The role of transforming growth factor-β signalling in the patterning of the proximal processes of the murine dentary



The evolution of the novel mammalian jaw articulation has resulted in an increased complexity of the dentary bone, reflecting the multiple roles it now fulfils as the primary bone of the mandible. Signalling through the Tgf-β type II receptor is important in the development and patterning of the proximal dentary processes, especially the angular process, and secondary cartilages. We show that expression of Tgf-β2 is associated with the developing angular process, and that the connective tissue marker Scleraxis is co-expressed with Tgf-β2. Scleraxis expression is lost around the angular process of Tgfbr2 conditional knockouts and Tgf-β signalling can induce Scleraxis expression in explant culture. Induction of secondary cartilages in explant culture can be prevented by inhibition of Tgf-β signalling. This study suggests that the proper development of the processes and their secondary cartilages relies on both Tgf-β signalling and mechanical forces working in concert. Developmental Dynamics 237:1604–1613, 2008. © 2008 Wiley-Liss, Inc.


The lower jaw, or mandible, performs multiple functions: it forms the jaw joint with the upper jaw, carries the dentition, and serves as an attachment site for the muscles of mastication. In non-mammalian quadrupeds, the mandible is a compound structure made up of a number of membranous and endochondrally derived bones. In mammals, however, a single membranous bone, the dentary, serves these varied functions. The multiple functions of the mammalian dentary is reflected in its division into a number of different morphological and functional units (Klingenberg et al.,2003). Most commonly the units described are: the mandibular body, the alveolar bones of the molar and incisor teeth, the mental/rostral process, or “chin,” and the three proximal processes, the coronoid, condylar, and angular (Atchley et al.,1985; Atchley and Hall,1991) (Fig. 1A).

Figure 1.

Development of the articular portion of the mouse mandible. A: Schematic of adult mouse mandible. B: Parasagittal section through the first branchial arch of an e13.5 mouse showing expression of the early bone marker Runx2 by DIG-labelled in situ hybridisation. Outline shows the patterning of condensing mesenchyme in the three processes of the mammalian mandible. C–F: Alizarin red/ alcian blue skeletal preparations of embryonic mouse dentaries between e14.5 and p0. (C) e14.5, (D) e15.5, (E) e16.5, (F) p0. Bone is stained in red, cartilage is stained blue. Black arrows in C indicate the presence of mandibular processes with no secondary cartilage. Blue arrows (D–F) indicate processes terminating in secondary cartilage.

The posterior processes of the mandible terminate in secondary cartilages, the function of which is to facilitate growth and to enable the articulation of the dentary with the cranial base at the squamosal (or squamosal portion of the temporal) bone (Frommer,1964; Beresford,1981; Depew et al.,2002). The exact nature of the development of the secondary cartilages is still in debate. Such cartilages have been reported to grow in contradictory manners in closely related species, for in the mouse they have been reported to develop continuously with the bony process, whilst in the rat and hamster, they develop as a sesamoid, an apparently distinct element separate from the developing bone (Vinkka,1982; Vinkka-Puhakka and Thesleff,1993).

The morphology of the mandible, and the manner of its articulation with the cranial base, is one of the defining criteria of mammals and so when coupled with the knowledge that a number of genetic disorders include mandibular abnormalities (Suri et al.,2006; Jakobsen et al.,2007; Mueller and Callanan,2007; Nezarati and Aftimos,2007), the importance of the study of the development of the mandible becomes apparent. A number of molecules have been demonstrated to be important in the development and patterning of the mandible, one group of which is the Tgf-βs. The Transforming growth factor β (Tgf-β) isoforms, Tgf-β 1, 2, and 3, represent a three-member group of the Tgf-β superfamily of signalling molecules; this superfamily contains more than 30 members, including the Tgf-βs, Bone Morphogenetic Proteins (Bmps), and activins, and is important for normal craniofacial development (Dudas and Kaartinen,2005). In vivo and in vitro investigations have demonstrated the importance of Tgf-β in chondrogenesis and osteogenesis (Alvarez et al.,2002; Janssens et al.,2005; Mukherjee et al.,2005). Knockout studies of Tgf-β 1, 2, and 3 demonstrate that the phenotype of Tgf-β 2 −/− mice have a number of unique phenotypes. These include hypoplasia of the mandible, including a loss of the angular process and diminished condylar and coronoid processes, but an apparent retention of the secondary cartilages (Sanford et al.,1997). This suggests an important role of Tgf-β 2 in the development of the proximal portion of the dentary. Targeted deletion of Tgfbr2, the common type II receptor for all three isoforms of Tgf-β (Kitisin et al.,2007), in the Wnt1 expressing neural crest cells using the cre recombinase system (Tgfbr2 wnt1-cre fl/fl) produces a similar phenotype in the dentary, although in this case the secondary cartilage on the angular fails to form and the condylar cartilage fails to develop mature chondrocytes or undergo ossification (Ito et al.,2003; Oka et al.,2007).

