SEARCH

SEARCH BY CITATION

Keywords:

  • Cyp26b1;
  • retinoic acid;
  • palatogenesis;
  • Cyp26a1;
  • Fgf10;
  • Tbx1;
  • cleft palate;
  • tongue

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Background: In previous studies, we investigated the effects of excess retinoic acid (RA) during palatogenesis by RA administration to pregnant mice. In the present study, we deleted Cyp26b1, one of the RA-degrading enzymes, to further study the effects of excess RA in the normal developing palate and to understand how endogenous levels of RA are regulated. Results: Excess RA, due to the absence of Cyp26b1, targets cells in the bend region of the palatal shelves and inhibits their horizontal elevation, leading to cleft palate. An organ culture of Cyp26b1−/− palatal shelves after tongue removal did not rescue the impaired elevation of the palatal shelves. The expression of Fgf10, Bmp2, and Tbx1, important molecules in palatal development, was down-regulated. Cell proliferation was decreased in the bend region of palatal shelves. Tongue muscles were hypoplastic and/or missing in Cyp26b1−/− mice. Conclusions: We demonstrated that CYP26B1 is essential during palatogenesis. Excess RA due to the lack of Cyp26b1 suppresses the expression of key regulators of palate development in the bend region, resulting in a failure in the horizontal elevation of the palatal shelves. The regulation of RA signaling through CYP26B1 is also necessary for the development of tongue musculature and for tongue depression. Developmental Dynamics 241:1744–1756, 2012. © 2012 Wiley Periodicals, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Palate development (palatogenesis) in mammalian embryos involves a sequence of critical events, such as growth, elevation, medial elongation, and the fusion of palatal shelves. A failure of any of these steps can result in cleft palate, one of the most common birth defects in humans (Ferguson, 1988; Gritli-Linde, 2007). Analyses of patients with cleft palate have revealed that nonsyndromic cleft palate is etiologically heterogeneous and involves multiple genetic and environmental factors (Dixon et al., 2011).

Experimentally numerous genes have been shown to be expressed in the developing palate and the disruption of some of these genes can result in cleft palate in rodents (Hilliard et al., 2005; Okano et al., 2006; Gritli-Linde, 2007). However, no single gene has been proven to be responsible for the majority of human cleft palate cases, suggesting that human cleft palate is genetically quite heterogeneous. On the other hand, some environmental agents have been associated with oral clefts in humans such as retinoic acid (RA) and anticonvulsant drugs (Lammer et al., 1985; Jentink et al., 2010).

RA is the biologically most active compound among retinoids and plays important roles in various biological phenomena (Ross et al., 2000). Clinically, RA is widely used for the treatment of cancers and other diseases (Mi, 2011). In addition, vitamin A and its derivatives (retinoids) are essential for embryonic development (Rhinn and Dolle, 2012). On the other hand, retinoids are potent teratogens and excess RA disrupts normal development and induces developmental abnormalities both in rodents and humans (Ross et al., 2000). Pregnant women exposed to 13-cis-RA in utero have an unusually high risk of craniofacial, cardiac, and central nervous system malformations in their fetuses (Lammer et al., 1985). The developing palate appears to be a major target of teratogenic doses of RA and cleft palate is among the major malformations induced by RA (Ackermans et al., 2011). Although the teratogenicity of RA is widely accepted, its pathogenesis and the underlying molecular events are incompletely understood, partly because it is difficult to increase RA levels selectively in target tissues under experimental conditions.

RA is synthesized by retinaldehyde dehydrogenases (RALDHs) (Niederreither et al., 1999) and degraded by the cytochrome P450 family 26s (CYP26s) (Abu-Abed et al., 2001; Sakai et al., 2001, 2004). Therefore, endogenous RA levels are determined by the balance between the activity of these two groups of enzymes. RA binds to RA receptors (RARs) or retinoid X receptors (RXRs) and enters the nucleus to activate or suppress target genes (Rhinn and Dolle, 2012). Targeted disruption of these receptors in mice produces developmental abnormalities that recapitulate the defects caused by RA deficiency. Cleft palate occurs in RARα/RARγ double knockout mice indicating an endogenous role of RA in palatal shelf development (Lohnes et al., 1994; Ross et al., 2000).

To investigate the mechanism underlying RA-induced cleft palate, we aimed to elucidate the roles of Cyp26b1, a RA-degrading enzyme, during palate development. First, we analyzed the expression pattern of Cyp26s in the developing palate and alterations in their expression after RA treatment. Cyp26b1 was found to be expressed at critical stages of early palatogenesis and showed a spatially and temporally specific expression pattern. RA treatment up-regulated Cyp26b1 expression in the fetal palate. Next, we analyzed the RA levels and the expression of possible down-stream genes of RA signaling in the developing palate. This study has provided new findings concerning the regional heterogeneity in the developing palate which is closely related to RA-induced cleft palate as well as the molecular networks that play critical roles in normal and abnormal palatogenesis.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Spatial and Temporal Expression Pattern of Cyp26b1 and Cyp26a1 in the Developing Palate

We investigated the expression pattern of the members of the Cyp26s family, i.e., Cyp26a1, -b1, and -c1 in the developing palate by in situ hybridization on sections.

