Development of the lip and palate involves a complex series of events that requires the close co-ordination of programmes for cell migration, growth, adhesion, differentiation, and apoptosis (Ferguson, 1988). Failure of these processes results in orofacial clefting (OFC), which has a prevalence of 1 in 500–2,500 live births depending on geographic origin, racial and ethnic variation, and socio-economic status (Vanderas, 1987; Murray et al., 1997). OFC results in considerable morbidity to affected individuals and their families as well as imposing a substantial financial and societal burden. Individuals who exhibit OFC may experience problems with eating, speaking, and hearing, which can be corrected to varying degrees by surgery, dental treatment, and speech therapy. In addition, the altered facial appearance may lead to psychosocial consequences. The complex aetiology of OFC does, however, afford significant opportunities to identify genes and gene-environment interactions, and to dissect the fundamental mechanisms underlying craniofacial development and how these are disrupted in OFC (Murray, 2002).
On the basis that the lip/primary palate and the secondary palate have distinct developmental origins, OFC can be subdivided into cleft lip occurring either with or without cleft palate (CLP) and isolated cleft palate (CPO). The observation of CLP and CPO segregating within the same family has been termed the mixed clefting type (Neilson et al., 2002). Although OFC may occur as part of a syndrome, over 70% of cases of CLP and 50% of cases of CPO arise in the absence of other abnormalities and are collectively classified as non-syndromic (Jones, 1988). Van der Woude syndrome (VWS; OMIM 119300) is the most common form of syndromic OFC, accounting for 2% of all cases. This autosomal dominant condition, being characterised by the mixed clefting type, paramedian lower lip pits, and hypodontia, has the phenotype that most closely resembles the non-syndromic forms. Recently, mutations in the gene encoding the transcription factor interferon regulatory factor 6 (IRF6) have been demonstrated to underlie VWS (Kondo et al., 2002). IRF6 belongs to a family of transcription factors that share a highly conserved penta-tryptophan, helix-turn-helix DNA binding domain, and a less-well-conserved protein-binding domain (Taniguchi et al., 2001); however, the function of IRF6 during facial development remains unknown. To provide insights into the role of IRF6 during development of the lip and palate, we have analysed the expression pattern of Irf6 during facial development in mouse and chick embryos.
Development of the lip and primary palate in both the mouse and chick closely parallels that observed in man. The first signs of overt development of the primary palate occur on embryonic day (E) 9.5 in the mouse/Hamilton Hamburger (HH) stage 19 in the chick with the formation of the frontonasal prominence, the paired maxillary processes, and the paired mandibular processes that surround the primitive oral cavity. The formation of the nasal placodes subsequently divides the lower portion of the frontonasal prominence into paired medial and lateral nasal processes. The merging of the facial processes results in the upper lip becoming continuous by E11.5 in the mouse/HH stage 28 in the chick. The first sign of development of the secondary palate in the mouse occurs on E12 with the outgrowth from the maxillary processes of paired palatal shelves, which initially grow vertically down the sides of the developing tongue (E13). At E14, the shelves rapidly re-orientate to a horizontal position above the tongue and contact. The medial edge epithelia (MEE) of the apposed shelves adhere to form an epithelial seam (E14.5) that subsequently degenerates via a combination of apoptosis and epithelial-mesenchymal transformation to allow mesenchymal continuity across the palate (E15) (Griffith and Hay, 1992; Cecconi et al., 1998; Martinez-Alvarez et al., 2000; Cuervo et al., 2002). In contrast, although chick palatal shelves grow towards one another above the tongue and contact at HH stage 30–35 (E7-E9), they remain cleft rather than fusing (Shah and Crawford, 1980). While the mouse, therefore, provides an excellent system in which to study normal development of the secondary palate, the chick has the advantage of mirroring the pathology seen in cleft palate. The spatio-temporal expression patterns of Irf6 during development of the lip and palate in the mouse and chick were, therefore, analysed using a combination of in situ hybridisation and immunohistochemistry.
