THE AMERICAN JOURNAL O F ANATOMY 188:401-408 (1990)
Expression of the Epidermal Growth Factor Receptor in Developing Fetal Mouse Palates: An lmmunohistochemical Study KOHEI SHIOTA, SHIGEYUKI FUJITA, TETSU AKIYAMA, AND CHISATO MORI Congenital Anomaly Research Center (K.S.) and Departments o f Anatomy (C.M.) and Oral and Maxillofacial Surgery (S.F.), Faculty of Medicine, Kyoto University, Kyoto 606, Japan, and Institute of Medical Science, University of Tokyo, Minatoku, Tokyo 108, Japan (T.A.)
ABSTRACT Epidermal growth factor (EGF) stimulatesthe growth of various tissues and, therefore, EGF receptor expression in fetal tissues may play a key role in organogenesis. We have examined immunohistochemicallythe ontogeny and localization of the EGF receptor in the fetal mouse palate during in vivo and in vitro palatogenesis using the anti-human EGF receptor rabbit antibody. Immunoreactive products against the EGF receptor were observed in the palatal tissue examined on days 12,13, and 14 of gestation. On days 12 and 13, the immunoreactive products were predominantly positive on the oral and medial edge epithelia but were minimal on the epithelium of the vertical shelf. The EGF receptor immunoreactivity was less intense in the posterior palate as compared with the midpalatal region. In the fusing palate of day 14 fetuses, the cells forming the midline epithelial seam were continuously positive for EGF-Rimmunoreactivity.The mesenchyme of palatal shelves also showed regional heterogeneity and temporal sequence in EGF receptor expression. The localization of the EGF receptor in fetal mouse palates cultured in a serumless medium generally simulated that observed in vivo.
On the other hand, EGF receptors (EGF-R) have been detected in the developing palatal tissue of mouse and rat embryos (Nex@e t al., 1980; Abbott and Pratt, 1988; Abbott et al., 1988). Autoradiographic studies have demonstrated the presence of EGF-R binding sites in the palatal epithelium of mouse embryos by day 13 of gestation (Nex@e t al., 1980). Recently, Abbott et al. (1988) showed by immunoperoxidase staining that the medial epithelium of fetal mouse palates contains EGF-R a t least from day 12 of gestation. Therefore, EGF-R expression in the palatal tissue appears to be related to mammalian palatogenesis, but the details remain to be elucidated. The present study was undertaken to demonstrate immunohistochemically the localization of EGF-R in the fetal mouse palate and their temporal sequence during in vivo and in vitro palatogenesis. MATERIALS AND METHODS
Mature female ICR mice (SLC, Shizuoka) were mated overnight with a male mouse, and the day on which a vaginal plug was found was designated as day 0 of gestation. Between 1000 and 1200 h r on days 12, 13, and 14, pregnant females were killed by cervical dislocation. The fetuses were removed from the uterus and their maxillary region was dissected. The fetal palates were fixed in 4% paraformaldehyde, embedded in paraffin, and serially sectioned. INTRODUCTION For in vitro studies, fetuses were aseptically removed The formation of the secondary palate in mammalian from pregnant ICR mice on the morning of day 13, and fetuses involves a complex series of developmental their maxillary region was dissected with sterile scalsteps including the growth, elevation, and fusion of pal- pels and cultured in a chemically defined medium by a atal shelves. The proliferation, differentiation, and in- suspension culture technique. Briefly, palatal explants teractions of epithelial and mesenchymal cells are all were cultivated in a 50 ml penicillin bottle with 8 ml of considered to play important roles in palatogenesis the culture medium. The details of the composition of (Greene and Pratt, 1976; Pratt and Christiansen, 1980; the culture medium have been described elsewhere (Shiota et al., 1990). No serum or antibiotics were Ferguson, 1988). Experimental studies have reported that epidermal added to the medium. Three or four explants were put growth factor (EGF), a n embryonicifetal hormone re- into one bottle, and the bottle was sealed air-tight with sponsible for the growth of specific epithelial tissues, a rubber stopper and a metal clamp. The bottle was may be involved in palatal differentiation. Fetal pala- then flushed with a gas mixture of 50% O,, 5% CO,, tal shelves could be successfully cultured in a serum- and 45% N, using syringe needles, and was incubated free medium containing EGF, and selective removal of a t 37°C on a roller device a t a speed of 25 rpm. After in EGF from the medium resulted in reduced growth and death of palatal epithelial cells (Pratt, 1987). It has also been shown that EGF inhibits the degeneration of the medial palatal epithelium of the rat fetus and promotes hypertrophy and keratinization of these cells Received August 7, 1989. Accepted January 29, 1990. (Hassell, 1975). Yoneda and Pratt (1981) showed that Address reprint requests to Kohei Shlota, M.D., Congenital Anommesenchymal cells from the human embryonic palate aly Research Center, Faculty of Medicine, Kyoto University, Kyoto are highly responsive to the stimulatory effect of EGF. 606, Japan. 8 1990 WILEY-LISS, INC.
