Disintegration of the medial epithelial seam: Is cell death important in palatogenesis?


Author to whom all correspondence should be addressed.
Email: s.iseki.emb@tmd.ac.jp


During palatogenesis, the palatal medial edge epithelium (MEE) forms the medial epithelial seam (MES) on adhesion of the opposing palatal shelves. The MES eventually disappears, leading to mesenchymal confluence of the palate and completion of palatogenesis. Failure of these processes results in cleft palate, one of the most common congenital anomalies in human affecting around one case in 500–2500 live births. The cell fate of MEE has been controversial for more than 20 years. Recent studies suggest that the disappearance of MES is a complex process involving cell death, epithelial-mesenchymal transition (EMT) and epithelial migration. Interestingly, transforming growth factor-β3 (Tgf β3) expression in MEE and the tip epithelium of the nasal septum begins just before palatal shelf reorientation and lasts until MES disruption, and several works including targeted disruption of the gene have indicated that the process appears to be regulated mainly by the TGFβ3-TGFβR signaling. However, how MEE cells choose their fate and how the cell fate is altered in response to cellular environment remains to be elucidated.


The mammalian palate is a structure that separates the oral and nasal cavities, and facilitates breezing and swallowing. It consists of two parts, the frontonasal process-derived primary palate and the maxillary process-originated secondary palate, and its development involves complex dynamic morphogenesis and cellular differentiation (reviewed in Ferguson 1988). The secondary palate begins as a protrusion from the maxillary process within the oral cavity at around embryonic day (E) 12 in mouse and 8 weeks in human pregnancy. The bilateral palatal shelves first grow vertically alongside of tongue (Fig. 1a), next the palatal shelves elevate in the horizontal position above the tongue (Fig. 1b). Consequently, the bilateral shelves establish contact at midline at the tip epithelium named the medial edge epithelium (MEE), forming the medial epithelial seam (MES; Fig. 1c), and disruption of the latter results in mesenchymal continuity (Fig. 1d). During the formation of the secondary palate, the elevated palatal shelves fuse with the primary palate, by E15 and 12 weeks in mice and human, respectively. Disturbance in any of these processes can result in cleft palate, one of the most common human congenital anomalies affecting around one case in 500–2500 live births (Schutte and Murray 1999).

Figure 1.

 Development of the secondary palate (a–d). (a) At embryonic day (E)13.5, in the stage of vertical growth, bilateral palatal shelves are growing along side of the tongue. (b) At E14.0, just after shelf elevation, palatal shelves are present above the tongue in the horizontal direction. The arrow indicates medial edge epithelium (MEE). (c) At E14.5, during fusion process, medial epithelial seam (MES) forms at the site of fusion. (d) At E15.0, formation of the palate has completed resulting in palatal mesenchyme continuity. P, palatal shelf; t, tongue. (e) High magnification view of MEE at E14.0. The epithelium consists of two layers; the outer layer of flat periderm cells and the inner layer of cuboidal basal MEE cells. (f) E14.5 MES. MES is one layer structure and accumulation of epithelial cells form the epithelial triangle (asterisks) at the oral and nasal ends of the MES. (g) Ventral view of the complete palate at E15.0. The dotted line divides the palate between the primary palate and the secondary palate. About anterior two-thirds of the palate is the hard palate with the rugae on the surface and the bone beneath, the rest of the posterior part is soft palate, which is flanked by the solid lines. Scale bar; 200 μm (a–d), 10 μm (e,f), 1 mm (g).

The mammalian palate is highly specified during evolution (reviewed in Ferguson 1988). In birds, the secondary palatal shelves are derived from the maxillary process as in mammals, but they grow horizontally from the beginning and they contact each other but do not fuse, so that birds have a natural cleft palate. In amphibians and some reptiles, the secondary palatal shelves do not appear and the primary palate grows posteriorly instead. The crocodilians is one group of reptiles with an intact and fused mammal-like secondary palate; however, the anterior palatal shelf grows horizontally above the tongue and the very posterior part forms the valve. The lower vertebrates such as fish, lack a secondary palate as a roof of the mouth.