Previous work used the Tgfbr2 wnt1-cre fl/fl mouse to investigate the role of Tgf-β signalling in chondrogenesis and osteogenesis during mandible development (Oka et al.,2007). In this study, it was found that the mandible hypoplasia observed in these mice was not due to a failure in migration of the neural crest cells into the 1st branchial arch but due to a reduction in proliferation in Meckel's cartilage and the presumptive dentary mesenchyme. Msx1, a potential Tgf-β-controlled inducer of osteoprogenitor cell proliferation, was diminished in the undifferentiated mesenchyme of the dentary. Whilst the above study revealed the role of Tgf-β on the proliferation and differentiation of the cartilages of the mandible, its role in patterning the mandible was not addressed. In light of this, we aim here to investigate the role of Tgf-β signalling in patterning the mandible, specifically the three proximal processes and the secondary cartilage.


Morphology of the Developing Mandible

In mouse development, the ossified dentary is first detectable by alizarin red at e14.5 as a thin element lateral to Meckel's cartilage. At this initial stage the dentary already has its characteristic three-processed form, with distinct coronoid, condylar and angular outgrowths. The condylar process secondary cartilage is first detected with whole mount alcian blue staining at e15.5, and the angular process cartilage at e16.5, with both cartilages prominent at P0 (Fig. 1C–F).

In order to resolve the pattern of the developing dentary before ossification, alkaline phosphatase staining and in situ hybridisation for Runx2 were carried out in sagittal sections of e13.5 mouse mandibles. Runx2 is expressed in the mesenchyme of the future dentary in a form similar to that of the ossified element at e14.5, suggesting that the patterning of the dentary into three processes occurs early in development (Fig. 1B). NBT/BCIP staining for alkaline phosphatase reveals a similar pattern (data not shown). At e12.5, neither alkaline phosphatase nor Runx2 expression demonstrated the pattern of the future dentary (data not shown).

Expression of Tgf-β 1, 2, 3 in and Around the Developing Dentary

Tgf-β 1 is expressed throughout the dentary as it starts to ossify (e13.5–e15.5, Fig. 2A,B, and data not shown). In contrast, Tgf-β 2 and 3 are expressed in the surrounding mesenchyme (Fig. 2C,D). From e13.5 to e15.5, Tgf-β 2 is expressed specifically medially and posterior to the angular process (Fig. 2C and data not shown). When viewed in sagittal section, this region extends posteriorly from the angular process in a triangular shape. The localisation of Tgf-β 2 to the tissue around the angular process mirrors the phenotype of the knockout mouse, where the angular is lost (Sanford et al.,1997).

Figure 2.

Expression of the three isoforms of transforming growth factor β. Serial sagittal sections of a mouse mandible at e14.5. A: Haematoxylin and eosin stain. Dotted line highlights developing dentary bone. D, dentary; A, angular process of dentary. The coronoid process is out of the plane of section. B–D: Expression by radioactive in situ hybridisation of Tgf-β1 (B), Tgf-β2 (C), and Tgf-β3 (D). Image shown serial to boxed region in A. B: Tgf-β1 is highly expressed in the developing bone of the dentary. C: Tgf-β2 is expressed around the developing angular process of the dentary. D: Tgf-β3 is more generally expressed in the undifferentiated mesenchyme around the proximal end of the dentary.

