Figures
Abstract
The central importance of BMP signaling in the development and homeostasis of synovial joint of appendicular skeleton has been well documented, but its role in the development of temporomandibular joint (TMJ), also classified as a synovial joint, remains completely unknown. In this study, we investigated the function of BMPRIA mediated signaling in TMJ development in mice by transgenic loss-of- and gain-of-function approaches. We found that BMPRIA is expressed in the cranial neural crest (CNC)-derived developing condyle and glenoid fossa, major components of TMJ, as well as the interzone mesenchymal cells. Wnt1-Cre mediated tissue specific inactivation of BmprIa in CNC lineage led to defective TMJ development, including failure of articular disc separation from a hypoplastic condyle, persistence of interzone cells, and failed formation of a functional fibrocartilage layer on the articular surface of the glenoid fossa and condyle, which could be at least partially attributed to the down-regulation of Ihh in the developing condyle and inhibition of apoptosis in the interzone. On the other hand, augmented BMPRIA signaling by Wnt1-Cre driven expression of a constitutively active form of BmprIa (caBmprIa) inhibited osteogenesis of the glenoid fossa and converted the condylar primordium from secondary cartilage to primary cartilage associated with ectopic activation of Smad-dependent pathway but inhibition of JNK pathway, leading to TMJ agenesis. Our results present unambiguous evidence for an essential role of finely tuned BMPRIA mediated signaling in TMJ development.
Citation: Gu S, Wu W, Liu C, Yang L, Sun C, Ye W, et al. (2014) BMPRIA Mediated Signaling Is Essential for Temporomandibular Joint Development in Mice. PLoS ONE 9(8): e101000. https://doi.org/10.1371/journal.pone.0101000
Editor: Songtao Shi, University of Southern California, United States of America
Received: April 7, 2014; Accepted: May 31, 2014; Published: August 5, 2014
Copyright: © 2014 Gu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files.
Funding: WW was supported by a fellowship from ZhongShan Hospital, FuDan University. LY was supported by a scholarship from the China Scholarship Council (No. 201208440191). This work was supported by the National Institutes of Health grants (R01 DE14044, DE17792) to YC. The funders had no role in this design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
As an evolutionary creature, the temporomandibular joint (TMJ) is a unique synovial joint generated only in mammals and is involved in food capture and intake, speech, as well as maturation of the facial contour [1]. It is made of specific components originated from the skull base and the low jaw including the glenoid fossa, condyle, articular disc, ligaments, and joint capsule. Although defined as a synovial joint, the developmental process of TMJ differs significantly from the joints of appendicular skeletons that are generated by cleavage or segmentation within a single skeletal condensation [2]. The TMJ develops from two distinct mesenchymal condensations, the glenoid fossa blastema that ossifies primarily through intramembranous bone formation, and the condylar blastema that undergoes endochondral ossification. These two primordia are initially separated widely by intervening mesenchyme that was thought to later contribute to the articular disc and capsule, as well as the synovial lining of joint cavity [3], [4] Subsequently, the condylar primordium, arising from the periosteum of the mandibular bone and therefore classified as secondary cartilage [5], [6], grows rapidly towards the glenoid fossa, and meanwhile, the articular disc forming from a condensed stripe flanking the apex of the condyle and subsequently separating from the latter, divides the interzone into the upper and lower joint cavities [7]. In mice, the mesenchymal condensation of condyle appears at embryonic day 13.5 (E13.5) and the glenoid fossa at E14.5 [8]. At E15.5, the shape of glenoid fossa and condyle has been established, and at E16.5, the upper synovial cavity becomes discernible with a disc beginning to form. Subsequently at E17.5, the lower joint cavity appears as a definite articular disc separates from the apex of the condyle. This intricate multi-step developmental process is regulated by intrinsic and extrinsic factors.
As for intrinsic constituents, the significance of genetic factors has attracted the attention of the field. Gene targeting studies have revealed essential roles for a number of transcription factors and growth factors in TMJ development, as evidenced by the absence of condylar cartilage in mice carrying mutations in Sox9, Runx2, or Tgfbr2, and by the abnormal development of mice carrying mutations in Shox2 or Spry1 and Spry2 [8]–[14]. Ihh, which plays a pivotal role in long bone development and digit joint formation [15], has been implicated in TMJ development by initiating the formation of articular disc and instructing the disc to undergo proper morphogenesis and to separate from the condyle, as well as in maintaining proper structure and function of the TMJ after it forms [16]–[18]. Lack of Ihh or its downstream effector Gli2 results in missing of a distinct disc in the TMJ [16], [17]. In addition, extrinsic factors such as biomechanical force also contribute to TMJ development [2].
Bone morphogenetic proteins (BMPs) exert diverse biological functions during development and postnatal homeostasis. BMP signals are transduced into cells through the type I and type II transmembrane serine/threonine kinase complexes by activating Smad-dependent (canonical) pathway, as well as Smad-independent (non-canonical) pathway via activation of the mitogen-activated protein kinase (MAPK) signaling [19]. Extensive studies have established critical roles for BMP signaling in skeletal development and joint morphogenesis, particularly in joint formation of long bones. Joint formation in the appendage skeletons begins with the formation of a condensed cell stripe known as interzone in the developing cartilage template [20]. Cells in the edges of the interzone give rise to the articular cartilage that covers the ends of the adjacent skeletal elements, while cells in the middle of the interzone undergo programmed cell death, leading to physical separation of the contiguous cartilage element and formation of joint cavity [21]. Several members of BMP family, including Bmp2, Bmp4, Gdf5, Gdf6, Gdf7, are expressed in the interzone along with BMP antagonists Chondin and Noggin [22]–[26]. Mice carrying mutations in Gdf5 or Gdf6 exhibit lack of joint formation at specific locations [25], [27], demonstrating a direct action of BMP signaling in joint morphogenesis. On the other hand, elevated BMP signaling also blocks joint formation, as manifested by failure in joint formation in the limbs of Noggin mutant mice [28]. These loss-of- and gain-of-function studies indicate an essential role for tightly regulated BMP activity in synovial joint formation. Furthermore, BMP signaling is also involved in postnatal joint homeostasis and tissue remodeling [26], [29].
