SEARCH

SEARCH BY CITATION

Keywords:

  • salivary gland;
  • branching morphogenesis;
  • hepatocyte growth factor (HGF);
  • c-met;
  • epithelial-mesenchymal interaction;
  • tongue;
  • real-time RT-PCR

Abstract

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

We investigated the involvement of hepatocyte growth factor (HGF) in salivary gland (SG) branching morphogenesis. The mouse submandibular gland (SMG) starts to develop at embryonic day 11.5–12 (E11.5–E12), and branching morphogenesis occurs in the area between the mandibular bone and tongue between E14 and E16.5. Real-time reverse transcriptase-polymerase chain reaction showed that the expression of the c-met/HGF receptor gene in SMG increased and peaked between E14 and E16.5, concomitant with epithelial branching, and high levels of HGF mRNA were detected in the surrounding mesenchyme at E14–E15.5. Although strong expression of the HGF and c-met transcripts was observed in the tongue muscles, this expression was limited at E13.5–E14.5. Serum-free organ cultures were established, in which SG rudiments that contained SMG and sublingual gland (SLG) primordia (explant 1) and SMG/SLG rudiments with peripheral tissue that included part of the tongue muscle (explant 2) were isolated from E13.5 or E14 embryos. Mesenchyme-free SMG epithelium was obtained by the removal of mesenchymal tissue from explant 1. In the explant 1 and 2 organ cultures, SMG/SLG rudiments showed growth and branching morphogenesis, while mesenchyme-free epithelium failed to grow. When E13.5 or E14 mesenchyme-free epithelium and a recombinant human HGF (rh-HGF) -soaked bead were placed on Matrigel, the epithelium migrated toward the bead and formed branches, while the E13 epithelium failed to branch. The exogenous application of rh-HGF and anti-HGF antibody to the SMG/SLG rudiment cultures resulted in stimulation and inhibition, respectively, of branching morphogenesis. However, the response of E13.5 SMG to rh-HGF was very weak, while the branching of E14 SMG was enhanced strongly by rh-HGF. The branching morphogenesis of SMG was also inhibited by the addition of either antisense HGF or c-met oligodeoxynucleotides to the cultures. The development of SMG in explant 2, which was significantly better than in explant 1, was comparable to that seen in vivo. Moreover, the expression of both HGF and c-Met in the SMG of explant 2 was higher than in the SMG of explant 1. These findings provide the first demonstration that the branching morphogenesis of SMG is regulated by interactions with the surrounding mesenchyme-derived HGF and c-met expression in SMG, which occur concomitant with epithelial branching. The present data also suggest that the HGF that is released transiently from tongue muscles may contribute to the rapid development of SMG at the branching stage. Developmental Dynamics 228:173–184, 2003. © 2003 Wiley-Liss, Inc.


INTRODUCTION

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

Many epithelial organs start to form when primitive epithelium interacts with specific mesenchymal tissues at defined stages of development, thereby producing organ-specific structures and functions. Branching morphogenesis, defined as growth and branching of epithelial buds (Saxen and Sariola, 1987), is a fundamental embryologic process in many developing organs: salivary gland (SG), lung, pancreas, mammary gland, and kidney (Kratochwil, 1969; Saxen and Sariola, 1987; Hieda and Nakanishi, 1997; Shannon et al., 1998; Hogan, 1999; Warburton et al., 2000). This type of organogenesis requires epithelial–mesenchymal tissue interactions, in which the mesenchymal function may be either permissive or instructive. These interactions are mediated by growth factors and cytokines, which act as paracrine signals to direct specific gene responses, thus regulating cell division and differentiation. The extracellular matrix is also involved in the process of branching morphogenesis (Grobstein and Cohen, 1965; Banerjee et al., 1977; Spooner and Faubion, 1980; Thompson and Spooner, 1983; Nakanishi et al., 1986a, b; Fukuda et al., 1988; Hardman and Spooner, 1992; Kadoya et al., 1995, 1998; Hosokawa et al., 1999).

Hepatocyte growth factor (HGF) was originally purified and cloned as a potent mitogen for mature hepatocytes in culture (Nakamura et al., 1984, 1989). HGF was shown to be identical to scatter factor, which was independently identified as a fibroblast-derived motility factor for epithelial cells (Stoker et al., 1987; Weidner et al., 1991). Recent cumulative evidence indicates that HGF is a pleiotropic factor that stimulates mitogenesis, motogenesis, and morphogenesis in a wide range of cellular targets, including epithelial and endothelial cells, neurons, muscle cells, and hepatocytes (Brinkmann et al., 1995; Woolf et al., 1995; Ebens et al., 1996; Takayama et al., 1996). HGF is a heterodimeric molecule that is composed of a 69-kDa α-subunit with four kringle domains and a 34-kDa β-subunit with serine protease-like domain (Nakamura et al., 1986). Recent reports suggest that HGF is a humoral mediator of morphogenic epithelial–mesenchymal interactions. The biological responses to HGF are mediated by its cell receptor c-Met, which is encoded by the proto-oncogene c-met (Bottaro et al., 1991; Naldini et al., 1991). C-Met is a 190-kDa glycoprotein that consists of two disulfide-linked protein chains: an extracellular 50-kDa α-chain and a membrane-spanning 140-kDa β-chain, which has a cytoplasmic domain with tyrosine kinase activity (Giordano et al., 1989).