Tgfbr2 wnt1-cre fl/fl Phenotype

It has been previously reported that the conditional loss of Tgfbr2 in neural crest–derived tissues results in a hypoplastic mandible, including the loss of the angular process and secondary cartilages (Ito et al.,2003; Oka et al.,2007). This previous work demonstrated the phenotype of the dentary bone by alcian blue/alizarin red skeletal preparation at e16.5, by which time the dentary is well developed. To investigate this phenotype further, and to investigate how this defect arises, we carried out histological analysis of the dentary in the Tgfbr2 wnt1-cre fl/fl conditional knockout at earlier stages from e13.5 to e16.5. At e14.5, the patterning of the ossified dentary of the fl/fl mouse does not differ greatly from that of the wild-type littermates in terms of the presence or absence of the posterior dentary processes (Fig. 3C,D). This is also reflected at e13.5, where the pattern of Runx2 expression in the non-ossified dentary does not significantly differ between wild type and fl/fl littermates (Fig. 3A,B). At e15.5, the dentary of the fl/fl mouse is small compared to wildtype littermates, with no secondary cartilages and reduced bony processes. The angular process is particularly reduced relative to the coronoid and condylar processes (Fig. 3E,F). This is even more marked at e16.5, with the angular process reduced to a small spur on the condylar process, which is itself much reduced (Fig. 3G,H). At no stage were mature cartilage cells observed on the proximal mandibular processes. In addition to the loss of hard tissue, there appears to be a loss or reduction of some soft tissues. The attachment of the mandibular muscles to the angular was absent (Fig. 3E–H, arrows). This effect was noticeable at E14.5, before the patterning defect in the proximal dentary.

Figure 3.

Comparison of the development of the dentary of Tgfbr2 wnt1-cre fl/fl conditional knockout mice and wild type littermates. A,B: Expression by radioactive in situ hybridisation of Runx2 in sagittal sections of the dentary anlage of wildtype (A) and Tgfbr2 wnt1-cre fl/fl (B) mice. The basic three-pronged pattern of the dentary is seen in both the wildtype and conditional knockout. C–H: Alcian blue stain of the dentary of wildtype (C, E, G) and Tgfbr2 wnt1-cre fl/fl (D, F, H) mice. By e15.5 (E, F), the angular process is reduced to a spur on the inferior aspect of the condylar process in the conditional knockout, and no secondary cartilage is observed by alcian blue staining. Red asterisks indicate the position of the angular process in wildtype mice. Closed arrow indicates muscle attachment at the angular process of wild type mandible. Open arrow indicates absence of muscle attachment at the angle of conditional knockout.

Cultures: Timed Inhibition of Secondary Cartilage Induction and Maintenance

In order to recapitulate the phenotype of the Tgfbr2 conditional knock out, and to determine the stages at which Tgf-β signalling is critical for secondary cartilage development, explant cultures of mouse half mandibles were carried out in the manner previously reported by Glasstone (Glasstone,1968,1971). Initially, explant cultures were carried out to confirm the ability of this system to develop secondary cartilages. Following culture in BGJb medium plus 10% serum, for 4 or 5 days, and staining with alcian blue/alizarin red, e15.5 mouse mandibular explants (n = 42) show both an increase in the condylar process secondary cartilage in 93% of explants, and development of an angular secondary cartilage in 60%. A similar pattern of secondary cartilage induction was observed with explants cultured from e14.5. In each case, the newly forming cartilage developed as a sesamoid, distinct from the ossified dentary (Fig. 4C).

Figure 4.

The effect of SB431542, a small molecule inhibitor of Type I receptor targets of Tgfbr2, on the expression of phosphorolated Smad2/3 protein and the development of secondary cartilages. A: The control explant cultured with DMSO, sectioned and stained using antibodies to pSMAD2/3. Phosphorylated Smads 2 and 3 are found in the perichondrium (PC) of Meckel's cartilage. B: Explant cultured with 10 μM SB431542. Expression of pSmads 2 and 3 in the perichondrium is lost. C,D: e14.5 +5 day mandibular explant cultures in serum-free medium. C: Control e14.5 +5 day explants cultured with DMSO stained with alcian blue and alizarin red. Cdy, Condylar secondary cartilage; Ang, angular secondary cartilage. D: e14.5 +5 day explants cultured with 10 μM SB431542, stained with alcian blue and alizarin red. No secondary cartilages are observed. E: Chart showing loss of secondary cartilage after treatment with inhibitor at e14.5 (n = 16) and e15.5 (n = 42).

Having established our culture conditions, the experiments were repeated using an inhibitor of the downstream signalling of Tgf-βs. SB431542, a small molecule inhibitor of the type I Tgf-β receptors Alk 4, 5, and 7, was chosen as these receptors are the partners of Tgfbr2 in the Tgf-β signalling cascade (Inman et al.,2002; Kitisin et al.,2007). The cultures were carried out at e13.5, e14.5, and e15.5. At e13.5 and e14.5, no secondary cartilage is observed before culture using alcian blue staining, while at e15.5 the condylar secondary cartilage is clearly visible. Half of the head was cultured with inhibitor, while the other half was cultured in control medium, and the pair was compared for alteration in the cartilage pattern. To confirm loss of Tgf-β signalling in the treated cultures, the level of Smad activation was investigated. Smad activity was shown at high levels in control cultures, particularly around the developing perichondrium (Fig. 4A). In the inhibitor-treated cultures, however, the level of Smad activity dropped considerably, confirming repression of this pathway (Fig. 4B).