During normal development, Cyp26b1 begins to be expressed robustly in the mesenchyme of the bend region of palatal primordia by E11.75, as well as in the medial region of the tongue (Fig. 1A, C). By E12.5 and E13.5, its expression has shifted ventrally and it could be detected close to the oral epithelium (Fig. 1E, G). The expression of Cyp26a1 at E11.5 was weakly restricted on the lateral sides of palatal primordia but it was strongly expressed at the later stages in the ventrolateral epithelium of the developing palate and in the tongue (Fig. 1I, K, M). Cyp26c1 was not present in the developing palate.

thumbnail image

Figure 1. Cyp26a1 and Cyp26b1 respond rapidly to RA-treatment during palatogenesis. A, C, E, G: In situ hybridization reveals that Cyp26b1 begins to be expressed in the bend region of palatal primordia (black arrowheads) and shifts ventrally prior to palatal shelf elevation. B, D, F, H: RA administration results in marked up-regulation of Cyp26b1 expression in the developing palate (double arrows). Note the persistent Cyp26b1 induction in the bend region 24 hr after RA treatment (white arrowheads), while it is not present in the bend region of the control embryo (E). I, K, M: Cyp26a1 is expressed in the ventrolateral epithelium of the developing palate (arrows). J, L, N: Cyp26a1 expression is significantly induced in palatal shelves by RA-treatment at E11.5 (white arrows), though its expression is similar between RA-treated and control fetuses by E13.5. T, tongue. Scale bar = 100 μm.

Download figure to PowerPoint

Since the developing palate is susceptible to excess RA (Abbott et al., 1989; Abbott and Pratt, 1991), we asked whether the expression of Cyp26a1 and/or Cyp26b1 responds to an increase in RA levels by gavage administration of RA to WT mice at gestation day 11.5, a stage at which RA is known to cause cleft palate (Okano et al., 2007). Exogenous RA caused a dramatic up-regulation of Cyp26b1 in the mesenchyme of palatal and tongue primordia 3 hr after RA treatment (Fig. 1B). The up-regulation of Cyp26b1 persisted in the bend region of palatal shelves at E12.5 (Fig. 1D, F). The single dose of RA did not maintain Cyp26b1 expression in the bend region of the E13.5 developing palate, and the morphology of RA-treated palatal shelves was different from those of control palatal shelves (Fig. 1G, H). Excess RA also led to up-regulation of Cyp26a1 in the epithelium and ectopic expression in the mesenchyme of palatal primordia 3 and 6 hr after RA treatment but its expression by E13.5 was comparable to that in the normal developing palate (Fig. 1J, L, N).

In conclusion, both Cyp26a1 and Cyp26b1 are expressed in the developing wild-type (WT) palate and their expression is rapidly increased following RA administration.

Cyp26b1−/− Mouse Fetuses Have Cleft Palate

Since the response to excess RA is more dramatic for Cyp26b1 than for Cyp26a1 and craniofacial anomalies have not been reported in Cyp26a1−/− fetuses (Abu-Abed et al., 2001; Sakai et al., 2001), we analyzed the developing palate in Cyp26b1−/− mice. All Cyp26b1−/− fetuses examined had a complete cleft of the secondary palate (30 of 30 Cyp26b1−/− fetuses at E18.5). Cyp26b1−/− mice at E18.5 exhibit a wide palate defect and die after birth, as previously reported (Yashiro et al., 2004; Maclean et al., 2009) (Fig. 2A). Based on the heterogeneity of the development of different regions of the palatal shelves, we divided palatal shelves into the anterior, middle, and posterior parts (Hilliard et al., 2005; Okano et al., 2006) and observed the morphology. Normally, palatal shelves grow vertically beside the tongue, elevate horizontally around E14.5, and fuse completely by E15.5 to separate the nasal and oral cavities. However, in Cyp26b1−/− fetuses, the middle and posterior palatal shelves remained unelevated, though the anterior palatal shelves elevated as in WT embryos (Fig. 2B). Cyp26a1 expression was up-regulated in the mesenchyme of Cyp26b1−/− palatal shelves (Fig. 2C). Histological analysis revealed that in WT mice the palatal shelves elevate horizontally and make contact at the midline by E14.5 from the anterior to the posterior palate (Fig. 3A, C, E). By contrast, Cyp26b1−/− palatal shelves remained vertically positioned at E14.5 (Fig. 3B, D, and F). Of note, the height of the tongue in Cyp26b1−/− mice was significantly higher in the middle and the posterior regions than in WT mice, suggesting that tongue withdrawal, necessary for palatal shelf elevation, might have been disturbed (Ferguson, 1988) (Fig. 3G). In addition, H&E staining at E15.5 showed tongue muscle anomalies as well as shortened and unelevated palatal shelves of the middle and posterior parts (Fig. 4). The origin of the genioglossus muscle, which is normally located at the symphysis of the mandible, was not detectable in Cyp26b1−/− mice (Fig. 4D), and the muscle body appeared hypoplastic compared with those of WT mice (Fig. 4C–F). We could not identify the bodies of the geniohyoid and mylohyoid muscles, and the anterior belly of the digastric muscle was lacking in Cyp26b1−/− mice (Fig. 4F). These results indicate an aberrant differentiation of the tongue muscles, consistent with previous findings on RA-treated fetuses (Okano et al., 2008).

thumbnail image

Figure 2. The absence of Cyp26b1 leads to cleft palate. A: E18.5 palatal views of wild-type (WT) and Cyp26b1−/− mice. *The defect. B: Six anterior palatal shelves, one middle palatal shelf, and one posterior palatal shelf were elevated out of eight E15.5 Cyp26b1−/− (KO) mice. **Statistically significant difference compared with WT by Chi-square test (P < 0.01). C: Cyp26a1 up-regulation in Cyp26b1−/− palatal shelves (arrows). T, tongue; M, molar. Scale bar = 100 μm.

