DEVELOPMENTAL DYNAMICS 217:401– 414 (2000) Positionally-Dependent Chondrogenesis Induced by BMP4 Is Co-Regulated by Sox9 and Msx2 ICHIRO SEMBA, KAZUAKI NONAKA, ICHIRO TAKAHASHI, KATSU TAKAHASHI, RALPH DASHNER, LILLIAN SHUM, GLEN H. NUCKOLLS, AND HAROLD C. SLAVKIN* Craniofacial Development Section, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Maryland ABSTRACT Cranial neural crest cells emigrate from the posterior midbrain and anterior hindbrain to populate the first branchial arch and eventually differentiate into multiple cell lineages in the maxilla and mandible during craniofacial morphogenesis. In the developing mouse mandibular process, the expression profiles of BMP4, Msx2, Sox9, and type II collagen demonstrate temporally and spatially restrictive localization patterns suggestive of their functions in the patterning and differentiation of cartilage. Under serumless culture conditions, beads soaked in BMP4 and implanted into embryonic day 10 (E10) mouse mandibular explants induced ectopic cartilage formation in the proximal position of the explant. However, BMP4-soaked beads implanted at the rostral position did not have an inductive effect. Ectopic chondrogenesis was associated with the up-regulation of Sox9 and Msx2 expression in the immediate vicinity of the BMP4 beads 24 hours after implantation. Control beads had no effect on cartilage induction or Msx2 and Sox9 expression. Sox9 was induced at all sites of BMP4 bead implantation. In contrast, Msx2 expression was induced more intensely at the rostral position when compared with the proximal position, and suggested that Msx2 expression was inhibitory to chondrogenesis. To test the hypothesis that over-expression of Msx2 inhibits chondrogenesis, we ectopically expressed Msx2 in the mandibular process organ culture system using adenovirus gene delivery strategy. Microinjection of the Msx2-adenovirus to the proximal position inhibited BMP4-induced chondrogenesis. Over-expression of Msx2 also resulted in the abrogation of endogenous cartilage and the down-regulation of type II collagen expression. Taken together, these results suggest that BMP4 induces chondrogenesis, the pattern of which is positively regulated by Sox9 and negatively by Msx2. Chondrogenesis only occurs at sites where Sox9 expression is high relative to that of Msx2. The combinatorial action of these transcription factors appear to establish a threshold for Sox9 function and thereby restricts the position of chondrogenesis. Dev Dyn 2000;217:401–414. Published 2000 Wiley-Liss, Inc.† Key words: BMP4; Sox9; Msx2; type II collagen; Meckel’s cartilage; chondrogenesis; † Published 2000 WILEY-LISS, Inc. This article is a US government work and, as such, is in the public domain in the United States of America. mouse embryo; mandibular process; bead implantation; organ culture; whole-mount in situ hybridization; Alcian Blue staining; competitive RT-PCR; adenovirus gene delivery INTRODUCTION Bone morphogenetic factor was used to describe the induction of cartilage and bone formation when semipurified bone matrix was introduced into the subectodermal and myogenic tissues of the adult rat and rabbit (Urist, 1965). Thereafter, these factors were characterized (Wozney et al., 1988) and revealed more than 15 related proteins belonging to the family of bone morphogenetic proteins (BMPs); which in turn belongs to an even larger group of growth and differentiation factors, the transforming growth factor beta (TGF␤) superfamily (see recent reviews Reddi, 1997; Vortkamp, 1997; Urist, 1997; Wozney and Rosen, 1998). BMPs signal through a system of type I and II transmembrane receptor serine/threonine kinases, which are then modulated by cytoplasmic Smad proteins (see reviews Derynck et al., 1998; Kawabata et al., 1998). In the adult, elevated expression of BMP4 has been associated with the pathological condition afflicting the skeleton, fibrodysplasia ossificans progressiva (Shafritz et al., 1996; Kaplan and Shore, 1998; Lanchoney et al., 1998). In the embryo, BMP2 and BMP4 have been implicated to function during skeletogenesis (see review Hall and Miyake, 1995), as well as in tissues in which epithelial-mesenchymal interactions mediate morphogenesis and differentiation (Nuckolls et al., 1998). Despite these significant contributions, the molecular mechanisms of bone or cartilage induction by any member of the family remain elusive, partly due to limitations associated with using either in vivo systems The first two authors contributed equally to this project and are considered co-first authors. Grant sponsor: National Institutes of Health; Grant number: Z01AR41114. Ichiro Semba’s present address is Department of Oral Pathology, Kagoshima University Dental School, Kagoshima, 890-8544, Japan. *Correspondence to: Dr. Harold C. Slavkin, Craniofacial Development Section, NIAMS, NIH, 6 Center Drive, Building 6, Room 324, MSC-2745, Bethesda, MD, 20892-2745. Received 16 August 1999; Accepted 7 January 2000 402 SEMBA & NONAKA ET AL. or serum-containing in vitro systems in which the interactive effects of many other growth and differentiative factors cannot be isolated from the direct actions of BMP itself. Further, the functions of BMP2 or BMP4 in the embryo cannot as yet be fully studied in genetically engineered mouse models because targeted disruption of either molecule or their type I receptor resulted in early embryonic lethality rendering the role of BMPs during organogenesis, including skeletogenesis uninformative (Mishina et al., 1995; Winnier et al., 1995; Zhang and Bradley, 1996). Finally, BMPs are a diverse group of proteins with pleiotrophic functions in multiple differentiation programs and developmental stages (see recent reviews Hogan, 1996; Graff, 1997). The precise microenvironmental milieu regulating a specific biological action of a specific BMP molecule is often difficult to identify. Therefore, in order to address what are the positional and temporal controls required for BMP-induced skeletogenesis, and what specific molecular events are associated with the developmental process, we employed a unique in vitro culture system in which embryonic mouse mandibular processes were explanted under serum-free conditions and subjected to a variety of micromanipulations. We report that exogenous BMP4 delivered by a protein-soaked bead implantation method induced ectopic cartilage formation, yet the extent of chondrogenesis was dependent on the dose of BMP4 administered, the developmental stage of the embryonic tissue, and the position within the developing mandible. These temporal and positional tissue responses were attributable to the induction and combinatorial expression of two transcription factors, Sox9 and Msx2. The combinatorial and antagonistic actions of these molecules appeared to restrict ectopic cartilage formation during mandibular morphogenesis. RESULTS Exogenous BMP4 Induced Ectopic Cartilage Formation in Embryonic Mandibular Process Cultured Under Serumless Conditions Mouse embryos at embryonic day 10 (E10) (Fig. 1a), corresponding to Theiler stage 18 with 40 – 44 somite pairs, are at a stage of development during which the anlagen for major organs are shaped and overt cytodifferentiation is initiated. The mandibular process of the developing first branchial arch was isolated by microdissection and explanted into an organ culture system supported by serumless, chemically-defined medium. The mandibular process (Fig. 1b) consisted of bilateral prominences of epithelium enclosing ectomesenchymal cells, derivatives of cranial neural crest cells that emigrated from the posterior midbrain and anterior segments of the hindbrain (rhombomeres 1 and 2) (Lumsden et al., 1991; Sechrist et al., 1993; OsumiYamashita et al., 1994), and head mesoderm. The serumless and chemically-defined culture system facilitates the investigation of time- and positionaldependent autocrine and paracrine growth and differentiation factors, without the numerous con- Fig. 1. Embryonic mandibular process explant culture. Embryonic day 10 (E10) mouse embryo (a) was microdissected to isolate the mandibular process (arrow) and placed in serumless, chemically-defined explant culture. At the beginning of the culture period (b), the mandibular process consisted of two halves fused at the midline. Gross morphogenesis and cytodifferentiation occurred by day 6 of culture (c). Meckel’s cartilage was identified by Alcian blue staining (d). Protein-soaked beads (arrowheads) were placed within the developing tissue to deliver exogenous proteins (b, c, d). Exogenous BMP4 delivered by this method induced ectopic cartilage formation as revealed by Alcian blue staining (d). Scale bar for (a) ⫽ 1mm, and for (b– d) ⫽ 150m. Sox9 and Msx2 CO-REGULATE BMP4-INDUCED CHONDROGENESIS Fig. 2. BMP4-soaked beads induced ectopic cartilage formation only when placed at the proximal position of the developing mandibular process. 