The interaction of HGF with its receptor c-Met is important for development, organogenesis, and tissue regeneration. Mice with deficiencies in either HGF or c-met gene functions show embryonic lethality, due to severe defects in the liver and placenta (Schmidt et al., 1995; Uehara et al., 1995). The analysis of HGF and c-met gene expression suggests that HGF/c-met acts in a paracrine manner during development of various epithelial organs (Defrances et al., 1992; Sonnenberg et al., 1993; Andermarcher et al., 1996). Moreover, it has been demonstrated that HGF/c-Met interaction contributes to the development, branching morphogenesis, and regeneration in various organs, including lung (Ohmichi et al., 1998), kidneys (Santos et al., 1994), and mammary glands (Niranjan et al., 1995).

Murine submandibular gland (SMG) organogenesis begins in the E11.5–E12 embryo as an outgrowth of the oral floor epithelium, which extends into the underlying mandibular mesenchyme (Jaskoll and Melnick, 1999). Initially, the outgrowth of the primordium shows a single epithelial stalk with a bulb at its distal end. Clefts subdivide the initial solid epithelial bulb, thus initiating the branching process. At E14–E16.5, the SMG forms a bush-like structure, which consists of a network of elongated epithelial branches and terminal epithelial buds, by repeated branching of the epithelial buds. These branches and buds hollow out to form the ductal system and presumptive acini, respectively. Information on the functional contribution of HGF to salivary gland development is generally lacking, except that HGF stimulates the proliferation and expression of the activin-β A-subunit mRNA in cultured primary epithelial cells from the rat SMG (Furue and Saito, 1998). Moreover, the kinetics of HGF and c-Met expression during salivary gland morphogenesis remain to be elucidated.

In this study, we examined the role of HGF in the epithelial–mesenchymal interaction during mouse salivary gland development. Our results suggest that mouse salivary gland development, which includes branching morphogenesis, is regulated by mesenchyme-derived HGF and is accelerated by HGF that is released transiently from tongue muscles.

RESULTS

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

Development of the Mouse Salivary Gland In Vivo

The results of the present histologic study are consistent with those presented in previous reports (Jaskoll and Melnick, 1999). The mouse SMG and sublingual gland (SLG) primordia were identified at embryonic days E11.5 and E12.5, respectively, as outgrowth of the oral floor epithelium at the linguogingival junction (Fig. 1a), which then migrated into the underlying mandibular mesenchyme. In the area between the mandibular bone and tongue, the SMG developed progressively and rapidly from E14 to E16.5. The SMG at E13.5–E14 was mainly at the pseudoglandular stage, in which the solid cord of epithelium elongated and formed the presumptive duct and terminal end buds (Fig. 1b). This branching epithelium was surrounded by closely packed mesenchyme. At E14.5, there was an increase in the number of epithelial lobules due to branching morphogenesis. The presumptive ducts, which consisted of multilayered epithelial cells, showed distinct lumina, and the surrounding mesenchyme loosened (Fig. 1c). The SMGs at E15.5–E16 were mainly at the terminal bud stage, in which ducts were clearly identified and the terminal bud lumina increased in size (Fig. 1d). The expression of HGF and c-Met in SG tissues and adjacent tissues was examined by immunohistochemistry. At E12.5, diffuse and faint staining for HGF and c-Met was observed in entire areas of the frontal sections, and weak HGF immunoreactivity was found in the tongue muscle cells and the mandibular mesenchymal cells, which were mainly located at the submucosa (Fig. 1e). The tongue muscle cells were weakly positive for c-Met (Fig. 1i). At E13.5, moderate immunoreactivity for HGF was seen in the mandibular mesenchymal cells and the tongue and mylohyoid muscle cells (Fig. 1f). The SGs stained faintly with both the anti-HGF and anti–c-Met antibodies (Fig. 1f,j). Distinct localization of HGF in the surrounding mesenchyme and of c-Met in the SGs was apparent at E14.5–E15.5 (Fig. 1g–l), whereas the SGs were also weakly positive for HGF at E14.5. The c-Met staining was pronounced in the cell membranes of the SG epithelial cells, and especially strong staining was observed at the luminal side of the presumptive ducts. During these embryonic days, the tongue muscles showed higher staining intensity for HGF and c-Met than the SG epithelia and surrounding mesenchyme. The SMGs from E12.5, E13.5, E14.5, and E15.5 embryos had 1–3, 5–7, 40–60, and numerous lobules, respectively (data not shown).

thumbnail image

Figure 1. Morphology of the developing embryonic salivary glands. Sections of the salivary glands at embryonic day (E) 12.5 (a,e,i), E13.5 (b,f,j), E14.5 (c,g,k), and E15.5 (d,h,l) were stained with hematoxylin and eosin (HE, a–d), anti–hepatocyte growth factor (HGF) antibody (e–h), or anti–c-Met antibody (i–l). In e–l, each inset shows a higher magnification of the area indicated by the dotted square. Scale bars = 300 μm in a–l, 100 μm in insets.

