Neural crest (NC) is a transient embryonic cell population that have a broad differentiation potential beginning with a well orchestrated gene regulatory control that has been extensively studied and reviewed (Mayor et al., 1999; Aybar and Mayor, 2002; Meulemans and Bronner-Fraser, 2004; Steventon et al., 2005; Sauka-Spengler and Bronner-Fraser, 2008; Betancur et al., 2010). However, an emerging body of evidences suggests that posttranscriptional and epigenetic contributions also play an important role in NC development. Here, we give a brief overview of NC development and then death into the current state of knowledge regarding epigenetic and miRNA contributions in this field.
NC cells are first induced in the neural plate border by WNT, BMPs, and FGFs signals as neurulation begins. These “inducing signals” further activate Zic, Msx, Dlx, and Pax family members to establish the neural plate border. The neural plate border then elevates to form the neural folds, where a subpopulation of precursor cells at the dorsal aspect of the fold is imbued with a NC potential. Spatial and temporal expression of “neural crest specifier genes” such as Snail2, FoxD3, Sox9, Sox10, AP-2, Id, and c-Myc specify their bona fide NC fates (Meulemans and Bronner-Fraser, 2004; Sauka-Spengler and Bronner-Fraser, 2008; Betancur et al., 2010). Later NC cells undergo an epithelial-to-mesenchymal transition (EMT) to migrate out of the neural tube and give rise to diverse derivatives such as craniofacial cartilage and bone, melanocytes, smooth muscle, and peripheral and enteric neurons and glia (Sauka-Spengler and Bronner-Fraser, 2008). NC cells are often referred to as the fourth germ layer and have been central to vertebrate evolution. In addition, defects in NC cell development are associated with several congenital defects, collectively known as neurocristopathies. These include disorders or craniofacial and organ development as well diseases such as neuroblastoma and melanoma. Consequently, understanding the normal epigenetic mechanisms of NC cell development will provide important clues regarding the mistakes that may lead to abnormal development or loss of the differentiated state.
The term epigenetic was initially proposed by Waddington (1957) as “the branch of biology which studies the causal interactions between genes and their products, which bring the phenotype into being.” He described cellular differentiation as a process largely governed by changes in the “epigenetic landscape” rather than alterations in genetic inheritance. In this context, epigenetics is actually defined as “the study of any potentially stable and, ideally, heritable change in gene expression or cellular phenotype that occurs without changes in Watson-Crick base-pairing of DNA” (Goldberg et al., 2007). Recently, there has been a myriad of publications dedicated to epigenetics, which is gaining recognition as a key factor in the fine regulation of gene expression and cell differentiation. NC development is not an exception and nowadays there is much emerging data in this promising and exciting field.
Epigenetic research is converging on the study of covalent and noncovalent modifications of DNA and histone proteins and the mechanisms by which such modifications influence overall chromatin structure that in turn regulates gene expression. These covalent and noncovalent modifications include posttranscriptional histone modifications, histone variants, DNA methylation, SUMOylation and noncoding RNAs.
MicroRNAs (miRNAs) comprise species of short noncoding RNAs that regulate gene expression posttranscriptionally. Recent studies have demonstrated that epigenetic mechanisms, like DNA methylation and histone modifications, not only regulate the expression of protein-encoding genes, but also miRNAs. In an opposite manner, miRNAs are also involved in controlling the expression of important epigenetic regulators, including DNA methyltransferases, histone deacetylases, and polycomb group genes (Sato et al., 2011).
In this review, we will describe the actors involved in the “writing,” “erasing,” and/or “reading” of the “epigenetic code,” as well as miRNA participation necessary to achieve the cellular memory that dictates gene expression programs in NC development.
Gene expression occurs in a chromatin context and therefore the wrapping of the genome around chromatin is now known to be a critical feature of gene regulation mechanism. Histones are one of the oldest families of proteins closely associated with DNA molecules, responsible for chromatin structure and playing an important role in the regulation of gene expression. The histones are organized in units called nucleosomes that are comprised of an octamer containing two molecules each of the four histones (H2A, H2B, H3, and H4), around which is wrapped 147 bp of DNA. The core histones are highly conserved basic proteins with globular domains and flexible N-terminal “tails” that protrude from the nucleosome susceptible to a variety of posttranslational modifications (Berger, 2007; Kouzarides, 2007; Gibney and Nolan, 2010). Acetylation and methylation of core histones, notably H3 and H4, were among the first covalent modifications to be described, and were long proposed to correlate with positive and negative changes in transcriptional activity. Since the pioneering studies of Allfrey and coworkers (1964), 130 sites for posttranscriptional covalent histone modifications have been identified and characterized; these include histone propionylation, butyrylation, formylation, phosphorylation, ubiquitylation, sumoylation, citrullination, proline isomerization, ADP ribosylation, tyrosine hydroxylation, and lysine crotonylation (Tan et al., 2011). Even though all of these histone modifications have relevant implications in transcriptional regulation, only methylation and acetylation were studied during NC development.