Expression of Irf6 in the developing mouse primary palate was observed in the ectoderm covering the first and second pharyngeal arches at E10.5 with high levels of expression also observed in the line of fusion between the apposed mandibular processes (Fig. 1A,E,I). At E10.75, Irf6 was expressed in the epithelium of the maxillary and mandibular processes of the first pharyngeal arch with strong expression being detected in the epithelial fusion zone between the medial and lateral nasal processes and the maxillary process (Fig. 1B,F,J). At E11.5, following fusion of the medial and lateral nasal processes, Irf6 was expressed in the epithelium of the intermaxillary segment (Fig. 1C,K). Irf6 was also expressed strongly in the nasolacrimal groove (Fig. 1G,K). During this period of facial development, Irf6 was almost exclusively restricted to the intra-oral epithelium; however, Irf6 expression was also detected in the ectoderm surrounding the developing eyes (Fig. 1K). At E12.5, the facial processes had completed fusion and residual Irf6 expression was detected in nares (Fig. 1D,L). Strong expression of Irf6 was also observed in the developing incisor tooth germs, the vibrissae, and supra and infra-orbital hair follicles (Fig. 1D,H).
At E12.5 of mouse development, as the palatal shelves began to extend from the maxillary processes, there was little evidence of Irf6 expression in the palatal epithelium. In contrast, Irf6 was expressed at a high level in the epithelium lining the floor of the mouth, particularly close to the base of the tongue, and in the invaginating molar tooth germs. This expression pattern was maintained during E13.5 as the palatal shelves extended vertically alongside the lateral border of the developing tongue, until immediately prior to palatal shelf elevation (Fig. 2A). At this point, Irf6 expression was detected in the epithelium at the tip of the palatal shelf extending along the presumptive nasal palatal epithelium but was absent from the presumptive oral palatal epithelium (Fig. 2D,G). As the palatal shelves elevated and came into contact during E14, Irf6 was expressed at high levels in the medial edge epithelia (Fig. 2B,E,H). Strong expression of Irf6 was also observed in the epithelia at the junction of the primary and secondary palate and along the base of the nasal septum (Fig. 2E,H). As fusion progressed during E14.5, Irf6 was expressed strongly in the medial edge epithelial seam (Fig. 2C,E,F) with increasing levels of expression being detected in the nasal and oral palatal epithelia as the midline seam degenerated. Throughout the period of development of the secondary palate, Irf6 expression was also detected in the epithelial components of the developing incisor and molar tooth germs with the highest levels being detected in the enamel knot and enamel organ close to the cervical loop (Fig. 2D,F). Irf6 expression was also detected in the vibrissae (Fig. 1). Notably, at all the stages examined, the localisation of Irf6 RNA and protein were precisely correlated.
To compare the expression of Irf6 in the mouse and the chick, we initially identified the chick Irf6 sequence using a combined bio-informatics and 5′ RACE approach. This allowed us to demonstrate that chicken Irf6 encodes a 460–amino acid protein that displays 86% identity with murine IRF6 and 85% identity with human IRF6, thereby confirming their orthologous nature (Fig. 3). The sequence conservation was highest in the DNA binding domain, where the chicken and murine/human sequences exhibit 93% identity with the penta-tryptophan repeat and those residues that are predicted to contact DNA being completely conserved (Fig. 3). The protein interaction domain was also highly conserved with the chicken and murine/human orthologues displaying 85% identity (Fig. 3). Further support for the genes being orthologous, rather than homologous, was obtained from studying their chromosomal locations; chicken IRF6 maps to chromosome 26 in a region of conserved synteny with human chromosome 1q32-q41 and mouse chromosome 1 (data not shown).
The expression patterns of Irf6 during early craniofacial development in the chick, as assessed by in situ hybridisation, mirrored those observed in the mouse. At HH stage 26 (E6.5), Irf6 expression was observed in the ectoderm surrounding the maxillary processes and frontonasal and lateral nasal processes with the highest levels of expression being detected in the fusion zones between the processes (Fig. 4A). Irf6 expression was also detected in the ectoderm overlying the second and third pharyngeal arches (Fig. 4D). High levels of Irf6 expression continued to be observed throughout HH stages 26–28 as the facial processes fused together to form the primary palate with strong expression also being detected in the paired mandibular processes, but the expression levels were down-regulated as fusion progressed. By HH stage 30 (E7), the primary palate and beak had completed fusion and expression of Irf6 was down-regulated at the fusion zones (Fig. 4B). The palatal shelves had also commenced their outgrowth from the maxillary processes at HH stage 30 but did not express Irf6. Irf6 transcripts were, however, expressed in the epithelium covering the dorsum of the tongue and in the nasal passages (Fig. 4E). At HH stage 32 (E8), the palatal shelves had approximated, and while Irf6 expression was observed in the nasal and oral palatal epithelia, Irf6 could not be detected in the medial edge epithelium (Fig. 4C,F). This pattern was maintained through HH stages 34 and 35 as the shelves made contact but failed to fuse. At these stages, high levels of Irf6 expression were also evident in the developing feather germs (Fig. 4G). It was not possible to determine the expression of IRF6 protein in the chick as the antibody utilised in this study did not cross-react with chick IRF6 as assessed by immunostaining (data not shown).
