Drs. Wang and Liu contributed equally to this work.
Tissue interaction is required for glenoid fossa development during temporomandibular joint formation
Article first published online: 26 SEP 2011
Copyright © 2011 Wiley-Liss, Inc.
Volume 240, Issue 11, pages 2466–2473, November 2011
How to Cite
Wang, Y., Liu, C., Rohr, J., Liu, H., He, F., Yu, J., Sun, C., Li, L., Gu, S. and Chen, Y. (2011), Tissue interaction is required for glenoid fossa development during temporomandibular joint formation. Dev. Dyn., 240: 2466–2473. doi: 10.1002/dvdy.22748
- Issue published online: 19 OCT 2011
- Article first published online: 26 SEP 2011
- Manuscript Accepted: 2 SEP 2011
- NIH. Grant Number: DE17792
- TMJ formation;
- glenoid fossa development;
- tissue interaction;
The mammalian temporomandibular joint (TMJ) develops from two distinct mesenchymal condensations that grow toward each other and ossify through different mechanisms, with the glenoid fossa undergoing intramembranous ossification while the condyle being endochondral in origin. In this study, we used various genetically modified mouse models to investigate tissue interaction between the condyle and glenoid fossa during TMJ formation in mice. We report that either absence or dislocation of the condyle results in an arrested glenoid fossa development. In both cases, glenoid fossa development was initiated, but failed to sustain, and became regressed subsequently. However, condyle development appears to be independent upon the presence of the forming glenoid fossa. In addition, we show that substitution of condyle by Meckel's cartilage is able to sustain glenoid fossa development. These observations suggest that proper signals from the developing condyle or Meckel's cartilage are required to sustain the glenoid fossa development. Developmental Dynamics 240:2466–2473, 2011. © 2011 Wiley-Liss, Inc.
The temporomandibular joint (TMJ) is a unique synovial joint that is essential for movement and function of the jaw in mammals. It is composed of a fibrous capsule that encloses the mandibular condyle, the glenoid fossa of the temporal bone, and an articulating disc separating the two articular facets lined with fibrocartilage instead of hyaline cartilage (Sperber,2001). The glenoid fossa forms a deep concavity in the temporal bone, receiving the mandibular condyle to make the hinge of TMJ function. Associated with the TMJ are also the tendon of the pterygoid muscle and various surrounding ligaments. The TMJ disorders affect a large population in human beings, causing chronic myofacial pains and difficulty in chewing function. While the structures and functions of the TMJ have been well documented, TMJ development remains poorly understood in terms of underlying cellular and molecular mechanisms in contrast to a wealth of information regarding the formation of synovial joints in the developing limb.
The TMJ develops from two distinct and separate mesenchymal condensations, the temporal and condylar blastemas. These two blastemas grow toward each other and undergo different ossification processes, forming the glenoid fossa and condyle, respectively. Despite that it is classified as synovial joint, the formation of TMJ differs significantly from the limb joints that develop by the cleavage or segmentation within a single skeletal condensation. In mice, TMJ development begins at embryonic day (E) 13.5 as evidenced by the formation of the condylar condensation and the expression of Sox9 in the condensed mesenchyme (Gu et al.,2008; Shibata and Yokohama-Tamaki,2008). At E14.5, the condensation of the glenoid fossa becomes visible as a triangle structure and begins to express Runx2 at the dorsal–caudal aspect of the condylar condensation. Subsequently, these two mesenchymal condensations grow toward each other, narrowing down the distance between them and assuming their position at E16.5. At the same time, the articular disc also begins to form, which develops from a separate flat-shaped mesenchymal condensation located between the apex of condyle and the glenoid fossa (Frommer,1964). By E17.5, all the major components of the TMJ are formed. The glenoid fossa has encompassed the apex of the definite condyle, and the upper and lower synovial cavities, which are separated by the compact articular disc, become discernible at this stage (Gu et al.,2008).