402
K. SHIOTA ET AL.
vitro culture for 24 or 48 hr, the explants were washed with saline solution and fixed in 4%paraformaldehyde. They were embedded in paraffin and sectioned. Paraffin sections were reacted immunohistochemically with the anti-EGF-R rabbit antibody against synthetic peptides corresponding to the amino acid residues 1173 to 1186 in human EGF-R (hereafter referred to as the anti-EGF-R-Ab) (Akiyama et al., 1988). After washing with phosphate-buffered saline (PBS, pH 7.21, immunohistochemical staining of the sections was done according to the avidin-biotin-peroxidase complex (ABC) method (Hsu et al., 1981) by using a Vectastain elite ABC kit (Vector Laboratories, Inc., California, USA). The bound peroxidase was visualized using the diaminobenzidine/H,O, reaction in Graham/Karnovsky’s reaction medium (Graham and Karnovsky, 1966). For the control immunohistochemical staining, the anti-EGF-R-Ab was replaced by normal rabbit serum to get a negative reaction. The nuclei in the sections were counterstained with 1% methyl green dissolved in veronal acetate buffer a t pH 4.2. In absorption studies using the amino acid residues 11731186 in the human EGF-R, immunoreactive products against the EGF-R were not detected in the sections from the specimens examined. RESULTS
EGF-R immunoreactivity was detected in all the fetal mouse palates examined on days 12, 13, and 14 of gestation. In most cases, EGF-R were expressed predominantly on the epithelial cells and less intensely in the mesenchymal tissue of the palatal shelf. No immunoreactivity was detected in control sections incubated with normal rabbit serum. In the midpalatal region of day 12 fetal palates, the oral and medial edge epithelia of the shelf were stained intensely for EGF-R, but the epithelium of the vertical wall showed only minimal immunoreactive EGF-R staining. The mesenchyme of palatal shelves was stained homogeneously for EGF-R (Fig. 1A). In the posterior region of the palate, however, both the epithelium and mesenchyme exhibited much less intense EGF-R immunoreactivity than in the midpalatal region (Fig. 1B). On day 13 when palatal shelves elevate and grow in a medial direction, the oral and medial edge epithelia showed intense immunoreactive EGF-R staining. However, EGF-R immunoreactivity was still minimal in the epithelium of the vertical wall. The loose mesenchyme of palatal shelves was intensely positive for EGF-R, but the mesenchymal cells closer to the epithelium exhibited less intense staining (Fig. 2A, B). In the posterior region of the palatal shelf, immunoreactive EGF-R staining had become more intense as compared with the day 12 palate. However, the immunoreactivity at the posterior palate was less intense than in the midpalatal region (Fig. 2C). On day 14 of gestation, palatal shelves have contacted each other along the median line, and, in many cases, an epithelial seam was observed between the fusing shelves. Figure 3 shows a section of the palate in which shelves have just contacted each other. Although rather intense immunoreactive EGF-R staining was recognized in the epithelium and mesenchyme, the immunoreactivity was significantly weaker on some epi-
thelial cells at the medial tip and in the underlying mesenchyme as compared with the neighboring tissues. EGF-R immunoreactivity was more intense on the oral epithelium than on the nasal epithelium. In the palates with a midline epithelial seam, the cells forming the epithelial seam were positive for the EGF-R immunoreactive products (Fig. 4). The EGF-R were expressed intensely on the oral epithelium, which was becoming stratified squamous. The nasal epithelium showed positive but relatively less intense EGF-R staining. The mesenchyme adjacent to the epithelial seam showed less intense EGF-R immunoreactivity than the deeper mesenchyme of the fusing shelves. Fetal mouse palates cultured in vitro for 24-48 hr also expressed EGF-R, and the localization and ontogenetic pattern of the immunoreactive products generally resembled those observed in fetal palates in vivo (Fig. 5). Although the mesenchyme adjacent to the fusion seam at first exhibited only weak EGF-R immunoreactivity (Fig. 5A,B), it became more immunoreactive when the epithelial seam was disappearing and the shelf mesenchyme was becoming confluent (Fig. 5C,D). DISCUSSION
It has been shown that EGF is a potent mitogen and induces accelerated growth and differentiation of a variety of cells (Carpenter and Cohen, 1979). In the adult, actively proliferating epithelia express EGF-R a t the cell surface; and once cells have diffcrcntiated into a nonproliferating component, EGF-R is reduced to undetectable levels or extinguished (Damjanov et al., 1986). Recent studies have demonstrated the presence of EGF-R in various fetal tissues of the human (Oliver, 1988), ovine (Johnson et al., 1989), and rodent (Nexg et al., 1980; Adamson et al., 1981; Adamson and Meek, 1984; Abbott and Pratt, 1988; Abbott et al., 1988). Adamson and Meek (1984) observed some temporal sequence in the EGF-R expression in tissues of the mouse embryo and hypothesized that EGF may function differently throughout development; it initially stimulates proliferation of embryonic cells and then stimulates differentiation as the tissues mature. In the present immunohistochemical study, EGF-R immunoreactivity was detected in the developing mouse palate on days 12, 13, and 14 of gestation, and the immunoreactivity in the epithelium and mesenchyme of the palate exhibited some regional heterogeneity and temporal sequence. Generally, the EGF-R staining was predominantly positive on the oral and medial edge epithelia and was less intense on the epithelium of the vertical wall. After palatal closure, EGF-R immunoreactivity was more intense on the oral epithelium than on the nasal epithelium. The significance of such regional heterogeneity of EGF-R expression is difficult to interpret, but the regionally specific expression of EGF-R may possibly be required for palate fusion and/or the differentiation of specific cells. Following palate fusion, the nasal epithelium differentiates into pseudostratified ciliated columnar cells, while the oral epithelia differentiates into stratified squamous cells. Since EGF is known to stimulate the division, hypertrophy, and keratinization of oral epithelial cells (Hassell, 1975; Grove and Pratt, 1984) and since EGF-R were expressed more intensely on the oral
EGF RECEPTOR IN MOUSE PALATOGENESIS
Fig. 1. Immunohistochemical staining of the day 12 fetal mouse palate using the anti-human EGF receptor rabbit antibody against synthetic peptides corresponding to the amino acid residues 1173 to 1186 in the human EGF receptor, with 1%methyl green as a counterstain. (A) Midpalatal region. Immunoreactive products against the EGF receptor (brown precipitates) are observed on the medial edge
403
epithelium (arrowheads), but the epithelium of the vertical wall show little immunoreactivity. The palatal mesenchyme is uniformly positive for the EGF receptor. (B) Posterior palate. Both the epithelium and mesenchyme show minimal or no immunoreactivity. Bars = 50 pm.