Formation of the secondary palate takes place towards the end of organogenesis period during development, E12–15 in mice and 8–12 weeks in humans. During normal palatogenesis, cell death seems to be mainly involved in MES disintegration, death of some MEE cells contributes to palatal mesenchyme continuity, and cell death is hardly found in the palate mesenchyme except that in close proximity to MES (Martínez-Álvarez et al. 2000; Goudy et al. 2010). In this review, several issues will be discussed starting with what has been known about the fate of MEE cells, including cell death, then the importance of Tgf-β signaling in MES disintegration, followed by the contribution of cell death to palatogenesis. Finally, some of the newly produced genetically modified mice in which pathological cell death could induce cleft palate will be introduced. Although cleft palate in the mutants appears not to be directly due to impaired MES disappearance, changes in mesenchyme properties could influence epithelial cell characters. Since the discussion mainly focuses on secondary palatal formation, the term, palate, is used to represent the secondary palate.

Multiple fates of medial edge epithelial (MEE) cells

There is controversy regarding MEE cell fate, especially how MES disappears during palate formation. Among possible mechanisms for the disappearance, cell death, epithelial-mesenchymal transition (transformation or transdifferentiation) and/or cell migration have been particularly considered (reviewed in Dudas et al. 2007; Nawshad 2008). Until recently, each separate view was in disagreement with the others. Before adhesion, MEE comprises two layers of epithelia, a flat periderm and a cuboidal basal MEE (Fig. 1e). After palatal shelf reorientation, the periderm cells start to change morphologically, they swell and show features of cell death then detach from the surface (Waterman & Meller 1974; Fitchett & Hay 1989). The basal MEE cells become exposed and ready to contact with MEE of the opposing palatal shelf. Contact of opposing MEE commences in the region of the second ruga (middle third of the palate) and extends to both the anterior and posterior ends of the palate. In the anterior part, the majority of periderm cells are shed by the time of contact, although some are trapped between the opposing basal MEE (Fitchett & Hay 1989). Accordingly, the MES is initially a two-layer thick structure but becomes thin to a single layer by intercalation of MES cells from both sides of the shelf (Fig. 1f; Tudela et al. 2002; Cuervo & Covarrubias 2004). The epithelial triangle, which represents accumulation of the epithelial cells of a triangle shape, is formed at the oral and nasal ends of the MES (Fig. 1f). As palatal height is increased dorso-ventrally, the seam breaks up into epithelial islands, and eventually mesenchymal confluence is achieved, leading to completion of MES disappearance (Fig. 1d). In general, when the cell fate of the MEE is discussed, that of basal MEE cells is the focus of the attention.

The concept that cell death plays a major role in MES disintegration was suggested about half a century ago (DeAngelis & Nalbandian 1968; Farbman 1968; Hayward 1969; Shapiro & Sweney 1969 and references therein) based on detailed histological analysis, and was generally accepted for about two decades. The notion of epithelial-mesenchymal transition (EMT) of basal MEE cells was first proposed by Fitchett & Hay (1989). They carried out a study based on the finding that suggested changes in the cellular state between epithelial and mesenchymal cells in Müllerian duct during development (Trelstad et al. 1982). During development, there are events in which epithelial-mesenchymal or mesenchymal-epithelial transition occurs. They carried out a histological study with transmission electron microscopy (TEM) and immunohistochemical staining and concluded that cell death occurs only in periderm cells while basal MEE cells transform to mesenchymal cells followed by expression of vimentin and degradation of the basal lamina (Fitchett & Hay 1989). In vivo labeling of the epithelium with a fluoresceinated lipid soluble marker, carboxydichlorofluorescein diacetate succinimidyl ester (CCFSE) by intrauterine injection in pregnant mice, identified labelled cells in the palate mesenchyme (Griffith & Hay 1992). This fluorescent dye diffuses across the membrane, where intracellular esterases cleave off the acetates to release fluorophore as a water-soluble compound that cannot diffuse out of the labeled cells. The CCSFE product is taken up by phagosome-like structures as isolation bodies. Isolation body-containing cells in the palate mesenchyme were observed long after the completion of mesenchymal confluence. Other in vivo and in vitro studies also identified epithelial cells in the palate mesenchyme stained with another fluorescent dye, DiI (1,1 dioctadecyl 3,3,3′,3′ tetramethyllindocarbocyanin perchlorate; Shuler et al. 1991, 1992; Kaartinen et al. 1997).