The mandibular condyle undergoes endochondral ossification and represents a major growth site in the mandible contributing to the elongation of the mandibular ramus (Sibbermann and Frommer, 1972). Although the condylar cartilage is classified as the secondary cartilage (Beresford,1975), it grows through the chondrogenic processes similar to the primary cartilage and expresses many genes that are important for the primary cartilage growth and differentiation in the long bones (Fukada et al.,1999; Rabie and Hägg,2002; Kuboki et al.,2003; Ogawa et al.,2003; Watahiki et al.,2003; Tang et al.,2004; Shibukawa et al.,2007; Shibata et al.,2006, Shibata and Yokohama-Tamaki,2008; Purcell et al.,2009). For example, Sox9 is expressed in all chondroprogenitors and differentiated chondrocytes, and is required for cartilage formation, including both the primary and secondary cartilages including the condyle (Bi et al.,1999; Akiyama et al.,2002; Mori-Akiyama et al.,2003). While distinct sets of genes are expressed in the developing limb joints and the TMJ, respectively, several genes known to be critical for joint formation are indeed expressed in these two types of synovial joints (Purcell et al.,2009). Among these genes is Ihh, a signaling molecule required for regulation of chondrocyte proliferation, maturation, and ossification. Ihh deficient in embryonic mouse leads to digit joint ablation as well as defective TMJ lacking the articular disc (St-Jacques et al.,1999; Shibukawa et al.,2007). Mice deficient in Gli2, a downstream effector of the hedgehog signaling, exhibit similar TMJ defect (Purcell et al.,2009). In addition, ablation of Ihh in cartilages of neonatal mice produces dysplastic TMJ associated with partial disc ankylosis (Ochiai et al.,2010), mimicking the phenotype in Shox2 mutant mice in which Ihh expression is down-regulated in the developing condyle (Gu et al.,2008).
Studies on TMJ development have been primarily focusing on condyle growth and disc formation, little is known regarding the mechanistic basis of glenoid fossa formation. It is widely accepted that growth factor-mediated tissue interaction regulates growth, differentiation, and patterning of many organs during vertebrate embryogenesis. In the present study, we asked a fundamental question as whether there exists tissue interaction between the developing condyle and glenoid fossa during TMJ formation. We used various knockout/transgenic mouse models to examine the influence of the condyle on the development of the glenoid fossa. Our results show that absence or dislocation of the developing condyle arrested the development of the glenoid fossa. In addition, substitution of the condyle by Meckel's cartilage could sustain glenoid fossa development.
Inactivation of Sox9 in the Cranial Neural Crest Leads to Absent Condyle and Arrested Glenoid Fossa Development
It was reported previously that ablation of Sox9 in the cranial neural crest resulted in a complete elimination of cartilages and endochondral bones derived from the cranial neural crest cells; however, all the mesoderm-derived skeletal elements and intramembranous bones were essentially not affected (Mori-Akiyama et al.,2003). Although both the condyle and glenoid fossa derive from the cranial neural crest cells (Gu et al.,2008), the glenoid fossa condensation does not express Sox9, as shown in Figure 1, and develops through intramembranous ossification. We thus took advantage of the Wnt1Cre-mediated Sox9 gene inactivation approach to examine the development of the glenoid fossa in the absence of the condyle.
To create Wnt1Cre;Sox9f/f mice, we first generated Wnt1Cre;Sox9f/+ mice. While Sox9 haploinsufficiency results in perinatal lethality due to a cleft palate defect (Bi et al.,2001), it was reported that most Wnt1Cre;Sox9f/+ mice were viable and developed into fertile mice with normal lifespan, despite the formation of the cleft secondary palate in a small percentage of these mice (Mori-Akiyama et al.,2003). In our studies, we found that approximately 75% Wnt1Cre;Sox9f/+ mice developed a cleft palate defect. Those Wnt1Cre;Sox9f/+ mice that survived perinatal lethality appeared and behaved almost indistinguishable from their littermate controls. Skeletal preparation of Wnt1Cre;Sox9f/f mice at postnatal day 0 (P0) indeed demonstrated a complete absence of the cranial neural crest-derived cartilages and endochondral bones, including the condyle (Fig. 2A,B; and data not shown), as reported previously (Mori-Akiyama et al.,2003). In situ hybridization assays revealed absent Sox9 expression in the site where the condyle supposed to form in Wnt1Cre;Sox9f/f embryos (Fig. 1A–D), indicating successful inactivation of Sox9 in the cranial neural crest-derived cells.