After 4–5 days in culture, no secondary cartilage was visible by alcian blue staining in the presence of the inhibitor in cultures started at e13.5, e14.5 (Fig. 4D), or e15.5 (summarised in Fig. 4E, e13.5 data not shown). Failure of secondary cartilages to develop in e13.5 and e14.5 explants suggests that Tgf-β signalling is vital in the initiation of secondary cartilages. The already established condylar cartilage of e15.5 explants disappeared after 4 days in culture, indicating the necessity of Tgf-β signalling in the maintenance of secondary cartilages. The control sides, however, developed well-formed secondary cartilage, as observed before (Fig. 4C).

Co-Localisation of Tgf-β 2 and the Connective Tissue Marker Scleraxis Around the Developing Angular

Tgf-β2 was shown to have strong expression around the developing angular process in a dense triangular-shaped structure visible using histological stains (Fig. 5A1–A3, B1–B3). In order to help characterise this tissue, which might be acting upon the angular process primordium, in situ hybridisation for Scleraxis (Scx) and Myf5 mRNA expression was carried out. Scx has been demonstrated to act as a marker for tendon and ligaments, whilst Myf5 is an early muscle marker.

Figure 5.

Expression of Tgf-β 2, tendon marker Scx, and muscle marker Myf 5 in the developing articular region of the mouse mandible at e13.5 and e15.5. A1–D1, A2–D2: Serial sagittal sections of e13.5 mouse mandibles at medial (A1–D1) and lateral (A2–D2) positions through the mandible. A3–D3: Serial sections through the dentary of e15.5 mouse mandible. A1–3: Sirrus red/Alcian blue staining. Mx, Maxilla; Md, Mandible; D, Dentary; Cdy, Condylar; Ang, Angular; Cr, Coronoid; MC, Meckel's cartilage. B1–3: Radioactive in situ hybridisation for Tgf-β 2. Expression around the angular process at e13.5 is found at medial and lateral positions. C1–3: Radioactive in situ hybridisation for Scx. At e13.5, Scx is co-expressed with Tgf-β2 around the angular process at medial positions (C1) but not laterally (C2), nor at e15.5 (C3). Scx is expressed at the muscle attachment sites on the coronoid and angular process from e13.5 (C2-3). D1–3: Radioactive in situ hybridisation for Myf 5. Arrowheads indicate dense triangle of Tgf-β2 expression cells around the angular process. Red asterisks indicate site of attachment of m. temporalis on the coronoid process.

From e13.5, before the dentary has ossified, Scx is expressed around the future sites of muscle attachment of the presumptive dentary at the coronoid and angular process (Fig. 5C1–C2). The expression of Scx, when observed in sagittal section, takes the form of a thin band along the border of the future bone; it is also strongly co-expressed with Tgf-β2 in the extended region posterior and medial to the future angular process. Expression of Scx, however, does not extend as laterally as Tgf-β2 (Fig. 5B1,2, C1,2). At this stage, Myf5 is weakly expressed in the developing extraocular and facial muscles, where as by e15.5 expression is strong in these tissues (Fig. 5D1–3). At the site of attachment for the temporalis muscle at the coronoid process, there is a close association of Myf5 expression and the expression of Scx at the interface of the muscle with the bone. However, there is no such relationship between Scx and Myf5 in the region around the angular process in which Scx shares its expression domain with Tgf-β2 (Fig. 5B1, C1, D1).

The dense triangular-shaped tissue around the developing angular process, which co-expresses Scx and Tgf-β2 mRNA in the wild type mouse (Fig. 6C, asterisk), was absent in the Tgfbr2 wnt1-cre fl/fl conditional knockout (Fig. 6A). In order to access whether the expression of Scleraxis was also specifically lost in this region, Scleraxis expression was investigated in the conditional knockout. Scleraxis expression was lost around the angular process at e14.5 in the mutant. However, expression was still observed in developing tendons in other parts of the proximal jaw (Fig. 6D). The loss of expression was, therefore, confined to those areas where Scleraxis and Tgf-β2 were coexpressed.

Figure 6.