Download figure to PowerPoint

HGF and c-met Gene Expression in Mouse Salivary Gland Tissues

The expression levels of the HGF and c-met genes were assayed during mouse SG development. SMG and tongue tissues were isolated on various embryonic days and subjected to real-time reverse transcriptase-polymerase chain reaction (RT-PCR). Amplified products that corresponded to the HGF and c-met transcripts were detected in all SMGs from E12.5 to E17.5 (Fig. 2A). The levels of HGF mRNA in SMGs increased gradually during the early days (E12.5–E15.5), and then decreased (E16.5–E17.5). The levels of c-met transcripts in SMG were very low at E12.5–E13.5, but increased rapidly to peak between E14 and E16.5 and decreased thereafter. When the SMG epithelium and surrounding mesenchyme were examined separately, the HGF and c-met mRNAs were found exclusively in the mesenchymal and epithelial compartments, respectively (Fig. 2B). Strong expression of both the HGF and c-met transcripts was found in the tongue muscles, but was limited at E13.5–E14.5 (Fig. 2C).

thumbnail image

Figure 2. Hepatocyte growth factor (HGF) and c-met gene expression during mouse submandibular gland development, as determined by reverse transcriptase-polymerase chain reaction (RT-PCR). A: Total mRNA was isolated from the submandibular glands at the indicated embryonic days. B: Total mRNA was isolated separately from the epithelium (E) and mesenchyme (M) of the submandibular gland at embryonic day 14.5 (E14.5). C: Samples that were isolated from tongues at the indicated embryonic days were subjected to RT-PCR. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Download figure to PowerPoint

Submandibular Gland Development in Organ Culture

Tissue explants were isolated from E13.5 embryos and prepared for organ culture. Both the SG rudiments (explant 1) and SG rudiments with peripheral tissue that included part of the tongue muscle (explant 2) were used. In addition to the surrounding mesenchyme, the SG rudiments included the SMG and SLG primordia, which typically possessed 5–7 epithelial lobules or 1 epithelial lobule, respectively. Mesenchyme-free SMG epithelium was obtained by removal of the surrounding mesenchyme from explant 1. These explants were cultured in BGjb medium without supplements, such as serum and growth factors. Within the first 12 hr of cultivation, clefts began to form on the lobules of explant 1, followed by branching morphogenesis. SMG development was assessed as increases in the number of lobules (Fig. 3). With increased cultivation time, the sizes of the SMG/SLG and the number of lobules increased, except in the case of the mesenchyme-free SMG epithelium. Under these culture conditions, the mesenchyme-free epithelium failed to grow in culture, and eventually the epithelial structure was destroyed, even when E14 epithelium was used. When the SMG/SLG rudiments were cultured directly with part of the tongue tissue (explant 2), the SGs showed more extensive growth and branching morphogenesis compared with those seen during in vivo SG development. On days 3 and 4, the numbers of SMG lobules in the explant 2 samples were 3.0- and 4.2-fold higher, respectively, than those in the explant 1 samples (mean branching ratio [final lobule number/initial lobule number]: 13.2 ± 5.4 and 21.2 ± 6.4 in explant 2 vs. 4.8 ± 2.0 and 5.1 ± 1.9 in explant 1 [n = 3], respectively). These findings suggest that growth and branching morphogenesis require epithelial–mesenchymal tissue interactions, during which certain soluble factors are released from the mesenchymal tissue.

thumbnail image

Figure 3. Development of the embryonic salivary gland in organ culture. The explants were isolated and prepared from the embryonic day (E) 13.5 mouse. A: Microphotographs of the organ culture. The E13.5 submandibular gland primordia usually possessed five to seven epithelial lobules (a). Explant 1 (salivary gland [SG] rudiments, b), explant 2 (SG rudiments with tongue muscle, c), and mesenchyme-free SG epithelium (d) were cultured in serum-free BGjb medium for 3 days. B: The lobules in the submandibular gland (SMG) were counted at various intervals. The data shown represent the mean ± standard error of the values for triplicate explants. The number of lobules of SMG in explant 2 (c) was significantly higher than that in explant 1 (b); *P < 0.01. The mesenchyme-free SMG epithelium (d) failed to grow under this condition. Scale bars = 300 μm in A. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

HGF Induces Branching of the Salivary Gland Epithelium

The mesenchyme-free SMG epithelium preserved its epithelial structure for several days but did not show branch formation, even when it was embedded in Matrigel and cultured in BGjb medium. Therefore, we examined the direct effects of HGF on the branching morphogenesis of mesenchyme-free epithelium. One lobule of the E13.5 SMG epithelium and a recombinant human HGF (rh-HGF) -soaked bead (approximately 150-μm diameter) were placed on a Matrigel-coated filter at a distance of 150 μm, and cultured in serum-free BGjb. During the first 24 hr, the epithelium extended toward the source of HGF, made contact with the bead, and formed two or three lobules by 48 hr (Fig. 4). The frequency of branch reproduction in this system was approximately 30% of the lobules of the E13.5 epithelium, whereas the frequency increased to approximately 75% when E14 epithelium was used. Migration and branching were negligible when E13 epithelium was used. The epithelium stayed at the initial position and did not show branching when a phosphate-buffered saline (PBS) -soaked bead was placed in the well. This finding suggests that HGF acts as an inducer or stimulator of SG branching morphogenesis through its chemoattractive and mitogenic properties.

thumbnail image

Figure 4. Hepatocyte growth factor (HGF) -soaked bead-induced migration and branching of the salivary gland epithelium. The mesenchyme-free submandibular gland epithelium was isolated from embryonic day 13.5 embryos by enzyme treatment and mechanical dissection. One lobule of the epithelium and an HGF-soaked bead (approximately 150-μm diameter) were placed on a Matrigel-coated filter at a distance of 150 μm and cultured in serum-free BGjb for the indicated time. The asterisks indicate the initial lobule position. The epithelium migrated toward the bead, made contact within 48 hr, and formed branches. PBS, phosphate buffered saline. Scale bars = 300 μm.

