Embryonic stem (ES) cells possess the unique property of being able to retain their capacity for self-renewal and potential to form cells of all three embryonic germ layers (endoderm, mesoderm, ectoderm). Understanding the factors that determine the ability of ES cells to maintain their self-renewal and pluripotency, and the interplay of these factors, is of utmost relevance for both developmental biology and stem cell research. Over the last two years, a trio of transcription factors (TFs) have emerged—OCT4, SOX2, and NANOG—which play a key role in determining the fate of ES cells [1
]. Many of these findings appear to hold both for human and murine systems, but some distinct differences exist with regard to target genes (TGs) of these transcriptional regulators and external factors. For example, LIF and BMP4 are critical external factors which maintain mouse, but not human, ES cells in vitro [5
]. In contrast, BMP4 induces trophoblast differentiation of human ES cells [7
], and this is accompanied by downregulation of OCT4. Factors involved in maintaining and propagating human ES cells are bFGF, Activin A, and the BMP antagonist Noggin [8
]. The wnt pathway is important for both murine and human ES cell maintenance [11
]. We consider networks that apply to both mouse and human. For reasons of simplicity, we use the symbols SOX2, OCT4, and NANOG for both species.
Recently, two genome-wide studies [3
] were performed, in which these TFs were localized at promoter regions of human and mouse ES cells. Several transcriptional bindings were discovered: NANOG binds to the promoter regions of OCT4 and SOX2, as well as to its own. In addition, the OCT4–SOX2 heterodimer regulates NANOG, OCT4, and SOX2 individually. What emerges is a compact network with self-regulation and positive feedback. As pointed out in [3
] and revealed by another large gene perturbation study with mouse ES cells [12
], OCT4, SOX2, and NANOG also regulate a number of other genes. Many of these TGs are themselves TFs, some are responsible for maintaining ES cells by controlling self-renewal and pluripotency, and others perform key developmental functions that include differentiation into extra-embryonic, endodermal, mesodermal, and ectodermal cell types [3
]. Such genes, which are crucial for development, are found to be repressed in ES cells. Henceforth, when we refer to differentiation genes,
we imply a set of genes that are upregulated upon differentiation of ES cells and are negatively regulated by NANOG, OCT4, and SOX2. Similarly, by stem cell genes
we mean those genes that are expressed in ES cells and are positively regulated by the three TFs. Thus, in this study we concentrate on TGs regulated by all three factors.
The three TFs OCT4, SOX2, and NANOG thus regulate genes with two distinct and opposing functions: self-renewal and differentiation. The TGs are regulated both by OCT4 and NANOG individually, and by the combined effects of OCT4–SOX2 and NANOG [3
]. Furthermore, since OCT4–SOX2 activates NANOG, the TGs are at the receiving end of a feedforward loop [13
]. We explore the consequences of this architectural feature after establishing the dynamics of the core of the network.
The core transcriptional network that emerges from [3
] is depicted in . We have investigated whether similar architectures exist in the gene regulatory networks of other organisms. Our study suggests that the architecture is unique to higher organisms, and it cannot be found outside the ES cell system from the available databases. In we include additional signals, denoted A+
, and B−
. Signals A+
positively and negatively regulate the expression of OCT4 and SOX2. The majority of Oct4 binding sites at TGs in ES cells are Sox–Oct composite elements [4
] and here we consider OCT4 and SOX2 as common TGs of signals A+
. Signals B+
positively and negatively regulate NANOG expression, respectively. We do not consider signals that act either positively or negatively on all three factors, since this situation is covered by the presence of A+
. Several growth factors triggering certain signal transduction pathways required for keeping ES cells in an undifferentiated and multipotent state as well supporting their self-renewal have been described in the literature as mentioned above. These factors can be different for human and mouse ES cells. However, a signal pathway positively controlling both mouse and human ES cells is, for example, the wnt pathway (A+
). It is currently not known which of the three TFs are directly regulated by wnt in ES cells. An example of a negative external signal (A−
) is BMP4, which induces trophoblast differentiation in human ES cells [7
]. In principle, signals A+
, and B−
can also represent internal factors such as p53. p53 can be induced by DNA damage and negatively regulates NANOG in mouse ES cells [15
]. It has been hypothesized that such a program gives the stem cell the opportunity to initiate differentiation, thereby reducing the impact of DNA damage [15
The Core Transcriptional Network of the ES Cell
We will discuss the dynamics of the transcriptional network as a function of the inputs A+, A−, B+, and B−. The operator sites on OCT4, SOX2, and NANOG at which these factors act are assumed to be far from the binding sites for OCT, SOX2, and NANOG, and hence no direct interactions between the factors and the trio of TFs occur.
We use the Shea–Ackers approach [16
] to construct a stoichiometric model, which describes the dynamics of OCT4, SOX2, and NANOG. Using these tools we can describe the regulation of downstream TGs by OCT4, SOX2, and NANOG and explore the switch between self-renewal/pluripotency and differentiation—two very different outcomes.
The transcriptional network of contains positive feedback loops, which will turn out to give rise to bistable switch-like behavior, for a wide range of dynamical model parameters. Through the feedback loops, the expression of OCT4, SOX2, and NANOG can be jointly triggered or blocked by the environmental signals. We will argue that this switch-like behavior is highly useful in the context of stem cells, as it gives a clear separation of two very different cell fates: either maintain a stem cell state and differentiation is blocked or self-renewal and pluripotency of the ES cell is lost and differentiation is initiated. In Boolean terms, the stem cell genes should be on and the differentiation genes off, or the other way around, depending on whether the trio of TFs all are on or off. The model explains, qualitatively, several experimentally observed results concerning the behavior of OCT4, SOX2, and NANOG and their downstream TGs for maintaining an ES cell and for differentiation with respect to upstream signals. We use the model to generate several predictions. These predictions could be used to further validate the model or to pinpoint missing components.