Being one of the two primary BMP type I receptors (BMPRIA and BMPRIB), BMPRIA plays crucial roles in skeleton patterning and development. In developing limb skeletons, BmprIa is expressed in the joint interzone, perichondrium, periarticular cartilage, and hypertrophic chondrocytes [26], [30]–[32]. Tissue specific deletion of BmprIa in cartilage lineage leads to chondrodysplasia attributed at least partially to the defective cell proliferation as well as the premature hypertrophy of chondrocytes associated with down-regulation of Ihh [33], [34]. While joint defect was not identified in mice carrying cartilage specific inactivation of BmprIa, mice carrying tissue specific inactivation of BmprIa in the interzone indeed exhibited missed joints in the ankles [26], indicating a requirement of BMPRIA mediated signaling in joint formation.
Despite a wealth of documents on BMP signaling in bone and joint formation of appendicular skeletons, little is known about its role in TMJ development. Thus far, the only line of evidence implicating a possible involvement of BMP signaling in TMJ formation is that Bmp2 and Bmp7 were found to be expressed in the developing condyle [17], [35]. To gain an insight into BMP signaling in TMJ development, in this study, we used transgenic loss-of- and gain-of-function approaches to investigate the function of BMPRIA mediated signaling in TMJ development.
Materials and Methods
Ethics statement
Experiments that involved use of animals in this study was approved by the Institutional Animal Care and Use Committee (IACUC) of Tulane University (protocol number: 0367R) and was in strict accordance with the recommendations in the Guide for Care and Use of Laboratory Animals of the National Institutes of Health.
Animal and sample collection
The generation and identification of transgenic and gene-targeted animals, including Wnt1-Cre, BmprIaf/f, and pMes-caBmprIa that carries a conditional constitutively active form (with Gln203 to Asp change) of BmprIa transgenic allele, have been described previously [15], [31], [36], [37]. Wnt1-Cre;BmprI f/f embryos were obtained by crossing Wnt1-Cre;BmprIaf/+ mice with BmprIaf/f line. Wnt1-Cre;pMes-caBmprIa embryos were generated by mating Wnt1-Cre mice with pMes-caBmprIa transgenic line. Ihh null mutants were harvested from intercross of Ihh heterozygous mice. Embryos with BmprIa deficiency in their neural crest cells (Wnt1-Cre;BmprIaf/f) die around E12.5 due to norepinephrine depletion [38], [39]. To prevent early embryonic lethality, pregnant females were administrated with the β-adrenergic receptor agonist from 7.5 postcoitum (dpc) on by supplementing drinking water of dams with 200 µg/ml isoproterenol, which would allow Wnt1-Cre;BmprIaf/f embryos to survive to term [40], [41]. Embryos were collected from the timed pregnant females, and head samples were dissected in ice cold PBS, fixed individually in 4% paraformaldehyde (PFA) or z-fix (ANATECH Ltd; #170) overnight at 4°C, and tail clip from each embryo was used for PCR-based genotyping, respectively. Mutant and control heads were positioned for serial coronal sections through the TMJ. Comparable sections through the apex of the condyle were picked up for histological, in situ hybridization, immunostaining, and cell apoptosis analyses.
Histology, in situ hybridization, and immunohistochemistry, and Tunel assay
For histological study, paraffin sections were made at 6 µm and subjected for standard Hematoxylin/Eosin staining or Azoncarmine G/Aniline blue staining, as described [42]. Five mutant samples at each stage examined were used to ensure consistency of the phenotype. For in situ hybridization analyses, sections were cut at 10 µm and pretreated with proteinase K and hybridized with appropriate probes. Transcripts were detected by color reaction using BM purple (Roche) as described previously [43]. For immunohistochemical staining, frozen sections, made at 8 µm, were blocked with 4% goat serum and then incubated with primary antibodies against BMPRIA (Abcam; ab38560), Lubricin (Santa Cruz; sc-9854), pSmad1/5, pJNK, pERK, and p-p38 (from Cell Signaling; #9516, #9255, #4370, and #9211), respectively, at 4°C overnight. After washing, samples were incubated with secondary antibodies (Alexa Fluor488 goat anti-rabbit IgG from Invitrogen; #A-11034), counterstained with DAPI, and visualized under fluorescent microscope. Negative controls without primary antibodies were included in parallel. At least three samples of each genotype were used for histology, in situ hybridization, and immunohistochemistry analyses. Terminal deoxynucleotidyl transferase dUTP nick end labeling (Tunel) assay was applied to detect apoptotic cells using In Situ Cell Death Detection Kit (Roche), as described previously [44], [45]. Three samples of each genotype were subjected to Tunel assays. Tunel-positive cells within the interzone were counted and presented as percentage of total cells within arbitrarily defined areas. Student's t-test was used to determine the significance of difference between wild type controls and mutants, and the results were presented as P value.
Results
Inactivation of BmprIa in CNC lineage leads to defective TMJ formation
To investigate the role of BmprIa in TMJ development, we began with examination of BMPRIA expression by immunohistochemistry. At E14.5 when both the primordial condyle and glenoid fossa become discernible, BMPRIA was found present in the developing condyle and glenoid fossa as well as the interzone, with a relatively low level in the condylar cartilage (Fig. 1A). At E15.5, BMPRIA expression retained in the condyle, glenoid fossa, and interzone, with an increased level in chondrocytes undergoing hypertrophy (Fig. 1C). This expression pattern is similar to that in the developing limb skeleton including the joints [26], [30]–[32], suggesting a role for BMPRIA in TMJ morphogenesis. Since the condyle, glenoid fossa, and interzone cells are all derived from CNCs [8], we inactivated BmprIa in CNC lineage using the Wnt1-Cre transgenic allele. Immunohistochemistry confirmed the absence of BMPRIA in the developing TMJ of Wnt1-Cre;BmprIaf/f mice (Fig. 1B).
(A–D) Immunohistochemistry shows expression of BMPRIA in the condylar cartilage, interzone, and glenoid fossa of E14.5 (A) and E15.5 (C) wild type animals, but a lack of staining on the Wnt1-Cre;BmprIaf/f TMJ (B) and on the negative control (D). Red arrows point to positive staining in the interzone, and white arrow points to the hypertrophic region where strong expression is detected. Abbreviation: C, condyle; G, glenoid fossa; IZ, interzone. Scale bar = 100 µm.