Download figure to PowerPoint

thumbnail image

Figure 3. Palatal and tongue development in E14.5 Cyp26b1−/− mice. A–F: Histological analysis of the anterior, middle, and posterior region of palatal shelves revealed a failure of horizontal elevation of the palatal shelves in Cyp26b1−/− mice, while, in WT mice, palatal shelves elevate in all regions. Scale bar = 100 μm. G: The measurement of tongue height at E14.5 showed a statistically significant difference in the middle and the posterior parts of Cyp26b1−/− mutants compared to WT embryos. *P < 0.05.

Download figure to PowerPoint

thumbnail image

Figure 4. The lack of Cyp26b1 results in the aberrant development of palatal shelves as well as of tongue musculature at E15.5. A–F: WT palatal shelves of all regions fuse with each other to separate the oral and nasal cavities, while Cyp26b1−/− palatal shelves orient vertically except in the anterior part. The body of genioglossus muscle (gg; an arrowhead in D) is not obvious in the middle part. The bodies of geniohyoid (gh), mylohyoid (mh), and anterior digastric muscle (adg) (arrowheads in F) appear hypoplastic in the posterior part of Cyp26b1−/− mice. m, Meckel's cartilage; sg, submandibular gland. Scale bar = 100 μm.

Download figure to PowerPoint

It is important to note that both intrinsic factors (e.g., development of palatal shelves) and extrinsic factors (e.g., mouth opening and tongue withdrawal) are essential for palatal elevation (Ferguson, 1988). To understand whether the intrinsic factors are impaired in Cyp26b1−/− fetuses, we performed fetal palate culture in suspension (Shiota et al., 1990) (Fig. 5A–D). After 20 hr of cultivation, WT palatal shelves recapitulated the normal horizontal elevation in vitro (3/3) (Fig. 5C); the Cyp26b1−/− middle palatal shelves, however, remained at the vertical position while the anterior palatal shelves elevated (2/3) (Fig. 5D, note PS and PS*). This result suggests that the intrinsic factors are not functional in the middle and the posterior parts of palatal shelves of Cyp26b1−/− mice.

thumbnail image

Figure 5. Tongue removal does not rescue the impaired palatal elevation of Cyp26b1−/− mice. A, B: E13.5 WT and Cyp26b1−/− heads before cultivation show palatal shelves at the vertical position. C, D: After 20-hr organ culture, WT palatal shelves have elevated horizontally (PS). In Cyp26b1−/− mice, the anterior palatal shelves have also elevated horizontally (PS). However, the middle and posterior palatal shelves remained unelevated (PS*). E, eye; NS, nasal septum, PS, palatal shelves; PS*, unelevated palatal shelves after culture.

Download figure to PowerPoint

Altogether, the lack of Cyp26b1 caused both intrinsic and extrinsic defects and resulted in impaired elevation of palatal shelves.

The Absence of Cyp26b1 Affects the Bend Region of Middle Palatal Shelves

RA distribution can be identified by the use of RA response element (RARE)-hsplacZ reporter mice, in which lacZ gene expression reflects the transactivation activity of RA (Rossant et al., 1991). Utilizing this transgenic mouse line, we determined the endogenous RA level in the developing palate. In normal palatal primordia, no lacZ-positive cells were found, which suggests RA does not contribute to the formation of palatal primordia (Fig. 6A). At E12.5, lacZ staining was observed on the lateral side of the nasal septum and in the anterior part of the maxillary primordia (Fig. 6C). LacZ-positive cells were also found in the developing tongue musculature and external ocular muscles, whereas there was no RA activity in the developing palate (Fig. 6 E, G, I).

thumbnail image

Figure 6. RARE-hsplacZ reporter mice revealed significant alteration of RA localization in the absence of Cyp26b1 in developing palate and tongue. A, C, E, G, I: RA signaling activity is not observed in sections of WT palatal primordia but is detected in the maxillary primordia, tongue, extraocular musculature, and retina. B: The absence of Cyp26b1 results in increased RA levels in the bend region of palatal primordia (arrowheads) and tongue (arrows). D, F, H: Increased RA activity is detected especially in the bend region of middle palatal shelves (arrowheads) and the tongue muscles (*). J: The strong RA activity persists in the middle of E13.5 palatal shelves (arrowheads). Scale bars = 100 μm.

Download figure to PowerPoint

The deletion of Cyp26b1 caused a dramatic alteration of RA signaling activity. With the emergence of palatal primordia, strong lacZ expression was evident in the bend region of palatal primordia as well as in the mesenchyme of the tongue (Fig. 6B). This ectopic RA activity persisted in the bend region of middle palatal shelves of Cyp26b1−/− fetuses until E13.5 (Fig. 6D, F, H, and J). To investigate the effects of excess RA in the middle palate, we analyzed this region for programmed cell death by immunostaining using an anti-caspase 3 antibody and for cell proliferation by BrdU incorporation and immunodetection, because excess RA is known to perturb palate development and differentiation (Okano et al., 2007). There was no significant difference in the distribution of caspase 3–positive cells between WT (mean=5.10±1.56 (SD)%) and Cyp26b1−/− palatal shelves (mean=4.25±2.05 (SD)%) (Fig. 7A, B). On the other hand, cell proliferation was decreased in the bend region of Cyp26b1−/− palatal shelves in comparison to WT palatal shelves (Fig. 7C, D). We assayed the percentage of BrdU-positive cells in the mesenchyme of the bend region as previously reported (Rice et al., 2004) and found it was decreased in Cyp26b1−/− fetuses (mean=12.0±2.94 (SD)%) in comparison to WT fetuses (25.7±3.60 %) (Fig. 7E). To evaluate the intrinsic force in the developing palate, we followed the method of Lan et al. (2004), who calculated the ratio of the percentage of BrdU-positive cells of the medial halves to that of lateral halves of palatal shelves. This ratio represents the intrinsic force of the palatal shelf to elevate horizontally and is over 1.0 in normal fetuses, because the medial halves of the palatal shelves grow faster than the lateral halves prior to shelf elevation (Ferguson, 1988). In Cyp26b1−/− fetuses, this ratio was 0.825±0.0581, which was significantly lower than the corresponding value in WT fetuses (1.50±0.361) (Fig. 7F).