100ng/l BMP4 was delivered by bead implantation method into the rostral (a, d) or proximal (b, e) positions and allowed to develop in vitro for 6 days. Chondrogenesis was assayed by whole-mount Alcian blue staining (a, b, c) and toluidine blue staining on histological sections 403 d, e, f). Only beads implanted into the proximal position induced ectopic cartilage formation (b, e). PBS-soaked bead had no effect on chondrogenesis (c). The morphology of BMP4-induced cartilage (e) was similar to endogenous Meckel’s cartilage (f); arrowheads indicating perichondrium. All beads are circled. Rostral orientation is top of panel in (a– c). Scale bar for (a, b, c) ⫽ 150m, and for (d, e, f) ⫽ 50m. Fig. 3. BMP4, Msx2, and Sox9 exhibited restrictive expression patterns. Whole-mount in situ hybridization for BMP4 (a– d), Msx2 (e– h) and Sox9 (i–l) on 0 (a, e, i), 2 (b, f, j), 4 (c, g, k), and 6 days (d, h, i) of culture. Whole-mount Alcian blue staining demonstrated Meckel’s cartilage formation at 4 (m) and 6 days (n) of culture. Rostral orientation is top for all panels. Scale bars for (a, e, i) and (b– d, f– h, j–n) ⫽ 100m. 404 SEMBA & NONAKA ET AL. TABLE 1. Frequency of Ectopic Cartilage Induced by BMP4-Soaked Beads in Mandibular Processes in Explant Cultures Concentration of BMP4 (ng/l) Rostral Proximal 10 0/15 (0%) 0/15 (0%) 50 0/15 (0%) 6/15 (40%) 100 0/18 (0%) 15/18 (83.3%) founding variables found in serum supplement and the maternal circulation. Growth and differentiation of the mandibular process in vitro are delayed but not retarded when compared with in vivo development (Slavkin et al., 1989). Further, the explant system supports micromanipulations such as bead implantation and the microinjections used in this study. Beads soaked in BMP4 were positioned efficiently and reproducibly in the mandibular process at the beginning of culture (Fig. 1b) and the biological effects observed over a period of six days. By six days in culture (Slavkin et al., 1999; Fig. 1c), differentiation and the formation of cartilage, osteoid, incisor and molar tooth buds and tongue were observed. Whole-mount Alcian blue staining demonstrated the formation of Meckel’s cartilage which was composed of a rostral triangular-shaped piece, bilateral rod-shaped pieces and proximal pieces representing the future middle ear ossicles (Fig. 1d). Beads soaked in 100ng/l BMP4 consistently induced ectopic cartilage formation (Fig. 1d). Ectopic cartilage was observed to fuse with endogenous cartilage formation patterns resulting in expansion or bifurcation morphology, or as observed as isolated ectopic cartilage pieces. Ectopic Chondrogenesis Induced by BMP4 Was Positional-, Temporal-, and Dose-Dependent Further experiments demonstrated that beads soaked in 100ng/l BMP4 only induced ectopic cartilage formation when the bead was placed in the proximal position of the developing mandibular process when compared with the rostral position (Fig. 2a, b; Table 1). Control beads soaked in PBS (Fig. 2c), or beads soaked in 100ng/l FGF8 (data not shown), did not induce ectopic chondrogenesis or disrupt endogenous cartilage formation. Because Alcian blue staining demonstrated positive reaction to sulfated proteoglycan suggestive of cartilage, we additionally performed routine histological evaluation on frozen sections of the developing mandiblular explants at the site of bead implantation. Consistent with the whole-mount Alcian blue staining results, cartilage was observed enveloping beads implanted at the proximal position (Fig. 2e), whereas cartilage was not detectable surrounding beads implanted at the rostral position (Fig. 2d). Cartilage induced by BMP4 has morphological and histological characteristics similar to that of normal cartilage (Fig. 2f), including the presence of an intact perichondrium and mature chondrocytes residing within lacunae of the cartilagenous matrix. No signif- icant apoptosis was detected associated with bead implantation. The incidence of ectopic chondrogenesis was directly dependent on the concentration of BMP4 used in these investigations (Table 1). At 10ng/l BMP4, no ectopic cartilage was induced, whereas at 100ng/l the incidence was 83.3% when beads were implanted at the proximal position. BMP4 Induced Different Levels of Sox9 and Msx2 Expression Dependent on Positional Information To explain positional dependency of BMP4 induced chondrogenesis, whole-mount in situ hybridizations, using several gene probes at various time points, was performed to identify candidate molecules associated with the phenomenon. Mandibular processes at the beginning of culture, and after 2, 4, or 6 days of development were fixed and processed. Endogenous BMP4 transcripts were detected in the E10 mandibular process, with a higher level of expression on the medial aspect when compared with the lateral (Fig. 3a); consistent with previous reports (Bennett et al., 1995; Neubüser et al., 1997). Thereafter, BMP4 exhibited complex localization patterns (Figure 3b– d), presumably due to its functions in multiple cell lineages within the developing mandibular process— chondrogenic and odontogenic. The localization pattern of Msx2 in the mandibular process in culture in part overlapped with that of BMP4 expression especially in the rostral area, yet was more restrictive (Fig. 3e– h). At day 2 of culture, Msx2 was expressed in rostral-medial and bilaterally symmetrical proximal-lateral sites, which appeared to delineate the rostral and caudal extent of future Meckel’s cartilage formation (Fig. 3f ). Further, at day 4 of culture Msx2 expression was observed along a boundary interface with sites destined to become cartilage (Fig. 3g). Sox9 has been shown to be a direct transcriptional activator of the cartilage-specific type II collagen gene (Lefebvre et al., 1997). The localization of endogenous Sox9 in the developing mandibular process was associated with Meckel’s cartilage development (Figure 3i–l). At day 2 of culture, prior to cartilage differentiation, Sox9 was localized in a rostral triangular-shaped site, bilateral sigmoidal-shaped sites and at the proximal lateral margins of the explants (Fig. 3j). This pattern resembled the Alcian blue staining pattern of Meckel’s cartilage at day 6 of culture (Fig. 3n), and presumably served as a morphogenetic template for this cartilage. Sox9 expression was somewhat decreased but still detectable at day 4 and 6 of culture (Fig. 3k, l) at which times cartilage differentiation was observed (Fig. 3m, n). Because the endogenous expression pattern of Msx2 overlapped with that of BMP4, and that it also appeared to define boundary interfaces suggestive of a patterning role during chondrogenesis, the expression of Msx2 in mandibular processes implanted with BMP4-soaked beads was examined after 24 hours of culture (Fig. 4a, b). Consistent with the responses of Sox9 and Msx2 CO-REGULATE BMP4-INDUCED CHONDROGENESIS many other developmental stages and tissues, BMP4 induced Msx2 expression in the immediate vicinity of the source. However, the level of induced Msx2 expression appeared to be higher when the BMP4-soaked bead was implanted in the rostral as compared with that in the proximal position. This expression of Msx2 in response to BMP4 was transient and down-regulated in the following 24 hours (Fig. 4c, d). Because Sox9 is a direct transcription activator of the type II collagen gene, we examined its expression in response to exogenous BMP4. One day after bead implantation, Sox9 expression was induced surrounding the BMP4-soaked bead (Fig. 4e, f ). Contrary to Msx2 induced expression, the amount of Sox9 was higher around the BMP4-soaked bead implanted at the proximal as compared with the rostral positions. Similar to the Msx2 response, induced Sox9 expression was decreased by day 2 of explant culture (Fig. 4g, h). The expression of type II collagen, presumably mediated by the activation of Sox9, was detected surrounding the BMP4-soaked bead at 48 hours of culture; when Sox9 expression was already down-regulated (Fig. 4j). This observation is consistent with previous reports of Sox9 expression preceding that of type II collagen (Ng et al., 1997; Takahashi et al., 1998a). The expression of