SEARCH

SEARCH BY CITATION

Keywords:

  • retinoic acid;
  • tongue development;
  • tongue muscle development;
  • cleft palate;
  • CYP26A1;
  • CYP26B1;
  • Tbx1;
  • congenital anomaly

Abstract

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

Excess retinoic acid (RA) during pregnancy can cause various developmental anomalies in both humans and rodents. We investigated the mechanisms underlying the aberrant differentiation of tongue muscles in fetal mice exposed to exogenous RA in utero. RA-degrading enzymes (Cyp26a1 and Cyp26b1) were expressed at early stages of normal tongue development, but exogenous RA perturbed their expression in the fetal tongue. RA is normally distributed in the developing tongue muscles but its localization was disrupted by exogenous RA. After RA treatment, myogenic determination factors were reduced and the differentiation was significantly suppressed in tongue muscles. Tbx1, a candidate gene of DiGeorge syndrome, was down-regulated in the fetal tongue in response to excess RA. Moreover, Tbx1 as well as myogenic determination factors were not observed in tongue muscle primordia of Cyp26b1−/− fetuses. Our study suggests that RA signaling may play an essential role in tongue muscle differentiation via the regulation of Tbx1. Developmental Dynamics 237:3059–3070, 2008. © 2008 Wiley-Liss, Inc.


INTRODUCTION

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

The tongue is an organ crucial to feeding, swallowing, and speech. In mice, the tongue develops from the inner lining of the ventral wall of the primitive oropharynx around embryonic day 11.5 (E11.5). The early tongue bud develops rapidly and differentiates into a muscular organ covered with a mucosal lining. The members of the Bmp and Fgf families are expressed during tongue muscle development (Nie,2005; Suga et al.,2007). However, the roles of the molecules associated with the early stages of tongue muscle development are yet to be determined.

Vitamin A and its derivatives (retinoids) play key roles in a variety of biological processes and are essential for normal embryonic development (Sporn et al.,1984; Ross et al.,2000). Retinoic acid (RA) is the biologically most active retinoid during embryogenesis. RA is synthesized by retinaldehyde dehydrogenases (RALDHs) (Zhao et al.,1996; Grun et al.,2000; Li et al.,2000; Suzuki et al.,2000; Lin et al.,2003; Liu et al.,2007), enters the nucleus, and binds to RA receptors (RARs) and retinoid X receptors (RXRs). Eventually, RA is degraded by cytochrome P450 enzymes-CYP26s (Fujii et al.,1997; White et al.,2000). Therefore, the endogenous RA level in cells is determined by the balance of the activity of RA-synthesizing enzymes (RALDHs) (Niederreither et al.,1999) and RA-degrading enzymes (CYP26s) (Abu-Abed et al.,2001; Sakai et al.,2001).

On the other hand, excess RA is highly teratogenic in both rodents and humans, and induces a variety of developmental anomalies, including those of the central nervous, cardiovascular, and craniofacial systems (Cohlan,1953; Langman and Welch,1967; Shenefelt,1972). Cleft palate is one of the major malformations induced by excess RA in mouse fetuses (Kochhar and Johnson,1965; Abbott et al.,1989). In a previous study (Okano et al.,2007), we showed that excess RA at gestational day (GD) 11.5 causes cleft palate in mouse fetuses by quite different mechanisms from those reported by Abbott et al. (1989) and Abbott and Pratt (1991) who treated mouse dams at GD10.5 or GD12.5. RA seems to act on the tongue and the epithelium and mesenchyme of palatal shelves in different fashions (Okano et al.,2007). In the present study, we observed some aberrant morphogenesis of the tongue in mouse fetuses exposed to exogenous RA, and investigated tongue muscle development.

A previous study showed that Tbx1 is down-regulated in the pharyngeal arch and outflow tract by RA in zebrafish (Zhang et al.,2006). Therefore, we focused on Tbx1 as a candidate of the downstream of RA signaling. Tbx1 is a member of T-box transcription factors and has been associated with DiGeorge/velocardiofacial syndrome (DGS/VCFS) in analyses of both mice and humans (Baldini,2002; Yagi et al.,2003). The syndrome is characterized by abnormalities of the aortic arch and outflow tract, thymic and parathyroid aplasia or hypoplasia, and craniofacial defects (Scambler,2000). Experimentally, mice lacking Tbx1 exhibit phenotypes, including a cleft palate, which are similar to those of DGS/VCFS patients (Jerome and Papaioannou,2001). In mouse fetuses, Tbx1 is expressed in the tongue epithelium and mesenchyme at the tongue bud stage, and later in the tongue muscle primordia and the epithelium of the tongue (Zoupa et al.,2006). Therefore, Tbx1 is expected to play an essential role during tongue development.