Expression of Scx in the dentary of Tgfbr2 wnt1-cre fl/fl. A: Sirrius red/alcian blue stain of sagittal section of the dentary at e14.5 in Tgfbr2 wnt1-cre fl/fl. Asterisk indicates loss of triangle of cells associated with the angular process. C: Sirrius red/alcian blue stain of sagittal section of the dentary at e14.5 in Tgfbr2 wildtype littermate. The triangle of cells that lead off from the angular process is evident. B: Expression of Scleraxis in sagittal section of the dentary at e14.5 in the Tgfbr2 wnt1-cre fl/fl knockout. Asterisks mark region of loss of Scleraxis expression around angular process. Arrow indicates regions of Scleraxis expression unaffected in the mutant. D: Expression of Scleraxis in sagittal section of the dentary at e14.5 in a wildtype littermate. Asterisks and arrow marks expression of Scleraxis expression around dentary.

Tgf-β signalling induces the expression of Scleraxis and inhibition of Tgf-β signalling results in a loss of endogenous Scleraxis expression.

Due to the co-expression of Tgf-β2 and Scx, and the specific loss of Scleraxis around the angular in the conditional knockout, we hypothesised that there is a relationship between the two. To test this, e13.5 whole mandibular arch explants were treated with beads soaked in either Tgf-β 1 (n = 6), Tgf-β 2 (n = 9), or BSA (n = 6). Beads were implanted into the mesenchyme around the forming Meckel's cartilage and then the explants were cultured for 24 hr. Following whole mount in situ hybridisation staining for the expression of Scx mRNA, it was found that, in all explants Tgf-β 2 was able to induce Scx expression in the mandibular arch (Fig. 7C). This is true for the regions around the endogenous Scx domain, and areas away from it. In addition, all explants cultured in the presence of Tgf-β 1 beads showed induction of Scx outside of endogenous regions (n = 6) (Fig. 7B). There was no evidence that Scx was up-regulated in the presence of BSA-soaked beads (n = 6) (Fig. 7A). In section, Scleraxis could be observed upregulated in the mesenchyme in an arc following the endogenous expression along the developing dentary (Fig. 7D). This upregulation was confirmed by qRT-PCR (Fig. 7E), which found a significant increase the relative expression of Scx in explants treated with Tgf-β 2 (P < 0.05).

Figure 7.

Effect of inhibition and induction of Tgf-β signalling on the expression of Scleraxis in e13.5 mandibular explants. A–E: Expression of Scx in response to the addition of Tgf-β protein. A: Whole mount in situ hybridisation showing ectopic Scleraxis expression is not induced when mandibular arches are cultured with beads loaded with BSA. B:Scleraxis induced around Tgf-β1-loaded beads. C:Scleraxis induced around Tgf-β2 loaded beads. D: Section through culture with Tgfβ bead stained for expression of Scx by wholemount in situ hybridisation. Black arrow shows endogenous Scleraxis expression, white arrow shows ectopic Scleraxis expression in the mesenchyme, in an arc around the bead, following the endogenous expression pattern. E: Quantitative analysis of Scx expression by qRT-PCR in mandibular explants treated with BSA beads (control) or beads loaded with Tgf-β2. Asterisk indicates statistical significance (P < 0.05) determined by Student's t-test. F,G: In situ hybridisation for the expression of Scleraxis mRNA on sections of e13.5. Culturing in the presence of Tgf-β signalling inhibitor SB431542 (F) results in loss of the expression of Scleraxis in the mesenchyme of the muscle attachment sites on the developing dentary bone (G).

To confirm that Tgfβ signalling was essential for expression of Scleraxis in the proximal dentary, the expression of Scleraxis was investigated after addition of SB431542 to mandibular explants (n = 5 for each stage). After 24 hr, expression was lost in explants cultured from e13.5 and e15.5 with inhibitor (Fig. 7F,G. e15.5 not shown). Tgfβ signals are, therefore, required for both the induction and maintenance of Scleraxis expression around the angular process.


The Initial Pattern of the Dentary Anlage Is Independent of Tgf-β Signalling

Expression of Runx2 and alkaline phosphatase in the mouse dentary anlage at e13.5 indicates that the dentary has the basic adult bone pattern, with three proximal processes, predetermined before ossification. The work here largely corroborates the earlier work of Miyake and colleagues who investigated the expression of alkaline phosphatase in the developing head (Miyake et al.,1997). However, in this study the proximal processes are first described at Theiler's stage 23, which corresponds to around e15 in the mouse strain used (C57BL/6), whereas we demonstrate this pattern to be present at e13.5, which corresponds to Theiler's stages 21–22. This may reflect subtle differences in mouse strain. Additionally, Miyake et al. (1997) demonstrated that development of the secondary cartilages occurs around stage 24, a similar stage to that which we suggest, although no difference in the emergence of the condylar and angular processes was reported. In culture, the secondary cartilages developed distinct from the bone, indicating that the secondary cartilages of the mouse, like the rat and hamster, develop as sesamoids.

