Hepatocyte growth factor is essential for migration of myogenic cells and promotes their proliferation during the early periods of tongue morphogenesis in mouse embryos

Authors

  • Osamu Amano,

    Corresponding author
    1. Department of Histology and Embryology, Kanazawa University Graduate School of Medical Science, Kanazawa, Japan
    2. Department of Oral Anatomy II, Meikai University School of Dentistry, Sakado, Japan
    • Department of Oral Anatomy II, Meikai University School of Dentistry, 1-1 Keyakidai, Sakado, Saitama 350-0283, Japan
    Search for more papers by this author
  • Akira Yamane,

    1. Department of Pharmacology, Tsurumi University School of Dental Medicine, Yokohama, Japan
    Search for more papers by this author
  • Mayumi Shimada,

    1. Departments of Oral and Maxillofacial Surgery, Kanazawa University Graduate School of Medical Science, Kanazawa, Japan
    Search for more papers by this author
  • Uichi Koshimizu,

    1. Molecular Regenerative Medicine, Course of Advanced Medicine, Osaka University Graduate School of Medicine, Suita, Japan
    Search for more papers by this author
  • Toshikazu Nakamura,

    1. Molecular Regenerative Medicine, Course of Advanced Medicine, Osaka University Graduate School of Medicine, Suita, Japan
    Search for more papers by this author
  • Shoichi Iseki

    Corresponding author
    1. Department of Histology and Embryology, Kanazawa University Graduate School of Medical Science, Kanazawa, Japan
    • Department of Oral Anatomy II, Meikai University School of Dentistry, 1-1 Keyakidai, Sakado, Saitama 350-0283, Japan
    Search for more papers by this author

Abstract

Temporal and spatial occurrence of hepatocyte growth factor (HGF) and its cognate receptor c-Met in the mouse mandibular development was investigated by immunohistochemistry and quantitative reverse transcriptase-polymerase chain reaction. HGF was first recognized in the mesenchymal cells of the first branchial arch at the 10th day of gestation (E10), before tongue formation, whereas HGF receptor (c-Met) -positive myogenic cells first appeared at E11 in the center of mandibles. By E12, HGF turned to be colocalized with c-Met in the differentiating tongue myoblasts. Between E14 and E16, HGF disappeared, whereas c-Met remained, in the tongue myoblasts. The levels of HGF mRNA in the developing tongue decreased in accordance with the increase of desmin mRNA levels from E11 to E17. These in vivo results strongly suggest that the HGF/c-Met system takes part in the earlier stages of tongue development. To elucidate this hypothesis, the antisense oligodeoxyribonucleotide (A-ODN) for mouse HGF mRNA was added to the organ culture system of mandible with serumless, defined medium. Mandibular arches from E10 mouse embryos were cultured at 37°C for 10 days in the absence or presence of A-ODN, control (sense) oligonucleotide (C-ODN), or A-ODN plus recombinant HGF. In the control mandibular explants cultured without HGF or ODN, the anterior two-third of the tongue derived from the first branchial arch was formed. It contained abundant desmin-positive myoblasts and was equivalent to the tongue of E14–E15. In contrast, in the presence of A-ODN in the medium, neither the swelling nor myogenic cells were found in the tongue-forming region of explants, and myogenic cells accumulated behind the tongue-forming region. Such dysplasia of tongue was never induced in the presence of C-ODN or A-ODN plus recombinant HGF in the medium. The effect of A-ODN appeared to be developmental stage-specific, because tongue dysplasia occurred when A-ODN was present during the earlier 4 days but not during the later 4 days of the culture. Furthermore, recombinant HGF added to the culture without ODNs during the earlier 4 days caused elevation in the number of mitotic myoblasts. These results suggest that HGF regulates both the migration and proliferation of myogenic cells during the earlier stages of tongue development. © 2002 Wiley-Liss, Inc.

INTRODUCTION

In tongue development, myogenesis is the central event consisting of migration, proliferation, and differentiation of myogenic cells derived from the occipital somites. Myogenesis in tongue muscle proceeds in several phases as follows: (1) in the first phase, early myoblasts in the somite increase in number, resulting in a larger population of mononucleated myoblasts that produce desmin; (2) in the next phase, these progenitor myoblasts migrate to the developing tongue primordia in the first branchial arch (FBA) and undergo proliferation and cytodifferentiation; (3) in the last phase, these cells withdraw from the cell cycle and fuse to become multinucleated myotubes (Wachtler and Jacob, 1986; Brand-Saberi and Christ, 1999, Yamane et al., 2000). Several studies have suggested that growth factors are involved in successive stages of the myogenesis (reviewed in Olson et al., 1991).