epithelium, it is possible that EGF-R are involved at against EGF-R were continuously detected in the midleast in the differentiation of oral epithelial cells. line epithelial seam during palate fusion (Figs. 4,5). Some epithelial cells were found to lose EGF-R im- Although the disappearance of the medial edge epithemunoreactivity at the medial tip of the fusing palate lium after palate fusion is often cited as a classical when the shelves contacted each other (Fig. 3 ) . With example of the programmed cell-death concept, i t has this regard, i t should be noted that on medial edge recently been proposed that the epithelial cells of the epithelial contact, the cells at the medial tip develop seam transform into mesenchymal cells during mesospecific cell adhesion molecules and desmosomes that dermal confluence (Fitchett and Hay, 1989). In their may be necessary for palate fusion (Ferguson, 1988). histological study on the fetal rat palate, Fitchett and Therefore, the localized diminution of EGF-R immu- Hay (1989) showed that the cells in the disappearing noreactivity may be related to such differentiation of seam were not dying but were healthy in appearance. specific epithelial cells, but this needs to be investi- Our data on the EGF-R immunoreactivity of the epigated in future. It is yet unknown and should be elu- thelial seam appear to support the persistence of epicidated whether the loss of EGF-R immunoreactivity thelial cells and their mesenchymal conversion during has anything to do with the cessation of DNA synthesis palate fusion. Our present study confirmed the previous finding of (Hudson and Shapiro, 1973; Pratt and Martin, 1975) and selective death (Smiley and Koch, 1972; Tyler and Abbott et al. (1988) that the day 12 fetal mouse palate Koch, 1974) in the medial edge epithelial cells prior to contains EGF-R. However, when our results were compared with their work, some differences were noted beshelf contact. In the present study, immunoreactive products tween the two studies. Although Abbott et al. (1988)
Fig. 2. Immunohistochemical staining of the day 13 fetal mouse palate using the anti-human EGF receptor rabbit antibody. (A) Midpalatal region. Immunoreactive products are recognized on the oral and medial edge epithelia (arrowheads) and in the mesenchyme. (B) Higher magnification of the oral epithelium and its underlying mesenchyme at the midpalatal region. Note that the deeper loose mesenchyme (asterisk)shows more intense immunoreactivity than the mesenchyme closer to the epithelium. (C) Posterior palate. Reaction products against the EGF receptor are recognizable on the oral epithelium (arrowheads) and deeper mesenchyme, but are minimal on the vertical wall epithelium. Bars = 50 pm.
Fig. 3. Immunohistochemical staining of the day 14 fetal mouse palate using the anti-human EGF receptor rabbit antibody. (A) Opposed palatal shelves immediately after contact. The nasal epithelium shows less intense immunoreactivity than the oral epithelium. (B) Higher magnification of the medial tips of the shelves. Note that the immunoreactive staining is relatively weak on the medial tip epithelium (arrowheads) and in the mesenchyme close to the epithelium. NC, Nasal cavity; OC, oral cavity. Bars = 50 pm.
Fig. 4. Immunohistochemical staining of a day 14 fetal mouse palate with a midline epithelial seam (arrowheads) using the anti-human EGF receptor rabbit antibody. Immunoreactive products are recognized on the oral epithelium and epithelial seam, and less intensely on the nasal epithelium. Although the deeper palatal mesenchyme exhibits EGF-R immunoreactivity, few immunoreactive products are recognizable in the mesenchyme close to the epithelial seam. OC, Oral cavity. Bar = 50 p,m.
EGF RECEPTOR IN MOUSE PALATOGENESIS
showed that EGF-R immunoreactivity was equally intense in the nasal and oral epithelia until after palatal closure, we found that the oral epithelium was significantly more immunoreactive than the nasal epithelium at all time periods. In addition, in the study of Abbott et al. (1988), the immunoreactivity disappeared in a much larger area of the medial epithelium and the cells in the fusion seam did not express EGF-R, findings which were not consistent with our results. Since we used mice of essentially the same strain as theirs, these differences between the studies cannot be attributed to the strain difference of animals. The observed differences might be the result of the different antibodies used for EGF-R localization or of the different method of immunostaining (they used immunoperoxidase staining using a polyclonal antibody against purified mouse liver EGF-R). The EGF-R immunoreactivity in the mesenchyme was also worthy of note. Although the mesenchymal cells in the midpalatal region exhibited apparently homogeneous EGF-R staining on day 12, the immunoreactivity diminished in the mesenchyme beneath the medial edge epithelium by day 13. The lack of EGF-R in the mesenchyme adjacent to the medial edge was also observed by Ferguson (1988).In the deeper mesenchyme with intense EGF-R immunoreactivity, cells were less densely populated and there appeared to be more extracellular matrix than in the mesenchyme closer to the epithelium. In this context, it is interesting to note the previous findings that one of the chief components of the intrinsic shelf elevating force is the regionally specific accumulation of glycosaminoglycans, predominantly hyaluronic acid, which are synthesized by mesenchymal cells (Ferguson, 19881, and that the synthesis of hyaluronic acid by palatal mesenchymal cells is stimulated by EGF (Turley e t al., 1985). The EGF-R immunoreactivity in the mesenchyme was more intense in the midpalate than posteriorly (Figs. 1,2). This also seems consistent with the fact that there is more hyaluronic acid in the anterior portion of the palate than i n the posterior palate (Knudsen et al., 1985; Brinkley and Morris-Wiman, 1987). Since hyaluronic acid is capable of binding up to ten times its own weight in water and the accumulation of hyaluronic acid results in swelling of the extracellular matrix (Brinkley and Bookstein, 1986), its regional accumulation may play some role in the elevation and remodelling of palatal shelves. In our study, mesenchymal cells adjacent to the epithelial seam were found almost devoid of EGF-R (Fig. 4). Recently, Ferguson (1988) showed that mesenchyma1 cells adjacent to the fusion seam expressed EGF-R in a day 14 mouse embryo. The lack of EGF-R in the mesenchyme close to the seam may be a transient phe-
Fig. 5. Immunohistochemical staining of fetal mouse palates cultured in vitro using the anti-human EGF receptor rabbit antibody. (A) After 24 hr in culture. ( E D ) . After 48 hr in culture, showing various stages of palate fusion. The localization of the immunoreactive products appears to be similar to that observed in vivo (Figs. 1-4). The nasal and oral cavities are a t the top and bottom, respectively, of each figure. Bars = 50 pm.