Carette & Ferguson (1992) suggested that MEE cells migrate either towards the nasal or oral sides of the palate by maintaining their epithelial features and formation of epithelial triangles; where they eventually integrate into the pseudostratified ciliated columnar cells (nasal) or stratified squamous keratinized epithelium (oral). The same authors also investigated the fate of MEE in vitro by terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) staining and replication-defective helper-free retroviral vector labeling, and suggested that the MEE disintegrate by epithelial-mesenchymal transformation, epithelial migration or cell death according to the MEE differentiated/cell cycle state (Martínez-Álvarez et al. 2000). Around this period, the concept of various types of MEE cell fates began to be accepted.

Cuervo & Covarrubias (2004) suggested that cell death is the major contributor to MES disintegration with very little contribution by EMT if any. They concluded that only the periderm cells migrate and form the epithelial triangle where most of them die. It was also suggested that periderm cell migration is required for the sequential events of MEE cell death and palatal growth in a dorso-ventral direction.

The cre-loxp-based in vivo genetic marking technique was later used to trace the cell lineage in vivo. For labeling oral epithelium, the expression of cre recombinase is driven by cytokeratin14 (K14) promoter, which is a well characterized epithelium-specific regulatory element (Vasioukhin et al. 2001). Keratin14 (K14) is a prototypic marker of dividing basal keratinocytes (Byrne et al. 1994) and its promoter is activated at around E12.0 in the oral epithelium (Dassule et al. 2000). By birth, most keratinocytes in the skin are positive for the targeted recombination event in the K14-cre transgenic mouse. By crossing this transgenic mouse with Rosa26 reporter (R26R) mouse (Soriano 1999) line, Vaziri Sani et al. (2005) traced the cell fate of MEE, and did not find any X-gal-positive cells in the palate mesenchyme. They concluded that epithelial-mesenchymal transition does not occur and all MEE cells die and are subsequently removed. Their conclusion is consistent with the work of Xu et al. (2006) in which X-gal-positive cells were not found in the palate mesenchyme in K14-cre/R26R mice. However, their results were challenged by Jin and Ding (2006), who observed several X-gal-positive cells in the palatal mesenchyme in K14-cre/R26R mouse. Jin and Ding explained that the intensity of the β-galactosidase activity in the palatal epithelium varied among the K14-cre/R26R embryos, from strong to almost no staining. These three studies used different K14-cre transgenic mouse lines, and none of them was proper “knock-in”K14-cre transgenic. The discrepancy among these studies might be caused by different timing of activation of K14 promoter transgene and/or amount of synthesized cre due to different insertion site of each transgene. That also might account for the varied intensity of the β-galactosidase activity in the mesenchyme of palatal shelves. The results should emphasize the need to generate K14-cre mouse line by knock-in system to reproduce real in vivo K14-cre activity.

So far even genetic marking has not provided a firm conclusion on the three proposed fates of MEE cells; migration to the oral and/or nasal sides, mesenchyme transition, or death, and further techniques are required to determine MEE cell fate. The contradictory results among the in vitro studies might be due to different culture conditions used in these studies (Table 1), which could restrict the behavior of MEE cells. Indeed, Gurley et al. (2004) found that physical proximity of opposing MEE during the fusion process in vitro affects the fate of MEE cells in vitro. It is therefore possible that MEE cells select the cell fate according to the environmental condition.

Table 1.   Differentiation of medial edge epithelium (MEE)
ReferencesStudy typeFateMethodsNote
  1. The studies that focused on cell fate of MEE are listed. CCFSE, carboxydichlorofluorescein diacetate succinimidyl ester; DiI, 1,1 dioctadecyl 3,3,3′,3′ tetramethyllindocarbocyanin perchlorate; DMEM, Dulbecco’s modified Eagle’s medium; EMT, epithelial-mesenchymal transition; TUNEL, terminal deoxynucleotidyl transferase-mediated nick and labeling.