Hepatocyte growth factor (HGF) is a potent mitogen for mature hepatocytes in culture, and may act as a trigger for liver regeneration (Nakamura, 1991; Matsumoto and Nakamura, 1991, 1992). HGF was originally discovered in the serum of partially hepatectomized rats (Nakamura et al., 1984), purified first from the rat platelets (Nakamura et al., 1986), and was proved to be identical to “scatter factor,” a molecule capable of causing dissociation and migration of epithelial cells by enhancing their motility (Stoker et al., 1987; Weidner et al., 1990). HGF is secreted by cells of mesodermal origin as a single-chain precursor, which is then processed to yield a heterodimer composed of a large subunit of 69 kDa and a small subunit of 34 kDa. Complementary DNAs of human, rat and mouse HGF have been cloned and their complete primary structures determined (Nakamura et al., 1989; Tashiro et al., 1990). The action of HGF is transmitted by its specific receptor c-Met, a protooncogene product (Bottaro et al., 1991; Naldini et al., 1991). The c-Met has a heterodimeric structure consisting of an extracellular α subunit of 50 kDa and a membrane-spanning β-subunit of 145 kDa, which has a tyrosine kinase domain in its cytoplasmic region (Giordano et al., 1989).

During development, HGF has been shown to play multiple roles as a mitogen, a motogen, and a morphogen in the organization of multicellular structures (Matsumoto and Nakamura, 1997). Knockout studies in the mouse have indicated that the HGF/c-Met system plays a crucial role in limb development by regulating the migration of myogenic cells from the somites to limbs (Yang et al., 1996; Dietrich et al., 1999). With regard to tongue development, there is evidence that HGF and c-Met are expressed in the mouse mandibles during embryonic development in association with tongue myogenesis (Sonnenberg et al., 1993), as well as with osteogenesis and chondrogenesis (Amano et al., 1999). Bladt et al. (1995) and Dietrich et al. (1999) have shown that agenesis of tongue occurs in the mice lacking HGF or c-Met. As with limb myogenesis, these authors suggest that HGF functions primarily in the migration of myogenic cells from the somites into tongue. Although the somites supply myogenic cells for both limb myogenesis and tongue myogenesis, the two processes differ in the origin of surrounding mesenchyme, i.e, the mesoderm in the former and the neural crest in the latter (Noden, 1983, 1986; Nichols, 1986). Taken together with consideration of the multifunctional roles of HGF, it is possible that the HGF/c-Met system is responsible for tongue myogenesis not only in migration but also in proliferation and/or differentiation of myogenic cells.

To clarify this issue, the following experiments were designed: (1) we determined the temporal expression and cellular localization of HGF and c-Met in relation to tongue myogenesis during embryonic mouse mandibular development by semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) and immunohistochemistry; (2) we examined the effects of lack of HGF or c-Met on tongue myogenesis by applying antisense oligodeoxyribonucleotide (A-ODN) to the organ culture system of embryonic mouse FBA; (3) we also examined the effects of exogenous HGF on tongue myogenesis by applying recombinant HGF to the culture system. The results provide evidence for crucial roles of the HGF/c-Met system in both the migration and proliferation of myogenic cells during embryonic mouse tongue development.

RESULTS

Immunolocalization of HGF and c-Met During Tongue Development

In the sections of E10 embryos subjected to immunohistochemically, mesenchymal cells in the bilateral oval areas subjacent to the ectoderm facing the primitive oral cavity diffusely exhibited moderate reactivity with anti-HGF antibody (Fig. 1a). When the antibody was preabsorbed with recombinant HGF, no specific immunoreaction was found in any cells or areas of mandibles (Fig. 1b). These results suggest that HGF is initially expressed in the mesenchymal cells located in and around the presumptive area of tongue development before the migration of myogenic cells. In contrast, c-Met immunoreactivity was not present in the HGF-positive areas but was found in the dorsolateral portions of mandibular processes at E10 (not shown).

Figure 1.