Histone methylation is one of the best studied histone modifications, associated with both transcriptional activation and repression. Methylation of certain residues of Histone 3 and combinations of thereof, are involved in the recruitment of various modifiers of chromatin and transcriptional activators or repressor, resulting in different effects on gene expression (Kouzarides, 2007). The trimethylation of H3K4 (H3K4me3), catalyzed by histone methyltranferases of the Trithorax group (TrxG) proteins, is associated with active transcription (Barski et al., 2007; Pan et al., 2007; Cheung et al., 2010); while trimethylation of H3K27 (H3K27me3), established by Polycomb group (PcG) proteins, is associated with transcriptional repression (Schwartz et al., 2006; Tolhuis et al., 2006; Liu et al., 2011). Similarly to H3K4me3, trimethylated H3 lysine 36 (H3K36me3) is also frequently located in transcriptionally active euchromatic regions, with the first predominantly found in the promoter and the second in the bodies of genes. On the other hand, trimethylated H3 lysines 9 and 27 (H3K9me3 and H3K27me3) are repressive marks associated with heterochromatin or euchromatic repressed genes (Simon and Kingston, 2009).
Coordination to establish the correct “epigenetic code,” removing or adding repressive or activating marks, plays a key role during development, and failures in either step cause dysregulation of gene expression, which has been strongly linked to different diseases (Swigut and Wysocka, 2007; Nottke et al., 2009; Herz and Shilatifard, 2010; Ho and Crabtree, 2010; Lindeman et al., 2011). However, there are few publications that associate changes in histone methylation and NC development. In our work, we reported that JmjD2A is responsible for the demethylation of both H3K9me3 and H3K36me3, and demonstrated that it is required for activation of several key NC specifier genes such us Sox9, Sox10, FoxD3, and Snail2 (Strobl-Mazzulla et al., 2010). Consistent with this role, JmjD2A is initially found throughout the neural plate, but as neurulation proceeds, its expression becomes restricted to the NC progenitor-forming region and eventually disappears from migrating NC cells. We also showed that JmjD2A is required for H3K9me3 demethylation at the Sox10 and Snail2 loci to activate their expression prior NC cells specification (Strobl-Mazzulla et al., 2010), thereby demonstrating its mechanism of action. These finding strongly support the idea that histone modifications are involved in setting up the developmental program for NC cells specification.
Recent studies in zebrafish, demonstrated that PHF8 (plant homeodomain finger protein 8) is a JmjC domain-containing protein implicated in H4K20me1 and H3K9me1/me2 demethylation that has an important role in craniofacial development (Qi et al., 2010). This role is in part a consequence of the direct regulation of the homeodomain transcription factor MSX1/MSXB expression, which function to refine the neural non-neural border and NC development (Phillips et al., 2006).
Histone acetylation and deacetylation of lysine residues are conducted by histone acetyltransferases (HATs; Carrozza et al., 2003), and histone deacetylases (HDACs; Hsieh et al., 2004). Acetylation of lysines neutralize their positive charge, reducing the strength of the binding between the histone with negative charge of the DNA, thus allowing the opening of chromatin structure and facilitating transcription of target genes (Ekwall, 2005; Wang et al., 2009). Conversely, deacetylation of histones generally result in DNA packaging into condensed chromatin and the silencing of gene expression.