In the present study, we found that Tbx1 expression in the tongue muscle primordia was significantly suppressed after RA treatment and was not detected in mutant mice lacking Cyp26b1. The relationship between the excess RA-induced down-regulation of Tbx1 and abnormal tongue morphogenesis is discussed.

RESULTS

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

Abnormal Tongue Development in RA-Treated Fetuses

In E16.5 mouse fetuses, the dorsal surface of the tongue is normally flat (Fig. 1A). The opposed palatal shelves have fused with each other in the midline and the nasal and oral cavities are separated from each other. In RA-treated fetuses, however, palatal shelves remained small in size and vertically oriented beside the tongue, as we previously reported (Okano et al.,2007). In the tongue of RA-treated fetuses, not only the upper median surface of the tongue was depressed but also abnormal notches were often observed on the lateral sides (Fig. 1B). In addition to the depression of the upper median surface of the tongue, transverse muscle fibers were observed to be abnormally bent at the midline. Such an aberrantly shaped tongue was never observed in control fetuses.

thumbnail image

Figure 1. SEM observation of the fetal tongue (E16.5) in the frontal plane. A, B: In control fetuses, opposed palatal shelves have fused with each other and the tongue is flat. In contrast, in RA-treated fetuses, palatal shelves remain in the vertical position and abnormal notches or depressions are observed in the midline and on the lateral sides of the tongue (arrowheads). The muscle fibers of the transverse intrinsic tongue muscle are bent at the midline (arrows). Scale bar = 500 μm. T, tongue; op, oropharynx; m, molar; gg, genioglossus muscle; nop, naso-oropharnx; PS, palatal shelf.

Download figure to PowerPoint

Perturbed Expression Patterns of Molecules Associated With RA Signaling

We examined the expression of genes for RA-degrading enzymes, Cyp26a1, Cyp26b1, and Cyp26c1, by in situ hybridization (ISH) on sections. In control fetuses, the expression of Cyp26a1 was not recognizable in the tongue primordia at early E11 but was observed in patches at late E11 (Fig. 2A). On the other hand, a robust expression of Cyp26a1 was observed in the tongue epithelium of RA-treated fetuses at early E11 (3 h after RA treatment). By late E11 (6 h after treatment), Cyp26a1 was more strongly expressed throughout the tongue epithelium as well as in the epithelium of the mandible in RA-treated fetuses. Ectopic expression was also observed on the sides of the tongue and in the mandibular mesenchyme. The expression of Cyp26a1 in RA-treated fetuses continued to be up-regulated until E12.5 in the dorsal epithelium of the tongue, as compared to control fetuses. At E13.5, the expression pattern of Cyp26a1 in the tongue was similar in RA-treated and control fetuses.

thumbnail image

Figure 2. Altered expression of Cyp26a1 and Cyp26b1, and slightly affected Raldh2 in the fetal tongue induced by exogenous RA. All sections are the frontal sections of the middle part of the tongue. A: Exogenous RA causes increased expression of Cyp26a1 at 3 hr after treatment (early E11), 6 hr after treatment (late E11) and E12.5. The expression of Cyp26a1 has expanded to the epithelium of the tongue and of the mandible (arrows), as well as to the mesenchyme of the tongue and the mandible (arrowheads). B:Cyp26b1 expression is also increased in the tongue mesenchyme from early E11 to E12.5 by exogenous RA. At early E11, it is up-regulated after RA treatment in the deep mesenchyme of the tongue (white arrowheads), while it is increased in the mesenchyme in the central part of the tongue jaw (black arrowheads). At E12.5, its expression expands to the mesenchyme on the lateral sides in the RA-treated tongue (arrows). C:Raldh2 expression is slightly decreased in RA-treated fetuses, as compared with controls from early E11 to E12.5, while at E13.5 the range of its expression expands in the RA-treated tongue. Scale bars = 100 μm.

Download figure to PowerPoint

Cyp26b1, another isozyme of CYP26s, began to be expressed at early E11 in the mesenchyme near the median groove of the tongue primordium in control fetuses and was also observed in the mandible at late E11 (Fig. 2B). In RA-treated fetuses, however, its expression increased significantly in the deep mesenchyme of the tongue at E11. Until E12.5, Cyp26b1 continued to be expressed in the mesenchyme beneath the upper surface of the tongue in control fetuses. After RA treatment, its expression in the mesenchyme increased and expanded to the lateral sides of the tongue. By E13.5, the expression of Cyp26b1 was observed only in the mesenchyme adjacent to the upper surface and was decreased as compared to earlier stages in both control and RA-treated tongues. The expression of Cyp26c1 was observed in the dorsal epithelium at the tongue bud stage and was not affected by RA (data not shown).