407
nomenon, since EGF-R immunoreactivity became homogeneously positive in the mesenchyme when the epithelial seam was disappearing (Fig. 5). Abbott et al. (1988) detected EGF-R in fetal mouse palates on days 12-15 of gestation and noted a developmental decline in EGF-R in the palatal epithelium, demonstrated by a decrease in the binding of antibody to EGF-R and a decrease in the binding of 1251-labeled EGF. It has also been suggested that early embryonic tissues have high-affinity EGF-R, which respond to very low concentrations of EGF and that later fetal tissues have a larger number of EGF-R with lower affinity (Adamson et al., 1981; Adamson and Warshaw, 1982; Adamson and Meek, 1984). Thus, EGF-R might function variably during palatogenesis, but further data are needed before palatal differentiation can be analyzed in connection with the temporal sequence of EGF-R expression. It has been demonstrated that EGF-R expression can be disturbed by exogenous agents and that it may play a role in the induction of cleft palate (Abbott et al., 1988). In their experiment, while the medial epithelium of control fetal mice showed a developmental decline in EGF-R, medial epithelial cells exposed to alltrans-retinoic acid, which is a potent teratogen and induces cleft palate in mice, continued to express EGFR and bind EGF. Such a disturbance of EGF-R expression in the fetal palate could possibly be involved in the production of cleft palate by various environmental agents. In our in vitro study, fetal mouse palates cultured in a chemically defined medium showed a similar pattern of immunoreactive EGF-R staining as that seen in vivo. Since the palates were cultured in a serum-free medium, the expression of EGF-R is considered to be intrinsically regulated. This new technique of culturing fetal mouse palates in a chemically defined, serumless medium (Shiota e t al., 1990) may be useful for studying the role of growth factors and other hormones in murine palatogenesis. ACKNOWLEDGMENTS
This work was supported by research grants to K.S. from the Japanese Ministry of Education, Science and Culture (02670007) and the Fujiwara Memorial Foundation. LITERATURE CITED Abbott. B.D.. and R.M. Pratt 1988 EGF receDtor in the develoDine tooth is altered by exogenous retinoic acih and EGF. Dev. Biol, 128:300-304. ~~. . ~ ~.~ . Abbott, B.D., E.D. Adamson, and R.M. Pratt 1988 Retinoic acid alters EGF receptor expression during palatogenesis. Development, 102.353-867. Adamson, E.D., and J. Meek 1984 The ontogeny of epidermal growth factor receptors during mouse development. Dev. Biol., 103t6270. Adamson, E.D., and J.B. Warshaw 1982 Down-regulation of epidermal growth factor receptors in mouse embryos. Dev. Biol., 90: 430-434. Adamson, E.D., M.J. Deller, and J.B. Warshaw 1981 Functional EGF receptors are present on mouse embryo tissues. Nature (London), 291t656-659. Akiyama, T., T. Saito, H. Ogawara, K. Toyoshima, and T. Yamamoto 1988 Tumor promoter and epidermal growth factor stimulate phosphorylation of the c-erbB-2 gene product in MKN-7 human adenocarcinoma cells. Mol. Cell. Biol., 8t1019-1026. Brinkley, L.L., and F.L. Bookstein 1986 Cell distribution during
408
K. SHIOTA ET AL.
mouse secondary palate closure. 11. Mesenchymal cells. J . Embryol. Exp. Morphol., 96t111-130. Brinkley, L.L., and J. Morris-Wiman 1987 Computer-assisted analysis of hyaluronate distribution during morphogenesis of the mouse secondary palate. Development, 1OOt629-636. Carpenter, G., and S. Cohen 1979 Epidermal growth factor. Annu. Rev. Biochem., 48~193-216. Damjanov, I., B. Mildner, and B.B. Knowles 1986 Immunohistochemical localization of the epidermal growth factor receptor in normal human tissues. Lab. Invest., 55t588-592. Ferguson, M.W.J. 1988 Palate development. Development, 103 Suppl.t41-60. Fitchett, J.E., and E.D. Hay 1989 Medial edge epithelium transforms to mesenchyme after embryonic palatal shelves fuse. Dev. Biol., 131r455-474. Graham, R.C., and M.J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney. J. Histochem. Cytochem., 14.291-302. Greene, R.M., and R.M. Pratt 1976 Developmental aspects of secondary palate formation. J . Embryol. Exp. Morphol., 36225-245. Grove, R.I., and R.M. Pratt 1984 Influence of epidermal growth factor and cyclic AMP on growth and differentiation of palatal epithelial cells in culture. Dev. Biol., 106t427-437. Hassell, J.R. 1975 The development of rat palatal shelves in uitro. An ultrastructural analysis of the inhibition of epithelial cell death and palate fusion by the epidermal growth factor. Dev. Biol., 45: 90-102. Hsu, S.-M., L. Raine, and H. Fanger 1981 Use of avidin-biotin peroxidase complex (ABC) in immunoperoxidase techniques: A comparison between ABC and unlabeled antibody (PAP) procedures. J . Histochem. Cytochem., 29.577-580. Hudson, C.D., and B.L. Shapiro 1973 A radioautographic study of deoxyribonucleic acid synthesis in embryonic rat palatal shelf epithelium with reference to the concept of programmed cell death. Arch. Oral Biol., 18t77-84. Johnson, M.D., M.E. Gray, G. Carpenter, R.B. Pepinsky, H. Sundell, and M.T. Stahlman 1Y89 Ontogeny of epidermal growth Liactur
receptorikinase and of lipocortin-1 in the ovine lung. Pediat. Res., 25535 -541. Knudsen, T.B., R.F. Bulleit, and E.F. Zimmerman 1985 Histochemical localization of glycosaminoglycans during morphogenesis of the secondary palate in mice. Anat. Embryol., 173r137-142. Nexe, E., M.D. Hollenberg, A. Figueroa, and R.M. Pratt 1980 Detection of epidermal growth factor-urogastrone and its receptor during fetal mouse development. Proc. Natl. Acad. Sci. U.S.A., 77; 2782-2785. Oliver, A.M. 1988 Epidermal growth factor receptor expression in human foetal tissues is age-dependent. Br. J . Cancer, 58t461463. Pratt, R.M. 1987 Role of epidermal growth factor in embryonic development. In: Current Topics in Developmental Biology, Vol. 22. R. Sawyer, ed. Academic Press, New York, pp. 175-193. Pratt, R.M., and R.C. Christiansen 1980 Current Research Trends in Prenatal Craniofacial Development. ElsevieriNorth-Holland, Amsterdam. Pratt, R.M., and G.R. Martin 1975 Epithelial cell death and cyclic AMP increase during palatal development. Proc. Natl. Acad. Sci. U.S.A., 72t874-877. Shiota, K., T. Kosazuma, S. Klug, and D. Neubert 1990 Development of the fetal mouse palate in suspension organ culture. Acta anat. 137t59-64. Smiley, G.R., and W.S. Koch 1972 An in vitro and in vivo study of single palatal processes. Anat. Rec., 173t405-416. Turley, E.A., M.D. Hollenberg, and R.M. Pratt 1985 Effects of epidermal growth factoriurogastrone on glycosaminoglycan synthesis and accumulation in vitro in the developing mouse palate. Differentiation, 28.279-285. Tyler, M.S., and W.S. Koch 1974 Epithelial-mesenchymal interactions in the secondary palate of the mouse. J. Dent. Res., 53 Suppl.t64. Yoneda, T., and R.M. Pratt 1981 Mesenchymal cells from the human embryonic palate are highly responsive to epidermal growth factor. Science, 213t563-565.