The results of the expression analyses detailed above indicated that there was considerable overlap between the expression pattern of IRF6 and transforming growth factor β3 (Tgfb3) during development of both the mouse and chick secondary palate (Fitzpatrick et al., 1990; Pelton et al., 1990; Sun et al., 1998). As Tgfb3 expression in murine palatal shelves commences at E13.5, preceding Irf6 expression and suggesting that IRF6 may be downstream of TGFβ3, the expression of Irf6 was analysed in Tgfb3−/− embryos, which exhibit cleft palate (Proetzel et al., 1995; Kaartinen et al., 1995). While the normal patterns of Irf6 expression were observed in wild-type embryos at all stages examined, Irf6 expression could not be detected in the MEE of the palatal shelves of Tgfb3−/− littermate embryos (Fig. 5). Irf6 expression was, however, detected in the tooth germs of the Tgfb3−/− mice providing an internal positive control (Fig. 5).
We are taking an integrated approach to dissect the complex pathways that underlie development of the lip and palate including genetic analysis of humans and mice, expression analyses, and experimental embryology using murine and chick embryos. We have demonstrated that mutations in IRF6 underlie Van der Woude syndrome, the most common form of syndromic cleft lip and palate (Kondo et al., 2002). More recently, Zucchero and co-workers have provided strong evidence that variation within IRF6 is responsible for approximately 12% of the genetic contribution to non-syndromic cleft lip or palate, at least in a subset of populations (Zucchero et al., 2004). However, although we have demonstrated that the sites of Irf6 expression mirror precisely the phenotype observed in VWS, the role that IRF6 plays during development of the lip and palate is unknown.
The mouse and the chick have played a central role in the dissection of the molecular pathways underlying facial development. In both species, development of the lip and primary palate closely parallels that observed in man. As the embryonic chick face is accessible for experimental manipulation in ovo, much of our knowledge concerning the development of the primary palate is derived from analysis of this species (Richman and Lee, 2003). Our knowledge of the development of the secondary palate, on the other hand, has been derived to a much greater extent from analysis of the mouse where the morphological events are essentially the same as those occurring in man with the palatal shelves fusing to form an intact palate (Ferguson, 1988). In contradistinction, although chick palatal shelves grow horizontally towards one another above the tongue and contact, they remain cleft rather than fusing (Shah and Crawford, 1980). While the mouse, therefore, provides an excellent system in which to study normal development of the secondary palate, the chick has the advantage of mirroring the pathology seen in cleft palate.
Irf6 is expressed at key stages of facial development in both murine and chicken embryos. Specifically, high levels of Irf6 expression were detected in the ectoderm covering the facial processes immediately prior to and during their fusion to form the lip and primary palate in both species. Intriguingly, while Irf6 was expressed strongly in the MEE of the secondary palate immediately prior to palatal shelf elevation and in the midline epithelial seam of the fusing palate in the mouse, a similar staining pattern was not observed in the chicken embryo. These observations strongly support a role for IRF6 in the fusion events that are central to the development of the lip and palate and provide a molecular explanation as to why the mixed clefting phenotype is observed in VWS.