Next, we examined whether the members of the Raldh gene family were affected by exogenous RA. There are four members in the Raldh gene family (Raldh1, -2, -3, and -4). Raldh3 was weakly detected in the tongue mesenchyme at early E11, and its expression was not affected by RA. The expression of Raldh1 and -4 was not observed at the tongue bud stage in either group (data not shown). On the other hand, Raldh2 was expressed symmetrically in the mesenchyme at the base of the tongue between early E11 and E13.5 in control fetuses (Fig. 2C). RA-treated fetuses showed a similar expression pattern but the intensity was slightly decreased between early E11 and E12.5. Conversely, at E13.5, Raldh2 expression appeared to increase in the RA-treated tongue and mandible, as compared with controls.

Alteration of RA Distribution in the Fetal Tongue by Exogenous RA

Next, we examined the distribution of RA in the fetal tongue, using RA response element (RARE)-hsplacZ reporter mice, lacZ gene of which reflects the transactivation activity of RA (Rossant et al.,1991). At E11.5, lacZ expression was not observed in the tongue of control fetuses (Fig. 3A), while numerous lacZ-positive cells were observed in the tongue mesenchyme of RA-treated fetuses (Fig. 3B). By E12.5, lacZ-positive cells began to be visualized in the transverse intrinsic tongue muscles in control fetuses (Fig. 3C). LacZ-positive cells were also observed in the primordial hyoglossus and genioglossus muscles (Fig. 3E). On the other hand, lacZ expression in RA-treated E12.5 fetuses exhibited a diffuse pattern in the mesenchyme of the anterior part of the tongue (Fig. 3D). In the posterior part of RA-treated tongue, lacZ-positive cells were distributed in the primordial genioglossus muscle, but were spread diffusely throughout the entire tongue (Fig. 3F). In E13.5 control fetuses, lacZ-positive cells were clearly observed in the primordia of all the intrinsic tongue muscles, i.e., the vertical, transverse, inferior longitudinal and superior longitudinal muscles (Fig. 3G). LacZ-positive cells were also found along the myofibers of the primordial extrinsic tongue muscles such as the genioglossus, palatoglossus, styloglossus, and hyoglossus muscles (Fig. 3I). In RA-treated E13.5 fetuses, lacZ was expressed in both the primordial intrinsic and extrinsic tongue muscles, similarly as in control fetuses, but the distribution pattern differed from that of controls (Fig. 3H, J). RA distribution in the primordial genioglossus muscle was more intense in RA-treated fetuses than in controls, and the genioglossus muscle was often attached abnormally to the inferior longitudinal tongue muscle. Normally, the inferior longitudinal tongue muscle originates at the inferior surface of the tongue root, and its myofibers never extend to the genioglossus muscle located on the medial side of the tongue (Fig. 3G, H). At E13.5, the top of the posterior tongue in control fetuses was found to be bulged, while the tongue surface in RA-treated fetuses remained flat (Fig. 3I, J).

thumbnail image

Figure 3. RA localization in the fetal tongue of RARE-hsplacZ reporter mice. A, B: LacZ-positive cells are not observed in the control fetal tongue at E11.5, but are spread diffusely in the mesenchyme of the RA-treated tongue. C–F: At E12.5, the alteration of RA distribution by exogenous RA is the most evident. G–J: At E13.5, in control and RA-treated fetuses lacZ-positive cells are distributed in the intrinsic muscle primordia of the tongue, but numerous lacZ-positive cells are diffusely distributed in the mesenchyme of RA-treated tongue (arrowheads). In addition, the abnormal attachment of the intrinsic muscle primordium to the genioglossus muscle primordium is observed in RA-treated fetuses, as compared to controls (arrows). Ant, anterior region of tongue; gg, genioglossus muscle primordium; hg, hyoglossus muscle primordium; mh, myohyoid muscle primordium; pg, palatoglossus muscle primordium; Post, posterior region of tongue; sg, styloglossus muscle primordium; tr, transverse intrinsic muscle primordium. Scale bar = 100 μm.

Download figure to PowerPoint

Perturbed Myogenesis in the Developing Tongue of RA-Treated Fetuses

To investigate the effects of RA on tongue muscle development, we examined the expression of Myf5 and MyoD, early and late myogenic determination markers, respectively, in the developing tongue. We also analyzed tongue muscle differentiation, using an anti-myosin heavy chain fast antibody. There was no difference in the gross morphology of the tongue at E12.5 between control and RA-treated fetuses. RA treatment decreased both Myf5 and MyoD in the tongue, as compared with controls (Fig. 4A, D, G, J, M, P, S, V). Immunostaining using an anti-myosin antibody showed that tongue muscles begin to differentiate at E12.5 in control fetuses (Fig. 4B, H, N, T). Myosin was distributed both in extrinsic and intrinsic tongue muscle primordia, and Myf5- and MyoD-positive cells were observed not only in tongue primordial muscles but also in the tongue mesenchyme in both control and RA-treated fetuses (Fig. 4C, F, I, L, O, R, U, X). However, myosin-positive cells were significantly decreased in the RA-treated tongue (Fig. 4E, K, Q, W). Such a difference in myosin distribution between control and RA-treated fetuses was confirmed by observing the genioglossus muscle at higher magnifications, and some primordial myotubes were found to have no lumen and were collapsed in the RA-treated group (see Fig. 6E, F). At E14.0, numerous cells positive for Myf5 and MyoD were observed in both control and RA-treated fetuses (Fig. 5). In RA-treated fetuses, however, myosin-positive cells were not observed in the core of the primordial genioglossus muscle, which suggests that the number of primordial myotubes was decreased (Fig. 5B, E, H, K, N, Q, T, W). The abnormal attachment of the genioglossus and intrinsic muscles in RA-treated fetuses as observed by lacZ staining (Fig. 3H) was confirmed by immunostaining with an anti-myosin antibody (Fig. 5K), which was never found in control fetuses (Fig. 5I).