Histological analyses showed that the initial condensation of the condylar anlage in mutant mice appeared comparable to littermate controls at E13.5 (Fig. 2A, 2B). At E15.5, the morphology of the glenoid fossa did not exhibit an obvious difference between mutants and controls, but the size of the mutant condyle was reduced as compared to controls (Fig. 2C, 2D). At E18.5, the control TMJ displayed distinct structures, including a definite articular disc, the upper and lower synovial cavities, and the fibrocartilage/synovial membrane on the articular surface of the glenoid fossa and condyle (Fig. 2E, 2G). However, at this stage, the mutant TMJ exhibited a number of severe defects, including a hypoplastic condyle, lack of a definite disc, failed formation of an upper joint cavity evidenced by the existence of loose connective tissue in the interzone, as well as the absence of the fibrocartilage/synovial membrane of the glenoid fossa (Fig. 2F). Close examination of the mutant condyle at E18.5 revealed the formation of a disc-like structure that failed to separate from the apex of the condyle, leading to an absence of the lower joint cavity (Fig. 2F, 2H). The lack of a synovial joint cavity in the Wnt1-Cre;BmprIaf/f TMJ was further confirmed by the absent expression of Lubricin, a key component of joint fluids [46], [47], as compared to its abundant expression in controls (Fig. 2I, 2J).
(A–H) H&E staining shows histology of the developing TMJ of wild type controls (A, C, E, G) and mutants (B, D, F, H). Note that the initial condensation of the condylar anlagen at E13.5 (A, B) and the morphology of the glenoid fossa at E15.5 (C, D) appear comparable between the controls and mutants. However, the size of the mutant condyle is reduced at E15.5 (D). At E18.5, distinct structures including a definite disc, the upper and lower joint cavities, and the articular surface of the glenod fossa are well present in the control TMJ (E, G). However, in mutants, while a disc-like compact layer could be identified closely associated with the apex of the condyle, it fails to separate to form a distinct disc. In addition, the interzone cells persist, and a fibrocartilage layer fails to form on the articular surface of the glenoid fossa (F, H). (I, J) Immunnohistochemistry reveals expression of Lubricin in the synovial membrane of the control TMJ (I), and the complete absence of Lubricin in the mutant TMJ (J). Arrows in (A, B) point to the condylar condensation, and in (G, H) point to the disc. Arrowheads in (E, F) point to the disc. Red arrowhead points to the articular surface in (G) and the synovial membrane in (I). Abbreviation: C, condyle; G, glenoid fossa; M, Meckel's cartilage; IZ, interzone; LC, lower cavity; UC, upper cavity; LPM, lateral pterygoid muscle. Scale bar = 50 µm.
Delayed chondrocyte maturation in the Wnt1-Cre;BmprIaf/f condyle
Despite being a secondary cartilage, the growth of condylar cartilage takes the similar chondrogenesis and endochondral ossification process as that in long bone formation. Since BMPRIA mediated signaling is known to regulate primary cartilage differentiation [33], [34], we set to examine chondrogenic differentiation process in the Wnt1-Cre;BmprIaf/f condyle. In the developing condyle, mesenchymal condensation appears at E13.5, and chondrogenic differentiation occurs at E14.5, and hypertrophy initiates at E15.5 [8]. We found that the timing of initial condensation of the condylar anlagen, as indicated by the expression of Sox9, and chondrogenic differentiation, determined by Col II expression, was comparable between wild type controls and Wnt1-Cre;BmprIaf/f mice (Fig. 3A–D). However, the mutant condyle exhibited a delayed terminal hypertrophy of chondrocytes, as assessed by the delayed Col X expression (Fig. 3E–H), and the longer distance between the apex and the beginning of hypertrophic zone in the mutant condyle at E18.5 (Fig. 3G, 3H). This phenotype differs from that in long bones where inactivation of BmprIa in condrocytes causes premature chondrocyte differentiation [34], likely due to different properties of primary v.s secondary cartilage.
(A, B) In situ hybridization shows Sox9 expression in the condylar condensation (arrow) of wild type (A) and mutant (B) at E13.5. (C, D) Col II expression exhibits comparable pattern in the condyle of wild type (C) and mutant (D) at E14.5. (E–H) In situ hybridization reveals Col X expression in the wild type condyle at E15.5 (E) and E18.5 (G). However, Col X expression is not detected in the mutant condyle at E15.5 (F), but is seen at E18.5 (H). The distance between the apex of the condyle and the beginning of hypertrophic zone is longer than that in the control condyle (G, H). Abbreviation: M, Meckel's cartilage. Scale bar = 100 µm.
Down-regulation of Ihh and inhibition of apoptosis in the Wnt1-Cre;BmprIaf/f TMJ
The similar TMJ phenotype between Wnt1-Cre;BmprIaf/f mice and the mice carrying mutations in Ihh or its downstream effectors [16], [17], particularly the failure of disc separation, persistent interzone cells, and lack of fibrocartilaginous articular surface layer of the glenoid fossa in both mutants (Fig. 2F, 2H; Fig. 4B, 4D), and the overlapped expression pattern of BmprIa with Ihh in the developing condyle (Fig. 1) [8], [17], prompted us to examine Ihh expression in the developing condyle of Wnt1-Cre;BmprIaf/f mice. In situ hybridization assay revealed a dramatic down-regulation of Ihh expression in the developing condyle of the mutants at E14.5 and E15.5, as compared to controls (Fig. 4G–J), consistent with the role of BMPRIA as a positive regulator of Ihh expression [34], [48]. BmprIa thus likely acts through Ihh to regulate TMJ formation.