Download figure to PowerPoint

Effects of HGF and Anti-HGF Antibody on the Development of Submandibular Gland in Organ Culture

We added rh-HGF or the anti-HGF antibody to SMG/SLG rudiments in conventional organ cultures. The responses to rh-HGF and the anti-HGF antibody were different for explants 1 and 2, and for the E13.5 and E14 SMG/SLG rudiments (Fig. 5). In the case of explant 1 that was isolated from E14, the number of SMG lobules and the SMG size were increased by the addition of rh-HGF at 20–100 ng/ml. On day 3 of the treatment, the number of lobules was approximately twofold higher in the test than in the control (mean branching ratio: 4.6 ± 0.1 vs. 2.7 ± 0.2 for HGF-treated vs. untreated explants [n = 3], respectively; P < 0.01). The addition of rh-HGF to E13.5 explant 1 and E13.5 explant 2 stimulated SMG development (the number of lobules in each test group was approximately 1.1-fold higher than in the untreated control), although the effect was very slight and not significantly different from the effect on the untreated explants. The addition of the anti-HGF antibody to explant 2 cultures resulted in significant inhibition of SMG development; the number of lobules in the treated group was approximately 20–25% of that in the untreated control (mean branching ratio: 21.1 ± 6.4 vs. 4.0 ± 0.8 for untreated- vs. anti–HGF-treated explants [n = 3], respectively; P < 0.01). In the case of E14 explant 1, the number of lobules produced from tissues that were treated with the antibody was 60% of the number in the untreated control (mean branching ratio: 2.7 ± 0.3 vs. 1.4 ± 0.2 for untreated- vs. anti-HGF-treated explants [n = 3], respectively; P < 0.01). In all of the inhibition tests with the anti-HGF antibody, we used nonimmune serum as an additional control, and found that the results were identical to those obtained in the absence of serum. As an interesting morphologic finding, the main duct dilated in response to the addition of the antibody.

thumbnail image

Figure 5. Effects of hepatocyte growth factor (HGF) and the anti-HGF antibody on the branching morphogenesis of the salivary gland. Embryonic day (E) 13.5 (a) and E14 (b) salivary gland (SG) rudiments (explant 1) or E13.5 SG rudiments with tongue muscle (explant 2) (c) were cultured in serum-free BGjb medium in the absence or presence of 20 ng/ml HGF or 1 μg/ml anti-HGF antibody for 3 days. A: The lobules in the submandibular gland were counted. The data shown represent the mean ± SE of the values for triplicate explants. The asterisks indicate statistically different values from the nontreated control at P < 0.01. B: Microphotographs in which E13.5 (a) and E14 (b) explant 1 samples were cultured in medium with HGF, and E13.5 explant 2 samples were cultured in the absence (c) or presence (d) of the anti-HGF antibody. Scale bars = 300 μm in B.

Download figure to PowerPoint

Expression of HGF and c-Met in SGs in Organ Cultures

The concentrations of HGF in media that were conditioned with the E13.5 explant 1 and E13.5 explant 2 tissues were measured by enzyme-linked immunosorbent assay (ELISA; Fig. 6A). A significant quantity of HGF was found in the medium that was conditioned with explant 2 (0.40 ± 0.02 ng/ml HGF on day 3). The levels of HGF were markedly reduced, as was the level of branching, when the SMG/SLG rudiments and tongue tissues were separated and then cocultured without contact in the same well. We did not detect significant amounts of HGF in medium that was conditioned with explant 1. To examine the changes in the levels of HGF and c-Met in the SMGs, the explants were cultured for 3 days, and the isolated SMG rudiments were lysed in buffer and analyzed by immunoblotting (Fig. 6B). HGF was identified as an immunostained band of 82 kDa. The c-Met protein appeared as 50-kDa and 140-kDa bands, which correspond to the extracellular α-chain and the membrane-spanning β-chain, respectively. The initial SMG rudiment contained very small amounts of the HGF and c-Met proteins. Incubation of the explants resulted in increased levels of both HGF and c-Met in the SMGs. The protein levels of both HGF and c-Met in the SMGs were higher in explant 2 than in explant 1 tissues. Sections of the explant 2 tissues were prepared for morphologic examination after 3 days of cultivation (Fig. 6C). The histologic and immunohistochemical findings resembled those of the E15.5 SMGs in vivo. HGF staining was characteristically intense in the surrounding mesenchyme. The c-Met staining was pronounced in the cell membranes of most of the epithelial cells.

