SEARCH

SEARCH BY CITATION

Keywords:

  • mandibular processes;
  • epithelial-mesenchymal interactions;
  • FGFs;
  • FGFRs;
  • BMP-7;
  • Msx genes

Abstract

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

The mandibular processes are specified as at least two independent functional regions: two large lateral regions where morphogenesis is dependent on fibroblast growth factor (FGF)-8 signaling, and a small medial region where morphogenesis is independent of FGF-8 signaling. To gain insight into signaling pathways that may be involved in morphogenesis of the medial region, we have examined the roles of pathways regulated by FGFs and bone morphogenetic proteins (BMPs) in morphogenesis of the medial and lateral regions of the developing chick mandible. Our results show that, unlike in the lateral region, the proliferation and growth of the mesenchyme in the medial region is dependent on signals derived from the overlying epithelium. We also show that medial and lateral mandibular mesenchyme respond differently to exogenous FGFs and BMPs. FGF-2 and FGF-4 can mimic many of the effects of mandibular epithelium from the medial region, including supporting the expression of Msx genes, outgrowth of the mandibular processes and elongation of Meckel's cartilage. On the other hand, laterally placed FGF beads did not induce ectopic expression of Msx genes and did not affect the growth of the mandibular processes. These functional studies, together with our tissue distribution studies, suggest that FGF-mediated signaling (other than FGF-8), through interactions with FGF receptor-2 and downstream target genes including Msx genes, is part of the signaling pathway that mediates the growth-promoting interactions in the medial region of the developing mandible. Our observations also suggest that BMPs play multiple stage- and region-specific roles in mandibular morphogenesis. In this study, we show that exogenous BMP-7 applied to the lateral region at early stages of development (stage 20) caused apoptosis, ectopic expression of Msx genes, and inhibited outgrowth of the mandibular processes and the formation of Meckel's cartilage. Our additional experiments suggest that the differences between the effects of BMP-7 on lateral mandibular mesenchyme at stage 20 and previously reported results at stage 23 (Wang et al., [1999] Dev. Dyn. 216:320–335) are related to differences in stages of differentiation in that BMP-7 promotes apoptosis in undifferentiated lateral mandibular mesenchyme, whereas it promotes chondrogenesis at later stages of development. We also showed that, unlike mandibular epithelium and medially placed FGF beads, medially placed BMP-7 did not support outgrowth of the isolated mesenchyme and at stage 20 induced the formation of a duplicated rod of cartilage extending from the body of Meckel's cartilage. These observations suggest that BMPs do not play essential roles in growth-promoting interactions in the medial region of the developing mandible. However, BMP-mediated signaling is a part of the signaling pathways regulating chondrogenesis of the mandibular mesenchyme. © 2002 Wiley-Liss, Inc.


INTRODUCTION

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

During normal development, the lower jaw is formed from the mandibular component of the first branchial arch. After their initial formation, the mandibular processes undergo considerable growth along all three axes, merge in the midline region, and give rise to the elongated triangular-shaped lower jaw. Studies in avian and mammalian embryos showed that the skeletal elements of the mandibular arch are derived from mixed population of cranial neural crest cells (CNCC) originating from posterior mesencephalon and multiple rhombomeres (R) (Couly et al., 1996, 1998; Kontges and Lumsden, 1996; Osumi-Yamashita et al., 1997). More recent labeling studies in avian and mammalian embryos showed contribution of non–CNCC-derived cells to mandibular skeleton (Sohal et al., 1999; Chai et al., 2000).

For many years, it was thought that the patterning of the mandibular skeleton is the result of morphogenetic prespecification in premigratory CNCC before their arrival in the mandibular arch (Noden, 1983; Gendron-Maguire et al., 1993; Rijli et al., 1993, 1998; Couly et al., 1998). However, more recent observations suggest that morphogenesis of the mandibular skeleton is also regulated by several evolutionary conserved signaling factors and transcription factors after the arrival of CNCC in the mandibular processes (Hunt et al., 1995, 1998; Saldivar et al., 1997).

Classic tissue recombination studies indicate that outgrowth of the mandibular processes is dependent on epithelial-mesenchymal interactions (Hall, 1982; Wedden, 1987; Coffin-Collins and Hall, 1989; Richman and Tickle, 1989; Hall and Coffin-Collins, 1990; Mina et al., 1994). After removal of the overlying epithelium, reduced outgrowth of the mandibular mesenchyme is seen, suggesting that signals derived from mandibular epithelium regulate growth of the underlying mesenchyme. In addition, tissue recombination studies in which epithelia and mesenchyme from different facial processes were exchanged suggest similarities in the signaling cascades regulating outgrowth of all the facial processes (Richman and Tickle, 1989). Outgrowth of the mandibular mesenchyme is supported to some degree by epithelia covering other facial processes, and mandibular epithelium supports the outgrowth of the mesenchyme from other facial processes. Although these recombination studies clearly indicate mitogenic roles of mandibular epithelium on the underlying mesenchyme, they did not identify specific regions in the epithelium or mesenchyme involved in regulating the outgrowth of the developing mandible. Nonetheless, in recent years there has been significant progress in understanding the putative regions and the signaling networks controlling the outgrowth and development of the mandibular processes (for review, see Francis-West et al., 1998; Mina, 2001a,b).

Our previous observations of the differences between the effects of medial vs. the lateral chick mandibular epithelium on the induction of Msx1 and Msx2 in the underlying mesenchyme suggested differences in the signaling cascade regulating morphogenesis of the medial vs. lateral regions of the developing mandible (Wang et al., 1998, 1999). Unlike lateral mandibular epithelium, epithelium from the medial region (that overlies the Msx expressing mesenchyme in vivo) is able to induce ectopic expression of Msx genes in chick lateral mandibular mesenchyme. These observations also indicate that signaling events induced by epithelium of the medial region are similar to those induced by early odontogenic epithelium but are different from those induced by the epithelium of the lateral regions of the mandible (Wang et al., 1998, 1999).

The most direct genetic evidence for differences in the signaling cascades in the medial vs. lateral regions comes from a recent study in which Fgf8 was inactivated in the epithelium of the first branchial arch of transgenic mice by using Cre/lox technology (Trumpp et al., 1999). Newborn mutants lack most first branchial arch-derived structures except those that develop from the most anterior regions of the maxillary arch and from the medial region of mandibular arch. In the mandibles of these mutants, the body of Meckel's cartilage, mandibular bones, molar teeth, and middle ear (except for malleus) were absent. However, the mandibular incisors and their associated bones and the symphysial portion of Meckel's cartilage were invariably present. The abnormalities in the lateral, but not medial regions, of the mandibles of these mutants provide clear evidence for at least two independent genetic pathways in mandibular morphogenesis. Morphogenesis of the large lateral region is dependent on and controlled by FGF-8, whereas the small medial region depends on signals other than FGF-8.

The lack of involvement of FGF-8 in morphogenesis of the medial region leaves open the question of which signal(s) are involved in morphogenesis of this region. The expression of several other signaling molecules and transcription factors in the epithelium and mesenchyme of the medial region (for review see Francis-West et al., 1998; Mina, 2001a,b) are suggestive of their roles in morphogenesis of the medial region. Based on their patterns of expression, members of bone morphogenic protein (BMP) and fibroblast growth factor (FGF) (other than FGF-8) families of signaling molecules are also likely candidates involved in regulating morphogenetic of the medial region. However, the roles of BMPs in outgrowth of the mandibular mesenchyme remain unclear. Our previous studies (Wang et al., 1999) indicated that, despite their ability to regulate cell proliferation and Msx expression, laterally or medially placed BMP-4 and BMP-7 do not support outgrowth of chick mandibular mesenchyme from stage 23. On the other hand, studies by others (Barlow and Francis-West, 1997; Ekanayake and Hall, 1997) indicate negative effects of BMP-2 and BMP-4 on growth of the chick mandibular process at earlier stages of development. Previous observations also suggested that FGFs may be involved in the outgrowth and morphogenesis of the facial prominences (Richman et al., 1997). However, in their study, the effects of FGF-2 and FGF-4 on growth of only the medial region of chick mandibular arch at stage 24 were examined.

In the present study, we have examined and compare the roles of FGF- and BMP-mediated signaling pathways in the morphogenesis of the medial region of the developing chick mandible at different stages of development. The effects of these signaling factors on growth and morphogenesis of the mesenchyme of the medial region were also compared with their effects of the mesenchyme of the lateral regions of the mandible.

RESULTS

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

Outgrowth of the Medial Region Is Dependent on Epithelial-Mesenchymal Interactions

To gain further insight into the differences between the regulatory networks controlling morphogenesis of the medial vs. lateral mandibular mesenchyme, the roles of epithelial-mesenchymal interactions in the outgrowth and morphogenesis of these two regions were examined. In these experiments, segments of tissues isolated from the medial and lateral regions (Fig. 1) of stage 22 chick mandibles with and without their associated epithelium were grafted for 1 week into graft sites prepared on the dorsal side of a stage 22 chick wing buds. Cartilage rods of similar lengths formed in explants of lateral mandibular mesenchyme grown with or without epithelium (Table 1, Fig. 2A,B). On the other hand, in contrast to the small rounded cartilages formed in the explants of isolated mesenchyme from the medial region, homotypic recombinations from this region formed long rod-shaped cartilages (Table 1, Fig. 2C,D), which were approximately 55% of the length of the cartilage rods formed in the homotypic recombinations of the whole mandible (Table 1), and 30% of the length of the cartilage formed in intact mandible of the host animal at stage 37 (8.5 ± 1 mm, n = 10). Grafting experiments using tissues from stages 20 and 24 mandibles revealed similar results (data not shown).

thumbnail image

Figure 1. Line drawings illustrating frontal views of the developing mandibular arch at early stages of development indicating the regions and measurements for isolation of tissue fragments used in different experiments in this study. The positions for fibroblast growth factor beads on the medial or lateral regions of half of the isolated whole mandibular mesenchyme used in bead implantation and chorioallantoic membrane cultures are shown in A by filled circles. The 300-μm medial region used for grafting in the chick limb buds and for micromass cultures is indicated by horizontal line pattern in (B). The 300 μm from the lateral regions used for grafting in the chick limb buds is indicated by an XX pattern in B. The whole lateral regions used for micromass cultures is indicated by oblique line pattern in B.

Download figure to PowerPoint

Table 1. Length of the Cartilage Rods in Medial and Lateral Mandibular Mesenchyme After Explantation Into the Chick Wing Buds
Region of the mandibleLength of cartilage in mm
300 μm medial
 Mesenchyme (n = 20)0.5 ± 0.2
 Mesenchyme + epithelium (n = 10)2.6 ± 0.5
300 μm lateral
 Mesenchyme (n = 6)2.06 ± 0.1
 Mesenchyme + epithelium (n = 10)2.08 ± 0.3
Whole mandible
 Mesenchyme (n = 10)2.7 ± 0.6
 Mesenchyme + epithelium (n = 10)4.7 ± 0.5
thumbnail image

Figure 2. Cleared whole-mounts of limbs containing grafts of medial and lateral regions of mandible. Cartilage rods of similar lengths (indicated between arrowheads) formed in explants of lateral mandibular mesenchyme grown with (A) or without (B) epithelium. C: Mesenchyme from the medial region grown with epithelium formed two relatively long rods of cartilage separated by the symphysis. D: A small rounded cartilage (indicated by arrow) formed in mesenchyme from the medial region grown without epithelium. Epi+, with epithelium; Epi-, without epithelium. Scale bar = 1 mm in C (applies to A–D).

Download figure to PowerPoint

These observations indicate that, unlike the growth of the lateral region, the proliferation and growth of the mesenchyme in the medial region is dependent on signals derived from its overlying epithelium. Furthermore, these observations suggest that the 300-μm medial region makes significant contributions to the overall growth of the developing mandible and extend results reported previously by Richman and Tickle, (1989).

Patterns of Expression of Fgfs and Fibroblast Growth Factor Receptors (Fgfrs) in the Developing Chick Mandible

To examine the possible involvement of FGF-mediated signaling in morphogenesis of the medial region, the temporal and spatial distributions of mRNAs for Fgfrs1–3, Fgf4, and Fgf8 were examined in the developing chick mandible by using in situ hybridization.

At early stages of development (stages 18–21) Fgfr1, Fgfr2, and Fgfr3 were expressed uniformly throughout the mandibular mesenchyme (data not shown). However, after stage 22/23, the expression of Fgfr2 and Fgfr3 but not Fgfr1 became restricted to specific regions (Figs. 3, 4). At these stages, Fgfr2 was expressed in the mesenchyme of the medial region and lower border of the mandibular process (Figs. 3A, 4A). Fgfr2 was also expressed in the hyoid mesenchyme and in the mandibular epithelium (Figs. 3A, 4A). At the time of initiation of chondrogenesis (stage 25), Fgfr2 was expressed in the condensing mesenchyme of Meckel's cartilage and in the mandibular epithelium (Fig. 3C). The Fgfr2 (cek3) probe used in this study recognizes both isoforms of Fgfr2 that are generated by alternative splicing around exons 8 and 9. However, the results of previous studies (Peters et al., 1992; Orr-Urtreger et al., 1993; Peters et al., 1993; Kettunen et al., 1998) have shown that the Fgfr2 isoform containing the IIIb exon (Fgfr2b) is expressed predominately by epithelia of various tissues, whereas Fgfr2 isoform containing the IIIc exon (Fgfr2c) is found mainly in mesenchyme adjacent to the epithelia that express Fgfr2b. These observations suggest that the hybridization signals in the mesenchyme of the developing mandible most likely reflects the abundant expression of Fgfr2c isoform of Fgfr2 whilst the signals in the epithelium reflect the abundant expression of Fgfr2b isoform.

thumbnail image

Figure 3. Expression of Fgfr2, Fgfr3, Fgf4 in the developing chick mandibular arch. Darkfield (A–D) and corresponding brightfield (A′–D′) autoradiographs of frontal sections through stage 22/23 (A,B), 26 (C), and stage 18 (D) chick lower face hybridized with chick Fgfr2 (A,C), Fgfr3 (B), and Fgf4 (D) riboprobes. A,A′: At stage 22/23, Fgfr2 is expressed in the entire mesenchyme of the hyoid arches, in the mesenchyme at the lower border and medial regions of the developing mandible, and in the mandibular epithelium. B,B′: At stage 22/23, Fgfr3 is expressed at high levels in the hyoid and mandibular mesenchyme surrounding the first branchial cleft. C,C′: At stage 26, Fgfr2 is also strongly expressed in the condensing region of the developing Meckel's cartilage and in the mandibular epithelium. D,D′: Note the expression of Fgf4 in mandibular epithelium, and epithelium of the first and second branchial clefts separating the hyoid arch from the mandibular and the third branchial arches (indicated by arrows) at stage 18. Scale bars = 200 μm in A′ (applies to A,A′), in B′ (applies to B,B′), in C′ (applies to C,C′), 100 μm in D′ (applies to D,D′).

Download figure to PowerPoint

thumbnail image

Figure 4. Whole-mount in situ hybridizations with antisense probes for Fgfr2 (A), Fgfr3 (B), Fgf8(C), Fgf4 (D), Msx1 (E), and Msx2 (F). A: At stage 23, Fgfr2 is expressed in the hyoid arch and in the lower border of the mandibular arch extending to the medial region of the developing mandible. B: Note that, at stage 23, the hybridization signals for Fgfr3 is localized to the hyoid and mandibular mesenchyme surrounding the first branchial cleft. (indicated by arrows) C: At stage 23, Fgf-8 expression is restricted to the epithelium covering the lateral part of the mandible and is excluded from the epithelium covering the medial region. D: At stage 20, Fgf-4 is expressed throughout the mandibular epithelium and the branchial clefts. At stage 23, Msx1 (E) and Msx2 (F) are expressed in the medial region of the developing mandible. mx, maxillary process; md, mandibular process. Scale bars = 250 μm in A–F.

Download figure to PowerPoint

At stage 22/23, Fgfr3 was expressed in the hyoid and mandibular mesenchyme surrounding the first branchial cleft (Figs. 3B, 4B) and later at stage 25 in the condensing mesenchyme of Meckel's cartilage (data not shown). The expression patterns of Fgfrs in our study are consistent with and extend the results of Wilke et al. (1997).

At early stages of development, Fgf4 (Fig. 3D) and Fgf8 (not shown) were expressed in epithelium of the first and second branchial clefts. Between stages 18 and 25, Fgf8 expression was restricted to the epithelium covering the lateral part of the mandible (Fig. 4C), whereas Fgf4 was expressed in the entire epithelium of the upper surface of the mandibular arch (Fig. 4D).

The patterns of expression of Fgf4 in the branchial clefts and Fgf8 in the developing chick mandible in our study are consistent with previously reported results (Niswander and Martin, 1992; McGonnell et al., 1998). However, in contrast to the previously reported restricted expression of Fgf4 in the epithelium covering the medial region of the chick mandibular arch (Barlow and Francis-West, 1997), our studies show that Fgf4 mRNA is expressed throughout the mandibular epithelium (Figs. 3D, 4D). The pattern of expression of Fgf4 mRNA in the mandibular epithelium is similar to the pattern of expression of FGF-2 protein (Richman et al., 1997).

FGFs Can Maintain Expression of Fgfr2 and Msx Genes in the Medial Mandibular Mesenchyme

High levels of expression of Fgfr2 in the medial region (Figs. 3A, 4A), and Fgfr2 and Fgfr3 in the mesenchyme in the lower border of the developing mandible (Figs. 3A,B, 4A,B) identify cells that can respond to FGF secreted by the overlying epithelium. The high level of expression of Fgfr2 in the mesenchyme of the medial region is correlated with the domains of expression of Msx1 and Msx2 (Fig. 4E,F; Mina et al., 1995), suggesting that FGFs secreted by the epithelium of the medial region, through regulation of Msx1 and Msx2, may be involved in morphogenesis of the medial region. In addition, it has been shown that expression of Msx genes and Fgfr2 in facial processes is dependent on epithelial-mesenchymal interactions (Takahashi et al., 1991; Brown et al., 1993; Mina et al., 1995; Matovinovic and Richman, 1997).

To investigate whether FGFs may be the epithelially derived signals that regulate expression of Msx and Fgfr2 genes, we examined the effects of beads soaked in different concentrations FGF-2, FGF-4, and FGF-8 on the expression of Fgfr2, Msx1, and Msx2 in medial mandibular mesenchyme isolated from stage 22 chick embryos and compared them with the effects of mandibular epithelium (Fig. 1). In the absence of epithelium, medially placed beads soaked in all concentrations of these FGFs, and medial mandibular epithelium, but not BSA, induced a translucent zone (10/10), cell proliferation (10/10), and maintained the expression of Msx1 (10/10), Msx2 (10/10), and Fgfr2 (10/10) in the medial mandibular mesenchyme (Fig. 5A–E, and data not shown; in each set, n = 10).

thumbnail image

Figure 5. Effects of fibroblast growth factor (FGF) -4 on the expression of Msx genes and Fgfr2 and cell proliferation in chick mandibular mesenchyme after 24 hr. Beads soaked in 500 ng/μl of FGF-4 (A–D,F–I), and 1 μg/μl of bovine serum albumin (BSA; E,J) were applied to the medial (A–E) and lateral region (F–J) of stage 22 chick mandibular mesenchyme. Explants were fixed after 24 hr and processed for whole-mount in situ hybridization by using antisense probes for Msx1 (A,F), Msx2 (B,G), Fgfr2 (C,E,H,J). Regions in which probes have hybridized are stained purple. Note that medially placed FGF-4 (A–D) maintained Msx-1 (A), Msx-2 (B), and Fgfr2 (C) expression in the mesenchyme of the medial region. On the other hand, laterally placed FGF-4 beads (F–I) induced the expression of Msx-1 (F) but not Msx-2 (G) or Fgfr2 (H) in the mesenchyme of the lateral region. Medially placed (D) and laterally placed (I) FGF-4 induced cell proliferation in the adjacent mesenchyme. Also note the lack of hybridization signal for Fgfr2 around the medially (E) and laterally (J) placed BSA beads. Scale bar = 200 μm in J (applies to A–J).

Download figure to PowerPoint

To distinguish between inductive vs. maintenance roles of FGFs on the expression of Msx and Fgfr2 genes in mandibular mesenchyme, FGF-2, FGF-4, and FGF-8 were also applied to the lateral region of the mandibular mesenchyme where Msx and Fgfr2 genes are not normally expressed (Figs. 1A, 4E,F). In the absence of epithelium, laterally placed FGF beads were also able to induce the formation of a translucent zone (10 of 10), cell proliferation (10 of 10), ectopic expression of Msx1 (5 of 10), but not Msx2 (0 of 10) or Fgfr2 (0 of 10) in the lateral mandibular mesenchyme (Fig. 5F–J and data not shown; in each set, n = 10). FGF-mediated ectopic expression of Msx1 in the lateral mandibular mesenchyme was only observed in 50% (5 of 10) of the explants and only after 24 hr of color reaction.

The extent of changes induced by FGFs in the mandibular mesenchyme was concentration dependent and changes were seen with concentrations as low as 10 ng/μl (data not shown). In addition, no apparent differences were observed between the effects of different FGFs on the underlying mesenchyme. The similarities in the response of mandibular mesenchyme to relatively high concentrations of different FGFs observed in our studies are consistent with the effects of different FGFs in most other functional studies. However, despite these similarities, based on their patterns of expression, FGFs secreted by the epithelium of the medial region, FGF-4 and FGF-2 (but not FGF-8), are likely endogenous ligands that are involved in morphogenesis of the medial region through interactions with FGFR2, Msx1, and Msx2.

Applications of FGFs to Medial but Not Lateral Mandibular Mesenchyme Support Outgrowth

Next, the ability of FGFs to support the outgrowth and elongation of Meckel's cartilage in the absence of epithelium was examined by using in vitro chorioallantoic membrane (CAM) cultures. Explants of stage 22 isolated whole mandibular mesenchyme (Fig. 1A) grown for 7 days in the absence of mandibular epithelium formed two short rods of cartilages, whereas explants grown in the presence of mandibular epithelium (homotypic recombinations) formed two long rods of cartilages (Fig. 6A,B; Table 2). The length of cartilage rods in the homotypic recombinations in CAM cultures were longer than those formed by isolated mandibular mesenchyme but shorter than the cartilage rods in intact mandibles at stage 37 (Table 2). Application of beads soaked in FGF-4 and FGF-2 to the mesenchyme of the medial region (Fig. 1A) resulted in significant outgrowth in the mandibular mesenchyme and elongation of Meckel's cartilage in the treated sides (Fig. 6C; Table 2). The lengths of the rods of cartilage formed in explants treated with 500 ng/μl of FGF-4 were similar to the lengths of those formed in homotypic recombinations (Table 2). Medially placed beads soaked in 100 ng/μl of FGF-4 also supported outgrowth of the mandibular mesenchyme and the length of the cartilages in these explants were 70% (3.1 ± 0.3) of those in homotypic recombinations (Table 2). Application of high (500 ng/μl) but not low (100 ng/μl) concentrations of FGF-8 to the medial mandibular mesenchyme did not supported significant outgrowth of the mandibular mesenchyme and Meckel's cartilage (Table 2).

thumbnail image

Figure 6. Effects of fibroblast growth factor (FGF) -2 on outgrowth of mandibular mesenchyme and Meckel's cartilage. Cleared whole-mounts of stage 22 mandibular mesenchymes grown for 7 days in CAM cultures. A: Two long rods of Meckel's cartilage of similar length are formed by mandibular mesenchyme grown in the presence of epithelium. B: Two short rods of Meckel's cartilage of similar length are formed by mandibular mesenchyme grown in the absence of epithelium. C: Note the elongated rod of cartilage in the treated side compared with the untreated side of explant in which a bead soaked in FGF-2 was placed on medial region on one side of the mandible. D: Also note the formation of two short cartilage rods in treated and untreated sides of explants in which a bead soaked in FGF-2 was placed on lateral region on one side of the mandible. Scale bar = 500 μm in C (applies to A–D).

Download figure to PowerPoint

Table 2. Region-Specific Effects of FGF-2 and FGF-4 on Elongation of Meckel's Cartilage in CAM Culturesa
Explant typeLength of the cartilage
  • a

    Beads were soaked in 500 ng/μl FGF-2, FGF-4, FGF-8, and 1 μg/μl BSA. Student t-test was used to compare the lengths of the cartilages. Length of the Meckel's cartilage in explants of mandibular mesenchyme in millimeters. FGF, fibroblast growth factor; CAM, chorioallantoic membrane; BSA, bovine serum albumin.

  • *

    P < 0.05.

Mesenchyme (n = 7)2.6 ± 0.5
Epithelium + mesenchyme (n = 7)4.5 ± 0.9
BSA on the medial mesenchyme (n = 7)2.6 ± 0.6
BSA + lateral mesenchyme (n = 10)2.55 ± 0.5
FGF-4 on the medial mesenchyme (n = 10)4.19 ± 0.9
FGF-8 on the medial mesenchyme (n = 5)3.1 ± 1
FGF-2 on the medial mesenchyme (n = 10)4.3 ± 0.5*
FGF-4 on the lateral mesenchyme (n = 10)2.4 ± 0.4
FGF-2 on the lateral mesenchyme (n = 10)2.6 ± 0.5
FGF-8 on the lateral mesenchyme (n = 5)2.7 ± 0.7

In contrast to their effects on the medial mesenchyme, application of beads soaked in similar concentrations of FGFs to the lateral regions of the isolated whole chick mandibular mesenchyme (Fig. 1A) did not affect the growth of the mandibular process or Meckel's cartilage in treated sides (Fig. 6D; Table 2, and data not shown). The lengths of the cartilage rods in treated sides were similar to those in the untreated sides and similar to the lengths of cartilages formed by mandibular mesenchyme grown in the absence of mandibular epithelium (Table 2). Experiments using tissues from stages 20 and 24 mandibles revealed similar results (data not shown).

Our observations of the effects of medially placed FGF-2 and FGF-4 are similar but not identical to the results reported by Richman et al. (1997), who showed that the length of the cartilage rods formed in FGF-treated grafts were 60% of the length of the cartilage rods formed in homotypic recombinations. The differences in the results of these studies most likely are related to the differences in grafting procedures. In the study by Richman et al. (1997), FGF beads were applied to the central 500 μm of stage 24 mandibular arch, and then grown for 1 week in the dorsal surface of a stage 22 chick wing bud. Therefore, it is possible that the lack of full outgrowth in the mesenchyme implanted with FGF beads in their study may be due to the inhibitory effects of the dorsal limb ectoderm on the mandibular mesenchyme and not related to the inability of FGFs to support full outgrowth. Our observations indicate that, in the absence of epithelium, medially but not laterally placed FGFs can support the outgrowth of the mandibular mesenchyme and the elongation of Meckel's cartilage, indicating that medial and lateral mandibular mesenchyme respond differently to exogenous FGFs.

FGFs Stimulate Proliferation of the Medial Mandibular Mesenchyme

The differences between the effects of FGFs on the lateral vs. medial mandibular mesenchyme were further investigated in high density micromass cultures. Micromass cultures have been used extensively to examine the chondrogenesis of the mesenchymal cells from a variety of sites, including limb and craniofacial regions (reviewed by Daniels et al., 1996).The sequence of events leading to chondrogenesis in these cultures has been shown to parallel the events in vivo with the formation of cellular aggregates or condensations followed by the formation of cartilage nodules (Daniels et al., 1996). Our previous observations, consistent with the results reported by others (Wedden et al., 1986; Mina et al., 1991; Daniels et al., 1996; Gluhak et al., 1996), indicated that when dissociated mandibular mesenchyme from stage 23 is placed in micromass cultures, distinct cell aggregates appear in the first 24 hr. After their formation, the focal aggregates differentiate into cartilage nodules. The Alcian-blue stainable cartilage nodules become apparent at 48 hr of culture and increase thereafter.

By using high-density micromass cultures, we have shown previously that the mesenchyme in the 300-μm medial region lacks in vitro chondrogenic potential, whereas extensive chondrogenesis is observed in the lateral mandibular mesenchyme (Figs. 7A, 8A, D; Mina et al., 1995). Here, we have examined the effects of FGF-2 and FGF-4 on proliferation (BrdU labeling), chondrogenesis (the amount of guanidinium HCl-extractable Alcian blue-stained matrix), and expression of mRNAs for type II collagen and aggrecan core protein (Northern blot hybridization) in micromass cultures derived from lateral and medial mandibular mesenchyme (Fig. 1).

thumbnail image

Figure 7. Region-specific effects of fibroblast growth factor (FGF) -4 on proliferation and chondrogenesis of mandibular mesenchyme. A: Quantification of the amount of cartilage matrix production as measured by Alcian blue extraction in 4-day-old micromass cultures derived from the medial and lateral mandibular mesenchyme in the presence and absence of various concentrations of FGF-4. After 4 days, cultures were fixed and stained with Alcian blue. Bound stain was extracted with 8 M guanidine chloride and quantitated spectrophotometrically. The data are means of values from 6 to 12 pairs of separate experiments ± SD. B: Percentage of BrdU-labeled cells in micromass cultures derived from the medial and lateral mandibular mesenchyme grown in control media and in media with 2.5 ng/ml of FGF-4. Each value represent the mean of 5–10 separate experiments ± SD. Cells were labeled for 1 hr before harvest.

Download figure to PowerPoint

thumbnail image

Figure 8. Region-specific effects of fibroblast growth factor (FGF) -2 and bone morphogenetic protein (BMP) -7 on in vitro chondrogenesis of mandibular mesenchyme. A–F: Whole-mount staining of micromass cultures derived from the medial (A–C) and lateral (D–F) mandibular mesenchyme grown in control media (CT) (A,D) and in media with 2.5 ng/ml of FGF-2 (B,C) and 2.5 ng/ml of BMP-7 (E,F). A: Note the absence of cartilage nodules after 4 days in control-untreated cultures derived from the medial mandibular mesenchyme. B,C: Alcian blue–positive nodules in 2-day- (B) and 4-day-old (C) cultures of medial mesenchyme treated with 2.5 ng/ml of FGF-2. D: Note the presence of discrete cartilage nodules in control-untreated culture derived from the lateral mandibular mesenchyme. E: Note the reduced number of cartilage nodules after 4 days in culture of lateral mesenchyme grown for 4 days in the presence of 2.5 ng/ml of BMP-7. F: Addition of 2.5 ng/ml of BMP-7 to cultures of lateral mesenchyme 30 hr after initiation of culture stimulated chondrogenesis. Scale bars = 1 mm in A–F.

Download figure to PowerPoint

Analyses of BrdU incorporation suggested that differences in the in vitro chondrogenic potential of medial vs. lateral regions are related in part to differences in the rates of cell proliferation and the resulting increase in cell density (Fig. 7B). In contrast to the increases in cell number in micromass cultures derived from the lateral region, there were gradual decreases in the number of labeled nuclei in micromass cultures derived from the medial region during the first 48 hr (Fig. 7B).

Exogenous FGFs modulated proliferation and chondrogenesis of the micromass cultures derived from lateral and medial mandibular mesenchyme in a dose-dependent manner (Figs. 7, 8). In contrast to control untreated cultures, micromass cultures of the medial region treated with FGF-2 and FGF-4 showed mesenchymal condensations and Alcian-blue–positive nodules after 2 days (Fig. 8B and data not shown). During the next 2 days, the cartilaginous nodules increased in size and stained more strongly with Alcian blue (Fig. 8C, and data not shown). The modulation of chondrogenesis in FGF-treated cells was preceded by changes in the number of BrdU-positive cells. As shown in Figure 7B, in contrast to the control-untreated cultures, cultures treated with 2.5 ng/μl of FGF-4 had continuous increases in the number of labeled cells during the 48 hr of culture. Northern blot hybridization indicates that stimulation of chondrogenesis in FGF-treated cultures of the medial region is accompanied by the appearance of mRNAs for chondrogenic markers (type II collagen and aggrecan) after 24 hr and that the levels of these mRNAs increase thereafter compared with control cultures (data not shown).

In contrast to their effects on the medial mandibular mesenchyme, treatment of cultures from the lateral region for 4 days with FGF-2 and FGF-4 resulted in dose-dependent inhibition of chondrogenesis and did not stimulate proliferation in the micromass cultures from the lateral region (Fig. 7, and data not shown). These studies show that FGF can stimulate proliferation of the mandibular mesenchyme in the medial but not lateral regions.

Our observations of the effects of FGF-2 and FGF-4 on proliferation and chondrogenesis are consistent with and extend the results of Richman and Crosby (1990). By using micromass cultures from the whole mandible, these authors showed that addition of various concentrations of bFGF (FGF-2) did not support significant proliferation and reduced the extent of chondrogenesis.

In Vivo Studies Examining the Effects of FGFs on the Behavior of Lateral and Medial Mandibular Mesenchyme

The effects of FGF on the expression of Msx genes and outgrowth of the mandibular processes were also examined in vivo after implantation of FGF-soaked beads at different stages (stages 20 and 23) into different regions of the developing mandible. After bead implantation, host embryos were reincubated for 1 additional day (to examine the gene expression) or 7 days (to examine the skeletal changes).

In vivo implantation of beads soaked in FGF-2 and FGF-4 in the lateral regions of the mandibular processes, in all cases (for each set n = 10) did not induce ectopic expression of Msx1 (Fig. 9B) and Msx2 (Fig. 9D) genes around the beads. However, laterally placed FGFs up-regulated Msx1 expression in the maxillary primordia and extended Msx1 expression in the region of the hyomandibular cleft in the treated side (Fig. 9B). Furthermore, laterally placed FGFs did not affect the growth of the mandibular mesenchyme (Fig. 10D). Implantation of FGF beads to the lateral region of stage 20 mandibles (but not stage 23) resulted in the loss of the cartilaginous elements (quadrate, the retroarticular process, and medial and lateral processes) in the articulating end of Meckel's cartilage in the treated side (Fig. 10E).

thumbnail image

Figure 9. Effects of bone morphogenetic protein (BMP) -7 and fibroblast growth factor (FGF) -4 on the expression of Msx1 and Msx2 in chick mandibular arch in vivo. Beads soaked in 500 ng/μl FGF-4 (A–D) and 100 ng/μl of BMP-7 (E–H) were implanted in vivo to medial (A,C,E,G) and lateral (B,D,F,H) regions of stage 20 chick mandibles. Embryos were fixed 24 hr after bead implantation and processed for whole mount in situ hybridization with antisense chick-specific Msx1 (A,B,E,F) and Msx2 (C,D,G,H) probes. Beads were implanted into the right half of the mandible (left sides of the pictures) with the left half undisturbed (right sides of the pictures). A: Medially placed FGF-4 extended the domain of Msx1 expression more laterally in the treated side. B: Laterally placed FGF-4 did not induce ectopic expression of Msx1 around the bead. Note that hybridization signal for Msx1 in the maxillary primordia in the treated side is more intense than and in the untreated side. Also note the extended Msx1 expression in the region of the hyomandibular cleft in the treated side as compared with the untreated side. C: Medially placed FGF-4 did not change the domain of Msx2 expression in the medial region. D: Laterally placed FGF-4 did not induce ectopic expression of Msx2 in the lateral mandibular mesenchyme. E: Medially placed BMP-7 extended domains of expression of Msx1 in the medial region and down-regulated Msx-1 expression in the region of the hyomandibular cleft. F: Laterally placed BMP-7 induced ectopic expression of Msx1 throughout the mandibular mesenchyme in the treated side. G: Medially placed BMP-7 also extended domains of expression of Msx2 in the medial region of the developing mandible more laterally in the treated side. H: Laterally placed BMP-7 induced ectopic expression of Msx2 throughout the mandibular mesenchyme in the treated side. Scale bar = 500 μm in H (applies to A–H).

Download figure to PowerPoint

thumbnail image

Figure 10. Region-specific effects of local application of bone morphogenetic protein (BMP) -7 and fibroblast growth factor (FGF) -4 on the outgrowth of the mandible primordia and Meckel's cartilage. Effects of local application of BMP-7 and FGF-4 on the development of mandibular primordia and Meckel's cartilage as seen in cleared and stained whole-mount preparations of chick heads. Beads soaked in 100 ng/μl (A,B) and 50 ng/μl (C) of BMP-7 were applied to lateral (A) and medial (B,C) regions of the developing mandible. Note that both laterally and medially placed BMP-7 at both concentrations inhibited growth of the developing mandible in the treated sides resulting in deviation of the mandible to one side. Also note that, whereas laterally placed BMP-7 beads inhibited the formation of Meckel's cartilage in the treated side (A), medially placed beads induce an ectopic rod of cartilage extending from the body the Meckel's cartilage (B,C). D: Overexpression of FGF-4 in the lateral region of stage 20 chick mandible for 7 days did not affect the growth of the mandibular mesenchyme and the formation of Meckel's cartilage but affected the cartilaginous elements in the articulating end of the Meckel's cartilage in the treated side. E: Note that, in contrast to the presence of the three processes in the articulating ends of the Meckel's cartilage in the untreated side (right side of the picture), only one process is present in the articulating end of Meckel's cartilage grown for 7 days with laterally placed FGF-4 (left side of the picture). Also note the significant reduction in the quadrate in the treated side. a, angular bone; l, lateral process of Meckel's cartilage; m, medial process of Meckel's cartilage; p, pterygoid bone; q, quadrate; r, retroarticular process of Meckel's cartilage; sa,; s, splenial; supra-angular bone. F,G: Nile blue staining of chick face showing the patterns of cell death 24 hr after application of 500 ng/μl of FGF-4 (F) and 100 ng/μl of BMP-7 (G) to the lateral region of stage 20 chick mandible. The dark blue staining show regions of cell death. Note that in contrast to lack of staining in mandible treated with FGF (F), application of BMP-7–induced cell death throughout the mandible in the treated side (G). Scale bars = 1 mm in A–G.

Download figure to PowerPoint

Beads soaked in FGFs implanted in vivo in a medial position at stages 20 and 23 extended Msx1 domains of expression laterally in the treated side (Fig. 9A, n = 10) but did not affect the domain of expression of Msx2 (Fig. 9C, n = 10). However, in contrast to the results of our in vitro studies, in vivo application of FGF-4 to more medial positions at stage 20 and 23 did not affect the growth of the mandibular mesenchyme (data not shown). One of the major differences between the in vivo and in vitro studies is the presence and absence of the mandibular epithelium. It is possible that the lack of stimulation of outgrowth by medially placed beads in vivo (in the presence of epithelium) is the result of the presence of a preexisting maximal proliferative signal from the overlying epithelium. The effects of various concentrations of FGF-2 on the expression of Msx genes and outgrowth of the mandibular processes were similar to that of FGF-4.

Effects of BMPs on Morphogenesis of the Mandible In Vivo Are Region- and Stage-Dependent

Our previous studies in the stage 23 chick mandible indicated the inability of medially or laterally placed BMP-7 and BMP-4 to affect the growth of the mandibular processes (Wang et al., 1999). Our observations on the effects of BMP-7 in stage 23 chick mandibular mesenchyme were different from previously reported effects of BMP-2 (Barlow and Francis-West, 1997; Ekanayake and Hall, 1997). We suggested that the differences between our observations and those reported by others may be related to differences in the developmental stages of the mandibles (Wang et al., 1999). To directly test this possibility, in the present study, we examined the effects of beads soaked in BMP-7 and BMP-4 (10, 50, and 100 ng/μl) on growth and morphogenesis of the lateral and medial mandibular mesenchyme at stage 20. In vivo implantation of BMP-7 soaked beads in the more medial positions of the stage 20 mandibles extended Msx1 and Msx2 domains of expression more laterally and down-regulated Msx1 expression in the region of the hyomandibular cleft (Figs. 9E, 8G). Implantation of BMP-7 beads to lateral regions of mandibular processes at this stage resulted in ectopic expression of both Msx1 and Msx2 throughout the mandibular processes in the implanted sides (Figs. 9F, 8H).

In contrast to its inability to affect the outgrowth of stage 23 mandibles (Wang et al., 1999), in vivo implantation of BMP-7 beads to the lateral region of stage 20 mandibles exerted dose-dependent negative effects on the growth of the mandibular process and Meckel's cartilage (Figs. 9E,F, 10A). Beads soaked in 100 ng/μl resulted in the complete absence of the mandibular process and of Meckel's cartilage on the implanted side (10 of 13) (Fig. 10A), whereas beads soaked in 50 ng/μl caused reduced growth of the mandibular process and partial absence of the Meckel's cartilage and its articulating elements on the implanted side (7 of 10, data not shown). Beads soaked at 10 ng/μl did not affect the growth of the mandibular process and in two of five cases induced only minor defects in the articulating ends of Meckel's cartilage in the implanted sides (data not shown).

Implantation of beads soaked in similar concentrations of BMP-7 to a more medial position of stage 20 mandibles induced the formation of an additional rod of cartilage extending from the body of Meckel's cartilage in addition to negatively affecting outgrowth and formation of the skeletal elements in the lateral region (3 of 5 for 100 ng/μl, 2 of 6 for 50 ng/μl, 0 of 10 for 10 ng/μl) (Fig. 10B,C). In vivo implantation of beads soaked in different concentrations of BMP-4 into lateral and medial region of stage 20 mandibles induced similar effects (data not shown). Our observations of region-specific effects of BMP-7 and BMP-4 on stage 20 mandibular processes are similar to previously reported effects of BMP-2 (Barlow and Francis-West, 1997; Ekanayake and Hall, 1997). In a small set of experiments, the ability of BMPs to support the outgrowth and elongation of Meckel's cartilage in the absence of epithelium was examined by using in vitro CAM cultures. Similar to our previous observations at stage 23, in these studies medially and laterally placed BMPs did not support outgrowth of the stage 20 isolated mandibular mesenchyme (data not shown). These observations suggest that medial and lateral mandibular mesenchyme respond differently to exogenous BMPs and that, in contrast to FGFs, BMPs do not play positive roles in the outgrowth of the mandibular mesenchyme.

Stage- and Region-Specific Effects of BMPs on Proliferation and Chondrogenesis of Mandibular Mesenchyme

Our results show that at early stages of development (stage 20) overexpression of BMPs in the lateral region of the developing mandible inhibited outgrowth of the mandibular primordia and the formation of Meckel's cartilage (Fig. 10A). On the other hand, similar treatment at later stages of development (stage 23) resulted in the formation of ectopic cartilage in the caudal region of Meckel's cartilage (Wang et al., 1999). To gain further insight into the possible stage-specific differences in the response of lateral mandibular mesenchyme to BMPs, the effects of BMP-7 on high-density micromass cultures of the lateral region were examined (Fig. 1B). In these experiments, exogenous BMP-7 was added to micromass cultures at two different time points. The addition of 2.5 ng/μl of BMP-7 to the micromass cultures from lateral regions at day 0, before the formation of distinct cell aggregates (Gluhak et al., 1996), resulted in inhibition of chondrogenesis preceded by decreases in the number of attached cells and the number of BrdU-positive cells (Fig. 8E; Table 3). On the other hand, addition of the same concentration of BMP-7 at 24–30 hr after initiation of culture (after the formation of the prechondrogenic aggregate), stimulated chondrogenesis (Fig. 8F; Table 3).

Table 3. Effects of Exogenous BMP-7 on Proliferation and Chondrogenesis of the Lateral and Medial Mandibular Mesenchyme in Micromass Culturesa
Region of the mandibleControlTime of BMP-7 addition (2.5 ng/μl)
0 hrAfter 24–30 hr
  • a

    Note that proliferation was quantitated by BrdU labeling in 2-day-old cultures and chondrogenesis by spectrophotometric analysis of bound Alcian blue in 4-day-old cultures. ND, not determined; BMP, bone morphogenic protein.

Lateral
 Proliferation70% ± 555% ± 7ND
 Chondrogenesis0.70 ± 0.20.40 ± 0.11.75 ± 0.3
Medial
 Proliferation10% ± 265% ± 5ND
 Chondrogenesis0.06 ± 0.0050.200 ± 0.07ND

Our in vivo observations also show that, unlike their negative effects on the growth and chondrogenesis of mesenchyme of the lateral region, beads soaked in BMPs implanted to a more medial position of stage 20 mandibles in vivo induced the formation of a well-organized duplicated rod of cartilage extending from the body of Meckel's cartilage (Fig. 10B,C), These observations are suggestive that medial and lateral mandibular mesenchyme respond differently to exogenous BMPs. To gain further insight into the region-specific differences of the effects of BMPs on lateral vs. medial mandibular mesenchyme, the effects of BMP-7 on high-density micromass cultures of the medial region were also examined (Fig. 1B). Unlike its effects on the lateral mandibular mesenchyme, addition of BMP-7 to the micromass cultures of the medial region at day 0 (but not 24–30 hr after initiation of culture) initiated chondrogenesis in a dose dependent-manner (Fig. 8F; Table 3, and data not shown). The effects of BMP-7 on stimulation of chondrogenesis in the mesenchyme of the medial region were similar but less profound than those of FGFs. In summary, these observations are suggestive of stage-and region- specific roles of BMPs in chondrogenesis of the mandibular mesenchyme.

Effects of FGF and BMPs on Cell Death

Our experiments showed that laterally placed BMP-7 was able to induce ectopic expression of Msx2 and exerted negative effects on the growth of the stage 20 mandibular processes. On the other hand, laterally placed FGFs did not induce ectopic expression of Msx2 and did not inhibit outgrowth. These observations indicate a correlation between the effects of laterally placed FGFs and BMPs on the outgrowth of the mandibular processes and their ability to induce ectopic expression of Msx2. Results of numerous studies indicate a correlation between the expression of Msx2 and apoptosis (Wedden, 1991; Graham et al., 1994; Mina et al., 1995; Macias et al., 1996; Maden et al., 1997; Shen et al., 1997; Ferrari et al., 1998; Takahashi et al., 1998). Studies in the developing limb (Johnson et al., 1994; Ganan et al., 1996; Macias et al., 1996; Martin, 1998, Buckland et al., 1998) and teeth (Vaahtokari et al., 1996) also suggest that, in contrast to BMPs, FGFs are survival factors that inhibit apoptosis and promote mesenchymal cell proliferation. These observations suggest that the negative effect of BMPs on outgrowth and chondrogenesis of stage 20/21 chick mandibular arches may be in part due to BMP-induced cell death in undifferentiated mesenchyme. Nile blue staining of embryos 24 hr after bead implantation showed areas of cell death around laterally placed BMP beads (100 ng/μl, n = 5) but not around FGF-4 beads (500 ng/μl, n = 5) (Figs. 10F, 10G).

DISCUSSION

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

Our observations show the essential roles of epithelially derived signals on proliferation and outgrowth of the medial but not lateral mandibular mesenchyme. Our studies in micromass culture show that, in the absence of exogenous BMP-7 and FGFs, mesenchyme in the 300 μm medial region lacks in vitro chondrogenic potential related to the lack of proliferation. Our grafting experiments show that, unlike the growth of the lateral region, the proliferation and growth of the mesenchyme in the medial region is dependent on signals derived from its overlying epithelium. Our grafting experiments also showed that the 300 μm medial region makes significant contributions to the overall growth of the developing mandible. These results are consistent with grafting experiments reported by Richman and Tickle (1989) and DiI labeling studies in chick mandibles that indicated that cells in the medial region make a greater contribution to the overall expansion of the developing mandible compared with cells in the lateral region (McGonnell et al., 1998). Cell proliferation studies showed a high percentage of S-phase labeled cells in the medial region of the developing mandible (Mina et al., 1995; Barlow and Francis-West, 1997; McGonnell et al., 1998). These observations suggest that the medial region is involved in regulating the directional outgrowth of the developing mandible.

Our observations show that, unlike in other organs, the patterns of expression of Fgf8 in the developing mandible is not correlated with regions that direct the outgrowth. Previous studies in mouse and chick embryos indicate that Fgf8 is expressed in regions of the embryos where epithelium is known to direct outgrowth (Crossley and Martin, 1995; McGonnell et al., 1998). Previous studies in mice suggest that FGF-8 may be involved in the establishment of the oral-aboral polarity of the developing mandible (Tucker et al., 1999). A recent study in which Fgf8 was inactivated in the epithelium of the first branchial arch of transgenic mice by using Cre/lox technology has provided direct evidence for roles of FGF-8 in the development and morphogenesis of the lateral (but not medial) regions of the developing mandible (Trumpp et al., 1999).

FGF-Mediated Signaling Is Part of the Signaling Pathway That Mediates the Growth-Promoting Interactions in the Medial Region

The lack of abnormalities in the medial regions of the mandibles of the Fgf8 mutants (Trumpp et al., 1999) provide clear evidence that morphogenesis of the medial region depends on signals other than FGF-8. Our functional studies together with our tissue distribution studies suggest that FGFs secreted by the epithelium of the medial region are a part of the signaling pathway that mediates the growth-promoting tissue interactions in the medial region and that this pathway involves interactions with FGFR2 and downstream target genes including Msx1 and Msx2. Our observations indicate that medially placed FGFs are able to support the expression of Msx1, Msx2, and Fgfr2 in the medial mandibular mesenchyme. In the absence of epithelium, FGF beads placed on the mesenchyme of the medial region support the outgrowth of the mandibular mesenchyme and elongation of Meckel's cartilage.

Our observations also indicate differences between the response of the lateral and medial mandibular mesenchyme to FGFs. Unlike their effects on medial mandibular mesenchyme, FGFs did not induce ectopic expression of Msx genes in the lateral mandibular mesenchyme. Furthermore, laterally placed FGF did not stimulate proliferation, apoptosis, or chondrogenesis of the lateral mandibular mesenchyme. Analyses of Fgf8;Nes-cre mutants (Trumpp et al., 1999) indicated that lack of FGF-8 did not affect the proliferation of the lateral mandibular mesenchyme and that phenotypic abnormalities in lateral region of the developing mandible mutants were in part related to increased apoptosis in the lateral mandibular mesenchyme and in part related to changes in the expression of Lhx6, Barx1, Et1, and Gsc in the lateral region of the mandibular arch.

Based on their patterns of expression, FGF-4 and FGF-2 (Richman et al., 1997) are likely endogenous ligands mediating the epithelial-mesenchymal interactions in the medial region. Although the Fgfr2 probe used in this study does not distinguish the patterns of expression of two isoforms of Fgfr2, the results of the previous studies suggest that hybridization signals in the mesenchyme of the medial region most likely reflect the abundant expression of Fgfr2c isoform. These observations suggest the FGFR2c mediates the effects of FGF signals originating from the epithelium of the medial region. Receptor binding studies (Ornitz et al., 1996; Szebenyi and Fallon, 1999) showed that FGF-2 and FGF-4 are among the ligands that can activate the FGFR2c.

Understanding the roles of FGF/FGFR2 pathway in mandibular morphogenesis through standard gene targeting approaches has been problematic. A null mutation encompassing both isoforms of Fgfr2 results in preimplantation lethality (Arman et al., 1998), and mice expressing a hypomorphic allele of Fgfr2 die around embryonic day (E) 10 (Xu et al., 1998). However, mutant animals specifically deficient for the Fgfr2b isoform (De Moerlooze et al., 2000) and mice expressing a soluble dominant negative form of Fgfr2b (Celli et al., 1998) had cleft palate, smaller faces, and abnormalities in many tissues for which morphogenesis has been shown to be dependent on inductive tissue interaction, including the developing teeth, inner ears, lungs, and limbs. This finding suggests that the FGFR2B isoform may also be involved in mandibular morphogenesis. This isoform has been shown to be expressed mainly by epithelia and is activated by four known ligands (FGF-1,-3,-7, and -10) which are synthesized predominately by mesenchyme in various tissues. The expression of Fgfr2 in the epithelium of the developing mandible most likely reflects the Fgfr2b isoform and is suggestive of the possible involvement of FGF/FGFR2b in morphogenesis of the medial region in vivo.

Our observations suggest the essential roles of FGF-mediated pathway in proliferation of the mesenchyme in the medial region that results in the outgrowth of the mandibular processes. These observations suggest that the medial region of the developing mandible may be functionally similar to the distal tip of the growing limb bud (Johnson et al., 1994; Martin, 1998) in that mesenchymal cells in the medial region may constitute a “progress zone” that is maintained by FGFs (other than FGF-8) secreted by overlying epithelium.

Unlike FGFs, BMPs Do Not Play a Positive Role in the Outgrowth of the Developing Mandible

Our observations show that medially or laterally placed BMP-7 and BMP-4 inhibited the outgrowth of mandibular processes at stage 20. The effects of medially and laterally placed BMP-7 and BMP-4 at stage 20 are similar to the previously reported effects of BMP-2 (Barlow and Francis-West, 1997; Ekanayake and Hall, 1997) and together with the inability of BMPs to support outgrowth of the developing mandible at later stages of development (Wang et al., 1999) indicate that BMP-mediated signaling does not play a positive role in mandibular outgrowth. These observations also suggest that, similar to their roles in limb outgrowth (Ganan et al., 1998; Dahn and Fallon, 1999; Merino et al., 1999; Pizette and Niswander, 1999), epithelially derived BMPs may be involved in limiting the outgrowth of mandibular process. The induction of a duplicated rod of cartilage in medial mandibular mesenchyme by medially placed BMP-7, BMP-4, and BMP-2 (present study and Barlow and Francis-West, 1997) suggests roles of BMP-mediated signaling in chondrogenesis and elongation of Meckel's cartilage in vivo.

Our results also indicate that medial and lateral mandibular mesenchyme respond differently to exogenous BMP-7. At early stages of development (stage 20) overexpression of BMP-7 and BMP-4 in the lateral region of the developing mandible, similar to BMP-2 (Barlow and Francis-West, 1997; Ekanayake and Hall, 1997), caused apoptosis, ectopic expression of Msx genes, and inhibited outgrowth of the mandibular primordia and the formation of Meckel's cartilage. On the other hand, similar treatment at stage 23 resulted in the formation of ectopic cartilage in the caudal region of Meckel's cartilage (Barlow and Francis-West, 1997; Wang et al., 1999). In contrast to these observations in the chick mandible, a recent study showed that, application of beads soaked in 100 ng/μl of BMP-4 for 6 days to the lateral region of organ-cultured mouse mandibles at early stages of development (E10) induced ectopic cartilage (Semba et al., 2000). However, medially placed BMP-4 did not induce ectopic cartilage (Semba et al., 2000). The differences between the effects of laterally and medially placed BMPs in chick and mouse mandibles are most likely related to the differences in culture systems.

Our observations suggest that the differences between the effects of BMPs on lateral mandibular mesenchyme at stage 20 vs. 23 are related to the differences in their stages of differentiation. BMPs promote apoptosis in undifferentiated mandibular mesenchyme at early stages of development (stage 20 and undifferentiated mesenchyme in micromass cultures), whereas at a later stage after the formation of prechondrogenic blastema (stage 23 and 30 hr after initiation of micromass cultures), they promote chondrogenesis. These observations are consistent with studies in the developing limbs (Macias et al., 1997; Buckland et al., 1998).

Therefore, based on these observations, we suggest that the inhibition of outgrowth by BMPs in the lateral region is related to their apoptotic-inducing effects on undifferentiated mesenchyme, whereas the formation of ectopic cartilage in this mesenchyme is related to the known chondrogenic promoting effects of BMPs. Recent studies have indicated that the chondrogenic promoting effects of BMPs are mediated by both Bmpr-IA and Bmpr-IB and the downstream transcription factor Sox9 (Kawakami et al., 1996; Yokouchi et al., 1996; Zou and Niswander, 1996; Zou et al., 1997; Enomoto-Iwamoto et al., 1998; Bi et al., 1999; Healy et al., 1999; McDowall et al., 1999). However, it is still unclear which receptors (BMPRIB or BMPRIA) are involved in BMP induced apoptosis (Kawakami et al., 1996; Yokouchi et al., 1996; Zou and Niswander, 1996).

Interactive Roles of FGF and BMP Signaling in Morphogenesis of the Developing Mandible

Our observations on the differences in the response of the lateral and medial mandibular mesenchyme to both FGF- and BMP-mediated signaling is suggestive of possible interactive roles of these two signaling factors in the morphogenesis of both regions. Our functional studies indicate that medially placed FGFs supported outgrowth of mandibular mesenchyme and elongation of Meckel's cartilage but did not induce ectopic cartilages. On the other hand, medially placed BMPs did not support outgrowth and induced a duplicated rod of cartilage extending from the body of Meckel's cartilage. We suggest that the differences in the effects of these growth factors on the medial region of the developing mandible may be related to differences in the developmental origins of the cells in the medial and lateral regions and/or reflect the specific and interactive roles of these signaling pathways in morphogenesis of the medial region in vivo. Fate-map studies indicate that the cells in the medial region are derived exclusively from CNCC originating from midbrain, whereas cells in the lateral regions are derived from CNCC originating from the midbrain and from R2 to R4 (Couly et al., 1996; Kontges and Lumsden, 1996).

Our observations suggest that in the medial region one of the essential roles of the FGF-mediated pathway is to maintain proliferation of the undifferentiated mesenchyme, whereas the BMP-mediated pathway may be involved in elongation of Meckel's cartilage. These observations also suggest that lateral to medial elongation of the developing mandible and Meckel's cartilage is regulated by the availability of the pool of proliferative cells in the medial region mediated by FGF/FGFR2 signaling, and the continuous gradual addition of these cells to the tips of prechondrogenic aggregate mediated by BMP signaling, respectively.

Our observations, consistent with previously reported results (Barlow and Francis-West, 1997; Ekanayake and Hall, 1997), indicate that, at early stages of development, overexpression of BMPs in the lateral mandibular mesenchyme where survival of mesenchymal cells and morphogenesis has been shown to be dependent of FGF-8–mediated signaling (Trumpp et al., 1999) induced apoptosis. On the other hand, later at stage 23, laterally placed BMPs induce ectopic cartilage. These observations suggest that undifferentiated mandibular mesenchyme in the lateral region at early stages of development is under the balanced influence of death and survival signals provided by epithelially derived BMPs and FGFs, respectively. Survival and proliferation occurs when influences of FGF signaling predominate over the apoptotic influences of BMP signaling. On the other hand, cell death/apoptosis occurs when BMP-mediated apoptosis predominates over the FGF-mediated cell survival. This possibility leads to several testable predications including (1) FGFs should modulate the apoptotic effects in the lateral mesenchyme induced by BMPs, and (2) the absence of FGF signaling in the lateral mandibular mesenchyme may result in increased BMP signaling and, thus, cell death. Thus, although not examined, it is possible that increased apoptosis in the lateral mandibular mesenchyme in Fgf8;Nes-cre embryos may be due to the increased effects of BMP signaling in the absence of FGF signaling. The antagonistic interactions between BMPs and FGFs in patterning events are relatively well documented. Studies in developing limbs suggest that the extent of proximal-distal outgrowth, apoptosis, and chondrogenesis are also mediated by a balance between the activities of the FGFs and BMPs (Niswander and Martin, 1992; Ganan et al., 1996; Buckland et al., 1998). Studies in the first branchial arch indicate that the restricted patterns of expression of many regulatory molecules in the mesenchyme of the medial vs. lateral regions of the mandible (e.g., Pax9, Barx1, Msx1, Msx2) are established by stage-specific antagonistic interactions of BMP-4 and FGF-8 signaling in the lateral and medial regions (Neubuser et al., 1997; Tucker et al., 1998, 1999; Barlow et al., 1999; Ferguson et al., 2000). More recent observations also suggest that a growing number of evolutionarily conserved secreted proteins that antagonize BMP function are also involved in regulating BMP functions. These antagonists bind specifically to BMPs and, thus, block their interaction with their cognate receptors (for reviews see Thomsen, 1997; Kawabata et al., 1998; Schmitt et al., 1999; Smith, 1999).

Morphogenesis of the Medial Region of the Developing Mandible

The lack of involvement of FGF-8 in morphogenesis of the medial region leaves open the question of which signal(s) are involved in morphogenesis of this region. The unchanged patterns of expression of markers of the medial region, including Bmp4, dHAND, eHAND, Msx1, and Msx2 in Fgf8;Nes-cre embryos (Trumpp et al., 1999) suggest that these regulatory molecules are involved in morphogenesis of the medial region. The possibility of the involvement of Msx1 and dHAND in morphogenesis of the medial region is supported by the abnormalities in the medial region of mice deficient for Msx1 or dHAND (Satokata and Maas, 1994; Srivastava et al., 1997; Thomas et al., 1998). Lack of Msx1 resulted in the truncation of the most medial part of the mandibular arch, including the mandibular incisors (Satokata and Maas, 1994). Deletion of the dHAND gene resulted in embryonic death at E10 secondary to cardiac failure, absence of second branchial arches, and severely hypoplastic mandibles (Srivastava et al., 1997; Thomas et al., 1998).

In addition to Msx1 and dHAND mutant mice, mice lacking other regulatory genes including genes encoding components of the endothelin pathway (ETA, ET1, ECE-1) (Clouthier et al., 1998, 2000; Yanagisawa et al., 1998; Kurihara et al., 1999), also exhibit abnormalities in the medial region of the developing mandible. The shared subset of abnormalities in the medial region of the developing mandible in various mutants indicate the essential roles of these gene products for correct morphogenesis of the medial region and suggest that the regulatory hierarchies controlling morphogenesis of the medial region of the developing mandible are complex and involve many molecules and possibly multiple pathways. However, it is currently not clear how many signaling pathways regulate the development of the medial region.

Analyses of dHAND-/-, ETA-/-, and ET1-/- mutants provide information for one of the genetic pathways, “ET-1-dHAND-Msx1 pathway,” that is involved in morphogenesis of the medial region of the mandible (Thomas et al., 1998; Clouthier et al., 2000). The ET1-HAND-Msx1 pathway suggests that ET-1/ETA signaling in the medial region is required for high levels of expression of dHAND (and eHAND) in the underlying mesenchyme, which in turn regulates expression of Msx1 (but not Msx2) in mesenchyme of the medial region (Thomas et al., 1998). These studies also showed that lack of dHAND did not affect expression of many other regulatory genes, including Mhox (Prx1), Msx2, and Dlx2 in the developing mandible, suggesting that these genes either act upstream of dHAND or belong to different or parallel genetic pathways regulated by other signaling molecules (Thomas et al., 1998). Our functional studies suggest that FGFs other then FGF-8 and BMPs are also part of signaling pathways that mediates morphogenesis of the medial region.

EXPERIMENTAL PROCEDURES

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

Tissue Isolation

Whole mandibular processes from White Leghorn chick embryos at various stages of development (Hamburger and Hamilton, 1951) were isolated in Hanks' balanced salt solution (HBSS, Sigma) containing 10% fetal calf serum (FCS, HyClone). Either whole mandibles or tissues from specific regions of the mandible (see Fig. 1) were used. Epithelial and mesenchymal components from whole mandible and isolated fragments were separated by using 0.8 U/ml Dispase at room temperature for 1 hr.

Preparation of Beads

Heparin acrylic beads (200 μm diameter, Sigma) were washed with PBS and soaked in solutions containing 10, 50, 100, and 500 ng/μl recombinant FGF-2, FGF-4, FGF-8 (R&D) or 1 μg/μl bovine serum albumin (BSA) and incubated at 37°C for 1 hr. Affi-Gel blue agarose beads (75–150 μm or 150–300 μm diameter, Bio-Rad, Hercules, CA) soaked in 10, 50, 100 ng/μl of recombinant BMP-7 (kindly provided by Dr. M. Charette, Creative BioMolecules, Hopkinton, MA) and recombinant BMP-4 (kindly provided by Dr. V. Rosen, Genetics Institute) were prepared as described previously (Wang et al., 1998, 1999). Protein-soaked beads were stored at 4°C for up to 2 weeks. All beads were washed in culture medium before implantation to various regions in chick mandibular processes.

In Vitro Bead Implantation

Bead implantation was performed according to previously described procedures (Vainio et al., 1993; Wang et al., 1998, 1999). Heparin acrylic beads soaked in different concentrations of recombinant FGF-2, FGF-4, FGF-8, or bovine serum albumin (BSA) were placed on the medial or lateral regions of half of the isolated whole mandibular mesenchyme (see Fig. 1).

All explants of isolated whole mandibular mesenchyme containing protein-soaked beads and epithelium were cultured for 24 hr on Nuclepore filters according to previously described methods (Wang et al., 1998, 1999). After 24 hr, explants were fixed and processed for whole-mount in situ hybridization and immunocytochemistry.

To examine the effects of FGFs on the outgrowth of the mandibular processes, explants of the whole mandibular mesenchyme (Fig. 1) containing protein-soaked beads and epithelium were grafted onto the CAM of 8-day-old chick embryos as described previously (Wang et al., 1999). After 7 days, the explants were removed and processed for whole-mount staining for bone and cartilage. After staining, the length of the Meckel's cartilage in both halves (treated and untreated halves) of the mandibles was measured under a dissecting microscope using an ocular grid.

Grafting in the Chick Limb Bud

For these experiments, tissues from specific regions of the mandible (see Fig. 1) were used. The 300-μm medial region, characterized by expression of Msx1 and Msx2 in the mesenchyme and Bmp4 in the overlying epithelium, and tissue fragments of similar size (300 μm) from the lateral regions were isolated from chick mandibles (Fig. 1B). After separation of epithelium from mesenchyme, isolated mesenchyme and homotypic recombinations were grafted for 1 week into a graft site prepared on the dorsal side of a stage 22 chick wing bud according to the previously described methods (Wedden, 1987; Richman and Tickle, 1989).

Micromass Cultures

Standard high-density micromass cultures were prepared from mesenchymal cells isolated from the 300-μm medial and the whole lateral regions (see Fig. 1) of stage 23 chick mandibles as previously described (Mina et al., 1995). Briefly, after separation of epithelium from the mesenchyme, a cell suspension was prepared by pipetting the tissue fragments vigorously through fire-polished glass pipettes. After counting, 10 μl of a cell suspension containing 2 × 105 single cells were placed in 16-mm well of a 4-well multidish (Nunc). After 2 hr, the defined medium with and without various concentrations of FGF-2, FGF-4, and BMP-7 (0.1, 1, 2.5, 5 ng/ml) was added to the cultures and replaced every other day. The defined medium consisted of 60:40 ratio of F12: DMEM, 2 mM L-glutamine 100 units/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg/ml Fungizone (1% antibiotic/antimycotic solution) 150 μg/ml ascorbic acid (Gibco/BRL, Grand Island, NY), and 5 μg/ml of insulin, 5 μg/μl of transferrin, 5 ng/ml of sodium selenite (insulin-transferrin-sodium selenite media supplement) (Sigma, St. Louis, MO).

Experimental Manipulation of Chick Mandibles In Vivo

Beads soaked in various concentrations of FGF-2, FGF-4, BMP-4, and BMP-7 into different regions of chick mandible at various stages of development (stage 20 and stage 23) were implanted according to previously described methods (Barlow and Francis-West, 1997; Wang et al., 1999). Briefly, protein soaked beads were placed in the lateral region or a more medial position (closer to the Msx expressing mesenchyme in the medial region) of the mandibular processes. Operations were carried out on the right half of the mandibular process, which lies uppermost. The left half of the mandibular processes acted as a control. One week after bead implantation, the surviving embryos were harvested, heads were removed and processed for whole-mount staining for bone and cartilage as previously described (Wang et al., 1999).

Cell Death Analysis

Twenty-four hours after bead implantation, surviving embryos were flooded by sub-vitelline injection of 0.01% filtered Nile blue solution (1:10,000 in H2O), after which the embryos were re-incubated for 30–60 min (Mina et al., 1995). Embryos were then removed from eggs, washed extensively with PBS, viewed, and photographed.

In Situ Hybridization to Whole-Mounts and Sections

Tissue fragments were fixed in freshly prepared 4% paraformaldehyde at 4°C overnight. Tissues were either embedded in paraffin and sectioned in 7-μm thickness or processed for whole-mount in situ hybridization. Whole-mount hybridization using digoxigenin-labeled RNA probes and hybridization to tissue sections by using 32P-labeled RNA probes were performed as previously described (Mina et al., 1995; Wang et al., 1998, 1999).

A 1,200-bp fragment of chick Fgf4 cDNA (kindly provided by Dr. L. Niswander) in Bluescript was digested with EcoRI or HindIII and transcribed with T7 or T3 RNA polymerase for antisense and sense probes, respectively (Niswander et al., 1994). An 800-bp fragment of chick Fgf8 cDNA (kindly provided by Dr. G. Martin) in Bluescript was digested with EcoRI or XhoI and transcribed with T7 or T3 RNA polymerase for antisense and sense probes, respectively (Crossley and Martin, 1995). The 2,930-bp fragment of chick Cek1 cDNA (kindly provided by Dr. E. Pasquale) in Bluescript was digested with XhoI or BamHI and transcribed with T3 or T7 RNA polymerase for antisense and sense probes for Fgfr1, respectively (Pasquale and Singer, 1989; Patstone et al., 1993). Chick Fgfr2 antisense and sense probes (approximately 1,500 bp) was generated from chick Cek3 cDNA (kindly provided by Dr. Pasquale) in Bluescript (Pasquale, 1990) by digestion with EcoRV and XbaI and in vitro transcription with T3 or T7 RNA polymerase respectively. This probe is similar to that used by (Wilke et al., 1997), corresponds to nucleotides 60-1519, spans the three extracellular immunoglobulin domains and recognizes both Fgfr2c (exon 9 corresponding to bases 1124-1242) and Fgfr2b (exon 8 corresponding to bases 1064-1187). A 713-bp antisense and sense probe for chick Fgfr3 was generated from chick Cek2 cDNA (kindly provided by Dr. E. Pasquale), in Bluescript (Pasquale, 1990) by digestion with BamHI or EcoRV and in vitro transcription with T7 or T3 RNA polymerase, respectively. This probe is also similar to that used by (Wilke et al., 1997) and spans the first two extracellular domains of the receptor corresponding to nucleotides 1-713. Antisense and sense probes for chick Msx1 and Msx2 were generated as described previously (Mina et al., 1995).

Northern Blot Hybridization

Total RNA was prepared by the guanidine thiocyanate/phenol-chloroform method. Ten micrograms of total RNA was denatured with formamide-formaldehyde, separated by electrophoresis on a 1% agarose gel, and transferred to nylon membranes. Membranes were prehybridized and then hybridized with 32P-labeled probes for Type II collagen, aggrecan core protein, and rRNA. The relative levels of mRNAs for aggrecan core protein, and Type II collagen on the filters were quantified by using a PhosphorImager (PhosphorImager SI, Molecular Dynamics).

Acknowledgements

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

We thank Dr. Edward J. Kollar for this helpful discussion and critical comments on the manuscript, Dr. J. Richman for help in grafting experiments into the dorsal surface of chick wing bud, Dr. L. Niswander for providing Fgf4, Dr. G. Martin for providing Fgf8, and Dr. E, Pasquale for Cek 1-3 cDNAs. We also thank Dr. M. Charette and Dr. V. Rosen for providing the recombinant BMP-7 and BMP-4 respectively.

REFERENCES

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