In Tgfbr2 wnt1-cre fl/fl mice, which develop with a hypoplastic mandible including an absent angular process, expression of Runx2 in the dentary anlage is by and large normal and displays three proximal processes. This suggests that the early patterning of the membranous dentary is independent of Tgf-β signalling. Tgf-β signalling, however, was crucial for the initiation and maintenance of the secondary cartilages.

Essential Role for Tgf-β Signalling in Initiation and Maintenance of Secondary Cartilages

Secondary cartilages fail to form and differentiate in the Tgfbr2 conditional knockout mouse. In cultures treated with a Tgf-β signalling inhibitor, the secondary cartilages also failed to be induced, and, in addition, if already formed, failed to be maintained. Tgf-β signalling is, therefore, required throughout early secondary cartilage development. Despite this essential role, secondary cartilages develop in the three Tgf-β ligand knockout mice. Tgf-β2 −/− have no angular process but are in possession of a secondary cartilage at the site of the process and there is no mandibular defect in the Tgf-β1 −/− and Tgf-β3 −/− mice (Sanford et al.,1997). This implies that the Tgf-β ligands act in concert or can compensate for each other's loss with respect to induction of secondary cartilage.

Tgf-β Signalling Is Involved in Tendon Development at the Angular Process

Loss of secondary cartilages in the Tgfbr2 conditional mutant may well play a major role in the later failure of the processes of the dentary to extend. However, secondary cartilages are not necessary for the basic pattern of the three proximal processes of the dentary. In the conditional knockouts of the BMP type I receptor, Alk2, secondary cartilages are absent but the three processes of the dentary are still in evidence, though slightly reduced in size (Dudas et al., 2004). In a similar vein, in the Tgf-β 2 knockout, the secondary cartilages are initiated as normal but the angular process is still lost. The formation of the angular is, therefore, not only reliant on the formation of secondary cartilage for its extension and development into a major muscle attachment site.

Tgf-β 2 is expressed around the developing dentary with high levels associated with a triangle of cells under the angular process. This group of cells also co-expresses Scleraxis, a tendon and ligament marker (Liu et al.,1996; Murchison et al.,2007). Previous studies have suggested that Scx is induced by Fgf8 (Brent and Tabin,2004; Manfroid et al.,2006). However, we show that Tgf-β signalling can also induce Scx expression, and that inhibition of Tgf-β signalling can interrupt Scx expression in the muscle attachment sites at the angular process on the dentary bone. Whilst earlier cell culture studies demonstrated an upregulation of Scx by Tgf-β treatment in a dose-dependant manner (Schweitzer et al.,2001; Brent et al.,2003; Brent and Tabin,2004), we have demonstrated this effect in explant culture.

A recent study by Murchison and co workers on the Scx −/− mice do not report any loss of the angular process, or indeed any other mandibular phenotype (Murchison et al.,2007). However, in this study it is established that although Scx is a good marker for all tendons, surprisingly the loss of Scx does not result in a loss of all tendons, only the intermuscular tendons and the tendons responsible for transmitting musculoskeletal force in the limbs, tail, and trunk. There is no effect in those tendons that anchor muscles to skeletal elements, such as the dentary. This study demonstrates that Scx is not a master controller of tendon development, and suggests that there is some as yet unknown factor inducing those tendons unaffected by Scx loss.

Our data indicate an important role for mechanical force in shaping the dentary, in particular the angular process. In the Tgfbr2 conditional knockout, the triangular group of cells expressing Scleraxis and Tgf-β 2 is lost as is the expression of Scx in this region at e14.5. This is correlated with a defect in muscle attachment in the conditional mutant at this stage. Such loss would then lead to a subsequent loss of mechanical stimulation. Fetal jaw movement has been shown to affect condylar cartilage development from e15.5, thus the loss of muscle attachment in the mutant could secondarily affect development of the secondary cartilages (Habib et al.,2005). Importantly, although muscle and tendon defects are visible in the conditional mutant at e14.5, no dentary defect is observed until e15.5. This is also before the advent of secondary cartilage. Thus, the initial defect in the dentary is likely to be driven by defects in muscle attachment.