thumbnail image

Figure 4. Immunohistochemical staining for myogenesis-associated markers at E12.5. A, D, G, J, M, P, S, V: Myf5 (an early muscle maker) and MyoD (a late muscle marker) are expressed in the developing tongue at E12.5, and appear to be decreased both in the anterior and posterior parts of RA-treated tongues, as compared with the controls. B, E, H, K, N, Q, T, W: Myosin-positive cells are distributed in intrinsic muscle primordia of RA-treated tongues less than control tongues. Primordial myotubes in not only the genioglossus muscle but also the geniohyoid and mylohyoid muscles of RA-treated tongues are not well organized (arrowheads). C, F, I, L, O, R, U, X: Myf5 or MyoD, myosin, and DAPI are merged in the tongue of control and RA-treated fetuses. Ant, anterior region of the tongue; Post, posterior region of the tongue; gg, genioglossus muscle primordium; gh, geniohyoid muscle primordium; mh, mylohyoid muscle primordium. Scale bar = 100 μm.

Download figure to PowerPoint

thumbnail image

Figure 5. Immunohistochemical staining for myogenesis-associated markers at E14.0. A, D, G, J, M, P, S, V: Myf5- and MyoD-positive cells are observed in the entire tongue in both control and RA-treated fetuses. B, E, H, K, N, Q, T, W: Myotubes are formed in control and RA-treated tongues, but myotubes in the core region of the genioglossus muscle primordium exhibit poor growth in RA-treated tongues. Aberrant attachment of the genioglossus muscle and intrinsic tongue muscle primordium is also observed in RA-treated tongues, compared with controls (arrowheads). C, F, I, L, O, R, U, X: Myf5 or MyoD, Myosin, and DAPI are merged in the tongue of control and RA-treated fetuses. Ant, anterior region of the tongue; Post, posterior region of the tongue; gg, genioglossus muscle primordium; gh, geniohyoid muscle primordium; mh, mylohyoid muscle primordium. Scale bar = 100 μm.

Download figure to PowerPoint

Next, we counted the number of primordial myotubes in a fixed area within the genioglossus muscle at E12.5 and E14.0, to evaluate the effects of RA on tongue muscle development. At E12.5 and E14.0, there was no significant difference in the total cell number between the control and RA-treated fetuses (Fig. 6A). The ratio of Myf5-positive cells to the total cell number significantly decreased in the genioglossus muscle of RA-treated fetuses at E12.5 (P < 0.01), but the difference was not significant at E14.0 (Fig. 6B). The ratio of MyoD-positive cells significantly decreased in RA-treated fetuses both at E12.5 (P < 0.01) and at E14.0 (P < 0.05). The ratio of primordial myotubes significantly decreased after RA treatment (P < 0.01), but there was no difference in the size of primordial myotubes between the control and RA-treated fetuses (Fig. 6C, D). A high magnification of primordial myotubes in thegenioglossus muscle revealed that the myotubes were formed at E14.0 in both groups (Fig. 6G, H). In RA-treated fetuses, the lumina in primordial myotubes were round, but their size was variable.

thumbnail image

Figure 6. Analysis of myogenesis in the developing genioglossus muscle. A: RA does not affect the cell number in a fixed area in the genioglossus muscle primordium at E12.5 or E14.0. The light gray bars represent the data for control tongues and the dark gray bars show those for RA-treated tongues. Values represent the mean ± SE. B: RA affects the percentage of Myf5- and MyoD-positive cells in the genioglossus muscle primordium. **P < 0.01, *P < 0.05. C, D: RA decreases the ratio of the number but does not affect the average of the area of primordial myotubes in the genioglossus muscle primordium. E–H: A high magnification of primordial myotubes (myosin-positive) in the genioglossus muscle at E12.5 and E14.0 in control and RA-treated fetuses. In E12.5 RA-treated fetuses, primordial myotubes with lumina are collapsed (arrowheads). Scale bar = 20 μm.