We next examined the development of the glenoid fossa by histological analyses. At E14.5 when the glenoid fossa condensation became visible in the wild-type control, the glenoid fossa condensation with the similar shape was also observed in the mutant (Fig. 2C,D). This observation was further confirmed by Runx2 expression in the glenoid fossa condensation at this stage (Fig. 1E,F). At E15.5 and E16.5, the glenoid fossa grew toward the developing condyle and became ossified in the controls (Fig. 2E,G); however, the mutant glenoid fossa showed a slight delay in ossification and exhibited a regression in its lateral branch (Fig. 2F,H). At E18.5, while the glenoid fossa has encompassed the apex of the condyle in the wild-type control (Fig. 2I), the mutant glenoid fossa has completely regressed, leaving a residual bone of the lateral branch (Fig. 2J). This halted development and regression of the lateral branch of glenoid fossa in the mutants was further confirmed by the down-regulation of Runx2 in the glenoid fossa condensation at E15.5 (Fig. 1H). Of interest, a disc-like structure seemed to form at the apex side of the mandibular ramus in the mutant TMJ (Fig. 2J). However, this structure is less compact and the cell shape appears irregular, as compared to the control (Fig. 2K,L), suggesting the formation of an abnormal disc-like structure. These observations demonstrate that, in the absence of the condyle, the glenoid fossa is able to initiate its developmental process, as evidenced by the appearance of mesenchymal condensation and the expression of Runx2. However, the glenoid fossa fails to sustain its developmental program, and becomes regressed eventually.
Dislocation of the Developing Condyle Leads to Arrested Glenoid Fossa Development
To further confirm a requirement of the condyle for normal glenoid fossa development, we used a different transgenic mouse model in which the developing condyle becomes dislocated away from the developing glenoid fossa. In a separate study on Wnt signaling in palate development, we compounded the K14-Cre allele to an exon3 floxed Catnb allele (CatnbF(ex3); He et al.,2011). The CatnbF(ex3) allele, up Cre-mediated recombination, produces a stabilized β-catenin, leading to ectopic activity of the canonical Wnt signaling (Harada et al.,1999). In addition to severe defects in ectodermal appendages, K14Cre;CatnbF(ex3) mice exhibited a wide open mouth phenotype due to abnormal skin development (Närhi et al.,2008; Zhang et al.,2008). Skeletal preparations of K14Cre;CatnbF(ex3) mice revealed an abnormal position of the mandibular arch almost perpendicular to the maxilla (Fig. 3A,B), and formation of ectopic bone that connects the maxillae to the mandible (Fig. 3D). Coronal sections of the transgenic embryos at E14.5 demonstrated an anterior shift of the developing condyle (Fig. 3E–H). In the wild-type controls at this stage, the developing condyle, identified by Sox9 and Runx2 expression (Fig. 3E,G), is always associated with the developing glenoid fossa that undergoes condensation (Fig. 3E) and expresses Runx2 (Fig. 3G). In contrast, in the transgenic animals, a condensed primordial glenoid fossa was not present at the same section plane of the developing condyle (Fig. 3F), but was found, as evidenced by the expression of Runx2, at a caudal level where the condyle was not present (Fig. 3H). Similar to the phenotype observed in Wnt1Cre;Sox9f/f mice, at E16.5, the glenoid fossa failed to develop, lacking the lateral branch (Fig. 3J). These observations indicate that the presence of the developing condyle and its proximity are essential for the normal development of the glenoid fossa.
Of interest, despite the lack of the glenoid fossa, the condyle continued to develop at its dislocated position in K14Cre;CatnbF(ex3) mice. While slightly smaller in size, the transgenic condyle appeared to develop and differentiate normally, as assessed by histology and the expression of molecular markers, including Ihh, Col II, and Col X, when compared with its control counterpart (Fig. 4). However, the articular disc never formed (Fig. 4G,H). It appears that the presence of the glenoid fossa is not a prerequisite for the development and differentiation of the condyle, but is required for the formation of the articular disc, indicating a requirement for tissue interactions between the condyle and the glenoid fossa during normal TMJ formation.