type II collagen at 24 and 48 hours showed a pattern that was predictable of the differentiated phenotype as detected by Alcian blue staining on day 4. Type II collagen expression was detected around the developing Meckel’s cartilage. In addition, expression was also detected around the BMP4-soaked bead implanted at the proximal position, resulting in an expansion of the domain (Fig. 4j, l). Type II collagen expression, though detected surrounding the BMP4soaked bead implanted at the rostral position at 24 hours of culture (Fig. 4i), was observed to be downregulated to non-existence by 48 hours (Fig. 4k). This change in expression level was repeatedly consistent with the phenotypic outcome at day 6, when Alcian blue staining did not detect any ectopic chondrogenesis when BMP4-soaked beads were introduced at this position. The molecular responsiveness of the tissue to BMP4 was restricted to the first 24 hours of culture. BMP4soaked beads implanted into the mandibular process after the explants had been in culture for one day showed no induction of Msx2 nor Sox9, and no ectopic cartilage (data not shown). Taken together, these data demonstrate that tissue response to exogenous BMP4 is positionally and temporally dependent. The observed difference in the expression of BMP4 inducible genes was further analyzed using quantitated morphometric (Table 2) and semi-quantitative RT-PCR approaches (Fig. 5). First, to minimize variations in developmental stage and thereby responsiveness among different E10 explants, two beads were implanted into the same piece of tissue; one in each of the rostral and proximal positions, respectively. Consistent with the single bead implantation experiments, the rostral position demonstrated a more intense re- 405 sponse to Msx2 (Fig. 4m), and a weaker signal for Sox9 (Fig. 4n), when compared with the proximal position. Type II collagen expression was detected at both sites of implantation (Fig. 4o). Control beads soaked in PBS had no effect on gene expression level or pattern of either Msx2 (Fig. 4p), Sox9 (Fig. 4q) or type II collagen (Fig. 4r). Subsequently, explants processed for wholemount in situ hybridization were subjected to morphometric analyses in which the area of induced gene expression was converted to the number of pixels positive for the hybridization signal (Table 2). With over nine explants analyzed for each gene pattern, the quantitative analysis was consistent with the qualitative analysis. Expression of Sox9 surrounding the BMP4-bead implanted at the proximal position was 2.6 times higher than that at the rostral position. In contrast, Msx2 expression was three-fold weaker, comparing the corresponding positions. Expression of type II collagen was 6.1-fold higher surrounding the proximally placed BMP4-beads than those placed rostrally. The differences in size of induced Sox9, Msx2, and type II collagen expression between the rostral and proximal positions were statistically significant (P ⬍ 0.01). Semi-quantitative RT-PCR results also supported positionally-dependent differential levels of induction of Sox9 and Msx2 by BMP4. Consistent with data from both whole-mount in situ hybridization and morphometric analyses, a three-fold elevation of Sox9 induced by BMP4 was detectable when beads were implanted proximally as compared with beads implanted rostrally, or controls using PBS-soaked beads or no beads (Fig. 5a). We could not detect a significant increase in Sox9 expression induced by BMP4-soaked beads placed at the rostral position, possibly due to the high endogenous level of Sox9 expression marking the morphogenetic template of Meckel’s cartilage. Significant induction of Msx2 expression by BMP4 was also detected at both rostrally or proximally implanted beads; more than three-fold and almost two-fold, respectively, when compared with controls (Fig. 5b). The difference between rostral and proximal expression levels was statistically significant (P ⬍ 0.05). Direct Over-Expression of Msx2 Inhibited Chondrogenesis Sox9 expression preceded that of type II collagen, yet type II collagen expression was not maintained around the BMP4-soaked bead implanted at the rostral position, resulting in no ectopic cartilage. This could be due to the expression of Msx2, which was much more heavily detected at the rostral rather than the proximal position. To test the hypothesis that Msx2 inhibits BMP4-induced chondrogenesis, a recombinant adenovirus was engineered to over-express the Msx2 gene (AdV-Msx2). The adenovirus was microinjected into the developing mandibular process and its actions evaluated against explants microinjected with control adenovirus bearing LacZ expression (AdV-LacZ). Consistent with previous reports (Takahashi et al., 1998b), microinjection of AdV-LacZ resulted in high levels of 406 SEMBA & NONAKA ET AL. Fig. 4. BMP4 induced Sox9 and Msx2 expression in developing mandibular process in explant cultures. Beads soaked in 100ng/l BMP4 were implanted into the rostral (a, c, e, g, i, k) or proximal (b, d, f, h, j, l) positions of mandibular processes and explanted into cultures. Wholemount in situ hybridization was performed for Msx2 (a– d), Sox9 (e– h), and type II collagen (i–l), after 24 (a, b, e, f, i, j), or 48 (c, d, g, h, k, l) hours in culture. The induced expression of Msx2 and Sox9 was detected after 24 hours, whereas the inductive effect was down-regulated after 48 hours in culture. Sox9 expression appeared higher surrounding the BMP4soaked bead placed at the proximal (f) than that at the rostral (e) position, whereas this response appeared to be reverse for that of Msx2. Type II collagen was only observed surrounding BMP4-soaked beads implanted at the proximal postition (j). In order to demonstrate that positionaldependent differential gene induction was not due to differences among individual mandibles, multiple BMP4-soaked beads were implanted at various positions within the same mandible and whole-mount in situ hybridization was performed for Msx2 (m), Sox9 (n) and type II collagen (o) after 1 day in culture. The results were similar to that of the single bead implantation experiments. Control PBS-soaked beads implanted at any position did not induce Msx2 (p), Sox9 (q) or type II collagen (r) after 24 hours in culture. All beads are circled. Rostral orientation is top for all panels. Scale bars ⫽ 100m. expression detectable 36 hours subsequent to injection (Fig. 6a). AdV-LacZ did not affect BMP4-induced ectopic cartilage formation in the proximal position of the mandibular process (Fig. 6b). When AdV-Msx2 was microinjected at the same site where BMP4-soaked bead was implanted, ectopic cartilage formation was reduced (Fig. 6c, d) or abolished (Fig. 6e). The cartilage matrix was weakly stained with Alcian blue and appeared vacuolated. In addition, we also tested whether over-expression of Msx2 inhibits endogenous cartilage formation. The recombinant adenovirus was injected into the medial or lateral side of the developing mandibular process. Similarly, high expression of lacZ was detectable at 36 hours subsequent to injection (Fig. 7a, b). AdV-LacZ injected explants had no deleterious effects and Meckel’s cartilage developed normally (Fig. 7c, d). In contrast, AdV-Msx2 injected into the lateral side of the mandibular process resulted in an abrogation of Meckel’s cartilage formation, whereas neither AdV-Msx2 injected into the medial part of the explant, nor the uninjected side produced Meckel’s cartilage dysmorphogenesis. Histological analyses confirmed that normal cartilage formed in the control non-injected side (Fig. 7g), yet no evidence of cartilage was observed at the AdV-Msx2 injected side of the explant (Fig. 7h). No significant apoptosis was detected in adenovirus injected specimens. Sox9 and Msx2 CO-REGULATE BMP4-INDUCED CHONDROGENESIS 407 TABLE 2. Size of Sox9, Msx2, and Type II Collagen Expression Domain Induced by 100 ng/l BMP4-Soaked Beads in Mandibular Processes in Explant Cultures Sox9 Msx2 Type II Collagen Rostral mean ⫾ SD (pixels) 919.8 ⫾ 460.2 2809.2 ⫾ 632.6 257.3 ⫾ 75.3 Proximal mean ⫾ SD (pixels) 2346.0 ⫾ 578.1* 890.1 ⫾ 207.3* 1572.2 ⫾ 463.4* Proximal/rostral ratio 2.55 0.32 6.11 *P ⬍ 0.01. The level of type II collagen expression subsequent to AdV-Msx2 microinjection was also analyzed by competitive PCR technique in explants after 6 days in culture (Fig. 7i). Consistent with the morphological analyses, there was no difference in type II collagen expression when Msx2 was locally infected in the medial aspect of the developing mandibular process. In