Class I HDACs include HDAC1, HDAC2, HDAC3, and HDAC8. Despite the fact that these proteins are widely expressed during embryogenesis, they play specific roles as modulators of early developmental processes and organogenesis (Montgomery et al., 2007; Knutson et al., 2008; Haberland et al., 2009; Ye et al., 2009; Dovey et al., 2010). Conditional deletion of HDAC8 by Wnt1-Cre mice, but not HDAC1 and HDAC2, has demonstrated an essential role in cranial NC cell differentiation into facial skeleton (Haberland et al., 2009). In this work, HDAC8 is required for the suppression of Otx2, Lhx1, and other homeobox transcription factors in cranial NC cells. This demonstrates a highly specific developmental function (Haberland et al., 2009). Consequently, it has been reported that mothers who take the HDAC inhibitor valproic acid during the first trimester of pregnancy have a significantly increased risk of their fetuses developing craniofacial abnormalities (Alsdorf and Wyszynski, 2005; Wyszynski et al., 2005). Even though HDAC1 seems to be no essential for cranial NC differentiation, studies in zebrafish demonstrated the requirement of this protein for the proper repression of the transcription factor FoxD3 necessary for the mitfa-dependent melanophore development (Ignatius et al., 2008). Moreover, recent works demonstrated that HDAC3 plays a critical and specific regulatory role in the neural crest-derived smooth muscle lineage and in the formation of the cardiac outflow tract (Singh et al., 2011).
HDAC4 is a member of the class II histone deacetylase that has been demonstrated to play a key role during development in humans. HDAC4 deficiency has been associated with nonsyndromic oral clefts and brachydactyly mental retardation syndrome (BDMR) with craniofacial abnormalities (Park et al., 2006; Williams et al., 2010). In addition, studies performed in zebrafish demonstrated that HDAC4 reduction results in embryos and larvae with shortened faces and skeletal reduction and/or clefting (Delaurier et al., 2012).
Histone acetylation also acts as a binding site for bromodomain proteins, which function to recruit transcriptional activators or repressors (Grunstein, 1997). Similarly, our recent work showed that a plant-homeodomain 12 (PHD12) protein, with two PHD domains and a bromodomain, directly interacts with the Sin3A/HDAC repressive complex, which in turn interacts with Snail2. Together, they form a complex at the Cad6b promoter. We found that knock-down of either PHD12 or Snail2 prevents Cad6b promoter H3 lysine deacetylation, necessary for it repression and for the subsequent NC epithelial-to-mesenchymal transition (Strobl-Mazzulla and Bronner, 2012).
Chromatin remodeling complexes are enzymes that transiently disrupt the association between DNA and histones in an ATP-dependent manner. This in turn may induce conformational changes in nucleosomes and control different degrees of the condensation state of chromatin A recent study has shown that CHD7 (chromodomain helicase DNA-binding domain), an ATP-dependent chromatin remodeler related to the Drosophila trithorax-group factor Kismet, is essential for activation of core components of NC transcriptional gene regulatory network, including Sox9, Twist, and Slug (Bajpai et al., 2010). This study found that CHD7 is able to associate with distant enhancer on those genes essential for their activation. Moreover, in neural crest cells induced from human embryonic stem (ES) cells, CHD7 was also found to associate with PBAF a SWI/SNF family chromatin-remodeling complex, and both co-occupy a neural crest specific Sox9 enhancer as well as a regulatory element upstream of Twist marked by H3K4me1. Similarly, studies made in zebrafish mutant or morpholino knock-down for Brg1, a member of the SWI/SNF complex, demonstrated a critical role during NC induction and differentiation (Eroglu et al., 2006).
Williams Syndrome Transcription Factor (WSTF) is one of ∼25 haplodeficient genes in patients with the complex developmental disorder Williams Syndrome (WS). This syndrome is an autosomal dominant disorder resulted from the deletion of ∼1.5 megabases on chromosome 7 (Lu et al., 1998). These individuals exhibit characteristic malformations of craniofacial, heart, and neural structures; infantile hypercalcemia; developmental delays; and distinctive cognitive and behavioral profiles. WSTF is a core component of two functionally distinct chromatin remodeling complexes, WICH and WINAC, and is expressed in a wide variety of tissues during development (Lu et al., 1998). Several studies support the idea that WSTF contributes to NC development. In Xenopus embryos, WSTF expression overlaps with NC markers (Cus et al., 2006), and WSTF-null mice show defects in heart and craniofacial skeleton (Yoshimura et al., 2009). Moreover, WSTF-depleted embryos evidenced normal NC induction/specification, but presented a severe defect on migration and/or maintenance of NC cells (Barnett et al., 2012). Taking together these results suggest the idea that NC defects, resulting from WSTF haploinsufficiency, may be one of the major contributors of WS. However, because WSTF is expressed in a wide variety of tissues, NC-specific knockout models will be needed to address this hypothesis.