Meckel's Cartilage Can Substitute for Condyle to Sustain Glenoid Fossa Development
Mice deficient in Noggin, a BMP antagonist, exhibit a spectrum of craniofacial defects, including a significantly enlarged Meckel's cartilage (McMahon et al.,1998; Bachiller et al.,2000; Stottmann et al.,2001; Anderson et al.,2006; He et al.,2010). A detailed analysis of the developing TMJ in Noggin mutants (Nog−/−) revealed a surprising phenotype. The rapid growth and expansion of Meckel's cartilage pushed the developing condyle away from its position in the proximity of the forming glenoid fossa condensation. Skeleton preparation of P0 wild-type and Nog−/− mice revealed a direct contact of Meckel's cartilage with the temporal bone (Fig. 5A,B). Histological analyses showed that, at E14.5, the enlarged Meckel's cartilage has assumed a position adjacent to the glenoid fossa condensation, evidenced by its morphology and the expression of Runx2 (Fig. 5D,F). At E16.5, the mutant glenoid fossa continued to grow and differentiate, encompassing the apex of the contacting Meckel's cartilage (Fig. 5H). At E17.5 when the articular disc formed in the wild-type control, no such disc-like structure was observed in the Meckel's cartilage-glenoid fossa complex, indicating a true TMJ did not form (Fig. 5I,J). Interestingly, the size of the glenoid fossa in the mutants appeared much larger than the controls, in accord with the larger size of Meckel's cartilage. These results demonstrate that Meckel's cartilage is able to substitute for the condyle to sustain glenoid fossa development, suggesting that Meckel's cartilage and the condyle produce similar secreted factors responsible for this process. However, it cannot replace the condyle for the formation of a normal TMJ, because a true joint with the articular disc did not form.
Tissue-tissue interaction plays an essential role in the embryonic development and organ formation. It involves a series of inductive and permissive interactions, regulating multiple processes during organogenesis such as determination, differentiation, and pattern formation. It is well documented that tissue interaction is mediated by diffusible signaling molecules also known as growth factors. The TMJ develops from two distinct mesenchymal condensations, the glenoid fossa and condylar blastemas, that undergo different ossification processes. Many studies have been conducted to examine gene expression and function during TMJ development, particularly in the development of the condyle and the articular disc. However, glenoid fossa development has received little attention, and its underlying cellular and molecular mechanisms are basically unknown. In the present study, we looked into the fundamental question as whether or not tissue interaction is required for glenoid fossa development during TMJ formation. We present evidence demonstrating that in the absence of the condyle, the glenoid fossa condensation fails to sustain its development, leading to the lack of a definite glenoid fossa structure. Similar developmental defect was also observed in the glenoid fossa when the developing condyle is dislocated. These results suggest that the presence and its proximity of the developing condyle are essential for the development of the glenoid fossa during TMJ formation. It appears that the initial development of the glenoid fossa is independent on the condyle, but its continuous development relies on interaction with the condyle. In Ihh, Gli2, or Shox2 mutants, while the condyle is hypoplastic, the glenoid fossa forms anyway (Shibukawa et al.,2007; Gu et al.,2008; Purcell et al.,2009), suggesting that even a hypoplastic condyle is able to support glenoid fossa development. Because the condensations of the condyle and glenoid fossa do not make direct contact, tissue interaction between them is apparently mediated by diffusible factor(s). On the other hand, condyle development and differentiation do not rely on the presence of the glenoid fossa. One possibility is that the developing mandibular ramus, which is in a close proximity to the developing condyle, and similar to the glenoid fossa, undergoes intramembranous ossification and may provide signals to sustain condyle development.
The articular disc forms from the mesenchymal cells located between the condyle and the glenoid fossa (Frommer,1964). In the absence of Ihh, a normal disc fails to form (Shibukawa et al.,2007). In WntCre;Sox9f/f mice, although a disc-like tissue forms, its abnormal structure and cell shape exclude it a true disc, suggesting a requirement of the condyle for the formation of a normal functional disc. In contrast, despite that the condyle develops and differentiates relatively normal in the absence of the glenoid fossa in K14Cre;CatnbF(ex3) mice, the articular disc fails to form. This is very likely due to the dislocation of the condyle away from the mesenchymal cells that form the articular disc.
Our results also show that Meckel's cartilage can substitute for the condyle to sustain glenoid fossa development in Nog−/− mice. However, a true joint with the TMJ characteristics never formed in Noggin mutants. In fact, Meckel's cartilage is considered the primary cartilage, while the condyle is classified as the secondary cartilage, characterized by its rapid differentiation from progenitor cells to hypertrophic chondrocytes and its preosteoblastic characteristics (Shibermann et al.,1987; Shibata et al.,1997; Miyake et al.,1997; Fukada et al.,1999). The difference in these two types of cartilage explains why Meckel's cartilage cannot substitute for the condyle for the formation of a normal TMJ. On the other hand, the condyle and Meckel's cartilage appears to produce similar diffusible signaling molecule(s) that acts to sustain glenoid fossa development. This is supported by the fact that the condyle does not only undergoes the similar chondrogenic processes, but also expresses many genes that are also present in the primary cartilage (Fukada et al.,1999; Rabie and Hägg,2002; Kuboki et al.,2003; Ogawa et al.,2003; Watahiki et al.,2003; Tang et al.,2004; Shibata et al.,2006; Shibukawa et al.,2007; Shibata and Yokohama-Tamaki,2008; Purcell et al.,2009). While the identity of the diffusible signaling molecule(s) is currently unknown, a survey for the candidate factors warrants future investigation and will be crucial for understanding the mechanisms of glenoid fossa development.