A number of other molecules are thought to be involved in development of the lip and palate. For example, SHH, MSX1, and MSX2, and the control of BMP signalling, have been implicated in the control of the fusion events underlying development of the lip/primary palate (Sun et al., 2000; Ashique et al., 2002). Similarly, a number of growth factors and cell adhesion molecules have been implicated in fusion of the secondary palate (Fitzpatrick et al., 1990; Pelton et al., 1990; Dixon et al., 1991; Kaartinen et al., 1995; Proetzel et al., 1995; Miettinen et al., 1999; Mogass et al., 2000; Suzuki et al., 2000). Of the latter molecules, the expression pattern of Tgfb3 during palatal development in both the mouse and the chick is very similar to that of Irf6. Expression analyses in the mouse have indicated that TGFβ3 is expressed specifically in the future MEE immediately prior to palatal shelf elevation at E13 and in the MEE itself at E14.5 (Fitzpatrick et al., 1990; Pelton et al., 1990). Conversely, TGFβ3 is not expressed in the MEE of the chicken secondary palate (Sun et al. 1998). These combined observations suggested an important role for TGFβ3 in palatal fusion (Fitzpatrick et al., 1990; Pelton et al., 1990; Cui et al., 1998). This hypothesis was confirmed by the demonstration that blockade of TGFβ3 in vitro or ablation of the gene in vivo prevents palatal fusion in the mouse and that this can be rescued by administration of exogenous TGFβ3 (Brunet et al., 1995; Proetzel et al., 1995; Kaartinen et al., 1995, 1997). The down-regulation of Irf6 expression in the MEE of Tgfb3-null embryos presented here strongly suggests that IRF6 functions down-stream of TGFβ3 in the molecular events driving palatal fusion. Intriguingly, administration of exogenous TGFβ3 to the palatal shelves of Tgfb3-null embryos or chick embryos in vitro reverses the cleft palate phenotype observed in these species (Sun et al., 1998). Future experiments will, therefore, be directed at analysing whether IRF6 expression is restored in the Tgfb3-null mice and induced in the chick subjected to this treatment regime. Similarly, it will be interesting to analyse whether the administration of exogenous IRF6, for example by adenoviral delivery, can also rescue the cleft palate phenotype observed in these embryos. These experiments will be complemented by the generation and analysis of mice in which the function of IRF6 has been ablated. Analysis of such mice will facilitate further dissection of the cascade of molecular events in which IRF6 functions during development of the lip and palate and will help us to dissect the molecular pathogenesis of Van der Woude syndrome.
Eggs and Embryos
MF1 mice were mated and the morning on which the vaginal plug was detected was designated embryonic day (E) 0.5. Pregnant female mice were sacrificed, the embryos dissected from the uterine decidua and fixed in 4% paraformaldehyde in phosphate buffered saline (PSB) overnight at 4°C. Embryos were washed in PBS and dehydrated into methanol for whole-mount in situ hybridisation or processed through a graded series of ethanols, cleared in chloroform, and embedded in paraffin wax for section in situ hybridisation or immunohistochemistry. Fertilised White Leghorn chick eggs were incubated in a humidified chamber at 38°C for 5–10 days. Embryos were dissected from the eggs, staged according to Hamburger and Hamilton (1951), and processed as above.
Isolation of Chick Irf6
Searching the chick EST database (www.chick.umist.ac.uk) using the murine Irf6 coding sequence (NM_016851) resulted in the identification of two overlapping chick cDNA clones (ChEST745h20 and ChEST58f7). Alignment with the human and mouse cDNA sequences indicated that the assembled chick cDNA sequence lacked the 5′ end of the coding sequence. RACE experiments performed using the 5′ RACE System for Rapid Amplification of cDNA Ends, Version 2.0 (Invitrogen, La Jolla, CA) resulted in the generation of a 480-bp product that contained the translation initiation codon. The entire chick Irf6 open reading frame was identified, translated, and aligned with human and mouse IRF6.
In Situ Hybridisation
The murine sense and antisense riboprobes were generated from a 1.6-kb region in the 3′ untranslated region of Irf6 and transcribed with the SP6 and T7 polymerase, respectively. The chick sense and antisense probes were generated from a 1.1-kb region in the coding region of Irf6 and transcribed with the T7 and T3 polymerases, respectively. Whole-mount in situ hybridisation using a digoxigenin-labelled riboprobe was performed as described by Nieto and co-workers (Nieto et al. 1996). Section in situ hybridisation on 5-μm sections of murine and chick embryos was performed as described by Wilkinson and co-workers (Wilkinson et al., 1987) using 33P-UTP-labelled riboprobes.
An anti-IRF6 polyclonal antibody was raised against the peptides EDELEQSQHHVPIQDTFPF (amino acid residues 147–165 of mouse IRF6) and SPEASWPKTEPLEMEV (amino acid residues 187–202). The antibodies were generated and affinity purified by Research Genetics (Huntsville, AL). Paraffin sections were dewaxed, rehydrated through a graded ethanol series, and blocked at room temperature for 6 hr in 10% normal goat serum (NGS) in PBS. The sections were incubated in 1:100 anti-rabbit IRF6 primary antibody in 10% NGS overnight at 4°C. The sections were washed in PBS and incubated in 1:100 goat-anti-rabbit biotinylated secondary antibody for 40 min at room temperature. To detect the secondary antibody, the ABC amplification system (Vector Labs, Burlingame, CA) was used followed by incubation in streptavidin-conjugated Cy3. The sections were stained with DAPI-DABCO and observed under a UV light.
We thank members of the Dixon Laboratory and Kathleen Maltby for technical assistance.