Mammalian development occurs by the progressive determination of cells from a pluripotent undifferentiated state through successive states of gradually restricted developmental potential, until the full complement of mature terminally differentiated cells has been specified. Embryonic development is a complex and highly orchestrated process during which multiple cell movements and changes in gene expression must be spatially and temporally coordinated to ensure that embryogenesis proceeds correctly. Complex genetic regulatory networks receive input in the form of extracellular signals and output instructions on the regulated expression of specific genes. The linchpins of the regulatory networks are the cis-regulatory elements that directly control gene expression through interpretation of the tissue-specific transcription factors (trans-elements). Embryonic stem cells are orientated across the dorso-ventral and the anterior-posterior axis of the early embryo. The orientation of progenitor cells along these two axes is thought to influence their fate by defining the identity and concentration of inductive signals to which they are exposed.In an effort to develop cell-based therapies, (i.e. for diabetes) experimental protocols aim to mimic the biological procedures that take place during embryonic development in order to differentiate embryonic stem cells towards specific cell types. One of the foremost challenges towards the development of cell therapies for diabetic people is to achieve the directed differentiation of cells capable of producing insulin. Elucidation of the genetic networks involved in the endocrine pancreas specification are thought to be essential for devising rational protocols to efficiently differentiate embryonic stem cells or pancreas progenitor cells into fully differentiated endocrine subtypes. Computational approaches allow the unravelling of complex regulatory networks including genomic (cis-cis) or proteomic (trans-trans) interactions or a combination (cis-trans) of both. In this study the genomic regulatory regions (cis elements) of several genes known and putative targets of the transcription factor NGN3 were analyzed. The NGN3 transcription factor is the major regulator of “insulin-producing cell” formation. Taking into account data from microarray experiments from pancreas progenitor cells, in which NGN3 has been induced, genes shown to be co-regulated (upregulated or downregulated) by this transcription factor were selected for analysis. Using a combination of sophisticated computational tools for exploiting and analyzing genomic data and developing the suitable algorithms, an extensive in silico analysis of the regulatory regions of these genes was performed.Evolutionarily conserved regions are linked with experimentally identified regulatory elements. Comparative genomics are commonly used in order to identify transcription factor binding sites, which are functionally important regions that are thought to be well-conserved. Analysis of genomic regulatory regions included not only genes corregulated by NGN3, but also their orthologs in several species including the most phylogenetically distant species (fish), which have pancreas. In parallel, housekeeping genes, like B-ACTIN, and those not expressed in embryos and stem cells, like B-GLOBIN, were used as negative controls. Regulatory region analysis revealed the presence of a highly conserved regulatory element, where many transcription factors with established involvement in pancreas development bind, in all the orthologs of several genes co-regulated by NGN3. Furthermore, motif identification in separate clusters of the regulatory elements of either upregulated or downregulated genes revealed the presence of additional binding motifs for the factor AP4 only in downregulated genes. In parallel, the regulatory region analysis of the entire mouse genome and the statistical analysis of the upcoming results showed that both types of regulatory elements (with and without AP4) were non-randomly identified inside the regulatory regions of genes whose transcription is controlled by NGN3. Moreover the selective presence of the AP4 binding sequence into this region renders it a highly specific suppressor found in only a small number of genes downregulated by NGN3. Taking into account that both these regulatory elements were identified at considerable distances from each gene’s transcription start site, it was assumed that they represent enhancers, and those capable of binding AP4 were considered silencers. This conclusion was enforced by the compositional analysis of these regions showing low GC levels, similarly to the majority of the regulatory regions implicated in embryonic development, something that has not been reported for promoter sequences. Moreover, analysis of protein-protein interactions showed that some of the transcription factors, predicted to bind onto these elements, together with other non-specific transcription factors, constitute a core transcription control complex. This protein complex interacts with the remaining members of the predicted cluster of transcription regulators and works either as an inducer or a suppressor of transcription. This is determined by the presence of a HAT and/or an HDAC in this protein complex assumed to locally control chromatin acetylation. Based on these data, we constructed a model of the complex regulatory network that describes how through the transcriptional regulation of the analyzed genes mainly guided by ΝGN3 the gradual differentiation of cells capable of producing insulin takes place.
Synthetic and genomic regulatory elements reveal aspects of cis-regulatory grammar in mouse embryonic stem cells