thumbnail image

Figure 7. Unaltered apoptotic cell distribution but decreased proliferative cell distribution in the bend region and the altered ratio of proliferative cells in the medial half to those in the lateral half of palatal shelves in Cyp26b1−/− mice. A, B: Immunohistochemical staining with an anti-caspase 3 antibody shows that there is no difference in the magnitude and distribution of programmed cell death between WT and Cyp26b1−/− palatal shelves. Asterisks show autofluorescence in blood cells. C, D: The percentage of BrdU-positive cells is decreased in the bend region (white boxed areas) and the ratio of the percentage of BrdU-positive cells in the medial half compared to the percentage in the lateral half of palatal shelves (surrounded area by white solid lines, dot lines and the tip of palatal shelves) was compared between WT and Cyp26b1−/− fetuses. Scale bar = 100 μm. E: The percentage of BrdU-positive cells in a fixed area of the bend region of palatal shelves is decreased in Cyp26b1−/− fetuses (**P < 0.01). F: The ratio of the percentage of BrdU-positive cells in the medial half to that in the lateral half of the palatal shelves is significantly decreased in Cyp26b1−/− fetuses (**P < 0.01). Values represent as mean±SD.

Download figure to PowerPoint

Thus, deletion of Cyp26b1 increases endogenous RA levels in the bend region of the palatal shelves, and results in significantly weaker intrinsic forces that are required for the palatal shelves to elevate horizontally.

The Increase of Endogenous RA Levels in the Bend Region of Middle Palatal Shelves Correlates With Down-Regulation of Fgf10

Based on the fact that palatal shelf elevation proceeds in an anterior-to-posterior direction (Ferguson, 1978), the results described above raised the possibility that the elevation of palatal shelves is impaired specifically in the middle region in Cyp26b1−/− mice. In other words, an unelevated middle palate is always followed by a failure of posterior palate elevation. Therefore, one aspect of the pathogenesis of cleft palate in Cyp26b1−/− fetuses may be the impaired development of the middle palate. To test this hypothesis, we focused on the middle palatal shelves to determine if molecules required for the elevation of palatal shelves are affected in the absence of Cyp26b1.

In Fgf10−/− fetuses, palatal shelves fail to elevate and cell proliferation is decreased in the bend region as well (Rice et al., 2004). Given this finding, we assessed whether the expression of Fgf10 was affected in the palate of Cyp26b1−/− fetuses. Coinciding with the emergence of palatal primordia, Fgf10 expression was observed in the mesenchyme (Fig. 8A). By contrast, in Cyp26b1−/− fetuses, no expression of Fgf10 was observed in the palatal mesenchyme, and Fgf10 expression was also reduced in the tongue and the mandibular primordia (Fig. 8B, D). As the palatal shelves grow downward, in WT mice, we found Fgf10 to be expressed in the bend region as well as in the mesenchyme of middle palatal shelves as previously reported (Rice et al., 2004; Alappat et al., 2005) (Fig, 8E, G). In Cyp26b1−/− fetuses, Fgf10 was expressed weakly in the ventrolateral mesenchyme but was absent in the bend region of the palatal shelves (Fig. 8F, H), although its expression was similar in the submandibular gland of WT and Cyp26b1−/− fetuses (compare Fig. 8E with F, and 8G with H). We collected anterior, middle, and posterior palatal shelves tissues to evaluate Fgf10 expression quantitatively (Fig, 8I). Real-time PCR showed that Fgf10 expression was comparable in the anterior palatal shelves of WT and Cyp26b1−/− mice, but was decreased in the middle and the posterior parts of palatal shelves in Cyp26b1−/− mice (Fig. 8I). This result correlates with the horizontal elevation of the anterior palatal shelves but impaired elevation of the middle and posterior palatal shelves in Cyp26b1−/− mice.

thumbnail image

Figure 8. Fgf10 expression is absent in the bend region of middle palatal shelves in Cyp26b1−/− mice. A–D: Fgf10 begins to be weakly expressed in the mesenchyme of palatal primordia and tongue in WT fetuses, whereas its signal is absent in palatal primordia of Cyp26b1−/− fetuses. Fgf10 is decreased in the Cyp26b1−/− mandibular primordia at E11.75. E–H: Fgf10 is expressed in the bend region (arrows) and in the ventral mesenchyme of palatal shelves, tongue, submandibular gland, and mandibular primordia from E12.5 to E13.5 in WT. However, Fgf10 expression is undetectable in the bend region of Cyp26b1−/− palatal shelves (arrowheads). Fgf10 expression in the submandibular gland primordia (sgp) and submandibular gland (sg) is used as internal controls between WT and Cyp26b1−/− mice. Scale bar = 100 μm. I: E13.5 palatal shelves were divided into anterior (AP), middle (MP), and posterior (PP) parts. Real-time PCR for each region showed decreased Fgf10 expression in middle (MP) and posterior (PP) palatal shelves but not in anterior palatal shelves (AP). Values (indicated as red) are means + SD, relative to WT palatal shelves in each part (set as 1). *P < 0.05.