Download figure to PowerPoint

Down-Regulation of Tbx1 by Exogenous RA

To examine whether the expression of Tbx1 was affected by exogenous RA, we performed ISH on fetal tongues. In control fetuses, Tbx1 expression was evident in the epithelium and mesenchyme of the median part of the tongue at early E11 and became more intense in the tongue mesenchyme by late E11 (Fig. 7A, C). In RA-treated fetuses, however, Tbx1 expression disappeared in the epithelium of the tongue and was down-regulated in the mesenchyme at early E11 (Fig. 7B). Its expression in the mesenchyme was further decreased at late E11 (Fig. 7D). At E12.5 and E13.5, Tbx1 was expressed in the tongue epithelium and was in the intrinsic and genioglossus muscle primordia in control fetuses (Fig. 7E, G). After RA treatment, Tbx1 expression in the tongue muscle primordia and the epithelium was less intense than in controls (Fig. 7F, H). We carried out real-time RT-PCR to quantify the Tbx1 expression level at E12.5 and E13.5. It was not possible to detect an altered expression in early and late E11 tongues, possibly because Tbx1 was expressed in the tongue bud and the mandible from E11.5. A significant decrease in Tbx1 expression was confirmed at 24 h (at E12.5) and 48 h (at E13.5) after RA treatment, as compared with controls (P < 0.01) (Fig. 7I). The level of Tbx1 mRNA was 0.74±0.075 (SE) and 0.34±0.12 times, respectively, the corresponding values for the control group.

thumbnail image

Figure 7. Altered Tbx1 expression after RA treatment. A–D:Tbx1 expression is down-regulated both in the epithelium and the mesenchyme by RA between early and late E11. E–H: In E12.5 and E13.5 RA-treated fetuses, Tbx1 expression is relatively weak in the genioglossus muscle primordium and the tongue epithelium. Scale bar = 100 μm. I: The results of real-time quantitative PCR are shown using fetal tongues after RA treatment. The decrease was significant at E12.5 and E13.5 in RA-treated tongues, as compared with the corresponding values in control tongues (*P < 0.01). The dark gray bars plus SE show the Tbx1 expression level in RA-treated tongues relative to the corresponding controls (light gray bars).

Download figure to PowerPoint

Tongue Muscle Development in Cyp26b1−/− Fetuses

Mutant mice lacking Cyp26b1 exhibit a significantly higher endogenous RA levels than wild-type (WT) mice during craniofacial development (Okano et al., unpublished data). We investigated the tongue muscle development in Cyp26b1−/− fetuses and found that Tbx1 was down-regulated in the tongue epithelium and was undetectable in tongue muscle primordia of E13.5 Cyp26b1−/− fetuses (Fig. 8A, B). In the anterior part of the tongue of E14.0 Cyp26b1−/− fetuses, immunostaining with an anti-myosin antibody showed a similar pattern to that in WT fetuses (Fig. 8C). However, myosin was not observed in the core of the primordial genioglossus muscle similarly as in E14.0 RA-treated fetuses (Fig. 8D). The alignment of mylohyoid muscle myofibers was aberrant, as compared to controls (compare Fig. 5N, T). It was also noted that the shape of the tongue was similar in Cyp26b1−/− and RA-treated fetuses. Double immunostaining with an anti-Myf5 or an anti-MyoD and with an anti-myosin antibodies revealed that there were few Myf5- or MyoD-positive cells in the genioglossus muscle primordia in Cyp26b1−/− fetuses, although they were abundantly observed in the genioglossus muscle primordia of WT fetuses (Fig. 8E–H).

thumbnail image

Figure 8. Impairment of myogenesis in the developing tongue of Cyp26b1−/− fetuses. A, B:Tbx1 was expressed in the tongue epithelium and mesenchyme of WT fetuses, but not detectable in the mesenchyme of the tongue of E13.5 Cyp26b1−/− fetuses. C, D: Myosin distribution in the tongue of E14.0 Cyp26b1−/− fetuses. In the posterior part, the core region of the genioglossus muscle was not myosin-positive. Myofibers of mylohyoid muscle showed a disorganized alignment (arrowheads). E–H: Double immunostaining with an anti-Myf5 or an anti-MyoD antibodies and an anti-myosin antibody revealed that there were no Myf5- and MyoD-positive cells in genioglossus muscles of Cyp26b1−/− fetuses. Ant, anterior region of the tongue; Post, posterior region of the tongue. Scale bar = 100 μm in B and D.