thumbnail image

Figure 6. Expression of hepatocyte growth factor (HGF) and c-Met in mouse salivary gland (SG) rudiments in organ culture. Explant 1 (SG rudiments, a) and explant 2 (SG rudiments with tongue muscle, b) from E13.5 mice were cultured in serum-free BGjb medium for 3 days. In a separate explant 2 sample (c), the SG rudiment and the tongue tissues were separated and then cultured without contact in the same well. A: The concentration of HGF in the culture medium was measured by an enzyme-linked immunosorbent assay. The bars represent the mean ± SD. B: Before (a) and after (b) cultivation of explant 1 or explant 2 (c), equal amounts of protein from the total lysates of SG were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and analyzed by immunoblotting with the anti-HGF and anti-c-Met antibodies. C: After cultivation for 3 days, sections of explant 2 were prepared and stained with hematoxylin and eosin (a), the anti-HGF antibody (b), or the anti–c-Met antibody (c). ND, not determined. Scale bars = 300 μm in C. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Effects of Antisense HGF Oligodeoxynucleotides or Antisense c-met Oligodeoxynucleotides on Submandibular Gland Development in Organ Culture

We examined the effects of the antisense HGF oligodeoxynucleotides (ODNs) and antisense c-met ODNs, which block the translation of the endogenous HGF and c-Met, respectively. Figure 7 shows the effects of the ODNs on explant 2. The HGF and c-Met protein expression levels in SMGs were estimated by immunoblotting, and the results showed that the antisense HGF ODNs and antisense c-met ODNs inhibited endogenous HGF and c-Met expression, respectively. SMG development was inhibited by the addition of the antisense HGF ODNs or antisense c-met ODNs to the culture, whereas treatment of the explants with the sense HGF ODNs or sense c-met ODNs did not produce significant changes in protein expression. The numbers of lobules were reduced by approximately 35% and 38% by the antisense HGF and antisense c-met ODNs, respectively. The mean branching ratio was 11.4 ± 2.2 for the untreated explants and 4.0 ± 1.0 vs. 8.8 ± 3.1 for the antisense HGF- and sense HGF-treated explants (n = 3), respectively (P < 0.01), and 4.4 ± 1.9 vs. 9.8 ± 2.1 for the antisense c-met- and sense c-met-treated explants (n = 3), respectively (P < 0.01).

thumbnail image

Figure 7. Effects of antisense hepatocyte growth factor (HGF) and antisense c-met oligodeoxynucleotides (ODNs) on submandibular gland development. Embryonic day (E) 13.5 submandibular gland (SMG) rudiments with associated tongue tissue (explant 2) were cultured for 3 days in serum-free BGjb medium either without ODNs (a), or with the antisense HGF ODNs (b), sense HGF ODNs (c), antisense c-met ODNs (d), or sense c-met ODNs (e). A: The lobules in the submandibular gland were counted. The data shown represent the mean ± SE of the values for triplicate explants. Asterisks indicate statistically significant differences between the antisense and sense treatments at P < 0.01. B: Microphotographs. C: Immunoblot analysis of HGF and c-Met expression in the parenchyma. Scale bars = 300 μm in C.

Download figure to PowerPoint

DISCUSSION

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

The regulation of SG differentiation has been studied by using neoplastic SG cell lines and primary cells that were isolated from SGs (Shirasuna et al., 1986, 1988; Okura et al., 1993; Hiraki et al., 2002). However, there is little information on the development of SGs, particularly during the initial stages of the development cycle. SGs develop from outgrowth of the oral mucous epithelium, followed by branching morphogenesis and finally the emergence of specialized cells, such as acinar cells, from basal or ductal cells at the terminal ends. Branching morphogenesis is a fundamental embryologic process in many developing organs. Branching morphogenesis of SGs (Grobstein, 1953), as well as that of other organs, such as the lung, mammary gland, and kidney, does not proceed without the mesenchyme. The SG mesenchyme supports the branching of lung and mammary gland epithelia, while the mesenchyme that is derived from other organs, such as the lung, also induces branching morphogenesis in the SG epithelium (Lawson, 1974). Some investigators have suggested that the mechanical force of the mesenchymal cells is involved in embryonic morphogenesis (Oster et al., 1983; Nakanishi et al., 1987). In organ cultures that used chambers consisting of double compartments that were separated by a membrane filter, Nogawa and Takahashi (1991) demonstrated that the SMG epithelium of an E13 mouse showed growth and branching morphogenesis when embedded in Matrigel and cultured with the mesenchyme, which had been placed on the opposite side of the filter. However, the epithelium showed little growth, and no branching morphogenesis occurred without the Matrigel or the mesenchyme. These previously reported results suggest that SG branching morphogenesis requires an appropriate extracellular matrix and certain mesenchymal factors that diffuse through the filter. Our experimental results confirm this hypothesis. Mesenchyme-free SMG epithelium failed to grow or form branches, and the epithelial structure was eventually destroyed. Moreover, the present study suggests that HGF is one of the mesenchymal factors that stimulate the growth and development of the SG epithelium. When E13.5 or E14 mesenchyme-free SMG epithelium and rh-HGF–soaked bead were placed on Matrigel and cultured in serum-free medium, the epithelium migrated and formed branches. This finding suggests that HGF acts as an inducer or stimulator of SG branching morphogenesis through its chemoattractive and mitogenic activities. Serum-free culture of the explant that contained the SG rudiments and surrounding mesenchyme showed tissues with branching morphogenesis, and the addition of the anti-HGF antibody, antisense HGF ODNs, or antisense c-met ODNs to the culture resulted in the inhibition of branching morphogenesis. These findings suggest that endogenous HGF is present in the explant cultures. Furthermore, our results suggest that the interaction between mesenchyme-derived HGF and c-Met plays an important role in SG branching morphogenesis.