contrast, when Msx2 was over-expressed in the lateral side of the mandible, the amount of type II collagen expression was reduced to only 5% of control value (p ⬍ 0.001). DISCUSSION The induction of bone and cartilage by BMPs has been well documented in both the embryo and the adult. In the embryo, endogenous actions of BMPs contribute to skeletogenesis (Lyons et al., 1989; FrancisWest et al., 1994; Dewulf et al., 1995; Reddi, 1994, 1995; Hall and Miyake, 1995; Frenz et al., 1996; Kawakami et al., 1996; Zou et al., 1997) and exogenous applications result in ectopic bone and cartilage formation (Duprez et al., 1996a; Francis-West et al., 1996; Barlow and Francis-West, 1997; Ekanayake and Hall, 1997; Macias et al., 1997; Nifuji et al., 1997). Consistent with the literature, we first confirmed that such osteo-chondro-inductive potential of BMP4 also functions in the embryonic mandibular mesenchyme (Fig. 1). We extended these observations to the transcription level, and identified and characterized the mechanism of BMP4 action using a serumless, chemically-defined explant culture system model (Figs. 1 and 2). We observed that ectopic cartilage formation induced by BMP4 was preceded by the up-regulation of Sox9 and Msx2 (Fig. 4). However, overt chondrogenesis was only defined by the relative high expression level of Sox9 when compared with that of Msx2 (Figs. 4 and 5). In contrast, elevated levels of Msx2 inhibited ectopic cartilage induction by BMP4 despite the concomitant expression of Sox9. Because Sox9 has been shown to be a direct activator of the type II collagen promoter, we hypothesize that Msx2 is a repressor of chondrogenesis. Indeed, over-expression of Msx2 by adenoviral gene delivery in the mandibular explants inhibited BMP4induced chondrogenesis (Fig. 6) and also resulted in the down-regulation of type II collagen expression and abrogation of Meckel’s cartilage formation (Fig. 7). The combinatorial actions of these two transcription factors regulate the extent of chondrogenesis and appear to define the temporal and positional information of tissue responses to BMP4. Fig. 5. BMP4 induced different levels of Sox9 and Msx2 expression according to positional information in developing mandibular process in explant cultures. Relative expression of Sox9 (a) or Msx2 (b) was assayed by semi-quantitative RT-PCR on mandibular processes implanted with 100ng/l BMP4-soaked beads at the rostral or proximal position, and then compared with controls which consisted of non-implanted or PBSsoaked bead implanted specimens. *P ⬍ 0.05 compared with control, ✣ P ⬍ 0.05 compared with proximal value. N ⫽ 8 or 9 for all groups. 408 SEMBA & NONAKA ET AL. Fig. 6. Over-expression of Msx2 inhibited BMP4induced chondrogenesis. Adenovirus expressing LacZ (a, b) or Msx2 (c, d, e) was microinjected at the same site where BMP4-soaked beads were implanted at the rostral position of the mandibular process. Beta-galactosidase staining of mandibular processes after 36 hours in culture confirmed the site and spread of lacZ expression (a). Whole-mount Alcian blue staining of infected mandibular processes after 6 days demonstrated that ectopic chondrogenesis induced by BMP4 was unaffected by lacZ expression (b) but was reduced (c, d) or abolished (e) by Msx2 expression. The affected cartilage appeared vacuolated and weakly stained with Alcian blue (arrows). Scale bar ⫽ 100 m. Combinatorial Signaling Contributes to the Diverse Functions of BMPs Due to the pleiotropic nature of BMPs, an in vitro serum-free culture system is invaluable for studies designed to identify and isolate the direct actions of BMP without confounding complications from interacting serum factors. Previous attempts in using serum-free systems to study chondrocyte maturation have provided initial phenotypic and biochemical evidence to BMP-induced chondrogenesis, but the molecular basis has yet to be elucidated (Leboy et al., 1997; Lee and Chuong, 1997). Our studies have confirmed BMP-induced chondrogenesis in serumless conditions. In addition, we identified the early molecular responses of cartilage initiation and demonstrated that Sox9, a marker of chondrogenesis, is induced by BMP4. Because BMP can also induce the expression of Cbfa1 (Ducy et al., 1997), marker for osteogenesis, it is likely that either the two skeletal cell lineages are segregated at an early time point of differentiation or that the action of BMP is temporally and positionally dependent. It is conceivable that biological signaling molecules interact synergistically or antagonistically with other molecules to specify the temporal and spatial differentiation program (Davidson, 1995). BMP4 exemplifies this concept due to its particularly diverse set of functions during embryogenesis. The action of BMP4 can be regulated at multiple hierarchical levels to achieve cell and tissue type specificity. At the ligand level, BMPs act antagonistically with FGFs to regulate Pax9 ex- pression (Neubüser et al., 1997), to specify tooth identity (Tucker et al., 1998), and to regulate limb outgrowth (Niswander and Martin, 1993; Buckland et al., 1998). BMP also antagonizes with Shh to regulate cartilage formation in the vertebrae (Watanabe et al., 1998). At the level of cytoplasmic signaling, BMP competes with EGF or HGF for activating versus inhibiting phosphorylation signal on Smad1 (Kretzschmar et al., 1997; Nonaka et al., 2000). Our results suggest that at the transcription level, BMP4 induces two antagonistic signals, which in turn regulate the chondrogenic pathway. This combinatorial regulatory sequence would appear to ensure the fidelity of responses elicited by a single extracellular signal. Threshold Level Exists for Sox9-Directed Chondrogenesis The Sry-type HMG-box containing transcription factor Sox9 is a direct transcriptional activator of the type II collagen promoter (Bell et al., 1997; Lefebvre et al., 1997); a signaling factor for chondrogenesis. Mutations in Sox9 gene in human are associated with the skeletal dysmorphology syndrome, campomelic dysplasia (Foster et al., 1994; Wagner et al., 1994). Expression of Sox9 is associated with chondrogenesis (Wright et al., 1995; Healy et al., 1996; Wheatley et al., 1996; Ng et al., 1997; Zhao et al., 1997). Misexpression of Sox9 results in ectopic cartilage formation (Kanzler et al., 1998; Healy et al., 1999) whereas teratomas with inactivated Sox9 contained no cartilage (Bi et al., 1999) suggesting that Sox9 is sufficient to induce or promote Sox9 and Msx2 CO-REGULATE BMP4-INDUCED CHONDROGENESIS 409 Fig. 7. Over-expression of Msx2 in developing mandibular processes inhibited Meckel’s cartilage formation. Adenovirus-expressing LacZ (a– d) or Msx2 (e, f) was microinjected to the medial (a, c, e) or lateral portions (b, d, f) of the right half of the E10 mandibular process as shown in the figure. Beta-galactosidase staining of mandibular processes after 36 hours in culture confirmed restrictive gene expression delivered by adenovirus microinjection and infection method (a, b). Whole-mount Alcian blue staining of infected mandibular processes after 6 days demonstrated no effect on Meckel’s cartilage formation by LacZ expression (c, d) or Msx2 expression to the medial portion (e). Expression of Msx2 at the lateral portion of the mandibular process inhibited Meckel’s cartilage formation (f). Toluidine blue staining of histological sections after 6 days revealed Meckel’s cartilage (outlined in white) on the uninjected side (g) but absence of chondrogenesis or excessive apoptosis on the Msx2 expressing side (h). Scale bars for (a–f) ⫽ 100m, and for (g, h) ⫽ 50m. Expression level of type II collagen was assayed by quantitative RT-PCR in mandibular processes infected with adenovirus expressing LacZ (hatched bars) or Msx2 (solid bars) at the medial or lateral portion (i). Consistent with morphological analyses, Msx2 expression at the lateral portion of the mandibular process inhibited type II collagen expression. *P ⬍ 0.001. N ⫽ 6. chondrogenesis. It has also been reported that Sox9 may bind and activate the promoters of two other cartilage markers, aggrecan and type X collagen (Harada et al., 1996; Sekiya et al., 1997). Despite evidence supporting the critical role of Sox9 during early chondrogenesis, further upstream regulatory elements of this pathway have been elusive. It has been implicated that the level of Sox9 expression is significant in determining whether chondrogenesis is to occur or not. High levels of Sox9 are observed to correlate with all sites of cartilage formation. Low levels of Sox9 