Rot-Nikcevic and co-workers report that mice lacking both Myf5 and MyoD, and hence lacking any muscle, develop a reduced and immature condylar cartilage, but no angular cartilage or process (Rot-Nikcevic et al.,2007). The pattern of the dentary of the Myf5 −/ −; MyoD −/− mouse, therefore, resembles very well that of the Tgfbr2 wnt1-cre fl/fl conditional knockout. Neither has an angular process, and the other processes are hypoplastic and the secondary cartilage is absent from the angular process and reduced in the condylar. This again strongly suggests that mechanical forces are important in the mandibular phenotype of the Tgfbr2 wnt1-cre fl/fl mouse.

Mechanical Force and Secondary Cartilage Formation

In our cultures, the dentary anlage formed secondary cartilages in the absence of mechanical stimulation, agreeing with previous experiments (Glasstone,1968,1971). The initiation of secondary cartilage is thus independent of muscle. These cartilages, however, did not mature and undergo secondary ossification, that is to say they did eventually not give rise to endochondral bone as occurs in vivo. This failure in secondary ossification is probably due to the lack of mechanical force, which is known to be important for the maturation of secondary cartilages (Herring and Lakars,1982; Hall and Herring,1990; Habib et al.,2005). It has been demonstrated that the culture of perinatal mandibular condyles in the absence of mechanical stimulation will result in a loss of the characteristic features of mature cartilage, and an increase in hypertrophic chondrocytes and deposition of bone (Silbermann et al.,1987). When condylar explants are cultured with functional loading, mimicking the mechanical stimulation of muscle action, or by electrical stimulation of the attached muscles, the increase in bone production is not seen, and the cartilage takes on the form of that seen in vivo (Kantomaa and Hall,1988; Pirttiniemi and Kantomaa,1996).

To summarise, we have shown here that the pattern of the proximal mammalian dentary into the coronoid, condylar, and angular processes is determined prior to embryonic day 13.5, at least 24 hr before the ossification of the bone. Furthermore, we show that this pattern is maintained at this stage in the Tgfbr2 wnt1-cre mouse, which is lacking proximal mandibular processes later in development. We show that Tgf-β signalling is essential for initiation and maintenance of secondary cartilages in culture, indicating that the different Tgf-β ligands are able to compensate for each other's loss in knockout mice. Loss of the angular process in the conditional Tgfbr2 knockout may be due in part to a combination of loss of secondary cartilage formation and loss of mechanical force. The region around the angular that expresses Tgf-β 2 also expresses the tendon marker Scleraxis, which can be induced by Tgf-β signalling. This region and the expression of Scleraxis is lost in the conditional knockout, resulting in a failure of muscle attachment. Finally, whilst it is apparent that the actions of muscles and other mechanical forces are important for the differentiation of secondary cartilages into the bones of the processes, and for the development of the angular process in particular, they may not be important for the initial induction of mouse secondary cartilages. The relationship between mechanical force, chemical morphogens, and other factors in development is complex and requires further study.


Embryo Collection and Dissection

CD1 female mice were mated overnight and the embryos collected at a range of stages. The age of the embryos was further assessed by using anatomical landmarks, such as development of the palate and vibrissae. Embryos were then processed for histology, wholemount analysis, or explant culture. Mutant tissue was generated as in Oka et al. (2007). Briefly, Tgfbr2fl/+;Wnt1-Cre mice were crossed with Tgfbr2fl/fl, resulting in progeny with Tgfbr2 Wnt1-cre fl/fl null allele, as detected by genotyping. Those progeny without the null allele were used as control litter mates.

Tissue Processing and Histological Staining

Following fixation in 4% paraformaldehyde at 4°C, wildtype and mutant embryonic mouse heads were dehydrated in an ethanol series and embedded in paraffin wax. Eight-micrometer sections were cut and serially laid out on superfrost plus slides. The first slide of each series was stained either with haematoxylin and eosin, or with sirrius red/alcian blue using standard techniques. Subsequent slides in each series were used for in situ hybridisation.

To stain for alkaline phosphatase, samples were fixed in 70% ethanol and dehydrated to 100% in a graded series. Dehydrated samples were embedded in molten polyester wax at 42°C. Ten-micrometer sections were cut using a conventional microtome. Sections were laid out on to slides coated with 1% bovine albumin and 1% bovine gelatine (Sigma) and left to dry overnight at 32°C. Sections were stained using NBT/BCIP.

In Situ Hybridisation

Radioactive probes for mouse Tgf-β 1, 2, and 3, Scx, and Myf5 RNAs were made and in situ hybridisations were carried out to detect the expression of these genes in sagittal plain cut sections of wildtype and Tgf β r2 wnt1-cre fl/fl mouse jaws using standard techniques (Mahmood and Mason,1999). Expression of Runx2 RNA and Scleraxis was detected using a dig-labelled probe using standard techniques (Pelton et al.,1991). Cultures were analysed in whole-mount and then sectioned.