HGF expression begins during embryogenesis and is distributed widely in the developing epithelium, limb buds, and neural tissues (Brinkmann et al., 1995; Woolf et al., 1995; Ebens et al., 1996; Takayama et al., 1996). Sonnenberg et al. (1993) clearly demonstrated that c-met transcripts were expressed in the developing epithelia of mouse organs, such as the lung, pancreas, SG, and kidney, whereas HGF transcripts were detected exclusively in the neighboring mesenchymal cells. However, these investigators did not mention the changes in gene expression during SG development. With the exception of this previous report, there is no information on the expression of the HGF and c-met genes during SG development. In the case of the adult rat, Amano et al. (1994) reported that epithelial cells of the granular convoluted tubules of the salivary gland expressed both HGF mRNA and protein, whereas other cell types did not express HGF. In the present immunohistochemical study, weak staining patterns for HGF as well as c-Met were observed in SG epithelia at E12.5–E13.5. The observed location of HGF in the cytoplasm of epithelial cells is thought to represent internalized HGF that is bound to the receptor (Sonnenberg et al., 1993). We detected HGF and c-met expression by using real-time RT-PCR, which is a valuable method for monitoring the kinetics of target gene expression. Separation of the tissues confirmed that the SMG epithelium and surrounding mesenchyme exclusively expressed the c-met and HGF genes, respectively. The levels of HGF mRNA increased gradually during the early stages of SG development (E12.5–E15.5). The levels of c-met transcript in the E12.5–E13.5 SMGs were very low, but they increased and peaked between E14.0 and E16.5, concomitant with increases in epithelial branching. The differences in levels of c-met expression between the E13.5 and E14 SMGs may influence the in vitro responses of SMGs to rh-HGF. The development of E14 SMGs was enhanced by the addition of rh-HGF, whereas the effect of rh-HGF on E13.5 SMGs was very slight. The E13.5–E14 mesenchyme-free SMG epithelium migrated toward the rh-HGF–soaked bead and formed branches, whereas the E13 epithelium did not. As seen in vitro, the development of the SG (including epithelial branching) was very slow at E12.5–E13.5, although HGF expression occurred at this time. From these findings, we conclude that the epithelial branching of SGs does not develop progressively until c-met expression in the SGs increases.

Although the present study demonstrates that the HGF/c-Met interaction contributes to the stimulation of SMG branching morphogenesis, it remains to be seen whether HGF is a crucial factor for the triggering of epithelial branching. Other growth factors and their receptors, such as epidermal growth factor (EGF)/EGF receptor and fibroblast growth factors (FGFs)/FGF receptors (FGFRs) are also important in the branching morphogenesis of the SG (Morita and Nogawa, 1999; Kashimata et al., 2000; Hoffman et al., 2002). FGF7 promoted stalk elongation of SMG epithelial explants that were cultured in the absence of mesenchyme (Morita and Nogawa, 1999). It has been suggested that FGFR1 signaling plays a central role in branching morphogenesis of SG by regulating the expression of FGFs and bone morphogenetic protein (BMP; Hoffman et al., 2002). There is increasing evidence that the development of various organs is regulated by the simultaneous actions of several growth factors, which include HGF, members of the FGF family, and members of the transforming growth factor-β (TGF-β) family, such as BMP and activin, which exert synergistic or antagonistic effects on target cells or organs (Niswander and Martin, 1993). Therefore, it will be very interesting to investigate the interactions of HGF/c-Met with candidate growth factors for SG development, in particular those involved in branching morphogenesis. Moreover, it seems important to define which signaling pathways downstream of c-met are involved in early epithelial branching of SG. Upon activation, several of the tyrosine residues in c-Met are phosphorylated, and some of these residues mediate the binding of signaling proteins, such as phosphatidylinositol 3-kinase (PI3K), phospholipase C-γ (PLC-γ), Grb2, and Gab1 (Graziani et al., 1993; Gual et al., 2000; Karihaloo et al., 2001). The sustained recruitment of PLC-γ to Gab1 plays a role in branching tubulogenesis (Gual et al., 2000), and Grb2 binding is required for the HGF-stimulated branching tubulogenesis of Madin–Darby canine kidney epithelial cells (Karihaloo et al., 2001). HGF-mediated cell migration and branching morphogenesis of renal epithelial cells requires the coordinate activation of PI3K, PLC-γ, and extracellular signal-regulated kinases 1 and 2 (ERK1/2), whereas EGF-mediated morphogenesis is dependent on PI3K, PLC-γ, and ERK5 but not on ERK1/2. In contrast, the ERK1/2 pathway is important for the EGF-stimulated branching morphogenesis of the fetal SMG (Kashimata et al., 2000). One important study has demonstrated that PI3K is involved in branching morphogenesis in mouse SMG and that its effects are mediated through signaling by means of phosphatidylinositol 3,4,5-triphosphate (PIP3) and possibly Akt (Larsen et al., 2003). It is very interesting that PI3K/PIP3 plays a role in initiation of cleft formation, the first step of branching morphogenesis.