DeAngelis & Nalbandian (1968), Farbman (1968), Hayward (1969), Shapiro & Sweney (1969)In vivoCell deathHistology
Electron microscopy
Fitchett & Hay (1989)In vivo
In vitro
EMT, cell death in the peridermElectron microscopy
Vimentin and keratin detection
Griffith & Hay (1992)In vivo
In vitro
EMTFluorescent labelling (CCFSE)Static culture of whole palate including the part of maxilla in agar. BGJb medium containing BSA and insulin transferrin selenium. 50% O2– 5% CO2– 45% N2
Shuler et al. (1991, 1992)In vivo
In vitro
EMTFluorescent labelling (DiI)Keratin and laminin detection. Paired shelves culture on filter. BGJb medium. 95% air – 5% CO2
Carette & Ferguson (1992)In vitroMigrationFluorescent labelling (DiI)Static culture of paired shelves on filter. Eagles MEM medium. No description of gas supply
Kaartinen et al. (1997)In vitroEMTFluorescent labelling (DiI)Static culture of paired shelves on filter. BGJb medium. 95% air – 5% CO2
Martínez-Álvarez et al. (2000)In vitroCell death, EMT, migration was not excludedTUNEL staining
Retroviral vector labelling
Static culture of paired shelves on filter. DMEM/Ham’s F12
Cuervo & Covarrubias (2004)In vitroCell death, very little EMT, migration of periderm cellsFluorescent labelling (CCFSE)
Adenoviral vector infection
Cytochalasin D
TUNEL staining
Floating culture of whole palate (including part of maxilla) on filter. Static culture of sliced whole palate in low-melting point agarose. McCoy 5a modified medium. 95% air – 5% CO2
Takigawa & Shiota (2004)In vitroCell death, EMT, migration, migration and cell death (periderm)Electron microscopy
Rotating suspension culture of maxilla. BGJb medium. 95% O2– 5% CO2
Gurley et al. 2004In vitroCell death, EMT, no migrationFluorescent labelling (CCFSE)
TUNEL staining
Static culture of paired shelves on filter with various spacing between the shelves. DMEM/Ham’s F12 medium. 95% air – 5% CO2
Jin & Ding (2006)In vivo
In vitro
Cell death, EMT, migrationK14-cre; R26R mouseStatic culture of paired shelves on the filter. BGJb medium. 95% O2– 5% CO2
Sani et al. (2005) Xu et al. (2006)In vivoCell deathK14-cre; R26R mouse 

Since palatogenesis takes place in the oral cavity, and thus is not visible externally, organ culture has been a tool to investigate various aspects of palate development. However, there is a need for a more careful examination to find out the best culture condition that can reproduce the in vivo situation. Takigawa & Shiota (2004) reported the condition for single palate shelf culture that is closely associated with palatal epithelium differentiation process in vivo; oral epithelial cells differentiated into mature squamous cells, nasal epithelial cells into ciliated columnar cells and non-ciliated bulgy cells, and MEE becomes distinguishable from both oral and nasal epithelia with multipolar fibroblast-like mesenchymal cells on the surface. With this culture system, cell death, EMT, and epithelial cell migration occurred at various stages of MEE cell disappearance. Interestingly, when the single palatal shelf was cultured with amniotic fluid, terminal differentiation of MEE cells was inhibited (Takigawa & Shiota 2004), suggesting that amniotic fluid contains certain factors that control MEE differentiation. These observations suggest that this culture system reproduces the in vivo condition to some extent. The results indicated that terminal differentiation of MEE cells is not necessarily dependent on palatal shelf contact, which contradicts the work of Griffith & Hay (1992) and Cuervo et al. (2002) that suggested the importance of shelf contact for initiation of MES disintegration.