The Adipocyte enhancer binding protein 2 (Aebp2), together with long noncoding RNAs, has been implicated in the recruitment of the Polycomb Repression Complex 2 (PRC2) to a subset of genomic loci (Kim et al., 2011). In developing mouse embryos, Aebp2 is mainly expressed within cells of NC origin and heterozygotes embryos display similar phenotypes to those observed in NC-derived disease, such as Hirschsprung and Waardenburg syndrome. Moreover, Aebp2 and PRC2 are direct upstream targets of several genes involved in NC specification and migration (Kim et al., 2011). This result suggests a role for Aebp2 in the regulation of NC development through the PRC2-mediated epigenetic mechanism.
DNA methyltransferases (DNMTs) are in charge of DNA methylation by recognizing CpG and catalyzing transfer of a methyl group to the cytosine residues on DNA (Cheng and Blumenthal, 2008). It has been well established in cancer and stem cells that high CpG methylation occurs at the promoter region correlates with inhibition of gene expression (Momparler and Bovenzi, 2000; Miranda and Jones, 2007; Altun et al., 2010). Moreover, there are several studies that highlight the importance of DNA methylation in disease and during normal development of organisms (Martin et al., 1999; Mhanni and McGowan, 2004; Linhart et al., 2007; Ehrlich et al., 2008). There are three DNMT families in vertebrates: DNMT1, DNMT2, and DNMT3 (Goll and Bestor, 2005). DNMT1 is implicated in maintaining existing methylation patterns as well as on the regulation of histone methylation (Rai et al., 2006). The functional importance of DNMT2 in vertebrates is largely unclear. However, a recent study in zebrafish demonstrated its involvement in cytoplasmic RNA methylation (Rai et al., 2006). On the other hand, both DNMT3A and 3B have been demonstrated to be essential for de novo methylation throughout embryonic cell differentiation (Chen et al., 2003; Goll and Bestor, 2005; Reik, 2007). Although DNMT3A and DNMT3B exhibit overlapping functions, each has distinct expression patterns and genomic targets during development. Dnmt3A and Dnmt3B knockout in mES cells have common as well as distinct DNA targets (Okano et al., 1999; Chen et al., 2003). DNMT3A-null mice die several weeks after birth and DNMT3B-null embryos have rostral neural tube defects and growth impairment (Okano et al., 1999). Moreover, mutations in human DNMT3B result in ICF (immunodeficiency, centromeric instablility, and facial anomalies) syndrome, in which patients exhibit facial abnormalities (Jin et al., 2008), suggesting a defect related to abnormal NC development. Furthermore, a recent report have shown that DNMT3B depletion in hES cells results in a hypomethylation of pericentromeric regions, rather than changes to promoters of specific dysregulated genes (Martins-Taylor et al., 2012). In addition, it causes a loss of H3K27me3 and the polycomb complex protein EZH2 at the promoters of early NC specifier genes (PAX3, FOXD3, SOX10, and SNAI2). This premature upregulation of NC specifier genes in hES cells without the addition of morphogens suggests that the knockdown of DNMT3B enhances and accelerates the differentiation of NC precursors. However, up to now the specific role of DNA methylation and DNMTs on NC cells development has not yet been reported.
Due to advances in high-throughput transcriptome analyses, several projects have demonstrated that mammalian transcriptomes have a large complement of noncoding RNAs (ncRNAs). In fact, only 1–2% of the total transcriptome codes for proteins, while the great majority are transcribed as ncRNAs (Amaral et al., 2008), highlighting the importance of these molecules.
The miRNAs are a specific subgroup of ncRNA that are involved in myriad cellular events including the balance between proliferation and differentiation during tumorigenesis and organ development. The miRNA are approximately 22-nt-long RNA oligonucleotides that derived from hairpin structures present in lncRNA precursors or introns of coding or noncoding genes. They are processed in two consecutive cleavage steps by Drosha and Dicer and the mature miRNAs base-pairs with target mRNAs to inhibit translation or direct mRNA degradation by means of the RNA-induced silencing complex (RISC; Gibney and Nolan, 2010).
Eberhardt et al. (2008), demonstrated for the first time that miRNA, in particular miR-140, affects cranial NC cells dispersion and modulates palatogenesis. After this pioneer work, several groups have demonstrated that miRNA are central regulators of NC development (Amaral and Mattick, 2008; Eberhart et al., 2008; Cordes and Srivastava, 2009; Xin et al., 2009; Ivey and Srivastava, 2010; Subramanyam and Blelloch, 2011).