(A–D) H&E staining reveals TMJ defects in the Ihh mutant TMJ (B, D), as compared to control (A, C) at P0. In mutant, a disc-like structure (arrowhead) forms but fails to separate from the apex of the condyle, the fibrocartilage layer fails to form on the articular surface of the glenoid fossa, and the interzone cells persist (B, D), as compared to the formation of distinct TMJ structures, including the disc (arrowhead), the fibrocartilginous articular surface, and the clear upper joint cavity, in controls (A, C). (E, E′, F, F′) Immunohistochemistry reveals significantly down-regulated BMPRIA expression in the condylar cartilage and interzone of Ihh mutant at E14.5 (F, F′), as compared to littermate control (E, E′). (G–J) In situ hybridization shows a dramatic down-regulation of Ihh in the condylar cartilage of Wnt1-Cre;BmprIaf/f mice at both E14.5 (H) and E15.5 (J), as compared to controls (G, I). Tunel assay reveals numerous apoptotic cells (arrowheads) in the interzone of wild type control at E15.5 (K, K′), but very few apoptotic cells in the interzone of the Wnt1-Cre;BmprIaf/fTMJ at the same age (L, L′). In contrast, some apoptotic cells (arrowheads) were observed in the mutant condyle (L). (M) Comparison of the percentage of apoptotic cells in the interzone of controls and mutants. Standard deviation values were indicated as the error bars, and the Student's t-test was used to determine the significance of difference between control and mutant, as presented as P value. Abbreviation: C, condyle; G, glenoid fossa; M, Meckel's cartilage; IZ, interzone; UC, upper cavity. Scale bar = 100 µm.
Because BmprIa is also expressed in the interzone mesenchymal cells that contribute to the articular disc and synovial membrane of the TMJ, the lack of fibrocartilage layer of the glenoid fossa and the persistence of interzone cells in the Wnt1-Cre;BmprIaf/f TMJ could be attributed either directly to the lack of BmprIa in these cells or indirectly to the significantly reduced level of Ihh in the condyle. To distinguish these alternatives, we conducted immunohistochemistry to examine BMPRIA expression in the Ihh-/- TMJ at E14.5. We found that although BMPRIA expression appeared comparable in the glenoid fossa of controls and mutants, its expression level was significantly reduced in the interzone and the condyle of the Ihh mutant TMJ (Fig. 4E–4F′).
While the mechanism of cavitation in TMJ development remains to be addressed, programmed cell death in the interzone is regarded as a critical cellular mechanism for joint cavity formation in long bones [21]. Since BMPRIA mediated signaling is required for programmed cell death in the limb, particularly in the interdigital region [26], [49], we wondered if the persistence of the loose connective tissue in the interzone of the Wnt1-Cre;BmprIaf/f TMJ is a consequence of reduced level of apoptosis. Tunel assay indeed revealed abundant apoptotic cells specifically in the interzone of the control TMJ at E15.5 (Fig. 4K, 4K′). In contrast, Tunel assay detected a significantly reduced level of apoptotic cells in the interzone as well as some apoptotic cells in the condyle of the mutant TMJ (Fig. 4L, 4L′, 4M). These observations suggest that similar to joint cavity formation in long bones, programmed cell death in the interzone also represents a critical cellular mechanism for joint cavitation during TMJ formation.
Augmented BMPRIA signaling in CNC lineage leads to TMJ agenesis
To further investigate the role of BMPRIA signaling in TMJ morphogenesis, we took a gain-of-function approach by transgenic expression of a constitutively active form of BmprIa (pMes-caBmprIa) [37] in CNC cells using the Wnt1-Cre allele. In situ hybridization revealed expression of BmprIa in the condensing condylar blastema, the forming site of glenoid fossa, and cells between them, but not in Meckel's cartilage of wild type embryo at E13.5 (Fig. 5A). In Wnt1-Cre;pMes-caBmprIa mice at the same stage, strong and wide spread expression of BmprIa was found in the TMJ forming region and its surrounding tissues including Meckel's cartilage, indicating successful transgenic expression of BmprIa in CNC lineage (Fig. 5B). We and others have shown previously that elevated BMPRIA mediated signaling in CNC cells leads to a spectrum of craniofacial bone defects, including cleft secondary palate, ectopic cartilage formation, and craniosynostosis [50], [51]. Histological analysis of the developing TMJ of Wnt1-Cre;pMes-caBmprIa mice identified unique TMJ developmental defects. Although the condensation of the condylar blastema occurred similarly to controls at E13.5 (Fig. 5C, 5D), the size of the condylar cartilage in transgenic animals became noticeably enlarged and the entire condylar cartilage appeared to become hypertrophic at E15.5, as compared to controls (Fig. 5E, 5F). Furthermore, unlike in controls that osteogenesis has begun in the glenoid fossa at this stage, the transgenic glenoid fossa failed to take osteogenic differentiation. At E17.5, the glenoid fossa became degenerated in transgenic animal (Fig. 5H). The failure of osteogenic differentiation in the glenoid fossa was further confirmed by the lack of bone formation in the glenoid fossa, assessed by Azocarmine G/Aniline blue staining, and by the absent expression of Runx2, a molecular marker for osteoblasts (Fig. 5I–L). Additionally, despite its expression in the mandibular bone, Runx2 expression is also down-regulated in the condylar cartilage of Wnt1-Cre;pMes-caBmprIa mice (Fig. 5L). By E18.5, both the condyle and glenoid fossa degenerated and became unrecognizable (data not shown).
(A, B) In situ hybridization shows expression of BmprIa in the condylar condensation, the future glenoid fossa forming site, and the interzone region, but not in Meckel's cartilage of an E13.5 wild type embryo (A), and an enhanced BmprIa expression in the TMJ forming site as well as the surrounding tissues including Meckel's cartilage in an E13.5 Wnt1-Cre;pMes-cBmprIa embryo (B). (C–H) H&E staining reveals initial condensation of condylar anlagen in control and transgenic animals at E13.5 (C, D), growth and differentiation into primary cartilage of the condylar cartilage and lack of osteogenesis in the glenoid fossa in the transgenic TMJ at E15.5 (F) and E17.5 (H). (I, J) Azocarmine G/Aniline blue staining reveals glenoid fossa degeneration, evidenced by lack of bone formation, in transgenic mouse (J), as compared to control (I). (K, L) In situ hybridization assay shows expression of Runx2 in the forming gelnoid fossa, perichondral region of the condyle, and mandibular bone of an E15.5 wild type control (K), but an absent expression of Runx2 in the glenoid fossa and a reduced expression in the condylar cartilage of an E15.5 transgenic animal (L). Note retention of Runx2 expression in the mandibular bone of transgenic mouse (L). Asterisk in (H, J) indicates the site of glenoid fossa degeneration. Open arrowheads in (K, L) point to Runx2 expression sites in the glenoid fossa. Abbreviation: C, condylar cartilage; G, glenoid fossa; M, Meckel's cartilage; EC, ectopic cartilage; LMP, lateral pterygoid muscle; Man, mandibular bone. Scale bar = 100 µm.