Download figure to PowerPoint

DISCUSSION

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

RA Signaling Plays a Significant Role in Tongue Development

It has been reported that Raldh2 is expressed at the tongue bud stage (E11.5) (Niederreither et al.,1997) and Cyp26a1, Cyp26b1, and Raldh3 at later stages (E14.5) (Abu-Abed et al.,2002; Niederreither et al.,2002), but it remains unknown whether RA signaling is involved in tongue development. In the present study, we found that Cyp26a1, Cyp26b1, and Raldh2 are expressed from the tongue bud stage (E11) to the later stage when tongue myoblasts are differentiating (E13.5). Cyp26a1 and Cyp26b1 are expressed in the epithelium and mesenchyme, respectively, and their expression never overlapped in the developing tongue. Our ISH study revealed that excess RA causes a significant up-regulation of Cyp26a1 and Cyp26b1 and disturbs their complementary expression patterns. After RA treatment, Cyp26a1 was up-regulated in the epithelium at early E11, before the normal onset of its expression and observed ectopically in the mesenchyme of the tongue at late E11.

The distribution of RA in embryonic tissues can be visualized by using RA-reporter mice (Rossant et al.,1991). We found that endogenous RA was localized mainly in the primordial intrinsic and extrinsic tongue muscles, while Raldh2 expression was observed at the base of the developing tongue. In other words, RA distribution was not localized in the region where RA-synthesizing enzyme was expressed. Such a discrepancy between Raldh expression and RA distribution was also shown in the forebrain development of mouse fetuses (Luo et al.,2004). They found that Raldh3 was expressed in the lateral ganglionic eminence, but lacZ expression was not observed there but in the hippocampus where no Raldh1-3 was expressed. In early embryos, the lacZ-positive region was overlapped with the region where some of Raldh is detectable, but it was not the case in later embryos. They concluded that RA may be supplied by diffusion and circulation to the forebrain at later stages. This hypothesis may also be applied to the developing tongue.

We observed lacZ expression along tongue muscle primordia. It seems that RA may be involved in myogenesis during tongue development. RA has been shown to promote myogenic differentiation of myoblasts in vitro (Alric et al.,1998), but the effects of exogenous RA on myogenesis were not entirely clear. After RA treatment, RA distribution was observed in the tongue bud where it is not detected normally. Such an aberrant distribution was more evident at E12.5. While lacZ-positive cells were localized in tongue muscle primordia in control fetuses, they were more abundant in the tongue mesenchyme than in the primordial tongue muscles in RA-treated fetuses. It is likely that the disturbance in the RA distribution in the developing tongue may perturb developmental programs and results in aberrant morphogenesis of the tongue.

Exogenous RA Down-Regulates Tbx1 in the Developing Tongue

Our present study showed that exogenous RA down-regulates Tbx1 expression in the fetal tongue. Tbx1 expression was not detected in the tongue muscle primordia of Cyp26b1−/− fetuses, in which the endogenous RA level in tongue muscles was significantly higher than in RA-treated fetuses (Okano et al., unpublished data). Roberts et al. (2005) grafted RA-soaked beads in the pharyngeal region of avian embryos and induced a down-regulation of Tbx1 expression. They showed that Tbx1 repression was dependent on the dose of RA in a cell line. Recently, Zhang et al. (2006) showed in zebrafish that excess RA suppressed Tbx1 expression in the pharyngeal arch and the developing outflow tract. These studies strongly indicate that RA is probably upstream of Tbx1. On the other hand, some studies have shown that RA signaling is affected by Tbx1. The expression of Cyp26a1 and Cyp26b1 was reduced in the pharyngeal tissues of Tbx1 null mutant mice (Roberts et al.,2006), and Guris et al. (2006) showed by examining RARE-lacZ expression that RA signaling was ectopically activated in Tbx1 null mutant mice. Therefore, RA signaling and Tbx1 seem to have negative reciprocal interactions with each other. It would be interesting to examine whether Tbx1 actually regulates RA signaling in the developing tongue using Tbx1 null mutant mice.

The Role of RA Signaling and Tbx1 in Tongue Muscle Development

We suggested previously that impaired tongue withdrawal at a proper timing may be a causal factor for cleft palate induced by RA given at GD11.5 (Okano et al.,2007). Normally, palatal shelves elevate and become horizontally oriented above the tongue by E14. After RA treatment, abnormal cell death was found to be induced in the genioglossus muscle primordium and palatal shelves remained unelevated. The failure of palatal shelves to elevate is one of the major pathogenetic events in cleft palate formation, as suggested by Ferguson (1988).

The genioglossus muscle originates in the mental spine of the mandible and functions to depress the tongue, in cooperation with the hyoglossus muscle. We observed that RA induces disorganized morphogenesis of primordial myotubes and their decrease in the fetal tongue. Immunostaining revealed that myosin-positive cells were not located in the core region of the genioglossus muscle and the ratio of primordial myotubes was significantly decreased in RA-treated fetuses. The abnormal attachment of the intrinsic tongue muscle in RA-treated fetuses may have produced the notches on the surface and the lateral sides of the tongue observed at E16.5. Therefore, the results of our present study strongly support our previous study that RA affects the development of the genioglossus and other tongue muscles and prevents tongue withdrawal at a proper timing (Okano et al.,2007).