TGF-β3 appears to play a major role in MES disintegration

Evidence suggests that transforming growth factor-β3 (TGF-β3) plays a major role in MES disintegration including cell death. Tgfb3 expression in the MEE appears around E13.5 before palatal shelf elevation and is maintained during MEE adhesion, then disappears after MES formation (Fitzpatrick et al. 1990; Pelton et al. 1990). Targeted disruption of Tgfb3 gene in mice resulted in impaired lung development and cleft palate; including both complete cleft and partially attached palate (Kaartinen et al. 1995; Proetzel et al. 1995). Cleft palate in these mice is due to impaired adhesion of bilateral palatal shelf MEE and failure of MES disintegration (Kaartinen et al. 1997; Taya et al. 1999). Since the Tgfb3 null mouse does not have any other discernible craniofacial defects, it was concluded that the cleft palate represented the direct effect of loss of TGF-β3 from the MEE.

TGFβs generally transduce their signals through heterotetrameric receptor complexes of two type II (TβR-II or TGF-βR2) and two type I (TβR-I or ALK5) receptors, resulting in phosphorylation of intracellular mediators, receptor-regulated rSmads (SMAD2 and SMAD3). The phosphorylated rSmads form a complex with SMAD4 and translocate into the nucleus where they modulate the transcription of various genes (Heldin et al. 1997; Massagué 1998).

In accordance with the results of targeted deletion of Tgfb3, epithelium-specific deletion of TGF-β type II receptor expression, Tgfbr2 (K14-cre;Tgfbr2fl/fl) and Alk5 (K14-cre;Alk5fl/fl), resulted in incomplete formation of the soft palate, the posterior part of the palate (Dudas et al. 2006; Xu et al. 2006). Both mouse lines showed adhered but persistent MES in the anterior part of the palate, and lack of cell death. In the K14-cre;Tgfbr2fl/fl mouse, the MES cells continued to proliferate and failed to die, which was not rescued by the addition of TGF-β3 to the palate culture medium. These results suggest that cell autonomous TGF-β 3-TGFβR signaling is required for MES disintegration. Deletion of the receptors from MEE is phenotypically milder than Tgfb3 null mouse suggesting TGF-β3 requirement in the palatal mesenchyme for normal development. The cell proliferation activity in palate mesenchyme is low in the Tgfb3 null mouse (Xu et al. 2006), indicating that TGF-β3 synthesized in the MEE plays a role in proliferation of palatal mesenchyme. Interestingly, overexpression of Smad2 in the Tgfb3 null mice partially rescued cleft palate phenotype, with extension of the MES formed area in the anterior-posterior direction and successful disintegration of MES is in the contact area (Cui et al. 2005). The above study, however, did not investigate the mechanism of MES disintegration. The role of TGF-β3 in MES disintegration is supported by studies on β ig-h3, a TGF-β-induced fibrillar extracellular matrix protein with four highly conserved drosophila fasciclin-1-like domains (fas-1) and a C-terminal arginine-glycine-aspartic acid (RGD) domain (Kim et al. 2003). β ig-h3 is cleaved after secretion, and the soluble C-terminal domain enters the cell and induces apoptosis probably by activating caspase-3. Tgfβ3 expression in MEE is followed by that of βig-h3 in vivo, and TGF-β3 induces the expression of β ig-h3 in HaCaT cells in vitro (Choi et al. 2009). Treatment of embryos in utero and statically cultured palate in vitro with β ig-h3 antisense oligonucleotides resulted in cleft palate and impaired fusion, respectively (Choi et al. 2009). These results suggest that β ig-h3 expression is induced by TGF-β3, and it is crucial for cell apoptosis and the disappearance of the MEE during palate fusion.

