and repressed promoters and enhancers in human ESCs. Actively transcribed genes occupied by one or two of the core regulators encode not only their own genes, but also transcription factors for components of signaling pathways including wnt and Tgf-ß pathways. Hence, the core factors promote not only their own expression by forming an interconnected autoregulatory loop but also promote the expression of genes encoding essential components of key signaling pathways as well as chromatin modifying proteins. Inactive genes co- bound by Oct4, Sox2 and Nanog are predominately genes involved in lineage commitment and differentiation. Taken together, the highly interconnected autoregulatory loop of the core transcription factors generates a
Introduction
26
metastable state in ESCs: on the one hand, the positive feedback loop promotes pluripotency if all three core factors are expressed at appropriate levels; on the other hand, perturbation of one factor by e.g. low expression unbalances the positive loop and might thereby favor entrance into a differentiation program (Boyer et al., 2005; Loh et al., 2006).
Although the three transcriptional regulators, Oct4, Nanog and Sox2, build the core of the pluripotency network, proteomic studies using affinity purification and mass spectrometry identified a large protein- protein interaction network including many transcription factors and chromatin modifying complexes that contribute to the maintenance of ESC properties.
Pluripotency related interacting proteins include Sal4, Rex1, Dax1, Klf4, Essrb, and Tcl1 as well as proteins linked to important signaling pathways to maintain pluripotency like Smad1, Stat3, and Tcf3 (Wang et al., 2006; Chen et al., 2008; Cole et al., 2008; Kim et al., 2008a, 2010; Mallanna et al., 2010; Pardo et al., 2010; van den Berg et al., 2010). Interacting proteins of chromatin modifying complexes involve components of the Swi/Snf (also called BAF) nucleosome remodeling complex, the NuRD/HDAC complex and polycomb complex 2 (Figure 9).
Figure 9. Overview of the oct4 centric module in ESCs
The oct4 centric module consists of a highly connected protein interaction network with Oct4, Nanog and Sox2 (circled in red in left picture) building important nodes within the network. The key regulators also directly or indirectly interact with various chromatin regulators (circled in green in left middle). Genome- wide binding data of the key regulators using chromatin immunoprecipitation (ChIP) followed by hybridization on a Chip or Sequencing revealed a complex transcriptional regulatory network (right picture). Many components of the network show autoregulatory but also interconnectivity mechanisms (adapted from Kim et al., 2010; Orkin and Hochedlinger, 2011).
Further studies investigating the genome- wide binding behavior of several key transcription factors in ESCs revealed a complex transcriptional regulatory network with many autoregulatory loops. Additionally, many transcription factors co- occupy target genes important for pluripotency, leading to a high interconnectivity among the components of the
Introduction
27
network (Figure 9). Consistent with this, Oct4 was shown to function as an anchor protein for the assembly of the multiprotein complexes on target genes (Chen et al., 2008; van den Berg et al., 2010).
By combining protein- protein and protein- DNA interaction studies, the various transcription factors active in ESCs cluster into two different modules: the Oct4- centric module, and the Myc- centric module (Chen et al., 2008; Kim et al., 2008a, 2010). The Oct4- centric module is also referred to as the core transcriptional network which consists aside from the master transcriptional regulators Oct4, Nanog and Sox2 of other transcription factors important for pluripotency (see also Figure 9). The Myc- centric module consists mainly of proteins associated with cell cycle regulation and metabolism, including c-Myc, n-Myc, Zfx, E2F1 and E2F4. Interestingly, components of the Myc regulated network include also various chromatin modifying enzymes like the HATs GCN5, p300 and Tip60-p400 complex (Kim et al., 2008a, 2010; Lin et al., 2009).
Interestingly, recent data suggest that the key regulators Oct4, Nanog and Sox2 also play crucial roles in cell fate choice and initiation of developmental processes. By integrating external differentiation signals which modulate Oct4 and Sox2 protein level and change their genome wide binding properties, lineage selection is induced without prior activation of any lineage specific markers (Boiani and Scholer, 2005; Thomson et al., 2011)