Tongue muscles are skeletal muscles derived from the progenitor cells, which migrate into the tongue primordia (Noden,1983). Myogenic Regulatory Factors (MRFs), the members of which are Myf5, MyoD, myogenin, and MRF4, have been well known to play important roles in skeletal muscle development (Buckingham et al.,2003). Dalrymple et al. (1999) reported that all MRFs were expressed in the developing tongue, Myf5 being most abundant. Hamade et al. (2006) showed in zebrafish that exogenous RA up-regulated MyoD in somites. Based on these findings, we asked whether Myf5 and MyoD were perturbed by exogenous RA in the developing tongue. After RA treatment, MyoD was significantly decreased both at E12.5 and E14.0 and Myf5 was significantly decreased at E12.5 but not at E14.0 in the genioglossus muscle. This is because Myf5 is an earlier marker of myogenic determination, and premature myoblasts expressing Myf5 at E14.0 may be free from the effect of RA given at GD11.5.

In E14.0 Cyp26b1−/− fetuses, myosin-positive cells were observed but neither Myf5 nor MyoD was expressed in their genioglossus muscle. This result indicates RA signaling is not necessary to initiate muscle development but is required to maintain it in the developing tongue. The analysis of the developing tongue of Cyp26b1−/− fetuses is ongoing in our laboratory to unravel further roles of RA signaling during tongue muscle development.

Tbx1 is expressed in the premyogenic mesoderm of the first and second pharyngeal arches from E9.5 before the onset of MRFs expression (Kelly et al.,2004). In Tbx1−/− mice, the expression of MyoD and Myf5 was lost in the pharyngeal arches, and mandibular arch–derived muscles such as the mylohyoid and anterior digastric muscles were absent or hypoplastic. Jerome and Papaioannou (2001) reported that Tbx1−/− fetuses have cleft palate. Their finding is interesting, because both the mylohyoid and anterior digastric muscles function to open the mouth and depress the mandible, and their impairment may prevent mouth opening, which is necessary for the elevation of palatal shelves (Ferguson,1988), resulting in cleft palate. Moreover, the genioglossus muscle showed poor growth and was morphologically abnormal in Tbx1−/− fetuses, which may be a secondary effect of the absence or hypoplasia of mandibular arch–derived muscles (see fig. 7 in Kelly et al.,2004). In the present study, we showed that after RA treatment, Tbx1 as well as Myf5 and MyoD were suppressed in the developing tongue. Our study has added to new evidence that RA signaling plays an important role in tongue myogenesis through the regulation of Tbx1 (Fig. 9). In the ventral part of the tongue, Raldh2 is expressed and synthesizes RA, which induces the expression of Cyp26a1 and Cyp26b1 is in the dorsal part of the tongue. RA is distributed in the tongue muscle primordia via the balance of RA synthesis by RALDH2 and its degradation by CYP26s. Excess RA down-regulates Tbx1, which is expressed in tongue muscle primordia, resulting in a significant decrease in muscle markers, Myf5 and MyoD, and finally causes a significant deficiency of myotubes in the developing tongue muscles.

thumbnail image

Figure 9. A model of the relationship between RA signaling and Tbx1 in tongue muscle development. Raldh2 is expressed at the base of the developing tongue, while the expression of Cyp26a1 and Cyp26b1 is observed in the dorsal region of the developing tongue. Excess RA down-regulates the expression of Tbx1 and reduces the expression of Myf5 and MyoD, the myogenic-determination markers, resulting in a decrease in the number of myotubes in the tongue muscles.

Download figure to PowerPoint

Tbx1 is a major candidate gene for DiGeorge/velocardiofacial syndrome (DGS/VCFS) (Baldini,2002) and 69–100% of DGS/VCFS patients have palatal anomalies, including cleft palate and submucous cleft palate (Kobrynski and Sullivan,2007). DGS/VCFS patients often have problems in feeding and swallowing, which may be due to poor coordination of the tongue, pharyngeal muscles, and esophageal muscles (Rommel et al.,1999). These clinical facts strongly support our assumption that Tbx1 perturbation can result in cleft palate and abnormal tongue muscle development.

Tongue and the palate are the organs in the oral cavity, and develop at the later stages of major organogenesis. However, tongue development has not been studied as extensively as palate development. There are even fewer studies on tongue muscle differentiation, as compared with those on taste papillae in the developing tongue. Because impaired tongue muscle development can result in various structural and functional problems after birth, further studies are needed to elucidate the molecular mechanisms underlying normal and abnormal tongue development.

EXPERIMENTAL PROCEDURES

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

Mice

RA (100 mg/kg) was given orally to pregnant ICR mice (SLC, Shizuoka, Japan) on GD11.5 as previously described (Okano et al.,2007). Cyp26b1+/− mice were generated and kindly provided by Dr. H. Hamada (Yashiro et al.,2004).