In vitro primary cell cultures of MES indicated that the immediate effect of TGF-β3 is cell cycle arrest, which is required for MES disintegration (Ahmed et al. 2007). Subsequently, TGF-β3 activates cell movement first by downregulation of E-cadherin and upregulation of vimentin and fibronectin. This EMT seems to be regulated by the balance between follistatin, a TGF-β antagonist, and TGF-β3 in MEE (Nogai et al. 2008). Finally, cell death occurs by inducing DNA fragmentation as well as the expression of caspase-3 and caspase-9, which are effector caspases that are activated at the end of the apoptosis process (see below). These findings suggest that EMT and cell death are independently regulated, which is supported by the findings that inhibition of apoptosis induced only partial MES degradation (Ahmed et al. 2007), and that inhibition of apoptosis by a caspase inhibitor did not particularly increase the number of MEE cells to initiate mesenchymal transition (Cuervo & Covarrubias 2004). Thus, the primary MES cell culture system is a useful tool for further analysis of MES differentiation. The importance of cell cycle arrest in the initiation of MES disintegration is demonstrated by studies in which inhibition of cell cycle arrest of the MEE cells with epidermal growth factor (EGF) resulted in maintained MEE (Yamamoto et al. 2003) and K14-cre;Tgfbr2fl/fl mouse persisted MEE cell proliferation (Xu et al. 2006). Exposure to exogenous retioic acid (RA) or inhibition of RA signaling affects cell death of MEE with altered expression pattern of Tgfb3 (Cuervo et al. 2002), which suggests involvement of RA in MES disintegration. However, the precise relationship between RA and Tgf-β3 signaling remains to be elucidated.

Contribution of cell death in MES degradation

Apoptotic cell death is evident by chromatin condensation, loss of mitochondrial membrane potential, plasma-membrane asymmetry, and detachment from the cellular matrix. It is generally accepted that activation of a series of caspases, a family of cycteine proteases that regulate several steps of the apoptotic cascade, is the hallmark of the apoptotic process. Among these proteins, caspase-3 is the major effector caspase that cleaves protein substrates within the apoptotic cell (Chipuk & Green 2005).

Ultrastructural analyses showed the presence of autophagic vacuoles, lysozomes and macrophages in the area of MEE (Shapiro & Sweney 1969; Smiley, 1970; Smiley and Koch, 1975). Other studies using TUNEL staining also reported DNA fragmentation and active caspase-3 localization in MEE (Cuervo & Covarrubias 2004; Takigawa & Shiota 2004; Vaziri Sani et al. 2005, Jin and Ding 2006). These results suggest that MEE cell death during normal palatogenesis is mainly due to apoptosis rather than other types of cell death.

Cuervo & Covarrubias (2004) reported that the application of z-VAD, a pan-caspase inhibitor, to cultured palatal shelves resulted in maintenance of MES and lack of disintegration. This suggests that apoptotic cell death is indispensable for MES degeneration. In contrast, Takahara et al. (2004) showed that in the presence of other caspase inhibitory factors, YVAD-CHO and DEVD-CHO, MES disappeared as in vivo while TUNEL staining showed successful inhibition of apoptosis. However, the authors did not investigate the mechanism of MES disintegration in the absence of apoptosis. Although these studies used different caspase inhibitors, their results are contradictory. Genetic approaches to address the same question also provided perplexing results. Mouse Apaf-1 gene is an orthologous gene of Caenorhabditis elegans CED-4 gene that encodes a protein to activate apoptotic downstream caspases, caspase-3 and caspase-9. The first study of targeted disruption of the Apaf-1 gene in mice reported inhibition of cell death in numerous tissues including brain and digits (Cecconi et al. 1998). The same study reported the formation and persistence of MES in the Apaf-1 null fetus (Cecconi et al. 1998). Interestingly, Jin and Ding (2006) demonstrated the complete disintegration of the MES in the homozygote, together with enlargement of the epithelial triangle and lack of apoptosis. The authors suggested that the finding was due to the use of a different line of Apaf-1 deleted mouse (Honarpour et al. 2000) compared with the one used in the study of Cecconi et al. (1998). In addition, MES retention was observed at E14.5 in the study of Cecconi et al. (1998), which is a stage that is too early to confirm MES disappearance. Deletion of a downstream caspase gene, caspase-3 or caspase-9, does not seem to result in cleft palate, since some of the null mice survive after birth (Kuida et al. 1996, 1998; Woo et al. 1998). The results of these studies suggest that apoptotic cell death is not indispensable for MES degeneration.