Following rehydration, slides were treated with H2O2 to quench endogenous peroxidises, washed in PBS, then blocked in 5% new born calf serum before being treated overnight with goat anti-pSMAD2/3 (Santa Cruz) at a dilution of 1/100 in block solution. Following washing in PBS and further blocking, the primary antibody was detected with biotinylated bovine anti-goat antibody (Santa Cruz), and the signal amplified with the ABC Vectastain system (Vector Laboratories), as per the manufacturer's instructions, and visualised using DAB. The slides were then dehydrated though an ethanol series, including a light counter-staining with alcoholic eosin in 95% ethanol, and mounted with DPX mounting medium.

Explant Culture

Dentaries from mice of embryonic day 14.5 (e14.5) and e15.5 were cultured in a manner modified after that of Glasstone (1971). After 4–5 days in culture, the explants were fixed either in 95% ethanol for alizarin red and alcian blue staining, or 4% paraformaldehyde for histological analysis.

In order to inhibit Tgf β signalling, explants of the developing dentary from e14.5 mice were cultured for 5 days with or without the addition of SB431542, dissolved as a stock of 10 mM in DMSO and used at a final molarity of 10 μM. Inhibitor was added to cultures of the right-hand side dentary of each embryo. The left-hand dentaries were cultured as controls in BGJb with equivalent concentration of DMSO. Cultures were then fixed in 95% ethanol and processed for alcian blue/alizarin red staining and the presence or absence of the angular and condylar secondary cartilages were recorded.

For the bead experiments, lower whole arch explants of e13.5 embryos were cultured as previously described (Tucker et al.,1998), but in the absence of serum. Affigel beads loaded with either Tgf-β 1 or Tgf-β 2 (R&D Systems) were then implanted lateral to Meckel's cartilage. Control beads were loaded with BSA. Additionally, e13.5–e15.5 whole arch explants were cultured for 24 hr in serum-free medium with the addition of either 10 μM SB431542 or 10 μl DMSO per ml of medium. These whole arch explants were then processed for in situ hybridisation.


Following explant culture in the presence of beads loaded with BSA or Tgf-β 2, levels of Scx RNA were measured by qRT-PCR (BioRad), along with a suitable housekeeping gene (βActin). Total RNA was extracted and cDNA generated by reverse transcription of the mRNAs using SuperScript™ First-Strand (Invitrogen Life Technologies) in four replicates. The following primers were used: forward Scx cccaaacagatctgcacctt, reverse Scx ggctctccgtgactcttcag, forward βActin agagggaatcgtgcgtgac, reverse βActin caatagtgatgacctggccgt. The real-time PCR was performed using a SYBR super mix kit (Bio-Rad), running with a hot start of 95°C for 15 min, then 35 cycles at 94°C for 15 s, 62°C for 25 s, and 72°C for 15 s. Melting curves were then analysed to determine PCR specificity. The relative gene levels were expressed by the difference of the Ct values of Scx and βActin (ΔCt = CtScx-CtβActin). Student's t-test was applied to determine statistical significance, where the significance threshold was set at P = 0.05. Values presented as mean ΔCt +1SD.

Alizarin Red–Alcian Blue Staining

e14.5 to p0 mouse embryos were fixed in 95% ethanol for 5 days. After 2 hr of fixation, the specimens of age e16.5 and above were skinned and eviscerated using fine forceps and the soft tissues discarded. Following fixation, the specimens were further dehydrated by immersion in acetone for 2 days. The acetone was then removed and the embryos allowed to dry before being placed in a freshly made stain solution made from 1 volume 0.3% alcian blue 8GS in 70% ethanol, 1 volume 0.1% Alizarin Red S in 95% ethanol, 1 volume 100% glacial acetic acid, and 17 volumes 70% ethanol for 3 days. The embryos were then washed briefly in distilled H2O, then macerated in 1% potassium hydroxide (KOH) for 2 days, or until the bones and cartilages were clearly visible. The clearing continued in increasing concentrations of glycerol in 1% KOH, for 1:3 through to 100% glycerol over 3 weeks. After analysis, the embryos were stored at 4°C.


Thanks are extended to H.L. Moses for Tgf-β 1, 2, and 3 plasmids, Cliff Tabin for the Scleraxis plasmid, and Susanne Dietrich for the Myf5 plasmid. This research was funded by the BBSRC.