Expression of the Epidermal Growth Factor Receptor in Developing Fetal Mouse Palates: An lmmunohistochemical Study KOHEI SHIOTA, SHIGEYUKI FUJITA, TETSU AKIYAMA, AND CHISATO MORI Congenital Anomaly Research Center (K.S.) and Departments o f Anatomy (C.M.) and Oral and Maxillofacial Surgery (S.F.), Faculty of Medicine, Kyoto University, Kyoto 606, Japan, and Institute of Medical Science, University of Tokyo, Minatoku, Tokyo 108, Japan (T.A.)
ABSTRACT Epidermal growth factor (EGF) stimulatesthe growth of various tissues and, therefore, EGF receptor expression in fetal tissues may play a key role in organogenesis. We have examined immunohistochemicallythe ontogeny and localization of the EGF receptor in the fetal mouse palate during in vivo and in vitro palatogenesis using the anti-human EGF receptor rabbit antibody. Immunoreactive products against the EGF receptor were observed in the palatal tissue examined on days 12,13, and 14 of gestation. On days 12 and 13, the immunoreactive products were predominantly positive on the oral and medial edge epithelia but were minimal on the epithelium of the vertical shelf. The EGF receptor immunoreactivity was less intense in the posterior palate as compared with the midpalatal region. In the fusing palate of day 14 fetuses, the cells forming the midline epithelial seam were continuously positive for EGF-Rimmunoreactivity.The mesenchyme of palatal shelves also showed regional heterogeneity and temporal sequence in EGF receptor expression. The localization of the EGF receptor in fetal mouse palates cultured in a serumless medium generally simulated that observed in vivo.
On the other hand, EGF receptors (EGF-R) have been detected in the developing palatal tissue of mouse and rat embryos (Nex@e t al., 1980; Abbott and Pratt, 1988; Abbott et al., 1988). Autoradiographic studies have demonstrated the presence of EGF-R binding sites in the palatal epithelium of mouse embryos by day 13 of gestation (Nex@e t al., 1980). Recently, Abbott et al. (1988) showed by immunoperoxidase staining that the medial epithelium of fetal mouse palates contains EGF-R a t least from day 12 of gestation. Therefore, EGF-R expression in the palatal tissue appears to be related to mammalian palatogenesis, but the details remain to be elucidated. The present study was undertaken to demonstrate immunohistochemically the localization of EGF-R in the fetal mouse palate and their temporal sequence during in vivo and in vitro palatogenesis. MATERIALS AND METHODS
Mature female ICR mice (SLC, Shizuoka) were mated overnight with a male mouse, and the day on which a vaginal plug was found was designated as day 0 of gestation. Between 1000 and 1200 h r on days 12, 13, and 14, pregnant females were killed by cervical dislocation. The fetuses were removed from the uterus and their maxillary region was dissected. The fetal palates were fixed in 4% paraformaldehyde, embedded in paraffin, and serially sectioned. INTRODUCTION For in vitro studies, fetuses were aseptically removed The formation of the secondary palate in mammalian from pregnant ICR mice on the morning of day 13, and fetuses involves a complex series of developmental their maxillary region was dissected with sterile scalsteps including the growth, elevation, and fusion of pal- pels and cultured in a chemically defined medium by a atal shelves. The proliferation, differentiation, and in- suspension culture technique. Briefly, palatal explants teractions of epithelial and mesenchymal cells are all were cultivated in a 50 ml penicillin bottle with 8 ml of considered to play important roles in palatogenesis the culture medium. The details of the composition of (Greene and Pratt, 1976; Pratt and Christiansen, 1980; the culture medium have been described elsewhere (Shiota et al., 1990). No serum or antibiotics were Ferguson, 1988). Experimental studies have reported that epidermal added to the medium. Three or four explants were put growth factor (EGF), a n embryonicifetal hormone re- into one bottle, and the bottle was sealed air-tight with sponsible for the growth of specific epithelial tissues, a rubber stopper and a metal clamp. The bottle was may be involved in palatal differentiation. Fetal pala- then flushed with a gas mixture of 50% O,, 5% CO,, tal shelves could be successfully cultured in a serum- and 45% N, using syringe needles, and was incubated free medium containing EGF, and selective removal of a t 37°C on a roller device a t a speed of 25 rpm. After in EGF from the medium resulted in reduced growth and death of palatal epithelial cells (Pratt, 1987). It has also been shown that EGF inhibits the degeneration of the medial palatal epithelium of the rat fetus and promotes hypertrophy and keratinization of these cells Received August 7, 1989. Accepted January 29, 1990. (Hassell, 1975). Yoneda and Pratt (1981) showed that Address reprint requests to Kohei Shlota, M.D., Congenital Anommesenchymal cells from the human embryonic palate aly Research Center, Faculty of Medicine, Kyoto University, Kyoto are highly responsive to the stimulatory effect of EGF. 606, Japan. 8 1990 WILEY-LISS, INC.