Scanning Electron Microscopy

At E16.5, fetuses were dissected and fixed in 2.0% glutaraldehyde in 0.1M phosphate buffer for 72 hr at 4°C and cut in the frontal plane. They were freeze-dried, mounted onto metal stubs with carbon-conductive paint, coated with a thin layer of gold using a sputter coater (Eiko IB-3), and viewed using a scanning electron microscope (SEM; Hitachi S-450).

In Situ Hybridization (ISH) on Sections and Immunohistochemistry

Fetuses were collected at intervals between E11.5 and E13.5, fixed in 4% paraformaldehyde at 4°C overnight, embedded in paraffin, and sectioned at 7 μm. ISH on sections was performed as described by Yoshida et al. (2001). The probes for Cyp26a1, Cyp26b1, Cyp26c1, and Raldh2 were kindly provided by Dr. H. Hamada (Uehara et al.,2007) and the probe for Tbx1, by Dr. P. J. Scambler (Roberts et al.,2006). Immunoflorescence staining was carried out on 7-μm paraffin-embedded sections that had been fixed in periodate-lysine-paraformaldehyde for 3 hr at 4°C. The following antibodies were used: rabbit anti-human Myf5 (1:500, Santa Cruz: sc-302), rabbit anti-mouse MyoD (1:50, Santa Cruz: sc-760), monoclonal mouse anti-myosin (skeletal, fast) (1: 100, Sigma), and secondary antibodies conjugated with antibody Alexa® 488 and Alexa® 568 (1: 200, Molecular Probes). All nuclei were counterstained with DAPI (Vector Laboratories). Three samples were examined for each control, RA-treated, or Cyp26b1−/− fetuses. As for the analysis of myogenesis in the developing genioglossus muscle (Fig. 6), a fixed area (120 × 120 μm) was defined in the genioglossus muscle at the boundary between the genioglossus and geniohyoid muscles and was used for cell counting using Image J. An unpaired t-test (two-tail) was used for statistical analysis of the data.

β-Galactosidase Staining

RARE-hsplacZ transgenic mice were mated with ICR or Cyp26b1+/− mice (Rossant et al.,1991). Genotyping was performed as previously reported (Yashiro et al.,2004) and the transgene-positive fetuses were selected and fixed with 0.4% glutaraldehyde for 25 min. Cryo-sections were made at 30 μm for β-gal (Wako, Kyoto, Japan) staining.

Semiquantitative and Real-Time RT-PCR

The tongues of control and RA-treated fetuses were dissected out at intervals between E11.5 and E13.5, and RNA was isolated using TRIZOL (Invitrogen, Tokyo, Japan). The first strand complementary DNA (cDNA) was synthesized with random primers using a Superscript First-Strand Synthesis System (Invitrogen, Tokyo, Japan). Semiquantitative RT- PCR was performed with HotStarTaq (QIAGEN, Tokyo, Japan). PCR conditions were: 30 sec at 94°C, 30 sec at 58°C, and 1 min at 72°C for 27 cycles (Cyp26a1 and Cyp26b1) and 24 cycles (EF-1). The products were electrophoresed on 1.5% agarose gel. The primers used were as follows: for Cyp26a1 (GenBank accession no. NM_007811), Forward: 5′-AGCAGCGAAAGAAGGTGATT-3′ Reverse: 5′-TAGCCACTGC- TCCCAGACAAC-3′ for Cyp26b1 (accession no. NM_175475), Forward: 5′-CGCTACCTGGACTGTGTCAT-3′ Reverse: 5′-GGATCTGGAAACCATCC- AGT-3′ for EF-1 (a housekeeping gene used as internal control).

PCR was repeated at least twice for a particular cDNA sample, and results were obtained from at least three independent cDNA preparations.

Real-time RT-PCR was carried out using a TaqMan 7300 thermocycler (Applied Biosystems, Tokyo, Japan). PCR products were made from each cDNA of control and RA-treated tongues at intervals using Taqman® Gene Expression Assays (Assay ID: mouse Tbx1, Mm00448948_m1; mouse Gapdh, Mm99999915_g1). Standard curves were performed using cDNA of the cranial half of E10.5 wild-type embryos with four calibration points: 1:4; 1:8; 1:16; 1:32. Gapdh was used for normalization, calculated using 7300 system SDS software. The Tbx1 expression level in RA-treated tongue relative to control was calculated at each interval. The reactions were run in triplicate, and the experiment repeated three times. An unpaired t-test (two-tail) was used to assess the significant expression level of Tbx1.

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. J. Rossant and Dr. H. Hamada for providing RARE-hsplacZ transgenic mice and Cyp26b1+/− mice, respectively. We also thank Drs. K. Yashiro and T. Okano for technical assistance, and Drs. H. Hamada, M. Uehara, and P. J. Scambler for the ISH probes.

REFERENCES

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