Generally speaking, TMJ development is an understudied field. Many studies in this field have been focusing on analyses of gene expression and function during development of the condyle and the disc. There is very little information of molecular and cellular regulatory mechanism available on glenoid fossa development. While several mutations have been shown to cause abnormal TMJ formation, most of the defects were confined to the condyle and disc (Shibata et al.,2004; Shibukawa et al.,2007; Gu et al.,2008; Purcell et al.,2009). The present study represents an initial effort to investigate how tissue-tissue interactions regulate the development of the TMJ.
The generation and genotyping of CatnbF(ex3), Nog+/−, Sox9f/f, K14-Cre, and Wnt1-Cre mice have been described previously (Danielian et al.,1998; McMahon et al.,1998; Harada et al.,1999; Akiyama et al.,2002; Andl et al.,2004). To generate Wnt1Cre;Sox9f/f mice, Wnt1-Cre mice were first crossed to Sox9f/f mice to generate Wnt1Cre;Sox9f/+ mice that were then further mated with Sox9f/f mice. Genotypes of mice or embryos were determined by PCR based genotyping method using genomic DNA extracted from tail or extra-embryonic membranes. Embryos collected from timed pregnant mice were fixed in 4% paraformaldehyde (PFA)/PBS at 4°C for overnight before being processed for histology or in situ hybridization.
Skeletal Preparation, Histology, and In Situ Hybridization
Skeletal staining was conducted by using Alcian blue /Alizarin red staining for nonmineralized cartilage and bone respectively, as described previously (Zhang et al.,2000). For histological examination, paraffin sections made at 10 μm were subjected to standard hematoxylin/eosin staining or Azon red/Anilin blue staining, according to the standard protocols (Presnell and Schreibman,1997). Nonradioactive section in situ hybridization was conducted as described before (St. Amand et al.,2000).
Y.P.C. was funded by a NIH grant.
- 2006. Endogenous bone morphogenetic protein antagonists regulate mammalian neural crest generation and survival. Dev Dyn 235: 2507–2520. , , , .
- 2004. Epithelial Bmpr1a regulates differentiation and proliferation in postnatal hair follicles and is essential for tooth development. Development 131: 2257–2268. , , , , , , , , , , , , , , .
- 2002. The transcription factor Sox9 has essential roles in successive steps of the chondrocyte differentiation pathway and is required for expression of Sox5 and Sox5. Genes Dev 16: 2813–1828. , , , , .
- 2000. The organizer factors Chordin and Noggin are required for mouse forebrain development. Nature 403: 658–661. , , , , , , , . , , .
- 1975. Schemes of zonation in the mandibular condyle. Am J Orthod 68: 189–195. .
- 1999. Sox9 is required for cartilage formation. Nat Genet 22: 85–89. , , , , .
- 2001. Haploinsufficiency of Sox9 results in defective cartilage primordial and premature skeletal mineralization. Proc Natl Acad Sci U S A 98: 6698–6703. , , , , , , .
- 1998. Modification of gene activity in mouse embryos in uterus by a tamoxifeninducible form of Cre recombinase. Curr Biol 8: 1323–1326. , , , , .
- 1964. Prenatal development of the mandibular joint in mice. Anat Rec 150: 449–461. .
- 1999. In situ hybridization study of type I, II, X collagens and aggrecan mRNAs in the developing condylar cartilage of fetal mouse mandible. J Anat 195: 321–329. , , , , .
- 2008. Shox2-deficiency leads to dysplasia and ankylosis of the temporomandibular joint in mice. Mech Dev 125: 729–742. , , , , .
- 1999. Intestinal polyposis in mice with a dominant stable mutation of the beta-catenin gene. EMBO J 18: 5931–5942. , , , , , , .