Download figure to PowerPoint

Increased RA Perturbs the Expression of Tbx1 and Bmp2 But Fgf10 Is Not Downstream of Tbx1 in the Bend Region of Developing Palate

In a previous study, we showed that Tbx1 expression is suppressed in the tongue muscle primordia of Cyp26b1−/− fetuses, which affects tongue muscle differentiation (Okano et al., 2008). Therefore, we asked whether Tbx1 expression is also perturbed in the palatal shelves in Cyp26b1−/− mice. Tbx1 was expressed in the entire epithelium of the normal palatal primordia and in both epithelium and mesenchyme of the tongue. However, in Cyp26b1−/− fetuses, Tbx1 was decreased in both the palatal primordia and tongue (Fig. 9A, B and insets shown at higher magnification), while its expression was similar in the developing forebrain of WT and Cyp26b1−/− fetuses (arrows in Fig. 9A, B). At E13.5, Tbx1 expression appeared to be decreased in the bend region of the palatal shelves in Cyp26b1−/− fetuses but it was difficult to quantify its expression with this technique (Fig. 9C, D). Real-time PCR confirmed the down-regulation of Tbx1 expression in the middle palatal shelves in Cyp26b1−/− fetuses (Fig. 9E).

thumbnail image

Figure 9. Altered Tbx1 expression in the bend region of palatal shelves in Cyp26b1−/− mice and the maintenance of Fgf10 expression in the bend region of middle palatal shelves of Tbx1−/− mice. A–D: Tbx1 is expressed in WT palatal primordia but is down-regulated in Cyp26b1−/− fetuses (black arrowheads and insets at higher magnification), while Tbx1 is expressed at similar levels in the forebrain (arrows). At E13.5, Tbx1 expression is still reduced in the bend region of Cyp26b1−/− palatal shelves. E: Real-time PCR validated the decreased Tbx1 expression in middle (MP) and posterior palatal shelves (PP). Values (indicated as red) are means + SD, relative to WT palatal shelves in each part (set as 1). AP, Anterior palatal shelves. *P < 0.05.

F–I: Fgf10 expression is similar in the bend region of middle palatal shelves (arrows) in WT and in Tbx1−/− mice. H and I are higher magnification of the top images. Scale bars = 100 μm in A–D and 200 μm in H and I.

Download figure to PowerPoint

Fgf10 is a direct target of Tbx1 in vitro and its expression is reduced in the developing heart of Tbx1 conditional mutant mice (Xu et al., 2004). Based on that background, we addressed the question of whether the loss of Fgf10 in the bend region in the absence of Cyp26b1 is via down-regulation of Tbx1 by analysing Tbx1−/− mice. Tbx1−/− mice exhibit cleft palate due to the failure of elevation of the palatal shelves (Goudy et al., 2010). In Tbx1−/− fetuses, Fgf10 expression was maintained in the bend region of palatal shelves (Fig. 9F–I). These findings suggest that excess RA suppresses Fgf10 and Tbx1 independently in the bend region of the developing palate, in contrast to the report in vitro (Xu et al., 2004).

Because BMP2 induces cell proliferation in palatal mesenchyme (Zhang et al., 2002), we examined whether Bmp2 expression was affected in Cyp26b1−/− mice. In WT mice, Bmp2 is expressed in the mesenchyme on the medial side of palatal shelves including the bend region (Fig. 10A). In Cyp26b1−/− fetuses, Bmp2 expression was absent in the corresponding region, while it was induced in the tongue mesenchyme (Fig. 10B). Finally, we evaluated Shh expression in the Cyp26b1−/− developing palate, based on the finding that Shh is a direct target of Fgf10 in the developing palate (Rice et al., 2004). In WT mice, we observed Shh expression in the oral ventrolateral and tongue epithelia but not in the bend region (Fig. 10C). In Cyp26b1−/− mice, Shh expression in the oral epithelium was not altered, but ectopic expression was detected in the tongue (Fig. 10D).

thumbnail image

Figure 10. Bmp2 expression is reduced in the bend region but Shh expression is maintained in Cyp26b1−/− palatal shelves

A, B: The deletion of Cyp26b1 results in loss of expression of Bmp2 in the medial half of palatal shelves (*) but Bmp2 expression is induced in tongue (**). C, D: Shh expression is not altered in the lateral epithelium of the palatal shelves but is up-regulated in tongue of Cyp26b1 mutants. Scale bar = 100 μm in A–D.

Download figure to PowerPoint

In summary, excess RA caused by the lack of Cyp26b1 affected Tbx1, Fgf10, and Bmp2 expression in the bend region of the developing palatal shelves.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

RA is one of the well-known teratogens to cause cleft palate in fetuses, but the mechanisms remain incompletely understood. Some of the reasons why the investigation of pathogenesis induced by excess RA is complicated include: (1) a high RA concentration cannot be maintained uniformly by exogenous RA administration because of the short in vivo half-life of RA (0.5–2 hr), (2) the teratogenic action of excess RA on palatal shelf development is dependent on the developmental stage. For instance, RA treatment of E10.5 fetuses leads to small palatal shelves that fail to contact, while RA treatment of E12.5 fetuses leads to normal-sized palatal shelves, which do contact but never fuse (Abbott et al., 1989; Abbott and Pratt, 1991). Mice lacking Cyp26b1, one of the RA-degrading enzymes, eliminated these problems and made it possible to analyze the effects of excess RA on the developing palate by increasing endogenous RA levels within the developing palatal primordia.