It is known that high-level transient expression of c-met and HGF occurs in muscle tissues, which suggests a role for the HGF/c-Met interaction in the development and/or patterning of muscle (Yang et al., 1996; Scaal et al., 1999). Amano et al. (2002) reported HGF and c-Met expression in mouse embryos during mandibular and tongue development, in which HGF was detected in the mesenchymal cells of the first branchial arch at E10 before tongue formation, whereas c-Met–positive myogenic cells were found in the mandible centers at E11. Strong coexpression of HGF and c-met was seen in differentiating tongue myoblasts at E12. Although HGF expression disappeared between E14 and E16, c-met expression persisted in the tongue muscles. Our findings differ slightly from those of Amano et al. (2002). The present immunohistochemical study showed coexpression of HGF and c-Met in the tongue muscles at E12.5–E15.5, although strong coexpression of HGF and c-met mRNA in the tongue muscles was confined to the period between E13.5 and E14.5. The discrepancy between our study and the previous report may be explained by the different experimental methods used. The RT-PCR technique that we used may be optimal for the quantification of gene expression. Between E14 and E15.5, SG branching morphogenesis advances progressively, while at the same time the tongue muscles express large amounts of HGF. This model appears to reflect the situation in the explant 2 organ cultures. When the SMG/SLG rudiments with tongue muscle were cultured, the development of SG was more extensive. The inhibition of branching morphogenesis by the addition of anti-HGF antibody was more pronounced in explant 2 than in explant 1 tissues, which suggests the presence of larger amounts of endogenous HGF in the explant 2 cultures. Indeed, the ELISA and immunoblot analyses indicated higher levels of the HGF and c-Met proteins in explant 2 cultures than in explant 1 cultures. Immunohistochemical examination of the explant 2 sections confirmed that the predominant accumulation of HGF protein in the surrounding mesenchyme provoked strong expression of c-Met in the epithelial cell membranes. SMG development in the explant 2 culture was comparable to that seen in vivo, which suggests accurate replication of the in vivo microenvironment. When the SMG/SLG rudiments and the tongue tissues were separated and cocultured without contact in the same well, the expression of HGF and c-Met was reduced, as was branching morphogenesis. These findings suggest that HGF production and c-Met expression may be regulated by an interaction between the SG tissues and adjacent tissues in vivo, and that SG development is supported not only by the surrounding mesenchyme but also by HGF released from the tongue tissues.

In conclusion, mouse SG development, which includes branching morphogenesis, is regulated by the interaction of mesenchyme-derived HGF and c-Met, which is expressed in SG concomitant with epithelial branching. HGF that is released transiently from tongue muscles may also contribute to the rapid development of the SG during the initial stages.

EXPERIMENTAL PROCEDURES

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

Reagents

The rh-HGF was obtained as described previously (Ohmichi et al., 1998). Synthetic ODNs, which included the antisense HGF ODNs and antisense c-met ODNs, were purchased from Sigma Genosys Japan, Ltd. (Hokkaido, Japan). RT-PCR primers were purchased from Invitrogen Life Technologies (Gaithersburg, MD). Rabbit polyclonal antibodies to mouse HGF and c-Met were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).

Tissue Preparation and Organ Culture

Pregnant time-mated ICR mice were obtained from SLC, Ltd. (Shizuoka, Japan). The plug day was considered to be the day of gestation initiation. Pregnant mice were anesthetized with diethyl ether and killed by cervical dislocation on days 12.5–17.5 of gestation (E12.5–E17.5). The embryos were dissected in cold PBS. The SG rudiments and peripheral tissues were isolated in Hanks' solution by using fine needles under a stereomicroscope.

The SG rudiments from E13.5 or E14 embryos were used for the organ cultures, and two types of tissue explant were generated: explant 1, SG rudiments; and explant 2, SG rudiments with peripheral tissue that included a part of the tongue muscle. Mesenchyme-free SMG epithelium was isolated by treatment with Dispase (100 protease units/ml in PBS; Godo Shusei Co., Tokyo, Japan) for 10 min, followed by removal of the surrounding mesenchyme and sublingual primordium with fine needles. These explants were placed on top of polyester Transwell filters (0.4-μm pore size; Corning-Costar, Becton Dickinson, Franklin Lakes, NJ), which were then placed in the individual wells of a 24-well tissue culture dish (Becton Dickinson) that contained 200 μl of culture medium. The medium consisted of BGjb (Gibco-BRL, MD) that was supplemented with 100 μg/ml of ascorbic acid (Wako Pure Chemical, Osaka, Japan), 50 U/ml penicillin, and 100 μg/ml streptomycin.