402
K. SHIOTA ET AL.
vitro culture for 24 or 48 hr, the explants were washed with saline solution and fixed in 4%paraformaldehyde. They were embedded in paraffin and sectioned. Paraffin sections were reacted immunohistochemically with the anti-EGF-R rabbit antibody against synthetic peptides corresponding to the amino acid residues 1173 to 1186 in human EGF-R (hereafter referred to as the anti-EGF-R-Ab) (Akiyama et al., 1988). After washing with phosphate-buffered saline (PBS, pH 7.21, immunohistochemical staining of the sections was done according to the avidin-biotin-peroxidase complex (ABC) method (Hsu et al., 1981) by using a Vectastain elite ABC kit (Vector Laboratories, Inc., California, USA). The bound peroxidase was visualized using the diaminobenzidine/H,O, reaction in Graham/Karnovsky’s reaction medium (Graham and Karnovsky, 1966). For the control immunohistochemical staining, the anti-EGF-R-Ab was replaced by normal rabbit serum to get a negative reaction. The nuclei in the sections were counterstained with 1% methyl green dissolved in veronal acetate buffer a t pH 4.2. In absorption studies using the amino acid residues 11731186 in the human EGF-R, immunoreactive products against the EGF-R were not detected in the sections from the specimens examined. RESULTS
EGF-R immunoreactivity was detected in all the fetal mouse palates examined on days 12, 13, and 14 of gestation. In most cases, EGF-R were expressed predominantly on the epithelial cells and less intensely in the mesenchymal tissue of the palatal shelf. No immunoreactivity was detected in control sections incubated with normal rabbit serum. In the midpalatal region of day 12 fetal palates, the oral and medial edge epithelia of the shelf were stained intensely for EGF-R, but the epithelium of the vertical wall showed only minimal immunoreactive EGF-R staining. The mesenchyme of palatal shelves was stained homogeneously for EGF-R (Fig. 1A). In the posterior region of the palate, however, both the epithelium and mesenchyme exhibited much less intense EGF-R immunoreactivity than in the midpalatal region (Fig. 1B). On day 13 when palatal shelves elevate and grow in a medial direction, the oral and medial edge epithelia showed intense immunoreactive EGF-R staining. However, EGF-R immunoreactivity was still minimal in the epithelium of the vertical wall. The loose mesenchyme of palatal shelves was intensely positive for EGF-R, but the mesenchymal cells closer to the epithelium exhibited less intense staining (Fig. 2A, B). In the posterior region of the palatal shelf, immunoreactive EGF-R staining had become more intense as compared with the day 12 palate. However, the immunoreactivity at the posterior palate was less intense than in the midpalatal region (Fig. 2C). On day 14 of gestation, palatal shelves have contacted each other along the median line, and, in many cases, an epithelial seam was observed between the fusing shelves. Figure 3 shows a section of the palate in which shelves have just contacted each other. Although rather intense immunoreactive EGF-R staining was recognized in the epithelium and mesenchyme, the immunoreactivity was significantly weaker on some epi-
thelial cells at the medial tip and in the underlying mesenchyme as compared with the neighboring tissues. EGF-R immunoreactivity was more intense on the oral epithelium than on the nasal epithelium. In the palates with a midline epithelial seam, the cells forming the epithelial seam were positive for the EGF-R immunoreactive products (Fig. 4). The EGF-R were expressed intensely on the oral epithelium, which was becoming stratified squamous. The nasal epithelium showed positive but relatively less intense EGF-R staining. The mesenchyme adjacent to the epithelial seam showed less intense EGF-R immunoreactivity than the deeper mesenchyme of the fusing shelves. Fetal mouse palates cultured in vitro for 24-48 hr also expressed EGF-R, and the localization and ontogenetic pattern of the immunoreactive products generally resembled those observed in fetal palates in vivo (Fig. 5). Although the mesenchyme adjacent to the fusion seam at first exhibited only weak EGF-R immunoreactivity (Fig. 5A,B), it became more immunoreactive when the epithelial seam was disappearing and the shelf mesenchyme was becoming confluent (Fig. 5C,D). DISCUSSION
It has been shown that EGF is a potent mitogen and induces accelerated growth and differentiation of a variety of cells (Carpenter and Cohen, 1979). In the adult, actively proliferating epithelia express EGF-R a t the cell surface; and once cells have diffcrcntiated into a nonproliferating component, EGF-R is reduced to undetectable levels or extinguished (Damjanov et al., 1986). Recent studies have demonstrated the presence of EGF-R in various fetal tissues of the human (Oliver, 1988), ovine (Johnson et al., 1989), and rodent (Nexg et al., 1980; Adamson et al., 1981; Adamson and Meek, 1984; Abbott and Pratt, 1988; Abbott et al., 1988). Adamson and Meek (1984) observed some temporal sequence in the EGF-R expression in tissues of the mouse embryo and hypothesized that EGF may function differently throughout development; it initially stimulates proliferation of embryonic cells and then stimulates differentiation as the tissues mature. In the present immunohistochemical study, EGF-R immunoreactivity was detected in the developing mouse palate on days 12, 13, and 14 of gestation, and the immunoreactivity in the epithelium and mesenchyme of the palate exhibited some regional heterogeneity and temporal sequence. Generally, the EGF-R staining was predominantly positive on the oral and medial edge epithelia and was less intense on the epithelium of the vertical wall. After palatal closure, EGF-R immunoreactivity was more intense on the oral epithelium than on the nasal epithelium. The significance of such regional heterogeneity of EGF-R expression is difficult to interpret, but the regionally specific expression of EGF-R may possibly be required for palate fusion and/or the differentiation of specific cells. Following palate fusion, the nasal epithelium differentiates into pseudostratified ciliated columnar cells, while the oral epithelia differentiates into stratified squamous cells. Since EGF is known to stimulate the division, hypertrophy, and keratinization of oral epithelial cells (Hassell, 1975; Grove and Pratt, 1984) and since EGF-R were expressed more intensely on the oral
EGF RECEPTOR IN MOUSE PALATOGENESIS
Fig. 1. Immunohistochemical staining of the day 12 fetal mouse palate using the anti-human EGF receptor rabbit antibody against synthetic peptides corresponding to the amino acid residues 1173 to 1186 in the human EGF receptor, with 1%methyl green as a counterstain. (A) Midpalatal region. Immunoreactive products against the EGF receptor (brown precipitates) are observed on the medial edge
403
epithelium (arrowheads), but the epithelium of the vertical wall show little immunoreactivity. The palatal mesenchyme is uniformly positive for the EGF receptor. (B) Posterior palate. Both the epithelium and mesenchyme show minimal or no immunoreactivity. Bars = 50 pm.