The Bend Region of Middle Palatal Shelves Is a Target of Excess RA in Developing Palate

By using RARE-hsplacZ reporter mice, we found that no endogenous RA signaling was detectable in the normal developing palate. The disruption of Cyp26b1 causes an increase of endogenous RA levels, which is restricted to the bend region of palatal shelves. In addition, our in situ hybridization results showed that the up-regulation of Cyp26a1 was not sufficient to compensate for the loss of function of Cyp26b1 in developing palate. These findings strongly suggest that Cyp26b1 is a master gene that maintains homeostasis of RA signaling during normal palatogenesis.

Excess RA decreased cell proliferation in the bend region of middle palatal shelves. The ratio of cell proliferation in the medial half to that in the lateral half of palatal shelves was lower in Cyp26b1−/− mice than in WT mice. These results suggest that the intrinsic force is impaired in Cyp26b1−/− mice. In the previous study, we demonstrated that RA administration to E11.5 fetuses causes cleft palate because of impaired elevation of palatal shelves by arresting mesenchymal cells in the G1 phase through up-regulation of p21 (Okano et al., 2007). Because Cyp26b1−/− palatal shelves are exposed to excess RA from the primordial stage (E11.5), the palatal mesenchymal cells may arrest in G1 phase as well.

BMP2-soaked beads were reported to induce cell proliferation in the palatal shelves in vitro, and the deletion of Bmpr1a leads to cleft palate due to a failure in outgrowth of the palatal shelves (Zhang et al., 2002; Liu et al., 2005). In Cyp26b1−/− mice, we could not detect Bmp2 expression in the medial half of palatal shelves, consistent with the decrease in cell proliferation. On the other hand, the absence of Cyp26b1 induced ectopic Bmp2 expression in the developing tongue. Since cross-talk between RA and Bmp2 has recently been reported in osteoblastic differentiation (Karakida et al., 2011), further studies are needed to investigate these relationships in palatal development.

RA Degradation by CYP26B1 Is Necessary for the Expression of Fgf10 in the Bend Region of Palatal Shelves

The bend region of palatal shelves is an important region where the expression of Fgf7, Fgf10, and their receptor, Fgfr2b overlaps (Rice et al., 2004). The deletion of either Fgf10 or Fgfr2b causes failure of the elevation of palatal shelves at E15.5, leading to cleft palate (Rice et al., 2004; Alappat et al., 2005). In both Fgf10−/ − and Fgfr2b−/ − mice, cell proliferation decreases in the bend region of the palatal shelves (Rice et al., 2004). Based on the cross-talk between RA signaling and Fgf signaling in various organs during embryogenesis (Wang et al., 2006; Ribes et al., 2009), we asked whether excess RA caused by the lack of Cyp26b1 affected Fgf signaling in the developing palate; we found that the increase of RA levels suppressed Fgf10 expression in the bend region of the middle palatal shelves, accompanied by decreased cell proliferation. This result suggests that RA degradation by CYP26B1 is required for Fgf10 expression in the bend region of the middle palatal shelves so that palatal cells proliferate properly. Altogether, RA signaling is linked to Fgf signaling during palatogenesis.

The molecular mechanism of the close connection between RA signaling and Fgf10 is at present unknown. The RARE consensus sequence is reported to comprise direct repeats (DR) of 5′-RGKTCA-3′ separated by 1, 2, or 5 nucleotides (Umesono et al., 1991). We could not find RARE sequences in the Fgf10 promoter by in silico analysis. This result, however, is not sufficient to assume that Fgf10 is not a direct target of RA signaling because a recent study showed that the majority of RA target genes contain anomalously spaced DRs rather than consensus DRs, implicating complexity in RA signaling (Delacroix et al., 2010).

Alternatively, Fgf10 may be regulated indirectly by RA signaling. Recently, Goudy et al. (2010) demonstrated that the palatal shelves remain unelevated in Tbx1−/− mice, resulting in cleft palate. Tbx1 has been shown to be a major candidate gene for the DiGeorge syndrome phenotypes, which is characterized by aortic arch anomalies, dysmorphic face, and hypoplasia of the thymus (Baldini, 2005). Roberts et al. (2005) showed that RA treatment represses Tbx1 expression in the developing pharyngeal arches in chicken embryos. In Tbx1−/− mice, Fgf10 expression is absent in the developing pharyngeal arches (Vitelli et al., 2002). Given those reports, we assessed whether RA signaling regulated Fgf10 expression via Tbx1 in the bend region of the palatal shelves by analyses of Tbx1−/− mice. We found a similar level of Fgf10 expression in the bend region of the middle palatal shelves in Tbx1−/− mice as that in WT mice. Thus, our study suggests that RA regulates Fgf10 and Tbx1 independently in the bend region of the developing palate, suggesting that Fgf10 is not downstream of Tbx1.