Enhanced BMPRIA signaling converts the condylar primordium from secondary cartilage to primary cartilage by ectopic activation of canonical signaling and inhibition of JNK signaling
As a secondary cartilage, the condylar cartilage expresses type I collagen (Col I), making it distinct from the primary cartilage [52]. Because of its aberrant differentiation, we wondered if the condylar cartilaginous element of Wnt1-Cre;pMes-caBmprIa mice retained its secondary cartilage characteristics. In situ hybridization assay revealed Col I expression in the control condylar cartilage and mandibular bone, but the absence of Col I in the transgenic condylar cartilaginous element despite its expression in the mandibular bone at E14.5 (Fig. 6A, 6B). However, the expression of Col II and Col X in the transgenic condylar cartilage confirmed its cartilage fate (Fig. 6C–F). We thus conclude that the Wnt1-Cre;pMes-caBmprIa condylar primordium adopts a fate of primary cartilage in response to an augmented BMPRIA-mediated signaling. Moreover, Tunel assay revealed an extensive apoptotic event in the transgenic condylar cartilage, beginning at E15.5, as compared to the lack of apoptosis in the control condyle at the same stage (Fig. 6G, 6H), which apparently contributes to the degeneration and disappearance of the condylar cartilage in transgenic animals.
(A–F) In situ hybridization detects Col I expression in the condylar cartilage and mandibular bone of an E14.5 control embryo (A), and in the mandibular bone of an E14.5 Wnt1-Cre,pMes-caBmprIa mice (B). However, Col I expression is not detected in the condylar cartilage of transgenic animal (B). Col II (C, D) and Col X (E, F) expression is observed in the condylar cartilage of control (C, E) and transgenic embryo (D, F) at E15.5. (G, H) Tunel assay reveals numerous apoptotic cells in the interzone but not in the condyle of the E15.5 wild type TMJ, but extensive cell death in the condylar cartilage of E15.5 transgenic embryo (H). Arrow in (F) points to Col X expression domain, and arrowheads in (G) point to apoptotic cells. Abbreviation: C, condylar cartilage; M, Meckel's cartilage; IZ, interzone. Scale bar = 100 µm.
We have shown previously that the expression of caBmprIa in CNC lineage induces ectopic activation of Smad1/5/8 signaling as well as p38 signaling in the developing palatal shelves [51]. We therefore set to examine alterations in BMP canonical and non-canonical signaling pathways in the condylar cartilage of Wnt1-Cre;pMes-caBmprIa mice by immunohistochemistry. Interestingly, we detected no activation of Smad-dependent as well as p38 and Erk1/2 pathways in the control condyle, as assessed by the lack of pSmad1/5, p-p38, and p-Erk1/2, but observed activity of p-JNK signaling (Fig. 7A, 7C,7E, 7G). In contrast, the condylar cartilage of Wnt1-Cre;pMes-caBmprIa mice exhibited an ectopic activation of pSmad1/5, but an absence of pJNK signaling, along with unaltered p38 and pEek1/2 pathways (Fig. 7B, 7D; 7F, 7H). These observations indicate that the switch between BMP canonical and non-canonical signaling pathways likely underlies the fate conversion from the secondary to primary cartilage, with the Smad-dependent signaling favoring the primary cartilage fate.
Immunohistochemistry shows absent pSmad1/5, p-p38, and pERK, but the presence of pJNK in the E14.5 control condyle (A, C, E, G), and the presence of pSmad1/5, but absent pJNK as well as p-p38 and pERK in the transgenic condylar cartilage (B, D, F, H). Abbreviation: C, condylar cartilage; G, glenoid fossa; M, Meckel's cartilage. Scale bar = 100 µm.
Discussion
Compared to synovial joint formation in the appendicular skeletons, TMJ development and the underlying molecular mechanisms are relatively under-studied. While the critical roles of BMP signaling in long bone joint development and homeostasis have been well documented [21], [26], [29], its role in TMJ formation remained completely unknown. In this study, we present evidence that BMPRIA mediated signaling is essential for TMJ morphogenesis, and overly activated BMPRIA signaling is detrimental to TMJ formation. Our results also reveal apoptosis in the interzone as a potential cellular mechanism for cavitation of the TMJ, similar to that in long bone joint formation.
A BMPRIA-Ihh positive regulatory pathway regulates TMJ development
It has been well established that BMP and Ihh signaling interact to regulate chondrocyte proliferation and hypertrophic differentiation [34], [53]–[55]. In the developing limb, BMP signaling, particularly the BMPRIA mediated pathway, positively regulates Ihh expression that could also activate in the perichondrium the expression of several BMP ligands, forming a BMP-Ihh positive feedback loop [34], [53], [54]. Although there is no evidence for an interaction of BMP and Ihh signaling in joint development, the fact that several BMP ligands and receptor are expressed in the interzone and that mutations in either Noggin or Ihh lead to joint defects including joint ablation in limbs implies the existence of such interaction [15], [21], [28]. Indeed, in the current study, we found that the ablation of BmprIa in CNC lineage produces TMJ defects resembling that in Ihh mutant. In both mutants, a functional TMJ failed to form, evidenced by the absent Lubricin expression (Fig. 2) [16]. In addition, both mutants displayed a lack of a distinct articular disc due to failed disc separation from the condyle, persistence of the interzone cells, as well as absent synovial membrane on the articular surface of the glenoid fossa. Consistent with Ihh function in disc formation and separation during TMJ morphogenesis [16], [17], we found a dramatic down-regulation of Ihh expression in the developing condyle of Wnt1-Cre;BmprIaf/f mice. This result also indicates that similar to its role in developing appendicular skeletons, BMPRIA mediated signaling also acts as a positive regulator of Ihh expression in the condylar cartilage. On the other hand, BmprIa expression was significantly reduced in the condyle and interzone of the Ihh-/- TMJ, suggesting the existence of a BMPRIA-Ihh positive feedback loop in the developing TMJ. However, the delayed hypertrophic differentiation observed in the condylar cartilage of Wnt1-Cre;BmprIaf/f mice appears to be opposite to the premature hypertrophic differentiation defect seen in the Ihh-/- condyle as well as in the long bones of mice carrying BmprIa deletion in chondrocyte lineage [16], [34]. Although the underlying mechanism is currently unknown, the discrepancy would likely be attributed to the residual Ihh expression in the Wnt1-Cre;BmprIaf/f condyle as well as the condyle's property as secondary cartilage.