The conditional loss of the Drosha cofactor Dgcr8, which codes for a double stranded RNA-binding protein that is central for miRNA biogenesis, displayed a wide spectrum of malformations in cardiac NC cells, including persistent truncus arteriosus (PTA) and ventricular septal defect (VSD). Moreover, a significant portion of the cardiac NC cells underwent apoptosis, causing a decrease in the pool of progenitors required for cardiac outflow tract remodeling (Chapnik et al., 2012).
The miR-143 and miR-145 have been described as regulators of the way out from the pluripotent state by targeting pluripotency factors such as Klf4, Sox2, and Oct4. Remarkably, introduction of miR-145, but not miR-143, into NC stem cells was sufficient to guide specific differentiation into vascular smooth muscle cells (Cordes and Srivastava, 2009), demonstrating that miR-145 can direct smooth muscle fate.
The miRNAs have been shown to be important not only for cardiac NC cells but also to play a central role in NC craniofacial structures. A recent cross-species analysis has identified many miRNAs expressed in cranial NC cells from three avian species (chicken, duck, and quail) before and after species-specific facial distinctions occur, demonstrating a remarkably dynamic regulation of the expressed miRNAs (Powder et al., 2012). Of interest, in Xenopus embryos, loss of Dicer or miR-200b, miR-96, and miR-196a, leads to severe cranial cartilage defects, and may also be involved in NC induction (Gessert et al., 2010).
Conditional deletion of Dicer from NC cell lineage does not prevent the initial induction and migration of NC. However, as development progress, massive cell death, and complete loss of NC cell-derived craniofacial structures was observed. On the other side, migration and patterning of cardiac NC cells, but not survival, were impaired resulting in a variety of cardiovascular abnormalities. (Huang et al., 2010b; Zehir et al., 2010; Nie et al., 2011). The observed phenotypes were, at least in part, mediated by miR-21 and miR-181a, which in turn, affects the MEK/ERK signaling pathway (Huang et al., 2010b). Similarly, deletion of Dicer in the NC cells resulted in malformation of the dorsal root ganglia, enteric nervous system, and sympathetic ganglia (Huang et al., 2010a). However, Dicer has miRNA-independent function related to cell survival that may contribute to the observed phenotypes (Kaneko et al., 2011). This possibility needs to be considered when interpreting the phenotypes of systems in which Dicer is dysregulated, and it is necessary to perform further studies targeting specific miRNAs.
Similarly, disruption of miRNA biogenesis in NC cells is essential for cranial and cardiac neural crest development, but also specifically affects the expression of Dlx2 in the mandibular component of the first pharyngeal arch (PA1; Sheehy et al., 2010). In the same study, NC cells miRNA profiling showed that miR-452 is one of the most enriched and sufficient to rescue proper expression of Dlx2 in PA1. Moreover, the miR-452 regulated epithelial–mesenchymal interactions by directly targeting Wnt5a in the NCC-derived mesenchyme (Sheehy et al., 2010).
These diverse studies are revealing essential roles for miRNAs during NC development, and exploring the existence of cross talk with epigenetic regulation to build a comprehensive view of the epigenetic influence into the NC gene regulatory network. However, most of the reviewed results lack the standards of proof demonstrating the expression of the miRNAs in the right time and at the right place to be able to regulate the expression of the target gene in NC cells. Moreover, miRNA knockdown and overexpression should affect the expression of the target gene to unambiguously demonstrate that the targets are real; otherwise we are only correlating results.
Many birth defects and diseases are associated with abnormal NC development. Although some of them have been clinically studied, the molecular mechanisms associated with these conditions are scarce. While we are just barely spotting the surface of the iceberg in understanding the contributions of chromatin regulatory mechanisms and miRNAs to the formation and development of NC cells (Fig. 1), it is clear that this will be an important and fertile area of future investigation. It will be of great interest the use of high throughput technologies, such us RNA-seq, and ChIP-seq, to map out the “transcriptome” and “epigenome” of the regulation network controlling NC cells development. Moreover, the knowledge regarding the epigenetic contributions toward directing a particular fate during NC differentiation has also far reaching clinical implications.
Taking together, understanding the epigenetic landscape and miRNA regulation of NC progenitors and later during their differentiation should shed important light on the general characteristics of “stemness,” the mechanisms that led to the evolution of new cell types like the NC in vertebrates, and therapeutic treatment on birth defects and diseases associated with aberrant development of the NC cells.
We thank Dr. Marianne E. Bronner for critical reading of the manuscript. P.H.S-M. was funded in part by grants from Fogarty-NIH, ANPCYT, and CONICET.