DiGeorge syndrome patients occasionally have cleft palate, while the frequency of cleft palate depends on strain background in Tbx1−/− mice (Vitelli et al., 2002). Fgf10 was up-regulated in the entire palatal shelves of Tbx1−/− mice (Goudy et al., 2010). The discrepancy between the studies by Goudy et al. (2010) and ours may be explained by: (1) the use of different mouse strains, (2) other data were obtained on the anterior palate (Fig. 6 C, D in Goudy et al., 2010), while we focused on the middle palate. Further analyses should be designed to determine whether RA signaling regulates Fgf10 directly or through secondary mediators in the developing palate.

The Effects of Deletion of Cyp26b1 on Extrinsic Forces

The cooperative actions of intrinsic and extrinsic forces are required for the palatal shelves to elevate horizontally at the proper time (Ferguson, 1988). The intrinsic force is generated by accumulation and hydration of hyaluronic acid (Ferguson, 1988). The extrinsic force is produced by mouth opening and tongue withdrawal, coinciding with palatal shelf elevation. A defect in either of these two forces can lead to cleft palate. Therefore, it is instructive to classify mutant mouse models with cleft palate into three groups in terms of causal factors for the failure of palatal shelf elevation: (1) the extrinsic force is defective, (2) the intrinsic force is defective, and (3) both the intrinsic and extrinsic forces are disrupted. The first group includes Hoxa2−/− and Gli3−/− mice (Barrow and Capecchi, 1999; Huang et al., 2008). In Hoxa2−/− fetuses, aberrant origin and/or insertion have been identified in seven of the eight muscles of the hyoid muscle complex (Barrow and Capecchi, 1999). While Gli3−/− mutant mice have a cleft palate due to the lack of palatal elevation in vivo, when the tongue was removed in vitro, the Gli3−/− palatal shelves elevate and fuse normally, indicating that extrinsic factors are the cause of cleft palate in Gli3−/− mice (Huang et al., 2008). Another example is Gad67−/− fetuses, in which the cleft palate was rescued by repeated mouth opening or tongue excision under ex utero surgery (Iseki et al., 2007). The second group includes Osr2−/− and Gsk3β−/− mice (Lan et al., 2004; He et al., 2010). In Osr2−/− palatal shelves, the ratio of BrdU-positive cells in the medial half compared to the lateral half of palatal shelves was lower than one (Lan et al., 2004). The specific deletion of Gsk3β in the palatal epithelium led to impaired elevation of palatal shelves, and an organ culture of Gsk3β−/− palatal shelves after tongue removal did not rescue the impaired elevation of palatal shelves (He et al., 2010).

In the present study, we demonstrated that Cyp26b1−/− mice could be classified into the third group, in which both the intrinsic and the extrinsic forces are defective. The considerable increase of the endogenous RA level in Cyp26b1−/− mice resulted in a failure of the elevation of the palatal shelves and an organ culture of Cyp26b1−/− palatal shelves after tongue removal did not rescue the impaired elevation of the palatal shelves in vitro. This finding strongly suggests that intrinsic forces are disrupted in the absence of Cyp26b1. On the other hand, aberrant differentiation in several tongue muscles was also observed (genioglossus, geniohyoid, mylohyoid, and anterior digastric muscle) in Cyp26b1−/− mice. The common feature of these muscles is that either their origin or insertion is located on the hyoid bone and all of them play an important role for depressing the tongue and mandibular primordia. Since the hyoid bone is absent in Cyp26b1−/− mice (Maclean et al., 2009), it is very plausible that the origin or insertion of these muscles is severely disrupted. The significantly greater height of the Cyp26b1−/− tongue compared to the WT tongue would be consistent with defective function of these tongue muscles in Cyp26b1−/− mice. This observation implies that excess RA due to the lack of Cyp26b1 leads to the defective extrinsic forces. Dysfunction of extrinsic factors may accelerate the pathogenesis of cleft palate in Cyp26b1−/− mice.

It is intriguing that Fgf10−/− mice can be classified into the third group as well. Rice et al. (2004) pointed out that cell proliferation is decreased in the bend region of Fgf10−/− palatal shelves. In addition, another study demonstrated abnormal adhesion of Fgf10−/− palatal shelves with the tongue and/or mandibular primordial, which inhibits the elevation of the palatal shelf, implicating a defective extrinsic force in Fgf10−/− mice (Alappat et al., 2005). Therefore, there are both extrinsic and intrinsic influences contributing to the cleft palate in Fgf10−/− mice.

Taking into consideration the idea that the pathogenesis of human nonsyndromic cleft palate is not due to a single factor in most cases (Beaty et al., 2011), the detailed investigation of the mouse models categorized into the third group may give new insight into the mechanism of nonsyndromic cleft palate in humans.

Although it is more than five decades since the developing palate was shown to be susceptible to RA excess, its teratogenic mechanism is poorly understood. In the present study, we demonstrated the importance of the regulation of RA levels via CYP26B1 for proper palate development. The deletion of Cyp26b1 leads to an increase of RA signalling specifically in the bend region of palatal shelves, which impairs the intrinsic forces required for palatal shelf elevation. Increased RA levels in the bend region of palatal shelves inhibits Fgf10 expression, a key regulator of palatal shelf development and results in the failure of the horizontal elevation of the palatal shelves.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Mice

Cyp26b1+/− mice were generated and kindly provided by Dr. Hamada (Yashiro et al., 2004). RARE-hsplacZ transgenic mice (Rossant et al., 1991) were mated with Cyp26b1+/− mice to generate Cyp26b1+/−; RARE-hsplacZ mice and genotyped as previously described (Yashiro et al., 2004). Tbx1+/− mice were generated as previously described (Jerome and Papaioannou, 2001). RA (100 mg/kg; Sigma-Aldrich, St. Louis, MO) was given orally to pregnant mice (SLC, Shizuoka, Japan) on gestational day 11.5 as previously described (Okano et al., 2007). All animal care and experimental protocols were approved by the Animal Care and Use Committee of Kyoto University.