Nevertheless, based on above mentioned observations and the established roles for BMP and Ihh signaling in limb development, we propose a model to summarize the function of the BMPRIA-Ihh regulatory pathway in regulating distinct steps during TMJ morphogenesis (Fig. 8). In this model, BMPRIA and Ihh regulate the expression of each other to coordinate chondrocyte proliferation and differentiation in the developing condyle. Meanwhile, Ihh, produced in the condyle, diffuses into the interzone to regulate disc separation and to maintain the expression of BMPRIA that in turn acts in a cell autonomous manner to regulate synovial membrane formation and to trigger apoptosis.
Abbreviation: C, condyle; D, disc; G, glenoid fossa; IZ, interzone.
Augmented BMPRIA signaling in CNCs converts the secondary cartilage of the condylar primordium to primary cartilage
Despite being a secondary cartilage, the condyle shares many similarities with primary cartilage in development, including the expression of genes known to be important for cartilage growth and differentiation such as Bmp2, Bmp7, Sox9, Runx2, Osterix, Ihh, Pthrp, Vegf, Col II and Col X [8], [16], [17], [35], [56], [57]. However, the condyle also differs from primary cartilage by its expression of Col I and Col II simultaneously and its capability of differentiating into either chondrocytes or osteoblasts [52]. Although the underlying mechanism for fate determination of primary v.s. secondary cartilage remains elusive, our results that augmented BMPRIA signaling is able to convert the condylar cartilage to a primary cartilage, evidenced by the lack of Col I expression, implicate BMP signaling in such fate decision. BmprIa is expressed in the cartilage condensations of both long bone and condyle (this study) [33], suggesting its role in the fate decision of both primary and secondary cartilages. It appears that a tightly tuned BMPRIA signaling is essential for fate determination of secondary cartilage. Accompanied with this fate conversion is the switch of the downstream BMP signaling pathway from the non-canonical JNK signaling to the Smad-dependent pathway in the condylar cartilage, suggesting that higher activity of BMPRIA signaling preferentially activates the Smad-dependent pathway, which favors primary cartilage formation. Indeed, the lack of Smad-dependent signaling in the condyle (this study) and the strong expression of pSmad1/5/8 in the limb cartilage condensation [33] further support this notion. Furthermore, consistent with the role of BMPRIA signaling in apoptosis in the developing limb [26], [32], overly activated BMPRIA signaling causes extensive apoptosis in the condylar cartilage and leads to condylar cartilage degeneration, indicating a detrimental effect on chondrocyte survival.
Despite an essential role for BmprIa in chondrogenesis and endochondral bone formation, in our current study, we found that inactivation of BmprIa did not affect glenoid fossa osteogenesis, suggesting that BmprIa may not be essential for intramembranous bone formation. However, elevated BMPRIA signaling instead inhibits osteogenesis in the glenoid fossa. Thus, although BMP signaling is generally accepted as a positive regulator of osteogenesis, elevated BMP signaling could have an opposite effect, depending on the tissue and cell types. Since a normal developing condyle is required to sustain the development of the glenoid fossa [9], the degeneration of the glenoid fossa in Wnt1Cre;pMes-caBmprIa mice could be the consequence of failed osteogenesis, or result from an abnormal condylar cartilage with altered property, or both.
Apoptosis as a cellular mechanism of TMJ cavitation
Cell death in the middle of the interzone is considered the cellular mechanism for physical separation of the contiguous cartilage elements during joint formation in long bones [21], [58]–[61]. In the developing TMJ, although the primordial condyle and glenoid fossa form independently and become approximately through condylar growth, disappearance of the interzone cells is necessary for the formation of a joint cavity. The interzone mesenchymal cells are believed to contribute to the articular disc, capsule, and the synovial membrane of the joint cavity [3], [4]. However, if apoptosis occurs in the interzone of the TMJ remains arguably. It was reported previously that in the rat TMJ at late developmental stage, apoptotic cells were found only at the subsurface of the condyle and in the region at which the lateral pterygoid muscle attached to the condyle, suggesting that apoptosis may be associated with the lower joint cavity formation of the TMJ [62]. However, in our studies, we found extensive cell death in the interzone of the control TMJ at E15.5, right before the upper cavity becomes discernibly at E16.5. However, in the Wnt1-Cre;BmprIaf/f TMJ, such extensive apoptosis was not observed, consistent with the pro-apoptotic role of BMPRIA mediated signaling. The discrepancy between our results and that by Matsuda and colleagues [62] could be attributed to the stage difference. Nevertheless, the lack of apoptosis appears to contribute to the persistence of interzone cells in the mutant TMJ. In addition, BMPRIA signaling is also required for organization of some interzone cells to become synovial lining layer, which could also contribute to the cavitation of the TMJ. We thus propose that the upper joint cavity of the TMJ is formed by organization of the interzone cells into synovial lining layer and capsule, and by removal of excessive cells via apoptosis.
In conclusion, our studies using transgenic loss-of- and gain-of-function approaches reveal the importance of BMPRIA mediated signaling in TMJ morphogenesis and establish a BMPRIA-Ihh positive regulatory pathway in controlling disc separation, synovial membrane formation, as well as joint cavity formation.
Acknowledgments
The authors thank Dr. Yuji Mishina of the University of Michigan for sharing BmprIaf/f mice.
Author Contributions
Conceived and designed the experiments: YC SG WW. Performed the experiments: SG WW CL LY CS WY XL JC. Analyzed the data: SG WW YC. Contributed reagents/materials/analysis tools: JC FL. Contributed to the writing of the manuscript: SG WW FL YC.
References
- 1. Kermack KA (1972) The origin of mammals and the evolution of the temporomandibular joint. Proc R Soc Med 65: 389–392.
- 2.
Gu S, Chen Y (2013) Temporomandibular joint development. In: Huang, G.T.-J., Thesleff, I. (Eds.), Stem Cells in Craniofacial Development and Regeneration. Wiley-Blackwell; John Wily & Sons, Inc. p71–85.
- 3.