are detected in a broad array of non-chondrogenic embryonic tissues such as notochord, neural tube, heart and lung (Ng et al., 1997; Zhao et al., 1997). It is yet unclear whether an absolute threshold level of Sox9 expression is required to induce chondrogenesis, nor how this level is determined and controlled. Our data suggests that a relative threshold level may exist. The expression of Sox9 precedes that of type II collagen expression, an indication that the responding mesenchymal cells can and have initiated the chondrogenic program regardless of the absolute expression level of Sox9. However, overt chondrogenesis only occurs at sites where Sox9 expression is high relative to that of the repressor Msx2. Therefore, the antagonism between Sox9 and Msx2 appears to establish a relative threshold for Sox9 function in cartilage induction, and that this relationship is site specific. This threshold may also be modulated positively or nega- 410 SEMBA & NONAKA ET AL. tively by the presence of other transcription factors such as other Sox proteins (Lefebvre et al., 1998). Msx2 Is a Repressor of Chondrogenesis The Msx family of transcription factors is multifunctional, although very little is known of Msx3. Msx1 and Msx2 are largely co-expressed in the developing embryo, and are thought to be redundant in functions (Davidson, 1995; Maas et al., 1996). Disruption of Msx functions results in multiple malformations including that of the craniofacial skeleton (Satokata and Maas, 1994; Liu et al., 1995; Foerst-Potts and Sadler, 1997; Winograd et al., 1997). In many cases, Msx genes mediate the function of BMPs including cranial suture morphogenesis (Kim et al., 1998), apoptosis (Marazzi et al., 1997; Takahashi et al., 1998b), tooth formation (Tureckova et al., 1995; Jernvall et al., 1998), dorsalization of midline structures along the neural tube (Takahashi et al., 1996), vertebral development (Monsoro-Burq et al., 1996; Watanabe and Le Douarin, 1996), and mammary gland development (Phippard et al., 1996). Further, Msx1 and Msx2 are considered to be general repressors of transcription (Catron et al., 1995; Semenza et al., 1995), directly modulating the function of the promoter for osteocalcin (Newberry et al., 1997), type I collagen (Dodig et al., 1996), and MyoD enhancer (Woloshin et al., 1995). Consistent with a repressor function of Msx, our investigations have provided direct evidence that expression of Msx2 is inhibitory in the chondrogenic pathway; including both endogenous as well as ectopic cartilage formation. This is supported by previous reports in which application of antisense oligonucleotides directed against Msx2 resulted in an increase in cartilage formation in the avian mandible (Mina et al., 1996). The repressor action of Msx2 is likely to be complex. For example, it is unclear how Msx2 overexpression eliminated endogenous cartilage from the lateral side of the mandibular process and not the medial side. We speculate that because the medial portion is a site of high endogenous expression of Msx1 and Msx2, our adenovirus-driven ectopic expression might not have achieved an additionally increased level. The endogenous ratio of Sox9 to Msx2 would then already determine the outcome of chondrogenesis. It is also possible that chondroprogenitor cells at the medial position are less committed in phenotype and less responsive to Msx2 than those at the lateral. The lateral part of Meckel’s cartilage was formed by day 4 of culture (Fig. 4m), whereas the medial portion was Alcian blue positive only by day 6 (Fig. 4n). Because adenovirus expression in our system is maintained for 48 –72 hours, there is the potential for compensatory growth when expression level declines. In addition, other results indicate that Msx1 or Msx2 can interact with other transcription factors and compete for DNA binding (Catron et al., 1995; Zhang et al., 1996; Wu et al., 1997; Zhang et al., 1997; Bendall et al., 1998). These additional regulatory networks may be different locally. Taken together, these results suggest that the regulation of a single extracellular signaling molecule by combinations of transcription factors would appear to be required to achieve a specific response in a specific developmental context. The Developing Mandibular Process Is Compartmentalized Compartmentalization of the developing mandibular mesenchyme is associated with early complex molecular patterning. The cranial to caudal and proximal to rostral axial plans are established and maintained by the expression of a sequence of homeobox-containing genes. The oral to aboral axis along the maxillary and mandibular processes of the first branchial arch is associated with the expression of Lim-domain homeoproteins, Lhx6 and Lhx7 (Grigoriou et al., 1998). The oral mesenchyme is thus competent to respond to epithelial odontogenic signals such as FGFs. These axes are further refined by homeobox containing transcription factors of the Msx and Dlx families, and may specify the formation of incisor versus molar teeth within the first branchial arch (Qiu et al., 1997; Thomas et al., 1997). Other domains have been suggested to limit the manifestation of other phenotypes, such as bone, cartilage and muscle within the embryonic craniofacial mesenchyme. We propose that further compartmentalization along the rostral-proximal and medial-lateral axes exists that are defined by the chondrogenic potential in the forming mandible. This assertion is supported by experiments in which meticulous microdissections followed by grafting, tissue recombination or micromass cultures of different segments of the developing mandible, identified varying potentials to form cartilage (Hall, 1982; Langille, 1994). Similar positional information also exists in the developing limb bud and has been defined by the proximodistal, dorsoventral, and anteroposterior axes (Cottrill et al., 1987; Ros et al., 1997). Although not assayed in this study, it is likely that BMP2, BMP7, and Msx1 may also be involved in developing mandibular cartilage based on their sequence homologies and overlapping expression pattern with BMP4 and Msx2, respectively. The specificity of these families of growth and transcription factors towards chondrogenesis versus odontogenesis remains to be characterized. During early development, progressively restrictive molecular patterning emerges from a sequence of transcriptional activations and inactivations. Our results suggest that within the presumptive myogenic region destined to be the tongue, chondrogenic potential exists and that mesenchymal cells can be induced to become chondrocytes in response to BMP4. The induction is dose-dependent on exogenous BMP4, and suggests that the endogenous developmental program consists of silencers to the chondrogenic pathway in the tuberculum impar. Indeed, the induction of cartilage by BMP may be the result of simultaneous regulation of two pathways; stimulation of chondrogenesis and inhibition of myogenesis. In chick limb bud micromass cultures, BMP2 specifically promotes chondrogenesis presum- Sox9 and Msx2 CO-REGULATE BMP4-INDUCED CHONDROGENESIS ably at the expense of myogenesis (Duprez et al., 1996b). In the developing somite, BMP signaling inhibits the activation of MyoD and Myf5 (Reshef et al., 1998). Further, in vitro studies support that endogenous Msx1 can bind to and repress the MyoD enhancer (Woloshin et al, 1995). These findings and our results indicate that BMP is instructive to both chondrogenesis as well as myogenesis in undifferentiated mesenchymal cells. CONCLUSION In summary, the evidence supports the concept that the emergence of the chondrogenic potential of BMP4 from the pleiotropic nature of this growth factor is mediated by the induced expression of Sox9 and Msx2. Sox9 and Msx2 function as positive and negative regulators of chondrogenesis, respectively. The combinatorial action of these transcription factors appears to define the extent of embryonic mouse mandibular chondrogenesis. EXPERIMENTAL PROCEDURES Mouse Embryo and Mandibular Process Explant Culture Timed pregnant Swiss Webster mice were obtained (Harlan Bioproducts for Science, Indianapolis, IN) and embryos were collected at embryonic day 10 (E10) or E12. Isolated embryos were further staged by external morphology and number of somite pairs according to Theiler staging system (Theiler, 1989). Organ culture of mandibular processes from E10 embryos was performed according to previously reported protocols (Shum et al., 1993; Slavkin et al., 1999). Briefly, the mandibular processes