epithelium, it is possible that EGF-R are involved at against EGF-R were continuously detected in the midleast in the differentiation of oral epithelial cells. line epithelial seam during palate fusion (Figs. 4,5). Some epithelial cells were found to lose EGF-R im- Although the disappearance of the medial edge epithemunoreactivity at the medial tip of the fusing palate lium after palate fusion is often cited as a classical when the shelves contacted each other (Fig. 3 ) . With example of the programmed cell-death concept, i t has this regard, i t should be noted that on medial edge recently been proposed that the epithelial cells of the epithelial contact, the cells at the medial tip develop seam transform into mesenchymal cells during mesospecific cell adhesion molecules and desmosomes that dermal confluence (Fitchett and Hay, 1989). In their may be necessary for palate fusion (Ferguson, 1988). histological study on the fetal rat palate, Fitchett and Therefore, the localized diminution of EGF-R immu- Hay (1989) showed that the cells in the disappearing noreactivity may be related to such differentiation of seam were not dying but were healthy in appearance. specific epithelial cells, but this needs to be investi- Our data on the EGF-R immunoreactivity of the epigated in future. It is yet unknown and should be elu- thelial seam appear to support the persistence of epicidated whether the loss of EGF-R immunoreactivity thelial cells and their mesenchymal conversion during has anything to do with the cessation of DNA synthesis palate fusion. Our present study confirmed the previous finding of (Hudson and Shapiro, 1973; Pratt and Martin, 1975) and selective death (Smiley and Koch, 1972; Tyler and Abbott et al. (1988) that the day 12 fetal mouse palate Koch, 1974) in the medial edge epithelial cells prior to contains EGF-R. However, when our results were compared with their work, some differences were noted beshelf contact. In the present study, immunoreactive products tween the two studies. Although Abbott et al. (1988)
Fig. 2. Immunohistochemical staining of the day 13 fetal mouse palate using the anti-human EGF receptor rabbit antibody. (A) Midpalatal region. Immunoreactive products are recognized on the oral and medial edge epithelia (arrowheads) and in the mesenchyme. (B) Higher magnification of the oral epithelium and its underlying mesenchyme at the midpalatal region. Note that the deeper loose mesenchyme (asterisk)shows more intense immunoreactivity than the mesenchyme closer to the epithelium. (C) Posterior palate. Reaction products against the EGF receptor are recognizable on the oral epithelium (arrowheads) and deeper mesenchyme, but are minimal on the vertical wall epithelium. Bars = 50 pm.
Fig. 3. Immunohistochemical staining of the day 14 fetal mouse palate using the anti-human EGF receptor rabbit antibody. (A) Opposed palatal shelves immediately after contact. The nasal epithelium shows less intense immunoreactivity than the oral epithelium. (B) Higher magnification of the medial tips of the shelves. Note that the immunoreactive staining is relatively weak on the medial tip epithelium (arrowheads) and in the mesenchyme close to the epithelium. NC, Nasal cavity; OC, oral cavity. Bars = 50 pm.
Fig. 4. Immunohistochemical staining of a day 14 fetal mouse palate with a midline epithelial seam (arrowheads) using the anti-human EGF receptor rabbit antibody. Immunoreactive products are recognized on the oral epithelium and epithelial seam, and less intensely on the nasal epithelium. Although the deeper palatal mesenchyme exhibits EGF-R immunoreactivity, few immunoreactive products are recognizable in the mesenchyme close to the epithelial seam. OC, Oral cavity. Bar = 50 p,m.
EGF RECEPTOR IN MOUSE PALATOGENESIS
showed that EGF-R immunoreactivity was equally intense in the nasal and oral epithelia until after palatal closure, we found that the oral epithelium was significantly more immunoreactive than the nasal epithelium at all time periods. In addition, in the study of Abbott et al. (1988), the immunoreactivity disappeared in a much larger area of the medial epithelium and the cells in the fusion seam did not express EGF-R, findings which were not consistent with our results. Since we used mice of essentially the same strain as theirs, these differences between the studies cannot be attributed to the strain difference of animals. The observed differences might be the result of the different antibodies used for EGF-R localization or of the different method of immunostaining (they used immunoperoxidase staining using a polyclonal antibody against purified mouse liver EGF-R). The EGF-R immunoreactivity in the mesenchyme was also worthy of note. Although the mesenchymal cells in the midpalatal region exhibited apparently homogeneous EGF-R staining on day 12, the immunoreactivity diminished in the mesenchyme beneath the medial edge epithelium by day 13. The lack of EGF-R in the mesenchyme adjacent to the medial edge was also observed by Ferguson (1988).In the deeper mesenchyme with intense EGF-R immunoreactivity, cells were less densely populated and there appeared to be more extracellular matrix than in the mesenchyme closer to the epithelium. In this context, it is interesting to note the previous findings that one of the chief components of the intrinsic shelf elevating force is the regionally specific accumulation of glycosaminoglycans, predominantly hyaluronic acid, which are synthesized by mesenchymal cells (Ferguson, 19881, and that the synthesis of hyaluronic acid by palatal mesenchymal cells is stimulated by EGF (Turley e t al., 1985). The EGF-R immunoreactivity in the mesenchyme was more intense in the midpalate than posteriorly (Figs. 1,2). This also seems consistent with the fact that there is more hyaluronic acid in the anterior portion of the palate than i n the posterior palate (Knudsen et al., 1985; Brinkley and Morris-Wiman, 1987). Since hyaluronic acid is capable of binding up to ten times its own weight in water and the accumulation of hyaluronic acid results in swelling of the extracellular matrix (Brinkley and Bookstein, 1986), its regional accumulation may play some role in the elevation and remodelling of palatal shelves. In our study, mesenchymal cells adjacent to the epithelial seam were found almost devoid of EGF-R (Fig. 4). Recently, Ferguson (1988) showed that mesenchyma1 cells adjacent to the fusion seam expressed EGF-R in a day 14 mouse embryo. The lack of EGF-R in the mesenchyme close to the seam may be a transient phe-
Fig. 5. Immunohistochemical staining of fetal mouse palates cultured in vitro using the anti-human EGF receptor rabbit antibody. (A) After 24 hr in culture. ( E D ) . After 48 hr in culture, showing various stages of palate fusion. The localization of the immunoreactive products appears to be similar to that observed in vivo (Figs. 1-4). The nasal and oral cavities are a t the top and bottom, respectively, of each figure. Bars = 50 pm.