Recent progress in the understanding of cell death allows us to investigate in more detail how cells actually die during development. It is now clear that failure of induction of apoptosis by caspases leads to the activation of the caspase-independent cell death (CICD; reviewed in Tait and Green 2008). CICD is classified into three categories, apoptosis-like cell death, autophagic cell death, and necrosis. Inhibition of caspases induces one of these alternative cell death pathways (reviewed in Kroemer & Martin 2005; Vandenabeele et al. 2006). Interdigital cell death is a good model of programmed cell death during development. This cell death is normally apoptotic in nature, and genetic absence of caspase activity in Apaf-1 mutant mice still induces necrotic interdigital cell death leading to successful removal of the interdigital web (Yoshida et al. 1998; Chautan et al. 1999). Therefore, lack of activation of the caspase pathway does not mean inhibition of cell death. During MES disintegration, activated caspase-3 is found in the epithelial triangle as well as MES with the presence of macrophages in close proximity to MES (Fig. 2a, b). One of the CICD pathway molecules, endonuclease G (EndoG; Bahi et al. 2006), is expressed in the oral epithelium including the MES, which could confirm the availability of alternative cell death pathway in the MES (Fig. 2c). Based on the results of previous studies on cell fate of MEE, there is a need to reconsider the roles of cell death in palatogenesis.

Figure 2.

 Localization of cell death associated molecules during the disintegration of medial epithelial seam (MES). Immunohistochemical detection of activated caspase-3 (Promega) (a), F4/80 antigen (a marker of macrophages) (Serotec) (b), and EndoG (Abcam) (c) during the period of MES (arrows) disintegration at E14.5. Some of the positively stained cells are indicated by arrowheads. Activated caspase-3 positive cells are present in the MES and epithelial triangles. Macrophages are located in the mesenchyme in the vicinity of disappearing MES and palate mesenchyme. EndoG is expressed in the MES cells and the oral and nasal epithelia. Please note that the epithelium of the nasal septum is not stained. ns, nasal septum. Scale bar, 100 μm.

First, with regard to the aforementioned three types of cell fate of MEE, we do not know the proportion of MEE destined for each type of cell fate during normal development. Second, we do not know how such proportions are altered upon culture of the palate in vitro. Third, it is not clear how the above proportions are altered in the lack of activation of the apoptotic pathway, both in vivo and in vitro. Fourth, it is not clear whether the lack of apoptosis of MEE induces CICD or whether other cell fates compensate the lack of apoptosis.

Based on the results of studies using the in vitro culture system (Table 1), the fate of MEE cells could be modified experimentally. For instance, the different results of Takigawa & Shiota (2004) and Cuervo & Covarrubias (2004) might be due to methodological differences, i.e., culture conditions used for palatal shelves (see Table 1). It is possible that treatment of suspension palate culture with caspase inhibitors induces cell migration and/or mesenchymal transition in the population of MEE that would have otherwise undergone apoptotic cell death, or CICD instead of apoptosis. On the other hand, the addition of caspase inhibitors to the medium of static palate cultures may not induce any mechanism that compensates for the loss of apoptosis and hence apoptosis plays a major role in this culture system. It would be also interesting to study how the fate of MEE cells is altered by targeted deletion of apoptosis-related genes. Combined with the fact that MEE cells have three possible cell fates, MEE disintegration is a much more flexible event than we think in terms of cell fate and type of cell death; representing a backup system to complete palate formation.

Pathological cell death in palate development

Recent genetic modification in mice has revealed that many genes are involved in secondary palate formation (Gritli-Linde 2007). Palatal shelf is a protrusion from the maxillary process, and its mesenchyme is derived from neural crest cells (Jiang et al. 2002, Yoshida et al. 2008). Therefore, not only defects in the epithelium but also those in neural crest formation, migration and proliferation could result in impaired palate development. Although these studies do not show the direct effects of the genes on MES disappearance, altered mesenchyme character could affect epithelial cell behavior.

Hand2 is one of the basic helix-loop-helix transcription factors and is expressed in developing palate epithelium and mesenchyme. Downregulation of the Hand2 gene in the epithelium induces periderm cell apoptosis, which decreases mesenchymal cell proliferation and pathological adhesion of the palate epithelium to the mandible in the anterior part and posterior part, respectively. This adhesion induces apoptosis of the fused basal epithelium and prevents palate shelves to elevate (Xiong et al. 2009).