Culture on Matrigel Using HGF-Soaked Bead

One lobule of mesenchyme-free SMG epithelium was used to test interactions with HGF-soaked bead. The protein-carrying beads were prepared as described by Weaver et al. (2000). Acrylic beads that carried immobilized heparin (Sigma) were rinsed with PBS three times, and beads approximately 125 to 175 μm in diameter were selected and soaked in 0.1 μg of purified recombinant HGF protein. The epithelium and HGF-soaked bead (approximately 150 μm diameter) were placed at a distance of 150 μm apart on the top of a Transwell filter that had been coated with 50 μl of Matrigel (Collaborative Research, Bedford, MA), and the entire system was cultured in BGjb medium for several days.

Antisense ODNs

We examined the effects of HGF and c-met antisense ODNs on SG development in culture. The following ODN sequences (Amano et al., 1999) were used: HGF antisense ODN, 5′-GGTCCCCCACATCAT-3′; c-ODN, 5′-GGACCCCCTCTACTA-3′; c-met antisense ODN, 5′-GGGGGCCTTCATTAT-3′; c-ODN, 5′-ATAATGAAGGCCCCC-3′.

ELISA and Western Blotting

The explants were cultured in BGjb medium for 3 days, and the concentration of HGF in the conditioned medium was measured by ELISA using the Rat HGF EIA Kit (Institute of Immunology Co., Ltd., Tokyo, Japan). The remaining explants were used for Western blotting and immunohistochemistry. The SMGs were isolated from the explants by using fine needles under a stereomicroscope, and Western blotting was performed as described previously (Hiraki et al., 2002). In brief, the sample was lysed by sonication in sodium dodecyl sulfate (SDS) lysis buffer (50 mM Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, 6% mercaptoethanol). The protein concentration was determined spectrophotometrically by using the Bio-Rad Protein Assay Kit (Bio-Rad, Hercules, CA). Samples that contained 30 μg of protein were electrophoresed on 7.5% SDS-polyacrylamide gels, and transferred electrophoretically to nitrocellulose membranes (Bio-Rad). After washing with TBS-Tween (25 mM Tris-HCl [pH 8.2], 144 mM NaCl, 0.1% Tween 20), the membranes were blocked with 5% nonfat skim milk in TBS-Tween at room temperature, and then incubated with either the anti-HGF or anti-c-Met antibody. Immunoreactive bands were visualized by using horseradish peroxidase-conjugated secondary antibodies and ECL reagents (Amersham Pharmacia Biotech). X-ray films were scanned into Adobe Photoshop with a CanonScan FB1200S and quantified by using the NIH Image software (National Institutes of Health, Baltimore, MD).

Total RNA Preparation and RT-PCR

Total RNA was extracted from E12.5–E17.5 SMGs or tongue tissues using TRIzol (Life Technologies, Gaithersburg, MD), according to the manufacturer's protocol. First-strand cDNA was synthesized from 3 μg of total RNA by using the Superscript II reverse transcriptase (Life Technologies) and random hexanucleotide primers. To ensure the fidelity of mRNA extraction and reverse transcription, all of the samples were subjected to PCR amplification with oligonucleotide primers that were specific for the constitutively expressed mouse gene for glyceraldehyde-3-phophate dehydrogenase (GAPDH) and normalized. All of the PCRs were performed by using the LightCycler-FastStart DNA Master SYBR Green I Kit (Roche). The following sets of primers were used: HGF forward primer, 5′-CCA TGA ATT TGA CCT CTA TGA-3′; HGF reverse primer, 5′-CTG AGG AAT CTC ACA GAC TTC-3′; c-met forward primer, 5′-AGA TGA ATG TGA ATA TGA AG-3′; c-met reverse primer, 5′-CAT ATG AGT TGA TCA TCA TAG-3′; GAPDH forward primer, 5′-CCA TCA CCA TCT TCC AGG AG-3′; GAPDH reverse primer, 5′-GCA TGG ACT GTG GTC ATG AG-3′. Amplification was performed for 33 cycles under the following conditions: denaturation at 95°C for 10 min for the first cycle, and 15 sec for subsequent cycles, annealing at 60°C (for HGF), 55°C (c-met) or 65°C (GAPDH) for 5 sec, and extension at 72°C for 20 sec. The PCR products were separated by electrophoresis on 1.8% agarose gels and then visualized by ethidium bromide staining with ultraviolet light illumination.

Histology and Immunohistochemistry

The mouse embryos at E12.5–E15.5 and SG explants were cultured for several days and then subjected to morphologic analyses. The specimens were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer, dehydrated in an ethanol series, and embedded in paraffin. The sections were subjected to either hematoxylin and eosin staining or immunohistochemistry. Immunohistochemical staining was performed as described previously (Nakayama et al., 2001). After the endogenous peroxidase was blocked by immersion in methanol plus 0.3% hydrogen peroxide, the sections were incubated with either the anti-HGF (diluted 1:10 in PBS) or anti-c-Met (diluted 1:100) antibody and then incubated with biotinylated anti-rabbit IgG, followed by staining with peroxidase-conjugated streptavidin (Histofine SAB-OP Kit; Nichirei Co., Tokyo, Japan).

Statistical Analysis

Results are expressed as the mean ± SE. Statistical comparison of the two mean values was carried out by using the unpaired t-test. All of the P values were analyzed on two sides.

REFERENCES

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