were isolated by microdissection and explanted into a Trowell-type organ culture system using serumless, chemically-defined BGJb medium (Life Technologies Inc., Gaithersburg, MA) supplemented with 100 g/ml ascorbic acid and 100 U each of penicillin and streptomycin, and allowed to develop at 37°C and 5% CO2. Bead Implantation Affi-Gel blue agarose beads (Bio-Rad Labs., Hercules, CA) at 100 –200 mesh, which correspond to a diameter of 50 –75 m were selected, washed twice in phosphate buffered saline (PBS), and soaked in 10, 50, or 100 ng/l human recombinant BMP4 (Genetics Institute Inc., Cambridge, MA), or PBS as control for 1 hr at room temperature. BMP4-soaked or PBS control beads were positioned by fine forceps and implanted into the mandibular process explants using a mouthcontrolled micropipette under the stereomicroscope (Slavkin., et al. 1999). Whole-Mount Alcian Blue Staining The presence of sulfated proteoglycans indicative of cartilage formation was detected by Alcian Blue staining as previously described (Shum et al., 1993), using 0.04% Alcian Blue 8GX (Sigma, St. Louis, MO) in acid 411 ethanol followed by tissue clearing with a graded series of potassium hydroxide and glycerol. Histology Specimens were fixed in 4% paraformaldehyde in 0.1 M PBS (pH 7.4) for 12 hr at 4°C. Subsequently, specimens were rinsed thoroughly in 0.01 M PBS, infiltrated through a graded series of sucrose up to 30% and finally in 10% acrylamide and 30% sucrose in 0.01 M PBS, embedded first in 10% acrylamide and then in Tissue Freezing Medium (Triangle Biomedical Sciences, Durham, NC). Seven micron-thick frozen sections were collected and stained with toluidine blue. Whole-Mount In Situ Hybridization Probes for BMP4 (GenBank X56848: 776-1248nt), Sox9 (Wright et al., 1995, 926-1683nt), and type II collagen (Metsaranta et al., 1991, GenBank M65161: 29648-31343nt) were obtained by RT-PCR method and confirmed by direct sequencing. Probe for Msx2 was an AluI to EcoRI fragment of the full length Msx2 cDNA (Bell et al., 1993, GenBank L11739: 251-708nt). Digoxigenin (DIG)-labeled sense and antisense riboprobes for mouse BMP4, Sox9, Msx2, and type II collagen were prepared by in vitro transcription of linearized pBlueScript (Stratagene, La Jolla, CA) or pCRII (Invitrogen Corp., Carlsbad, CA) phagemids containing cDNA insert using RNA Transcription Kit (Stratagene, La Jolla, CA) according to specifications from the manufacturer. Embryos and mandibular process explants were fixed overnight in 4% paraformaldehyde in PBS at 4°C, and processed according to the protocol described by Rosen and Beddington (1993) with minor modifications. Briefly, after fixation, the specimens were bleached for 1 hr with 6% hydrogen peroxide in PBS containing 0.1% Tween-20 at room temperature. Mandibular explants and E10 embryos were permeabilized with RIPA buffer and E12 embryos were treated with 10 g/ml proteinase K for 15 min at room temperature to enhance permeabilization. Following post-fixation and prehybridization, the specimens were hybridized overnight with 1 mg/ml DIG-labeled riboprobes at 70°C. The specimens were washed, blocked, and incubated with anti-DIG alkaline phosphatase conjugated antibody (Boehringer Mannheim Corp., Indianapolis, IN) at a dilution of 1:2000 for 90 min at 4°C. The specimens were washed extensively after which endogenous alkaline phosphatase activities were blocked by overnight incubation in 0.48 mg/ml levamisol (Sigma, St. Louis, MO). The color reaction was developed using nitro blue tetrazolium (NBT) and 5-bromo-4-chloro-3indolyl phosphate (BCIP) as substrates (Sigma, St. Louis, MO). The area of gene expression induced by BMP4-soaked beads and visualized by whole-mount in situ hybridization was determined by first photographing the processed specimens and then subjecting the images to morphometric analyses using NIH Image Version 1.6.1 (NIH, Bethesda, MA). Signal to background was determined by density slice method within 412 SEMBA & NONAKA ET AL. the software. Student’s t-test was performed on numerical data. Semi-Quantitative RT-PCR In order to maximize the sensitivity of detection, multiple PBS- or 100 ng/l BMP4-soaked beads were implanted at either the rostral or proximal positions on one side of the E10 mandibular process and allowed to grow in explant culture system. Twenty-four hr later, each mandible was bisected along the midline in order to separate the implanted half from the non-implanted half. Four halves of the same treatment were pooled as one sample and subjected to RT-PCR for Sox9, Msx2, and beta-actin (Stratagene, La Jolla, CA). Amplimers for Sox9 were 5⬘-AAGATAAGTTCCCCGTGTGC-3⬘ and 5⬘-GTAGTGAGGAAGGTTGAAGG-3⬘, and those for Msx2 were 5⬘-TGTTTTCGTCGGATGAGGAG-3⬘ and 5⬘-GTCGCTTAGGGTGACAATGC-3⬘. Thirty, 28, and 18 cycles were empirically determined for Sox9, Msx2, and beta-actin PCR, respectively to optimize for amplification linearity and signal. The expression levels of Sox9 and Msx2 were normalized against beta-actin. The results were subjected to Student’s t-test. Adenovirus and Microinjection Recombinant adenovirus carrying haemagglutin (HA) epitope-tagged mouse Msx2 gene (AdV-Msx2) (Takahashi et al., 1998b) or LacZ gene (AdV-LacZ) as control were prepared. Ten nl of the recombinant adenovirus solution (titer of 1 ⫻ 1010 pfu/ml) accompanied by 1% tetramethylrhodamine dextran (Molecular Probes Inc., Eugene, OR), which was used as a visible marker, were microinjected into the mandibular explants using a glass micropipette and a Transjector 5246 (Eppendorf, Hamburg, Germany) under a stereomicroscope. After 36 hr of culture, explants were processed for HA immunostaining and ␤-galactosidase staining as previously described (Takahashi et al., 1998b) to confirm the site and spread of infection. Quantitative RT-PCR for Type II Collagen mRNA Type II collagen mRNA expression levels were analyzed by quantitative RT-PCR method as previously described (Takahashi et al., 1998a). Six explants were pooled for each analysis. The amount of type II collagen mRNA in the adenovirus microinjected half of the mandibular explants was compared with the control uninjected half for each AdV-LacZ and AdV-Msx2 infected specimens after 6 days in culture and reported as their relative ratio. The data were subjected to statistical analyses using Student’s t-test. ACKNOWLEDGMENTS We are grateful to Dr. Yi-Hsin Liu (University of Southern California) for mouse Msx2 cDNA, Dr. Silvio Gutkind (National Institute of Dental and Craniofacial Research) for AdV-LacZ and Ms. Audra Wright for assistance with Figure 3. We also thank Dr. Hari Reddi and Dr. Lee Niswander for valuable suggestions to this project. Human recombinant BMP4 is a gift from Genetics Institute. REFERENCES Barlow AJ, Francis-West PH. 1997. Ectopic application of recombinant BMP-2 and BMP-4 can change patterning of developing chick facial primordia. Development 124:391–398. Bi W, Deng JM, Zhang A, Behringer RR, de Crombrugghe B. 1999. Sox9 is required for cartilage formation. Nat Genet 22:85– 89. Buckland RA, Collinsosn JM, Graham E, Davidson DR, Hill RE. 1998. Antagonistic effects of FGF4 on BMP induction of apoptosis and chondrogenesis in the chick limb bud. Mech Dev 71:143–150. Bell DM, Leung KKH, Wheatley SC, Ng LJ, Zhou S, Ling KW, Sham MH, Koopman P, Tam P PL, Cheah KSE. 1997. SOX9 directly regulates the type-II collagen gene. Nat Genet 16:174 –178. Bell JR, Noveen A, Liu YH, Ma L, Dobias S, Kundu R, Luo W, Xia Y, Maxson R. 1993. Genomic structure, chromosomal location, and evolution of the mouse Hox 8 gene. Genomics 16:123–131. Bendall AJ, Rincon-Limas DE, Botas J, Abate-Shen C. 1998. Protein complex formation between Msx1 and Lhx2 homeoproteins is incompatible with DNA binding activity. Differentiation 63:151–157. Bennett JH, Hunt P, Thorogood P. 1995. Bone morphogenetic protein-2 and -4 expression during murine orofacial development. Arch Oral Biol 40:847– 854. Catron KM, Zhang H, Marshall SC, Inostroza JA, Wilson JM, Abate C. 1995. Transcriptional repression by Msx-1 does not require homeodomain DNA-binding sites. Mol Cell Biol 15:861– 871. Cottrill CP, Archer CW, Hornbruch A, Wolpert L. 1987. The differentiation of the normal and muscle-free distal chick limb bud mesenchyme in micromass culture. Dev Biol 119:143–151. Davidson D. 1995. The function and evolution of Msx genes: pointers and paradoxes. Trends Genet 11:405– 411. Derynck R, Zhang Y, Feng XH. 1998. Smads: transcriptional activators of TGF-beta responses. Cell 95:737–740. Dewulf N, Verschueren K, Lonnyoy O, Moren A, Grimsby S, Vande Spiegle K, Miyazono K, Huylebroeck D, Ten Dijke P. 1995. Distinct spatial and temporal expression patterns of two type I receptors for bone morphogenetic proteins during mouse embryogenesis. Endocrinology 136: 2652–2663. Dodig M, Kronenberg MS, Bedalov A, Kream BE, Gronowicz G, Clark SH, Mack K, Liu YH, Maxon R, Pan ZZ, Upholt WB, Rowe DW, Lichtler AC. 1996. Identification of a TAAT-containing motif required for high level expression of the COL1A1 promoter in differentiated osteoblasts of transgenic mice. Biol Chem 271:16422–16429. Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G. 1997. Osf2/ Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89:747–754. Duprez D, Bell EJ, Richardson MK, Archer CW, Wolpert L, Brickell PM, Francis-West PH. 1996a. Overexpression of BMP-2 and BMP-4 alters the size and shape of developing skeletal elements in the chick limb. Mech Dev 57:145–157. Duprez DM, Colety M, Amthor H, Brickell PM, Tickle C. 1996b. Bone morphogenetic protein-2 (BMP-2) inhibits muscle development and promotes cartilage formation in chick limb bud cultures. Dev Biol 174:448 – 452. Ekanayake S, Hall BK. 1997. The in vivo and in vitro effects of bone morphogenetic protein-2 on the development of the chick mandible. Int J Dev Biol 41:67– 81. Foerst-Potts L, Sadler TW. 1997. Disruption of Msx-1 and Msx-2 reveals roles for these genes in craniofacial, eye, and axial development. Dev Dyn 209:70 – 84. Foster JW, Dominguez-Steglich MA, Guioli S, Kwok C, Weller PA, Weissenbach J, Mansour S, Young ID, Goodfellow PN, Brook JD, Schafer AJ. 1994. Campomelic dysplasia and autosomal sex reversal caused by mutations in an SRY-related gene. Nature 372:525– 530. Francis-West PH, Tatla T, Brickell PM. 1994. Expression pattern of the bone morphometric protein genes Bmp-4 and Bmp-2 in the developing chick face suggest a role in outgrowth of the primordia. Dev Dyn 201:168 –178. Francis-West PH, Richardson MK, Bell E, Chen P, Luyten F, Adelfattah A, Barlow AJ, Brickell PM, Wolpert L, Archer CW. 1996. The Sox9 and Msx2 CO-REGULATE BMP4-INDUCED CHONDROGENESIS effect of overexpression of BMPs and GDF-5 on the development of chick limb skeletal elements. Ann NY Acad Sci 785:254 –255. Frenz DA, Liu W, Capparelli M. 1996. Role of BMP-2a in otic capsule chondrogenesis. Ann NY Acad Sci 785:256 –258. Graff JM. 1997. Embryonic patterning: to BMP or not to BMP, this is the question. Cell 89:171–174. Grigoriou M, Tucker AS, Sharpe PT, Pachnis V. 1998. Expression and regulation of Lhx6 and Lhx7, a novel subfamily of LIM homeodomain encoding genes, suggests a role in mammalian head development. Development 125:2063–2074. Hall BK. 1982. Distribution of osteo- and chondrogenetic neural crestderived cells and of osteogenically inductive epithelia in mandibular arches of embryonic chicks. J Embryol Exp Morphol 68:127–136. Hall BK, Miyake T. 1995. Divide, accumulate, differentiate: cell condensation in skeletal development revisited. Int J Dev Biol 39:881– 893. Harada S, Rodan SB, Goodfellow PA, Rodan GA. 1996. SOX9, a developmental transcription factor, stimulates type X collagen promoter via unique recognition motifs. J Bone Mineral Res 11 (supp.1):145–145. Healy C, Uwanogho D, Sharpe PT. 1996. Expression of the chicken Sox9 gene marks the onset of cartilage differentiation. Ann NY Acad Sci 785:261–262. Healy C, Uwanogho D, Sharpe PT. 1999. Regulation and role of Sox9 in cartilage formation. Dev Dyn 215:69 –78. Hogan BL. 1996. Bone morphogenetic proteins: multifunctional regulators of vertebrate development. Genes Dev 10:1580 –1594. Jernvall J, Aberg T, Kettunen P, Keranen S, Thesleff I. 1998. The life history of an embryonic signaling center: BMP-4 induces p21 and is associated with apoptosis in the mouse tooth enamel knot. Development 125:373–382. Kaplan FS, Shore EM. 1998. Encrypted morphogens of skeletogenesis: biological errors and pharmacologic potentials. Biochem Pharmacol 55:373–382. Kanzler B, Kuschert SJ, Liu YH, Mallo M. 1998. Hoxa-2 restricts the chondrogenic domain and inhibits bone formation during development of the branchial area. Development 125:2587–2597. Kawabata M, Imamura T, Miyazono K. 1998. Signal transduction by bone morphogenetic proteins. Cytokine Growth Factor Rev 9:49 – 61. Kawakami Y, Ishikawa T, Shimabara M, Tanda N, Enomoto-Iwamoto M, Iwamoto M, Kuwana T, Ueki A, Noji S, Nohno T. 1996. BMP signaling during bone pattern determination in the developing limb. Development 122:3557–3566. Kim HJ, Rice D, Kettunen P, Thesleff I. 1998. FGF-, BMP- and Shh-mediated signalling pathways in the regulation of cranial suture morphogenesis and calvarial bone development. Development 125:1241–1251. Kretzschmar M, Doody J, Massague J. 1997. Opposing BMP and EGF signalling pathways converge on the TGF-beta family mediator Smad1. Nature 389:618 – 622. Lanchoney TF, Olmsted EA, Shore EM, Gannon FA, Rosen V, Zasloff MA, Kaplan FS. 1998. Characterization of bone morphogenetic protein 4 receptor in fibrodysplasia ossificans progressiva. Clin Orthop 346:38 – 45. Langille RM. 1994. In vitro analysis of the spatial organization of chondrogenic regions of avian mandibular mesenchyme. Dev Dyn 201:55– 62. Leboy PS, Sullivan TA, Nooreyazdan M, Venezian RA. 1997. Rapid chondrocyte maturation by serum-free culture with BMP-2 and ascorbic acid. J Cell Biochem 66:394 – 403. Lee YS, Chuong CM. 1997. Activation of protein kinase A is a pivotal step involved in both BMP-2- and cyclic AMP-induced chondrogenesis. J Cell Physiol 170:153–165. Lefebvre V, Huang W, Harley VR, Goodfellow PN, de Crombrugghe B. 1997. Sox9 is a potent activator of the chondrocyte-specific enhancer of the pro alpha1(II) collagen gene. Mol Cell Biol 17:2336 –2346. Lefebvre V, Li P, de Crombrugghe B. 1998. A new long form of Sox5 (L-Sox5), Sox6 and Sox9 are coexpressed in chondrogenesis and cooperatively activate the type II collagen gene. EMBO J 17:5718 – 5733. Liu YH, Kundu R, Wu L, Luo W, Ignelzi MR Jr, Snead ML, Maxson RE, Jr. 1995. Premature suture closure and ectopic cranial bone in 413 mice expressing Msx2 transgenes in the developing skull. Proc Natl Acad Sci USA 92:6137– 6141. Lumsden A, Sprawson N, Graham A. 1991. Segmental origin and migration of neural crest cells in the hindbrain region of the chick embryo. Development 113:1281–1291. Lyons KM, Pelton RW, Hogan BL. 1989. Patterns of expression of murine Vgr-1 and BMP-2a RNA suggest that transforming growth factor-beta-like genes coordinately regulate aspects of embryonic development. Genes Dev 3:1657–1668. Maas R, ChenYP, Bei M, Woo I, Satokata I. 1996. The role of Msx genes in mammalian development. Ann NY Acad Sci 785:171–181. Macias D, Ganan Y, Sampath TK, Piedra ME, Ros MA, Hurle JM. 1997. Role of BMP-2 and OP-1 (BMP-7) in programmed cell death and skeletogenesis during chick limb development. Development 124:1109 –1117. Marazzi G, Wang Y, Sasoon D. 1997. Msx2 is a transcriptional regulator in the BMP4-mediated programmed cell death pathway. Dev Biol 186:127–138. Metsaranta M, Toman D, de Crombrugghe B, Vuorio E. 1991. Mouse type II collagen gene. Complete nucleotide sequence, exon structure, and alternative splicing. J Biol Chem 266:16862–16869. Mina M, Gluhak J, Roger B. 1996. Downregulation of Msx-2 expression results in chondrogenesis in the medial region of the avian mandible. Connect Tissue Res 35:79 – 84. 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. Monsoro-Burq AH, Duprez D, Watanabe Y, Bontoux M, Vincent C, Brickell P, Le Douarin N. 1996. The role of bone morphogenetic proteins in vertebral development. Development 122:3607–3616. Neubüser A, Peters H, Balling R, Martin GR. 1997. Antagonistic interactions between FGF and BMP signaling pathways: a mechanism for positioning the sites of tooth formation. Cell 90:247–255. Newberry EP, Boudreaux JM, Towler DA. 1997. Stimulus-selective inhibition of rat osteocalcin promoter induction and protein-DNA interactions by the homeodomain repressor Msx2. Biol Chem 272: 29607–29613. Ng LJ, Wheatley S, Muscat GE, Conway-Campbell J, Bowles J, Wright E, Bell DM, Tam PP, Cheah KS, Koopman P. 1997. Sox9 binds DNA, activates transcription and coexpresses with type II collagen during chondrogenesis in the mouse. Dev Biol 183:108 – 121. Nifuji A, Kellermann O, Kuboki Y, Wozney JM, Noda M. 1997. Perturbation of BMP signaling in somitogenesis resulted in vertebral and rib malformations in the axial skeletal formation. J Bone Miner Res 12:332–342. Niswander L, Martin GR. 1993. FGF-4 and BMP-2 have opposite effects on limb growth. Nature 361:68 –71. Nonaka K, Shum L, Takahashi I, Takahashi K, Ikura T, Dashner R, Nuckolls GH, Slavkin HC. 1999. Convergence of the BMP and EGF signaling pathways on Smad1 in the regulation of chondrogenesis. Int J Dev Biol 43:795– 807. Nuckolls GH, Slavkin HC, Shum L. 1998. Bone morphogenetic protein signaling in limb and craniofacial development. In: Davidowitch Z, editor. Proceedings of the biological mechanisms of tooth eruption, resorption and replacement by implants. Birmingham, Alabama: Harvard Society for the Advancement of Orthodontics, EBSCO Media. p 39 – 47, Osumi-Yamashita N, Ninomiya Y, Doi H and Eto K. 1994. The contribution of both forebrain and midbrain crest cells to the mesenchyme in the frontonasal mass of mouse embryos. Dev Biol 164: 409 – 419. Phippard, DJ, Weber-Hall SJ, Sharpe PT, Naylor MS, Jayatalake H, Maas R, Woo I, Roberts-Clark D, Francis-West PH, Liu YH, Maxson R, Hill RE, Dale TC. 1996. Regulation of Msx-1, Msx-2, Bmp-2 and Bmp-4 during foetal and postnatal mammary gland development. Development 122:2729 –2737. Qiu M, Bulfone A, Ghattas I, Meneses JJ, Christensen L, Sharpe PT, Presley R, Pedersen RA, Rubenstein JL. 1997. Role of the Dlx homeobox genes in proximodistal patterning of the branchial arches: mutations of Dlx-1, Dlx-2, and Dlx-1 and -2 alter morphogenesis of proximal skeletal and soft tissue structures derived from the first and second arches. Dev Biol 185:165–184. 414 SEMBA & NONAKA ET AL. Reddi AH. 1994. Bone and cartilage differentiation. Curr Opin Genet Dev 4:737–744. Reddi AH. 1995. Cartilage morphogenesis: role of bone and cartilage morphogenetic proteins, homeobox genes and extracellular matrix. Matrix Biol 14:599 – 606. Reddi AH. 1997. Bone morphogenetic protein: an unconventional approach to isolation of first mammalian morphogens. Cytokine Growth Factor Rev 8:11–20. Reshef R, Maroto M, Lassar AB. 1998. Regulation of dorsal somitic cell fates: BMPs and Noggin control the timing and pattern of myogenic regulator expression. Genes Dev 12:290 –303. Ros MA, Piedra ME, Fallon JF, Hurle JM. 1997. Morphogenetic potential of the chick leg interdigital mesoderm when diverted from the cell death program. Dev Dyn 208:406 – 419. Rosen B, Beddington RSP. 1993. Whole-mount in situ hybridization in the mouse embryo: gene expression in three dimensions. Trends Genet 9:162–167. Satokata I, Maas R. 1994. Msx1 deficient mice exhibit cleft palate and abnormalities of craniofacial and tooth development. Nat Genet 6:348 –356. Sechrist J, Serbedzija GN, Scherson T, Fraser SE, Bronner-Fraser M. 1993. Segmental migration of the hindbrain neural crest does not arise from its segmental generation. Development 118:691–703. Sekiya I, Koopman P, Watanabe H, Ezura Y, Yamada Y, Noda M. 1997. SOX9 enhances aggrecan gene expression via the promoter region containing a single HMG-BOX-sequence in a chondrogenic cell line, TC6. J Bone Miner Res 12 (supp.1):P222–P222. Semenza GL, Wang GL, Kundu R. 1995. DNA binding and transcriptional properties of wild-type and mutant forms of the homeodomain protein Msx2. Biochem Biophys Res Commun 209:257–262. Shafritz AB, Shore EM, Gannon FH, Zasloff MA, Taub R, Muenke M, Kaplan FS. 1996. Overexpression of an osteogenic morphogen in fibrodysplasia ossificans progressiva. N Engl J Med 335:555–561. Shum L, Sakakura Y, Bringas P, Luo W, Snead ML, Mayo M, Crohin C, Millar S, Werb Z, Buckely S, Hall FL, Warburton D, Slavkin HC. 1993. EGF abrogation induced fusilli-form dysmorphogenesis of Meckel’s cartilage during embryonic mouse mandibular morphogenesis in vitro. Development 118:903–917. Slavkin HC, Bringas P, Sasano Y, Mayo M. 1989. Early embryonic mouse mandibular morphogenesis and cytodifferentiation in serumless, chemically-defined medium: A model for studies of autocrine and/or paracrine regulatory factors. J Craniofac Genet Dev Biol 9:185–205. Slavkin HC, Nuckolls GH, Shum L. 1999. Craniofacial development and patterning. In: Tuan RS, Lo CW, editors. Methods in molecular biology, vol. 136: Developmental biology protocols, vol. II. Humana Press Inc., Totowa, NJ. p 45–54. Takahashi I, Nuckolls GH, Takahashi K, Tanaka O, Semba I, Dashner R, Shum L, Slavkin HC. 1998a. Compressive force induces Sox9, Type II collagen and Aggrecan and inhibits IL-1␤ expression resulting in chondrogenesis in mouse embryonic limb bud mesenchymal cells. J Cell Sci 111:2067–2076. Takahashi K, Nuckolls GH, Tanaka O, Semba I, Takahashi I, Dashner R, Shum L, Slavkin HC. 1998b. Adenovirus mediated ectopic expression of Msx2 in even-numbered rhombomeres induces apoptotic elimination of cranial neural crest cells en ovo. Development 125:1627–1635. Takahashi Y, Tonegawa A, Matsumoto K, Ueno N, Kuroiwa A, Noda M, Nifuji A. 1996. BMP-4 mediates interacting signals between the neural tube and skin along the dorsal midline. Genes Cells 1:775– 783. Theiler K. 1989. The house mouse: atlas of embryonic development . New York: Springer-Verlag. p 73–77. Thomas BL, Tucker AS, Qui M, Ferguson CA, Hardcastle Z, Rubenstein JL, Sharpe PT. 1997. Role of Dlx-1 and Dlx-2 genes in pattering of the murine dentition. Development 124:4811– 4818. Tucker AS, Matthews KL, Sharpe PT. 1998. Transformation of tooth type induced by inhibition of BMP signaling. Science 282:1136 – 1138. Tureckova J, Sahlberg C, Aberg T, Ruch JV, Thesleff I, Peterkova R. 1995. Comparison of expression of the msx-1, msx-2, BMP-2 and BMP-4 genes in the mouse upper diastemal and molar tooth primordia. Int J Dev Biol 39:459 – 468. Urist MR. 1965. Bone: formation by autoinduction. Science 150:893– 899. Urist MR. 1997. Bone morphogenetic protein: the molecularization of skeletal system development. J Bone Miner Res 12:343–346. Vortkamp A. 1997. Defining the skeletal elements. Curr Biol 7:R104 – R107. Wagner T, Wirth J, Meyer J, Zabel B, Held M, Zimmer J, Pasantes J, Bricarelli FD, Keutel J, Hustert E, Wolf U, Tommeruo N, Schempp W, Scherer G. 1994. Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell 79:1111–1120. Watanabe Y, Le Douarin NM. 1996. A role for BMP-4 in the development of subcutaneous cartilage. Mech Dev 57:69 –78. Watanabe Y, Duprez D, Monsoro-Burq AH, Vincent C, Douarin NM. 1998. Two domains in vertebral development: antagonistic regulation by SHH and BMP4 proteins. Development 125:2631–2639. Wheatley S, Wright E, Jeske Y, McCormack A, Bowles J, Koopman P. 1996. Aetiology of the skeletal dysmorphology syndrome campomelic dysplasia: expression of the Sox9 gene during chondrogenesis in mouse embryos. Ann NY Acad Sci 785:350 –352. Winnier G, Blessing M, Labosky PA, Hogan BLM. 1995. Bone morphogenetic protein-4 is required for mesoderm formation and patterning in the mouse. Genes Dev 9:2105–2116. Winograd J, Reilly MP, Roe R, Lutz J, Laughner E, Xu X, Asakura T, vander Kolk C, Strandberg JD, Semenza GL. 1997. Perinatal lethality and multiple craniofacial malformations in MSX2 transgenic mice. Hum Mol Genet 6:369 –379. Woloshin P, Song K, Degnin C, Killary AM, Goldhamer DJ, Sassoon D, Thayer MJ. 1995. MSX1 inhibits myoD expression in fibroblast x 10T1/2 cell hybrids. Cell 82:611– 620. Wozney JM, Rosen V. 1998. Bone morphogenetic protein and bone morphogenetic protein gene family in bone formation and repair. Clin Orthop 346:26 –37. Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz RW, Hewick RM, Wang EA. 1988. Novel regulators of bone formation: molecular clones and activities. Science 242:1528 –1534. Wright E, Hargrave MR, Christiansen J, Cooper L, Kun J, Evans T, Gangadharan U, Greenfield A, Koopman P. 1995. The Sry-related gene Sox-9 is expressed during chondrogenesis in mouse embryos. Nat Genet 9:15–20. Wu L, Wu H, Ma L, Sangiorgi F, Wu N, Bell JR, Lyons GE, Maxson R. 1997. Miz1, a novel zinc finger transcription factor that interacts with Msx2 and enhances its affinity for DNA. Mech Dev 65:3–17. Zhang H, Bradley A. 1996. Mice deficient for BMP2 are nonviable and have defects in amnion/chorion and cardiac development. Development 122:2977–2986. Zhang H, Catron KM, Abate-Shen C. 1996. A role for the Msx-1 homeodomain in transcriptional regulation: residues in the N-terminal arm mediate TATA binding protein interaction and transcriptional repression. Proc Natl Acad Sci USA 93:1764 –1769. Zhang H, Hu G, Wang H, Sciavolino P, Iler N, Shen MM, Abate-Shen C. 1997. Heterodimerization of Msx and Dlx homeoproteins results in functional antagonism. Mol Cell Biol 17:2920 –2932. Zhao Q, Eberspaecher H, Lefebvre V, De Crombrugghe B. 1997. Parallel expression of Sox9 and Col2a1 in cells undergoing chondrogenesis. Dev Dyn 209:377–386. Zou H, Wieser R, Massague 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.