- 2010. Modulation of BMP signaling by Noggin is required for the maintenance of palatal epithelial integrity during palatogenesis. Dev Biol 347: 109–121. , , , , , , , .
- 2011. Epithelial Wnt/beta-catenin signaling regulates palatal shelf fusion through regulation of Tgfbeta3 expression. Dev Biol 350: 511–519. , , , , , , , , .
- 2003. Cbfa1/Runx2 gene expression in articular chondrocytes of the mice temporomandibular and knee joints in vivo. Arch Oral Biol 48: 519–525. , , , , , , , , .
- 1998. Noggin-mediated antagonism of BMP signaling is required for growth and patterning of the neural tube and somite. Genes Dev 12: 1438–1452. , , , , , .
- 1997. Stage-specific expression patterns of alkaline phosphatase during development of the first arch skeleton in inbred C57BL/6 mouse embryos. J Anat 190: 239–260. , , .
- 2003. Sox9 is required for determination of the chondrogenic cell lineage in the cranial neural crest. Proc Natl Acad Sci U S A 100: 9360–9365. , , , .
- 2008. Sustained epithelial β-catenin activity induces precocious hair development but disrupts hair follicle down-growth and hair shaft formation. Development 135: 1019–1028. , , , , , .
- 2010. Indian hedgehog roles in post-natal TMJ development and organization. J Dent Res 89: 349–354. , , , , , , , , , , , , .
- 2003. Localization and inhibitory effect of basic fibroblast growth factor on chondrogenesis in cultured mouse mandibular condyle. J Bone Miner Metab 21: 145–153. , , , , , , .
- 1997. Humason's Animal Tissue Techniques. 5th ed. Baltimore, MD: The Johns Hopkins University Press. , .
- 2009. Temporomandibular joint formation requires two distinct hedgehog-dependent steps. Proc Natl Acad Sci U S A 106: 18297–18302. , , , , , , , .
- 2002. Factors regulating mandibular condylar growth. Am J Orthod Dentofacial Orthop 122: 401–409. , .
- 1997. Immunohistochemistry of collagen types II and X, and enzyme-histochemistry of alkaline phosphatase in the developing condylar cartilage of the fetal mouse mandible. J Anat 191: 561–570. , , , .
- 2004. Runx2-deficient mice lack mandibular condylar cartilage and have deformed Meckel's cartilage. Anat Embryol (Berl) 208: 273–280. , , , , , , .
- 2006. An in situ hybridization study of Runx2, Osterix, and Sox9 at the onset of condylar cartilage formation in fetal mouse mandible. J Anat 208: 169–177. , , , , .
- 2008. An in situ hybridization study of Runx2, Osterix, and Sox9 in the anlagen of mouse mandibular condylar cartilage in the early stages of embryogenesis. J Anat 213: 274–283. , .
- 2007. Temporomandibular joint formation and condyle growth require Indian hedgehog signaling. Dev Dyn 236: 426–434. , , , , , , .
- 1972. The nature of endochondrial ossification in the mandibular condyle of the mouse. Anat Rec 172: 659–668. , .
- 1987. Further characterization of the extracellular matrix in the mandibular condyle in neonatal mice. J Anat 151: 169–188. , , , , , .
- 2001. Craniofacial development. Hamilton, Ontario: BC Decker Inc. .
- 2000. Antagonistisc signals between BMP4 and FGF8 define the expression of Pitx1 and Pitx2 in mouse tooth forming anlage. Dev Biol 217: 323–332. , , , , , , , .
- St. 1999. Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev 13: 2076–2086. , , .
- 2001. The BMP antagonists Chordin and Noggin have essential but redundant roles in mouse mandibular outgrowth. Dev Biol 240: 457–473. , , .
- 2004. Indian hedgehog: a mechanotransduction mediator in condylar cartilage. J Dent Res 83: 434–438. , , .
- 2003. Gene expression profiling of mouse condylar cartilage during mastication by means of laser microdissection and cDNA array. J Dent Res 83: 245–249. , , , , , .
- 2008. Activation of β-catenin signaling programs embryonic epidermis to hair follicle fate. Development 135: 2161–2172. , , , , , , , , , , , , .
- 2000. Targeted misexpression of constitutively active BMP receptor-IB causes bifurcation and duplication and posterior transformation of digit in mouse limb. Dev Biol 220: 154–167. , , , , , , .