407
nomenon, since EGF-R immunoreactivity became homogeneously positive in the mesenchyme when the epithelial seam was disappearing (Fig. 5). Abbott et al. (1988) detected EGF-R in fetal mouse palates on days 12-15 of gestation and noted a developmental decline in EGF-R in the palatal epithelium, demonstrated by a decrease in the binding of antibody to EGF-R and a decrease in the binding of 1251-labeled EGF. It has also been suggested that early embryonic tissues have high-affinity EGF-R, which respond to very low concentrations of EGF and that later fetal tissues have a larger number of EGF-R with lower affinity (Adamson et al., 1981; Adamson and Warshaw, 1982; Adamson and Meek, 1984). Thus, EGF-R might function variably during palatogenesis, but further data are needed before palatal differentiation can be analyzed in connection with the temporal sequence of EGF-R expression. It has been demonstrated that EGF-R expression can be disturbed by exogenous agents and that it may play a role in the induction of cleft palate (Abbott et al., 1988). In their experiment, while the medial epithelium of control fetal mice showed a developmental decline in EGF-R, medial epithelial cells exposed to alltrans-retinoic acid, which is a potent teratogen and induces cleft palate in mice, continued to express EGFR and bind EGF. Such a disturbance of EGF-R expression in the fetal palate could possibly be involved in the production of cleft palate by various environmental agents. In our in vitro study, fetal mouse palates cultured in a chemically defined medium showed a similar pattern of immunoreactive EGF-R staining as that seen in vivo. Since the palates were cultured in a serum-free medium, the expression of EGF-R is considered to be intrinsically regulated. This new technique of culturing fetal mouse palates in a chemically defined, serumless medium (Shiota e t al., 1990) may be useful for studying the role of growth factors and other hormones in murine palatogenesis. ACKNOWLEDGMENTS
This work was supported by research grants to K.S. from the Japanese Ministry of Education, Science and Culture (02670007) and the Fujiwara Memorial Foundation. LITERATURE CITED Abbott. B.D.. and R.M. Pratt 1988 EGF receDtor in the develoDine tooth is altered by exogenous retinoic acih and EGF. Dev. Biol, 128:300-304. ~~. . ~ ~.~ . Abbott, B.D., E.D. Adamson, and R.M. Pratt 1988 Retinoic acid alters EGF receptor expression during palatogenesis. Development, 102.353-867. Adamson, E.D., and J. Meek 1984 The ontogeny of epidermal growth factor receptors during mouse development. Dev. Biol., 103t6270. Adamson, E.D., and J.B. Warshaw 1982 Down-regulation of epidermal growth factor receptors in mouse embryos. Dev. Biol., 90: 430-434. Adamson, E.D., M.J. Deller, and J.B. Warshaw 1981 Functional EGF receptors are present on mouse embryo tissues. Nature (London), 291t656-659. Akiyama, T., T. Saito, H. Ogawara, K. Toyoshima, and T. Yamamoto 1988 Tumor promoter and epidermal growth factor stimulate phosphorylation of the c-erbB-2 gene product in MKN-7 human adenocarcinoma cells. Mol. Cell. Biol., 8t1019-1026. Brinkley, L.L., and F.L. Bookstein 1986 Cell distribution during
408
K. SHIOTA ET AL.
mouse secondary palate closure. 11. Mesenchymal cells. J . Embryol. Exp. Morphol., 96t111-130. Brinkley, L.L., and J. Morris-Wiman 1987 Computer-assisted analysis of hyaluronate distribution during morphogenesis of the mouse secondary palate. Development, 1OOt629-636. Carpenter, G., and S. Cohen 1979 Epidermal growth factor. Annu. Rev. Biochem., 48~193-216. Damjanov, I., B. Mildner, and B.B. Knowles 1986 Immunohistochemical localization of the epidermal growth factor receptor in normal human tissues. Lab. Invest., 55t588-592. Ferguson, M.W.J. 1988 Palate development. Development, 103 Suppl.t41-60. Fitchett, J.E., and E.D. Hay 1989 Medial edge epithelium transforms to mesenchyme after embryonic palatal shelves fuse. Dev. Biol., 131r455-474. Graham, R.C., and M.J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney. J. Histochem. Cytochem., 14.291-302. Greene, R.M., and R.M. Pratt 1976 Developmental aspects of secondary palate formation. J . Embryol. Exp. Morphol., 36225-245. Grove, R.I., and R.M. Pratt 1984 Influence of epidermal growth factor and cyclic AMP on growth and differentiation of palatal epithelial cells in culture. Dev. Biol., 106t427-437. Hassell, J.R. 1975 The development of rat palatal shelves in uitro. An ultrastructural analysis of the inhibition of epithelial cell death and palate fusion by the epidermal growth factor. Dev. Biol., 45: 90-102. Hsu, S.-M., L. Raine, and H. Fanger 1981 Use of avidin-biotin peroxidase complex (ABC) in immunoperoxidase techniques: A comparison between ABC and unlabeled antibody (PAP) procedures. J . Histochem. Cytochem., 29.577-580. Hudson, C.D., and B.L. Shapiro 1973 A radioautographic study of deoxyribonucleic acid synthesis in embryonic rat palatal shelf epithelium with reference to the concept of programmed cell death. Arch. Oral Biol., 18t77-84. Johnson, M.D., M.E. Gray, G. Carpenter, R.B. Pepinsky, H. Sundell, and M.T. Stahlman 1Y89 Ontogeny of epidermal growth Liactur
receptorikinase and of lipocortin-1 in the ovine lung. Pediat. Res., 25535 -541. Knudsen, T.B., R.F. Bulleit, and E.F. Zimmerman 1985 Histochemical localization of glycosaminoglycans during morphogenesis of the secondary palate in mice. Anat. Embryol., 173r137-142. Nexe, E., M.D. Hollenberg, A. Figueroa, and R.M. Pratt 1980 Detection of epidermal growth factor-urogastrone and its receptor during fetal mouse development. Proc. Natl. Acad. Sci. U.S.A., 77; 2782-2785. Oliver, A.M. 1988 Epidermal growth factor receptor expression in human foetal tissues is age-dependent. Br. J . Cancer, 58t461463. Pratt, R.M. 1987 Role of epidermal growth factor in embryonic development. In: Current Topics in Developmental Biology, Vol. 22. R. Sawyer, ed. Academic Press, New York, pp. 175-193. Pratt, R.M., and R.C. Christiansen 1980 Current Research Trends in Prenatal Craniofacial Development. ElsevieriNorth-Holland, Amsterdam. Pratt, R.M., and G.R. Martin 1975 Epithelial cell death and cyclic AMP increase during palatal development. Proc. Natl. Acad. Sci. U.S.A., 72t874-877. Shiota, K., T. Kosazuma, S. Klug, and D. Neubert 1990 Development of the fetal mouse palate in suspension organ culture. Acta anat. 137t59-64. Smiley, G.R., and W.S. Koch 1972 An in vitro and in vivo study of single palatal processes. Anat. Rec., 173t405-416. Turley, E.A., M.D. Hollenberg, and R.M. Pratt 1985 Effects of epidermal growth factoriurogastrone on glycosaminoglycan synthesis and accumulation in vitro in the developing mouse palate. Differentiation, 28.279-285. Tyler, M.S., and W.S. Koch 1974 Epithelial-mesenchymal interactions in the secondary palate of the mouse. J. Dent. Res., 53 Suppl.t64. Yoneda, T., and R.M. Pratt 1981 Mesenchymal cells from the human embryonic palate are highly responsive to epidermal growth factor. Science, 213t563-565.