Dixon AD (1997) Formation of the cranial base and craniofacial joints. In: Dixon, A.D., Hoyete, D.A.N., Rönning, O. (Eds.), Fundamentals of Craniofacial Growth. CRC Press LLC. p99–136.
- 4.
Sperber GH (2001) Craniofacial Development. Ontario, BC Decker Inc. Hamilton.
- 5. Miyake T, Cameron AM, Hall BK (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.
- 6. Shibata S, Suda N, Suzuki S, Fukuoka H, Yamashita Y (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.
- 7. Frommer J (1964) Prenatal development of the mandibular joint in mice. Anat Rec 150: 449–461.
- 8. Gu S, Wei N, Yu L, Fei J, Chen Y (2008) Shox2-deficiency leads to dysplasia and ankylosis of the temporomandibular joint in mice. Mech Dev 125: 729–742.
- 9. Wang Y, Liu C, Rohr J, Liu H, He F, et al. (2011) Tissue interaction is required for glenoid fossa development during temporomandibular joint formation. Dev Dyn 240: 2466–2473.
- 10. Purcell P, Jheon A, Vivero MP, Rahimi H, Joo A, et al. (2012) Spry1 and Spry2 are essential for development of the temporomandibular joint. J Dent Res 91: 387–393.
- 11. Mori-Akiyama Y, Akiyama H, Rowitch DH, de Crombrugghe B (2003) Sox9 is required for determination of the chondrogenic cell lineage in the cranial neural crest. Proc Natl Acad Sci USA 100: 9360–9365.
- 12. Shibata S, Suda N, Yoda S, Fukuoka H, Ohyama K, et al. (2004) Runx2-deficient mice lack mandibular condylar cartilage and have deformed Meckel's cartilate. Anat Embryol 208: 273–280.
- 13. Oka K, Oka S, Sasaki T, Ito Y, Bringas P Jr, et al. (2007) The role of TGF-beta signaling in regulating chondrogenesis and osteogenesis during mandibular development. Dev Biol 303: 391–404.
- 14. Oka K, Oka S, Hosokawa R, Bringas P Jr, Brockhoff HC 2nd, et al. (2008) TGF-beta mediated Dlx5 signaling plays a crucial role in osteo-chondroprogenitor cell lineage determination during mandible development. Dev Biol 321: 303–309.
- 15. St-Jacques B, Hammerschmidt M, McMahon AP (1999) Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev 13: 2072–2086.
- 16. Shibukawa Y, Young B, Wu C, Yamada S, Long F, et al. (2007) Temporomandibular joint formation and condyle growth require Indian hedgehog signaling. Dev Dyn 236: 426–434.
- 17. Purcell P, Joo BW, Hu JK, Tran PV, Calicchio ML, et al. (2009) Temporomandibular joint formation requires two distinct hedgehog-dependent steps. Proc Natl Acad Sci USA 106: 18297–18302.
- 18. Ochiai T, Shibukawa Y, Nagayama M, Mundy C, Yasuda T, et al. (2010) Indian hedgehog roles in post-natal TMJ development and organization. J Dent Res 89: 349–354.
- 19. Massagué J (2012) TGFβ signaling in context. Nat Rev Mol Cell Biol 13: 616–630.
- 20. Mitrovic DR (1977) Development of the metatarsophalangeal joint of the chick embryo: morphological, ultrastructural and histochemical studies. Am J Anat 150: 333–347.
- 21. Pacifici M, Koyama E, Iwamoto M (2005) Mechanisms of synovial joint and articular cartilage formation: recent advances, but many lingering mysteries. Birth Defects Res C 75: 237–248.
- 22. Storm EE, Kingsley DM (1996) Joint patterning defects caused by single and double mutations in members of bone morphogenetic protein (BMP) family. Development 122: 3969–3979.
- 23. Wolfman NM, Hattersley G, Cox K, Celeste AJ, Nelson R, et al. (1997) Ectopic induction of tendon and ligament in rats by growth and differentiation factors 5, 6, and 7, members of the TGF-beta gene family. J Clin Invest 100: 321–330.
- 24. Francis-West PH, Parish J, Lee K, Archer CW (1999) BMP/DGF-signaling interactions during synovial joint development. Cell Tissue Res 296: 111–119.
- 25. Settle SH Jr, Rountree RB, Sinha A, Thacker A, Higgins K, et al. (2003) Multiple joint and skeletal patterning defects caused by single and double mutations in the mouse Gdf6 and Gdf5 genes. Dev Biol 254: 116–130.
- 26. Rountree R, Schoor M, Chen H, Marks M, Harley V, et al. (2004) BMP receptor signaling is required for postnatal maintenance of articular cartilage. PloS Biol 11: e255.
- 27. Storm EE, Huynh TV, Copeland NG, Jenkins NA, Kingsley DM, et al. (1999) Limb alterations in brachypodism mice due to mutations in a new member of the TGF beta-superfamily. Nature 368: 639–643.
- 28. Brunet LJ, McMahon JA, McMahon AP, Harland RM (1998) Noggin, cartilage morphogenesis, and joint formation in the mammalian skeleton. Science 280: 1455–1457.
- 29. Lories RJU, Luyten FP (2005) Bone morphogenetic protein signaling in joint hoemostasis and disease. Cytok Growth Fact Rev 16: 287–298.
- 30. Dewulf N, Verschueren K, Lonnoy O, Moren A, Grimsby S, et al. (1995) Distinct spatial and temporal expression patterns of two type I receptors for bone morphogenetic proteins during mouse embryogenesis. Endocrinology 136: 2652–2663.
- 31. Mishina Y, Suzuki A, Ueno N, Behringer RR (1995) Bmpr encodes a type I bone morphogenetic protein receptor that is essential for gastrulation during mouse embryogenesis. Genes Dev 9: 3027–3037.
- 32. Zou H, Wieser R, Massagué J, Niswander L (1997) Distinct roles of type I bone morphogenetic protein receptors in the formation and differentiation of cartilage. Genes Dev 11: 2191–2203.
- 33. Yoon BS, Ovchinnikov DA, Yoshii I, Mishina Y, Behringer RR, et al. (2005) Bmpr1a and Bmpr1b have overlapping functions and are essential for chondrogenesis in vivo. Proc Natl Acad Sci USA 102: 5062–5067.