The T-Box family is defined as a group of transcription factors that share a common DNA binding domain known as the T-box (reviewed in Naiche et al. 2005). Among them Tbx1, 2, 3, and 22 have been known to be associated with craniofacial morphogenesis, mutations in TBX1 and 22 are involved in human congenital malformations, DiGeorge and X-linked cleft palate with ankyloglossia, respectively. Mouse studies show that both Tbx2 and 3 are expressed in palatal mesenchyme during the period of palate development, outgrowth, elevation and fusion (Zirzow et al. 2009). Targeted deletion of Tbx2 in Naval Medical Research Institute (NMRI) background mouse line exhibited complete cleft palate in 86% of embryos. Detailed analysis showed that outgrowth of the palatal shelves is impaired in the Tbx2 null mouse. In the mutant, increased cell death by TUNEL assay and active cell proliferation by BrdU incorporation were found in the anterior part of the shelf during the period of outgrowth with upregulation of Bmp4 and cyclin D1 expression (Zirzow et al. 2009), which could cause imbalance between cell proliferation and cell death.

Targeted deletion of another T-box family gene, Tbx1, also results in cleft palate (Goudy et al. 2010). Tbx1 is strongly expressed in facial epithelium including tongue and palate (Zoupa et al. 2006), and tongue muscle primordium and palate mesenchyme. The null mutant indicates that the palate length is shorter than wild type and fails to elevate. Cell proliferation and cell death analysis showed increased cell proliferation and cell death, which is partly the cause of cleft palate (Goudy et al. 2010).

Interestingly, static paired palatal shelves culture of the above three mutant mice demonstrated MES formation and disintegration although the cell fate was not investigated.

The entire mesenchyme except the vascular system in the palate is derived from cranial neural crest cells (Ito et al. 2003; Yoshida et al. 2008). It is therefore envisaged that quality and quantity of the neural crest cells in the region would affect the palate formation. For instance, conditional overexpression of Sprouty1 (Spry1) in cranial neural crest cells induced TUNEL staining-positive cell death and reduced cell proliferation in the neural crest-derived mesenchyme of branchial arch. Spry1 is highly expressed in the cranial neural folds at around E8.0, then neural crest-derived mesenchyme of craniofacial region until about E10.0 (Yang et al. 2010). Treacher Collin’s syndrome (TCS) is a congenital anomaly that is characterized by hypoplastic facial bones, particularly the zygomatic complex and mandible, cleft palate, and middle and external ear defects. Haploinsufficiency of Tcof1, which has been found to be a responsible gene for TCS, in mice resulted in cell death of the neuroepithelium and impaired cell proliferation of the cranial neural crest cells (Dixon et al. 2006). In both Spry1 and TCS cases, the authors attributed an insufficient amount of the neural crest cell-derived mesenchyme in the craniofacial region to an underlying cause of cleft palate. They did not investigate whether the cleft occurred simply because of impaired growth of the palatal shelves due to a decreased amount of neural crest-derived mesenchyme or failed tissue interaction between the MEE and the neural crest cell-derived mesenchyme. Although there is no direct evidence to support the notion that MEE of these mutants is impaired, as discussed above, palatal mesenchyme could influence MEE cell behavior, and therefore further studies are required to elucidate the effect of loss of these genes in neural crest derived palate mesenchymes.

All of the accumulated data from murine gene deleted models, conditional gene deleted models, and in vitro palate culture systems do not exclude any of the three possible cell fates of the MEE: apoptosis, epithelial-mesenchymal transition and cell migration, and the cell fate determination appears to depend on environmental conditions. Further studies would be required to understand the mechanism of cell fate determination including epithelial-mesenchymal interaction, and how and to what extent the cell fates can be altered in response to changes of the environmental conditions.


I thank Dr Shigeru Okuhara and Mr Norisuke Yokoyama for providing images. This work is supported by Grants-in-Aid from the Japanese Society for the Promotion of Science (#20390510) and by NIG Cooperative Research Program 2009-B7, 2010-B7.