Histological Analysis

For H&E staining, fetuses were fixed in 4% paraformaldehyde at 4°C overnight, embedded in paraffin, sectioned at 7 μm and stained. The evaluation of E14.5 tongue height was as described by Goudy et al. (2010). An unpaired t-test (two-tail) was used to assess the significance of the data. In situ hybridization on paraffin or frozen sections was performed as described by Sakai et al. (2004). The probes for Cyp26a1 and Cyp26b1 were kindly provided by Dr. Hamada (Uehara et al., 2007), Shh by Dr. McMahon (Rice et al., 2004), the Tbx1 by Dr. Scambler (Roberts et al., 2005), and Fgf10 by Dr. Itoh (Sakaue et al., 2002). The Bmp2 probe corresponds to the 207–1,407 nucleotide residues of the mouse Bmp2 cDNA (Genbank accession no. NM_007553.2). For analysis of RARE-hsplacZ expression, transgene-positive fetuses were selected by genotyping as previously described by Rossant et al. (1991) and fixed in 0.4% glutaraldehyde for 25 min; 30-μm cryo-sections were stained for β-gal (Wako, Kyoto, Japan).

RNA Isolation and Real-Time RT-PCR

E13.5 palatal shelves were dissected from WT and Cyp26b1−/− mice, and RNA was isolated using Nucleospin RNA XS (MACHEREY-NAGEL). The first strand complementary DNA (cDNA) was synthesized with oligo dT primers using an ImProm-II Reverse Transcription System (Promega, Madison, WI). Real-time PCR was performed in duplicates using the iQTM SYBR Green Supermix (Bio-Rad, Hercules, CA). The following primers were used: Fgf10-F, TCAAAGCCATC AACAGCAAC; Fgf10-R, GTGCTGCC AGTTAAAAGATGC; Tbx1-F, ACCTG CTGGATGACAATGG; Tbx1-R, CTGT CTTTTCGAGGGTCCAC; RPLP0-F, ATCAATGGGTACAAGCGCGTC; RP LP0-R, CAGATGGATCAGCCAGGAA GG.

Individual gene expression was normalized against the RPLP0 housekeeping gene. Three independent experiments were performed and representative data are shown. An unpaired t-test (two-tail) was used to assess significance of the data.

Proliferation and Apoptosis Assay

Immunofluorescence staining using an anti-rabbit caspase 3 antibody (1: 200, Cell Signaling, Danvers, MA) was carried out to detect apoptosis after the fetuses were fixed with periodate-lysine-paraformaldehyde for 3 hr at 4°C. Secondary antibodies were conjugated with antibody Alexa® 488 (1: 200, Molecular Probes, Eugene, OR). Total nuclei were counterstained with DAPI (Vector Laboratories, Burlingame, CA). For cell proliferation assays, pregnant mice were injected intraperitoneally with BrdU (50 mg/kg, Sigma-Aldrich), and were then killed after 1 hr. Embryos were processed into wax. Sections were incubated with anti-BrdU antibody (1: 50, Oxford Biotechnology), followed by the second antibody, Alexa® 568 (1: 200, Molecular Probes). The percentage of BrdU-positive cells to total mesenchymal cells in the fixed area was counted by two different methods as previously described (Lan et al., 2004; Rice et al., 2004). Briefly, the total number of mesenchymal cell nuclei as well as the number of BrdU-positive mesenchymal nuclei were counted in a fixed 0.015 mm2 area of the medial and the lateral half, respectively, near the tip of the palatal shelf. Sections were selected at the middle level of palatal shelves with comparable positions in WT and Cyp26b1−/− fetuses and cell counts were performed in the bilateral palatal shelves over five continuous sections of each fetus. For the analysis of the bend region, 0.009 mm2 in the bend region of palatal shelves was counted, following the method of Rice et al. (2004). The percentage of anti-caspase 3–positive cells to the total mesenchymal cells in 0.009 mm2 in the bend region of palatal shelves was also evaluated, Four samples were examined in each WT and Cyp26b1−/− group and an unpaired t-test (two-tail) was used to assess any significant difference.

Fetal Mouse Palate in Suspension Organ Culture

E13.5 palatal shelves were cultivated in suspension for 20 hr as previously described (Shiota et al., 1990). Briefly, they were placed into a sterile 15-ml glass bottle containing 1.5 ml of BGJ-b (Gibco-Life Technologies, Gaithersburg, MD), 6 mg/ml of BSA, and 150 mg/ml of ascorbic acid. The bottle was flushed for 2 min with a gas mixture of 95% O2, 5% CO2, and incubated at 37°C on a roller device (Taiyo Science Industrial Co., Tokyo, Japan). The explants were fixed in Bouin's solution and coronal planes were obtained.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We are indebted to Dr. Janet Rossant and Dr. Hiroshi Hamada for providing RARE-hsplacZ transgenic mice and Cyp26b1+/− mice, respectively. We are grateful to Dr. Alan Peterkofsky for critical reading of this manuscript. We also thank Drs. Hiroshi Hamada, Masayuki Uehara, Andrew McMahon, and Nobuyuki Itoh for the ISH probes.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES