DEVELOPMENTAL DYNAMICS 208:526–535 (1997) Characterization of the Fate of Midline Epithelial Cells During the Fusion of Mandibular Prominences In Vivo YANG CHAI,1 YASUYUKI SASANO,1 PABLO BRINGAS, JR.,1 MARK MAYO,1 VESA KAARTINEN,2 NORA HEISTERKAMP,2 JOHN GROFFEN,2 HAROLD SLAVKIN,3 AND CHARLES SHULER1* 1Center for Craniofacial Molecular Biology, School of Dentistry, University of Southern California, Los Angeles, California 2Department of Pathology, Surgery and Pediatrics, Childrens Hospital of Los Angeles Research Institute, Los Angeles, California 3Craniofacial Development Section, National Institute of Arthritis and Musculoskeletal and Skin Disease, National Institute of Health, Bethesda, Maryland ABSTRACT The fusion of the mandibular prominences along the midline is achieved with the absence of medial epithelial cells at the fusion site. Failure of fusion of the mandibular prominences results in median cleft of the lower lip and mandible. Cellular and molecular events controlling mandibular fusion were examined during the fusion process in mouse embryogenesis. Cell lineage analyses at the fusion site revealed that epithelial cells migrated to the surface and oral epithelia. DiI-labeled epithelial cells were not observed within the mandibular mesenchyme at any stage of fusion. Examination of the midline region did not reveal cells with ultrastructural changes characteristic of apoptotic cell death. An increase in lysosomal enzymes in the midline epithelial cells, which would be correlated with programmed cell death, was not observed. Mice lacking TGF-b3 did not have cleft mandible, but had clefting of the secondary palate as a feature of null mutation phenotype. We interpret our comparisons between wild type and homozygous TGF-b3 (2/2) mice to suggest that different developmental processes control palatal vs. mandibular fusion. We hypothesize that medial epithelial cells at the fusion site of mandibular prominences migrate to the surface epithelium during the fusion process and neither transdifferentiate into mesenchyme nor express apoptosis. Dev. Dyn. 208:526–535, 1997. r 1997 Wiley-Liss, Inc. Key words: mandible; morphogenesis; medial epithelial cells; TGF-b3 null mutation INTRODUCTION First branchial arch syndromes constitute a major proportion of craniofacial anomalies. These anomalies are the result of altered development of first arch derived structures such as palate, maxilla, and mandible. Although not common, several syndromes with mandible dysmorphogenesis (e.g., mandibular dysplasia, mandibulofacial dysostosis) have been described (Jones, 1988; Gorlin et al., 1990). Clinically, sporadic cases of median clefts of the lower lip and mandible have also been reported. The vertebrate facial morphor 1997 WILEY-LISS, INC. genesis requires five prominences to fuse according to exquisite time and positional information: (1) frontonasal prominence, (2) paired maxillary prominences, and (3) paired mandibular prominences. Median clefts of the lip and mandible result from a failure of fusion of the first branchial arch derived paired mandibular prominences, and/or failure of mesenchymal cell penetration into the midline. The severity of the mandibular cleft defects depends upon the developmental stage at which the fusion of mandibular prominences was disturbed (Oostrom et al., 1996). Significantly, the cellular, molecular, and developmental processes controlling fusion of mandibular prominences have not been determined. Complete fusion of the mesenchymal portion of the mandibular prominences requires the absence of a distinct epithelial component in the midline. In studies of medial edge epithelial (MEE) cell fate during palatal fusion, three different mechanisms have been proposed to account for the disappearance of the MEE from the midline position of the forming secondary palate. First, a long-standing hypothesis proposed that disappearance of MEE cells was a result of classical programmed cell death (Shapiro and Sweney, 1969; Pratt and Martin, 1975; Pratt et al., 1984; Clarke et al., 1993). Second, another hypothesis for the fate of the MEE stated that MEE cells underwent an epithelial-mesenchymal transformation and remained in the connective tissue of the secondary palate as viable mesenchymal cells (Fitchett and Hay, 1989; Griffith and Hay, 1992; Shuler et al., 1991, 1992; Shuler, 1995). Third, yet another hypothesis suggested that MEE cells retained both their viability and epithelial phenotype and migrated from the fusion midline position to join with either adjacent oral or nasal epithelia (Carette and Ferguson, 1992). A similar process may have a function in the disappearance of the epithelial cells from the midline of the fusing mandibular prominences. *Correspondence to: Dr. Charles Shuler, Center for Craniofacial Molecular Biology, School of Dentistry, University of Southern California, 2250 Alcazar Street, CSA 103, Los Angeles, CA 90033. Received 17 October 1996; Accepted 8 January 1997 FUSION OF MANDIBULAR PROMINENCES 527 To test the hypothesis that midline epithelial cell fate is neither epithelial-mesenchymal transformation nor apoptosis, we designed studies to examine the process of mandibular fusion during embryogenesis. The fate of epithelial cells at the midline of fusion were examined for evidence of programmed cell death, epithelialmesenchymal transformation, and epithelial cell migration. The results of these studies support the hypothesis that the developmental processes required for the fusion of mandibular prominences are significantly different than those required for secondary palate formation. RESULTS Three Stages of Mouse Embryogenesis and Mandibular Prominence Fusion Mandibular prominence fusion starts at E9.5 when the first branchial arch-derived, paired mandibular prominences approximate one another at the midline. Macroscopic orientation of whole mouse embryos and histology of mandibular morphogenesis from E9 through E11 are shown in Figures 1 and 2. Inspection of timed-pregnant mouse embryos indicated that the fusion of mandibular prominences was completed within 24–36 hr. At E11, 36 hr after the beginning of mandibular fusion, the mandible is completely fused at the midline. Histological evaluation revealed that midline epithelium was intact and two cell layers thick at E9 (Fig. 2A). At E9.5, the opposing medial epithelia were adherent at the midline, which marked the beginning of the fusion process (Fig. 2B). Twelve hours later, at E10, the opposing mandibular epithelia united by forming an epithelial cell mass 3 to 4 cell layers thick (Fig. 2C). Early disruption of midline epithelium was observed at E10.5 (Fig. 2D). Epithelial triangles at the two external surfaces were found in the mandibular prominences at E10.5 and E11. By E11, the fusion was complete with mesenchymal confluence across the midline and the absence of epithelial islands (Fig. 2E). Scanning electron microscopy (SEM) was used to examine the developmental stages of mandibular prominence fusion (Fig. 3). At E9 to E9.5, the opposing mandibular prominences came in close approximation (Fig. 3A,B). At E10, the external surfaces of the paired mandibular prominences were beginning to merge (Fig. 3C). Mandibular fusion was completed at E11 (Fig. 3D). Apoptosis Was Not Observed at the Mandibular Prominence Fusion Site At E10, a thick epithelial cell mass (3–4 layers thick) was seen at the fusion site in serial thin sections of the mandible (Fig. 4A). Interruption of extracellular matrix occurred as the fusion process was initiated (Fig. 4B,C). On adjacent sections from the same mandibular prominence, complete separation of oral and suboral epithelium was evident (Fig. 4D). Transmission electron Fig. 1. Embryonic mouse morphogenesis in vivo. A: Lateral view of E9 (13–20 somite pairs, Theiler stage 14) mouse embryo, which showed mandibular prominence (m). B: Lateral view of E10 (30–34 somite pairs, Theiler stage 16) mouse embryo, which showed both mandibular and maxillary (mx) prominences. C: Lateral view of E11 (40–44 somite pairs, Theiler stage 18) mouse embryo. Scale bar 5 1 mm. 528 CHAI ET AL. Fig. 2. Histological sections of mandibular prominences at the fusion site. A: At E9, the midline epithelium (me) was intact and two cell layers thick at the midline. B: At E9.5, the opposing medial epithelia were adherent at the midline (arrow). C: At E10, the opposing mandibular epithelia united by forming an epithelial cell mass 3 to 4 cell layers thick (arrow). D: At E10.5, early disruption of midline epithelium was observed (arrow). Epithelial triangles (double arrow) at the two external surfaces were found in the mandibular prominences. E: By E11, the fusion was complete with mesenchymal confluence across the midline and the absence of epithelial islands. Scale bar 5 100 µm. microscopy examination did not reveal intracellular changes characteristic for apoptosis in the epithelium at the mandibular fusion site (not shown). Another line of evidence commonly used to support ‘‘programmed cell death’’ as the mechanism for the disappearance of MEE in secondary palatal fusion, is FUSION OF MANDIBULAR PROMINENCES 529 the observation of increased lysosome activity in the midline seam (Mato et al., 1966). Lysosomal activity was noted, albeit faint, in the midline epithelium at the fusion site (Fig. 5). At E10, the midline seam was interrupted by mesenchymal cells, and the epithelial triangle and contacting midline epithelium showed a slight increased lysosomal activity (Fig. 5A,B). On completion of mandibular fusion at E11, lysosomal staining was present in only a limited area of oral surface epithelium (Fig. 5C,D). There was no lysosomal activity observed in the mesenchymal cells adjacent to the midline during the mandibular fusion. The absences of ultrastructural changes and lysosomal activity are not consistent with a programmed cell death or an apoptotic cell fate for the midline epithelia associated with mandibular prominence fusion. Mandibular Prominence Midline Epithelial Cells Did Not Transdifferentiate Into Mesenchymal Cells Fig. 3. Scanning electron microscopy showed the fusion stages of mouse mandibular prominences. A: At E9, the opposing mandibular prominences (m) came in close approximation. B: At E9.5, there was a broad contact between left and right mandibular prominences (arrow). C: At E10, the external surfaces of the paired mandibular prominences were beginning to merge. D: At E11, mandibular fusion was completed. Scale bar 5 100 µm. In the course of this study, fifty-three fetal amniotic sacs were injected with DiI and forty-one of the exposed fetuses remained viable for subsequent analysis. The injected fetuses were examined on days 9, 10, and 11 of gestation to determine the distribution of DiI labeled epithelial cells in the developing mandibular process. DiI labeled only the cells on the surface of the fetus that were in contact with the amniotic fluid. At E9, the DiI was incorporated only in the membranes of the epithelial cells covering the developing craniofacial prominences (Fig. 6A). There was no labeling of cells in the mesenchyme underlying the epithelium. Following the contact of the two opposing mandibular prominences there was a DiI-labeled bilayered epithelial seam along the line of fusion (Fig. 6B). There were no labeled cells in the underlying mesenchyme. The progress of fusion of the opposing mandibular prominences was associated with migration of the epithelial cells to the surfaces of the mandible and the lack of any migration of DiI labeled cells into the underlying mesenchyme (Fig. 6C). The migration of midline epithelial cells to the surface region was evident by the intense DiI-labeled epithelial triangle at those locations. The complete fusion of the mandible was associated with a mesenchyme without DiI-labeling in juxtaposition to a DiI-labeled surface epithelium. No epithelial cells originally labeled with DiI transdifferentiated to a mesenchymal phenotype (Fig. 6D). With regard to the DiI fluorescence intensity, variation in intensity is a reflection of cell shape and membrane surface area. In the midline, the epithelial cells are columnar with a large membrane surface area and, consequently, do not have a strong fluorescence, since the DiI is distributed throughout the membrane. On the other hand, at the adjacent surface epithelium, some squamous epithelial cells are present, which concentrate the DiI signal and increase fluorescent intensity. 530 CHAI ET AL. Fig. 4. Higher magnification of midline epithelia at the fusion site during different stages of mouse embryo morphogenesis. A: The united midline epithelia is 3 to 4 cell layers thick (me) at E10. B,C: Interruption of extracellular matrix occurred as the fusion process was initiated (arrow). D: On adjacent sections from the same mandibular prominence, complete separation of oral (oe) and suboral epithelium (soe) was evident (arrow). Scale bar 5 20 µm. 531 FUSION OF MANDIBULAR PROMINENCES Fig. 5. There was no significant increase in lysosome activity at the mandibular prominences fusion site. A,B: The midline seam was interrupted by mesenchymal cells and the epithelial triangle (arrow) and contacting midline epithelium (double arrow) showed a slight increased lysosomal activity. C,D: On completion of mandibular fusion at E11, lysosomal staining was present in only a limited area of oral surface epithelium (oe). There was no lysosomal activity observed in the mesenchymal cells adjacent to the midline during the mandibular fusion. Scale bar 5 50 µm. TGF-b3 Knock-Out Mice Have Normal Fusion of the Mandible the fusion site (data not shown). The TGF-b3 knock-out phenotype resulted in cleft of secondary palate but had no apparent effect on mandibular fusion. The absence of TGF-b3 has been shown to inhibit secondary palatal fusion; all pups homozygous for TGF-b3 null mutants presented a cleft of the secondary palate phenotype (Fig. 7C, arrow). Significantly, mandibular midline clefts were never observed in the TGF-b3 (2/2) mutants (Fig. 7B). Histologic analysis of serial-sections comparing wild type with homozygous TGF-b3 (2/2) embryos did not show any difference at DISCUSSION The present study reports evidence that supports the hypothesis that mandibular prominence fusion does not require either apoptosis or epithelial-mesenchymal transformation. The developmental process of mandibular prominence fusion starts at E9.5 (21–29 somites) 532 CHAI ET AL. Fig. 6. There was no transdifferentiation of midline epithelial cells into mesenchymal cells in the mandibular prominence. A: At E9, the DiI was incorporated only in the membranes of the epithelial cells covering the developing craniofacial prominences (arrows). There was no labeling of cells in the mesenchyme (mes) underlying the epithelium. B: Following the contact of the two opposing mandibular prominences there was a DiI-labeled epithelial seams along in the midline (arrow). There were no labeled cells in the underlying mesenchyme. B is at E9.5, which is equivalent to Figure 2B. C: There was migration of the epithelial cells to the oral (oe) and suboral (soe) epithelium and the lack of any migration of DiI labeled cells into the underlying mesenchyme. C and Figure 2D are at E10.5. D: The complete fusion of the mandible was associated with a mesenchyme without DiI-labeling in juxtaposition to a DiI-labeled surface epithelium. Scale bar 5 50 µm. and is completed within 36 hr. The in utero labeling of the mandibular prominences before fusion, and subsequent medial epithelial cell lineage analysis in vivo, provided the possibility for the examination of mandibular fusion. During this process, midline epithelial cells migrate to the oral surface of the fusing mandibular arch. This positional reassignment for midline epithelial cells is in striking contrast to the fate of MEE associated with secondary palate fusion. During palatal fusion, the basement membrane underlying the MEE cells is degraded and the epithelial cells migrated into the mesenchyme and transdifferentiated into a mesenchymal phenotype (Fitchett and Hay, 1989; Griffith and Hay, 1992; Shuler et al., 1991, 1992). Two different observations on the fate of the palatal MEE cells have been made that are particularly impor- tant, in vitro. One observation reported epithelialmesenchymal transdifferentiation (Ferguson, 1988; Fitchett and Hay, 1989; Shuler et al., 1991; Griffith and Hay, 1992). The other observation showed that the MEE cells migrate to adjacent epithelia (Carette and Ferguson, 1992). These results have led to the suggestion that the MEE cells along a fusion line have two possible fates: (1) epithelial-mesenchymal transformation, and (2) epithelial cells proximal to the oral surface migrate to the surface position and retain their epithelial phenotype (Shuler, 1995). During mandibular prominence fusion, the present investigation was unable to identify evidence for an epithelial-mesenchymal transformation fate for the midline epithelium covering the two opposing mandibular prominences. Rather, the DiI-labeled cells appeared to FUSION OF MANDIBULAR PROMINENCES 533 Fig. 7. TGF-b3 knock-out mice had normal fusion of the mandible. A: Frontal view of normal wild type new born mouse. B: Frontal view of TGF-b3 (2/2) mutant newborn mouse. There was no midline cleft in the mandible. C: Transverse section of newborn mice with normal wild type at the left and TGF-b3 (2/2) mutant (with arrow) on the right. All pups homozygous for TGF-b3 null mutants presented a cleft of the secondary palate phenotype (arrow). Scale bars 5 1 mm. be incorporated to a new position assignment within the oral surface. In comparing these results to those previously observed during palatogenesis in vivo, it would appear that the midline epithelial cells migrated to adjacent epithelial rather than undergoing phenotypic transformation. The size of the fusion area and the distance to the oral surface epithelium may be critical features of whether midline epithelial cells undergo either epithelial-mesenchymal transformation or migration to the oral surface position. The relatively small size of the fusion zone in the mandible may permit all of the midline epithelial cells to retain their phenotype and migrate to the surface. Another explanation that might account for the disappearance of MEE cells at the fusion site of secondary palate has been based on the possibility of programmed cell death. This fate has been supported by two lines of evidence. One data set indicates the appearance of intracellular ultrastructural change consistent with apoptosis. The second line of evidence in support of programmed cell death is an increase of lysosome activity at the fusion site during secondary palatal fusion (Mato et al., 1966). Recent studies, however, have shown that the slight increases in lysosomal activity might be the result of increased intracellular and extracellular remodeling by degradative enzymes, rather than programmed cell death (Carette and Ferguson, 1992). In contrast, the epithelia at the midline of the fusing mandibular prominences do not have increased lysosome activity and do not have intracellular changes consistent with apoptosis. During mandibular fusion, programmed cell death does not appear to play a role in the fate of the midline epithelia. Another potentially significant aspect of craniofacial morphogenesis is the suggested function of autocrine and/or paracrine controls by the growth factors and their cognate receptors for midline epithelium during either secondary palatal or mandibular prominence fusions. For example, TGF-b3 is identified and specifically expressed in MEE cells of pre-fusion palatal shelves and expression ceases shortly after the midline epithelial seam is formed (Fitzpatrick et al., 1990; Pelton et al., 1990). In vitro studies of depletion of TGF-b3 gene product by using either antisense oligodeoxynucleotides or neutralizing antibodies prevented secondary palatal fusion (Brunet et al., 1995). TGF-b3 534 CHAI ET AL. null mutant mice have both delayed pulmonary development and cleft palate (Kaartinen et al., 1995). These studies indicate an essential function of TGF-b3 ligand in signal transduction processed during secondary palate fusion. TGF-b3 is also identified and specifically localized during the formation of Meckel’s cartilage within mandibular development (Chai et al., 1994). However, the depletion of TGF-b3 using antisense oligodeoxynucleotides did not result in the clefting of Meckel’s cartilage, but rather reduced the anteriorposterior dimension of Meckel’s cartilage. In TGF-b3 (2/2) mutant mice, the mandibular prominence fusion was normal and no sign of cleft mandible was noted. Thus, the TGF-b3 mediated molecular mechanisms, which govern the behavior of palatal MEE cells and mandibular midline epithelial cells, would appear to be quite different. Craniofacial morphogenesis appears to have two different developmental processes for the fate of midline epithelial cells in the zones of fusion. In the case of relatively small fusion zones, as seen in mandibular prominence fusion, the epithelial cells retain their phenotype and become incorporated within the oral epithelia covering the mandible. When the fusion zone is larger, the epithelial cells closest to the surface may migrate to and merge with the surface epithelia. Epithelial cells which have lost contact with the surface, due to basement membrane degradation, become entrapped in the midline of the secondary palate and undergo epithelial-mesenchymal transformation. In the process of fusion, a very small number of epithelial cells may die; however, that does not provide evidence for programmed cell death. The presence of two alternative fates for the midline epithelial cells further suggests that the action of teratogens may be associated with differential effects on these two complementary yet different developmental processes. Cell lineage analyses can permit the cells to be identified and analyzed at subsequent stages of development to determine the molecular controls for these two processes. 5 µm intervals and stained with hematoxylin and eosin for morphological evaluation. EXPERIMENTAL PROCEDURES Mouse Embryos DiI-Labeled Cell Localization Techniques Swiss-Webster mice were obtained from Simonsen Labs (Gilroy, CA) and mated for 2 hr and the presence of a vaginal plug established day 0 and hour 0. Mouse embryos were dissected on either gestational days 9, 10, or 11 (E9, E10, and E11) to obtain the mandibular prominences during the midline fusion process. External developmental staging of the mouse embryos was evaluated according to Theiler (1989). The dissection methodology was previously described (Slavkin et al., 1989; Shum et al., 1993). Histology The mandibular prominences were fixed in 10% buffered formalin, dehydrated in a graded series of alcohols, cleared in xylene, and embedded in paraffin. Each mandibular prominence was serially sectioned at Scanning Electron Microscopy Isolated mandibles were fixed in 2.5% glutaraldehyde (Pelco, Inc., Tustin, CA) in 0.1M sodium cacodylate buffer, pH 7.2, at 4°C for 2 hr. Specimens were rinsed in buffer and post-fixed in 2% osmium tetroxide (Pelco, Inc.) in 0.1M cacodylate buffer for 1 hr at 4°C, dehydrated in graded alcohols and placed in perforated Beem capsules in acetone. Specimens were then dried in a critical-point drying bomb (Sorvall, Newtown, CT) using carbon dioxide and mounted in specimen studs, sputter coated with gold-palladium alloy in an argon atmosphere (Hummer V, Technics); pulsed for 7 min at 9 V and 10 mA. The specimens were viewed with a Cambridge S4-10 scanning electron microscope operating at 10 kV. DiI Labeling Methods Eight timed-pregnant Swiss-Webster mice were used. The amniotic sacs were injected with DiI at gestational age 8 days 22 hours. General anesthesia was induced with an intraperitoneal injection of sodium pentabarbitol (50 mg/kg). The pregnant bicornuate uteri were carefully exposed through a midline labarotomy and the amniotic sacs counted. Sterile glass micropipets were used for all injections. DiI working solution (0.025% in normal saline, 1% ethanol) was made immediately before injection from a 1003 DiI stock in 100% ethanol. Each amniotic sac was transilluminated to identify the fetus’s exact position in utero, and 10 µl of the DiI working solution was microinjected through the wall of the uterus into the amniotic fluid. The uterus was returned to the abdominal cavity, the wound irrigated and closed. The mice were subsequently sacrificed at selected stages of fetal development associated with the fusion of the mandibular processes of the first branchial arch. The fetuses were dissected and the heads quick frozen on dry ice for use in cryostat sectioning. Frozen sections, 8 µm thick, were prepared from the mouse embryonic heads with the sections made in a coronal plane. The sections were mounted on glass slides and immediately observed for DiI fluorescence with a Zeiss (Thornwood, NY) fluorescent microscope at an excitation wavelength of 546 nm and an emission wavelength of 563 nm. The DiI fluorescence localization patterns were recorded photographically on Ektachrome 400 film. Immunocytochemistry Specimens for light microscopy were fixed in Carnoy’s fixative overnight at 4°C, dehydrated in a graded series of ethyl alcohol, cleared in xylene, and embedded in paraffin. Sections were cut at 5 µm in the transverse plane. In preparation for immunohistochemistry, slides were cleared in xylene and hydrated to water through a FUSION OF MANDIBULAR PROMINENCES graded series of alcohols. A monoclonal antibody (1D4B) obtained from the Developmental Studies Hybridoma Bank (Johns Hopkins University, Baltimore) was used to detect regions of lysosomal activity during mandibular prominence fusion. Slides were then treated using the Histostain-Streptavidin Peroxidase kit (for mouse primary antibody) following the protocol provided by the vendor (Zymed Laboratories, Irvine, CA). Mice Deficient for TGF-b3 Mice lacking TGF-b3 were obtained by successful disruption of the TGF-b3 locus on exon 6, which encodes mature TGF-b3 (Kaartinen et al., 1995). RT-PCR analysis confirmed the absence of TGF-b3 transcripts in null mutants. Morphological analysis was done to compare the fusion of mandibular prominences between the wild type and homozygous knock-out of TGF-b3 (2/2) mice. ACKNOWLEDGMENTS The monoclonal antibody (1D4B) developed by J. Thomas August was obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, MD 21205, and the Department of Biological Sciences, University of Iowa, Iowa City, IA 52242, under contract N01-HD-2-3144 from the NICHD. This research was supported by NIDR Center grant DE-09165. REFERENCES Brunet, C.L., Sharp, P.M., and Ferguson, M.W.J. (1995) Inhibition of TGF-b3 (but not TGF-b1 or TGF-b2) activity prevents normal mouse embryonic palate fusion. Int. J. Dev. Biol. 39:345–355. Carette, M.J.M., and Ferguson, M.W.L. (1992) The fate of medial edge epithelial cells during palatal fusion in vitro: An analysis by DiI labeling and confocal microscopy. Development 114:379–388. Chai, Y., Mah, A., Crohin, C., Groff, S., Bringas, P. Jr., Le, T., Santos, V., and Slavkin, H.C. (1994) Specific transforming growth factor-b subtypes regulate embryonic mouse Meckel’s cartilage and tooth development. Dev. Biol. 162:85–103. Clarke, A.R., Purdie, C.A., Harrison, D.J., Morris, R.G., Bird, C.C., and Hooper, M.L. (1993) Thymocyte apoptosis induced by p53-dependent and independent pathways. Nature 362:849–852. Ferguson, M.W.J. (1988) Palate development. Development 103s: 41–60. Fitchett, J.E., and Hay, E.D. (1989) Medial edge epithelium transforms to mesenchyme after embryonic palatal shelves fuse. Dev. Biol. 131:455–474. 535 Fitzpatrick, D.R., Denhez, F., Kondaiah, P., Akhurst, R. (1990) Differential expression of TGF beta isoforms in murine palatogenesis. Development 109:585–595. Gorlin, R.J., Cohen, M.M. Jr., and Levin, L.S. (1990) ‘‘Syndromes of the Head and Neck,’’ 3rd ed. New York: Oxford University Press. Griffith, C.M., and Hay, E.D. (1992) Epithelial-mesenchymal transformation during palate fusion: Carboxyfluorescein traces cells as light and electron microscopic levels. Development 116:1087–1099. Jones, K.L. (1988) ‘‘Smith’s Recognizable Patterns of Human Malformation,’’ 4th ed. Philadelphia: W.B. Saunders Co. Kaartinen, V., Voncken, J.W., Shuler, C., Warburton, D., Bu, D., Heisterkamp, N., and Groffen, J. (1995) Abnormal lung development and cleft palate in mice lacking TGF-b3 indicates defects of epithelialmesenchymal interaction. Nature Genet. 11:415–421. Mato, M., Aikawa, E., and Katahira, M. (1966) Appearance of various types of lysosomes in the epithelium covering lateral palatine shelves during a secondary palate formation. Gunma J. Med. Sci. 15:46–56. Oostrom, C.A., Vermeij-Keers, C., Gilbert, P.M., and van der Meulen, J.C. (1996) Median cleft of the lower lip and mandible: Case reports, a new embryologic hypothesis, and subdivision. Plast. Reconstr. Surg., 97:313–320. Pelton, R.W., Hogan, B.L.M., Miller, D.A., and Moses, H.L. (1990) Differentiation expression of genes encoding TGFs b1, b2, b3 during murine palate formation. Dev. Biol. 141:456–460. Pratt, R.M., and Martin, G.R. (1975) Epithelial cell death and cAMP increase during palatal development. Proc. Natl. Acad. Sci. U.S.A. 72:874–877. Pratt, R.M., Kim, C.S., and Grove, R.I. (1984) Role of glucocorticoids and epidermal growth factor in normal and abnormal palatal development. Curr. Top. Dev. Biol. 19:81–102. Shapiro, B.L., and Sweney, L.R. (1969) Electron microscopic and histochemical examination of oral epithelial-mesenchymal interaction (programmed cell death). J. Dent. Res. 48:652–660. Shuler, C.F. (1995) Programmed cell death and cell transformation in craniofacial development. Crit. Rev. Oral Biol. Med. 6:202–217. Shuler, C.F., Guo, Y., Majumder, A., and Luo, R. (1991) Molecular and morphologic changes during the epithelial-mesenchymal transformation of palatal shelf medial edge epithelium in vitro. Int. J. Dev. Biol. 35:463–472. Shuler, C.F., Halpern, D.E., Guo, Y., and Sank, A.C. (1992) Medial edge epithelium fate traced by cell lineage analysis during epithelialmesenchymal transformation in vivo. Dev. Biol. 154:318–330. Shum, L., Sakakura, Y., Bringas, P. Jr., Luo, W., Snead, M.L., Mayo, M., Crohin, C., Millar, S., Werb, Z., Buckely, S., Hall, F.L., Warburton, D., and Slavkin, H.C. (1993) EGF abrogation induced fusilliform dysmorphogenesis of Meckel’s cartilage during embryonic mouse mandibular morphogenesis in vitro. Development 118:903– 917. Slavkin, H.C., Bringas, P. Jr., Sasano, Y., and 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. Theiler, K. (1989) ‘‘The House Mouse, Atlas of Embryonic Development.’’ New York: Springer-Verlag.