- 34. Yoon BS, Pogue R, Ovchinikov DA, Yoshii I, Mishina Y, et al. (2006) BMPs regulate multiple aspects of growth-plate chondrogenesis through opposing actions on FGF pathways. Development 133: 4667–4678.
- 35. Fukuoka H, Shibata S, Suda N, Yamashita Y, Komori T (2007) Bone morphogenetic protein rescues the lack of secondary cartilage in Runx2-deficient mice. J Anat 211: 8–15.
- 36. Danielian PS, Puccino D, Rowitch DH, Michael SK, McMahon AP (1998) Modification of gene activity in mouse embryos in uterus by a tamoxifen-inducible form of Cre recombinase. Curr Biol 8: 1323–1326.
- 37. He F, Xiong W, Wang Y, Matsui M, Yu X, et al. (2010) Modulation of BMP signaling by Noggin is required for the maintenance of palatal epithelial integrity during palatogenesis. Dev Biol 347: 109–121.
- 38. Stottmann RW, Choi M, Mishina Y, Meyers EN, Klingensmith J (2004) BMP receptor IA is required in mammalian neural crest cells for development of the cardiac outflow tract and ventricular myocardium. Development 131: 2205–2218.
- 39. Morikawa Y, Zehir A, Maska E, Deng C, Schneider M, et al. (2009) BMP signaling regulates sympathetic nervous system development through Smad-dependent and -independent pathways. Development 136: 3575–3584.
- 40. Morikawa Y, Cserjesi P (2008) Cardiac neural crest expression of Hand2 regulates outflow and second heart field development. Circ Res 103: 1422–1429.
- 41. Li L, Lin M, Wang Y, Cserjesi P, Chen Z, et al. (2011) BmprIa is required in mesenchymal tissue and has limited redundant function with BmprIb in tooth and palate development. Dev Biol 349: 451–461.
- 42.
Presnell JK, Schreibman MP (1997) Humason's Animal Tissue Techniques. Fifth edition. Baltimore, MD, The Johns Hopkins University Press.
- 43. St Amand TR, Zhang Y, Semina EV, Zhao X, Hu YP, et al. (2000) Antagonistic signals between BMP4 and FGF8 define the expression of Pitx1 and Pitx2 in mouse tooth-forming anlage. Dev Biol 217: 323–332.
- 44. Zhang Z, Song Y, Zhao X, Zhang X, Fermin C, et al. (2002) Rescue of cleft palate in Msx1-deficient mice by transgenic Bmp4 reveals a network of BMP and Shh signaling in the regulation of mammalian palatogenesis. Development 129: 4135–4146.
- 45. Alappat SR, Zhang Z, Suzuki K, Zhang X, Liu H, et al. (2005) The cellular and molecular etiology of the cleft secondary palate in Fgf10 mutant mice. Dev Biol 277: 102–113.
- 46. Swann DA, Silver FH, Slayter HS, Stafford W, Shore E (1985) The molecular structure and lubricating activity of lubricin isolated from bovine and human synovial fluids. Biochem J 225: 195–201.
- 47. Marcelino J, Carpten JD, Suwairi WM, Gutierrez OM, Schwartz S, et al. (1999) CACP, encoding a secreted proteoglycan, is mutated in camptodactyly-arthropathy-coxa vara-pericarditis syndrome. Nat Genet 23: 319–322.
- 48. Seki K, Hata A (2004) Indian hedgehog gene is a target of the bone morphogenetic protein signaling pathway. J Biol Chem 279: 18544–18549.
- 49. Zou H, Niswander L (1996) Requirement for BMP signaling in interdigital apoptosis and scale formation. Science 272: 738–741.
- 50. Komatsu Y, Yu PB, Kamiya N, Pan H, Fukuda T, et al. (2012) Augmentation of Smad-dependent BMP signaling in neural crest cells causes craniosynostosis in mice. J Bone Miner Res 28: 1422–1433.
- 51. Li L, Wang Y, Lin M, Yuan G, Yang G, et al. (2013) Augmented BMPRIA-mediated BMP signaling in cranial neural crest lineage leads to cleft palate formation and delayed tooth differentiation. PLoS ONE 8(6): e66107.
- 52.
Hall BK (2005) Bones and Cartilage: Developmental and Evolutionary Skeletal Biology. San Diego, CA, Elsevier Academic Press.
- 53. Pathi S, Rutenberg JB, Johnson RL, Vortkamp A (1999) Interaction of Ihh and BMP/Noggin signaling during cartilage differentiation. Dev Biol 209: 239–253.
- 54. Minina E, Wenzel HM, Kreschel C, Karp S, Gaffield W, et al. (2001) BMP and Ihh/PTHrP signaling interact to coordinate chondrocyte proliferation and differentiation. Development 128: 4523–4534.
- 55. Minina E, Kreschel C, Naski MC, Ornitz DM, Vortkamp A (2002) Interaction of FGF, Ihh/PThIh, and BMP signaling integrates chondrocyte proliferation and hypertrophic differentiation. Dev Cell 3: 1–20.
- 56. Fukada K, Shibata S, Suzuki S, Ohya K, Kuroda T (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 95: 321–329.
- 57. Kuboki T, Kanyama M, Nakanishi T, Akiyama K, Nawachi K, et al. (2003) Cbfa1/Runx2 gene expression in articular chondrocytes of the mice temporomandibular and knee joints in vivo. Arch Oral Biol 48: 519–525.
- 58. Mitrovic DR (1978) Development of the diathrodial joints in the rat embryo. Am J Anat 151: 475–485.
- 59. Nalin AM, Greenlee TK, Sandell LJ (1995) Collagen gene expression during development of avian synovial joints: transient expression of type II and X collagen genes in the joint capsule. Dev Dyn 203: 352–362.
- 60. Kimura S, Shiota K (1996) Sequential changes of programmed cell death in developing fetal mouse limbs and its possible role in limb morphogenesis. J Morph 229: 337–346.
- 61. Abu-Hijleh G, Reid O, Scothorne RJ (1997) Cell death in the developing chick knee joint. I. Spatial and temporal patterns. Clin Anat 10: 183–200.
- 62. Matsuda S, Mishima K, Yoshimura Y, Hatta T, Otani H (1997) Apoptosis in the development of the temporomandibular joint. Anat Embryo 196: 383–391.