Photomicrographs showing immunohistochemistry of the embryonic mouse tongues in the frontal sections at embryonic day (E) 10 (a,b) and E11 (c–f) with anti–hepatocyte growth factor (anti-HGF) antibody (a), anti-HGF antibody preabsorbed with recombinant HGF (b), anti-c-Met antibody (c,e), and anti-desmin antibody (d,f). a, c, and e are the consecutive sections of b, d, f, respectively. a: HGF immunoreactivity is visible diffusely in the bilateral oval area (asterisks) containing mesenchymal cells beneath the oral epithelium. o, primitive oral cavity; n, neural epithelium. b: No specific staining is detected in any portion of the mandible. c–f: A minor cell population in the mesenchymal mass of the presumptive tongue area (dotted circle) exhibits the immunoreactivity for both c-Met (c and higher magnification, e), and desmin (d, and higher magnification, f). Scale bars = 500 μm in a (applies to a,b), 100 μm in c (applies to c,d), 100 μm in e (applies to e,f).

At E11, HGF immunoreactivity remained localized in the mesenchymal cells, which were round in shape and distributed diffusely in the floor of the oral cavity (not shown). In contrast, c-Met immunoreactivity first appeared at E11 in the central area of mandible, where the bilateral lingual swellings had been fused at their medial margins (Fig. 1c,e). Desmin immunoreactivity was distributed in a pattern similar to that of c-Met immunoreactivity (Fig. 1d). Comparison of the consecutive sections suggested the colocalization of c-Met and desmin in the same cell population (Fig. 1e,f). The cells immunopositive for c-Met and desmin were small in size and spindle-like in shape (Fig. 1e,f). They were concentrated in the area of expected tongue formation but were also found solitarily in the dorsolateral portions of mandibular processes (not shown). These results indicated that c-Met is exclusively expressed in the myogenic cells destined to form tongue musculature.

In the tongue of E12 containing tuberculum impar, HGF immunopositivity was no longer present in the round mesenchymal cells but was found with weak to moderate intensity in the long spindle-like cells arranged radially and longitudinally to the lingual surface from the dorsal pharyngeal area (Fig. 2a). These cells were also intensely immunopositive for c-Met and desmin (Fig. 2b,c), suggesting that HGF had turned to be colocalized with c-Met in the myogenic cells.

Figure 2.

Photomicrographs showing immunohistochemistry of the developing embryonic mouse tongue in sagittal sections at embryonic day (E) 12 (a–c), frontal sections at E14 (d–f), and frontal sections at E16 (g–i) for hepatocyte growth factor (HGF) (a,d,g), c-Met (b,e,h), and desmin (c,f,i). At E12 and E14, the immunoreactivities for HGF (a,d), c-Met (b,e), and desmin (c,f) are colocalized in the myogenic cells of developing tongue. At E16, HGF immunoreactivity in myofibers is reduced in intensity (g), whereas the immunoreactivities for c-Met (h) and desmin (i) remain intense in myofibers. Scale bars = 200 μm in a (applies to a–c), 500 μm in d (applies to d–f), 500 μm in g (applies to g–i).;5>

At E14, immunoreactivity for HGF, c-Met, and desmin were diffusely colocalized in the myofibers of intrinsic tongue muscles as shown in the serial frontal sections of tongue (Fig. 2d–f). By E16, HGF immunoreactivity was markedly reduced in intensity, whereas c-Met and desmin immunoreactivities remained strong in the myofibers (Fig. 2g–i).

Expression of HGF and c-Met mRNA in Tongue Development

Figure 3 shows chronologic alterations in the relative abundance of mRNA for HGF and c-Met during embryonic tongue development in mice as analyzed by competitive RT-PCR. HGF mRNA was abundantly expressed in the tongue at E11. Subsequently, the level of HGF mRNA underwent continuous decrease, reaching 42% of that of E11 at E17 (P < 0.001). In contrast, the level of c-Met mRNA increased after E11, reached the maximum (138% that of E11) at E15 (P < 0.001), followed by rapid decrease to 62% of that of E11 at E17 (P < 0.001). These results, together with the immunohistochemical results, strongly suggests that the HGF/c-Met system plays roles in the earlier stages rather than in the later stages of tongue development.

Figure 3.

Graph showing changes in the relative amounts of hepatocyte growth factor (HGF) mRNA (a) and c-Met mRNA (b) in the developing mouse tongues at embryonic day (E) 11, 13, 15, and 17, as quantitated by competitive reverse transcriptase-polymerase chain reaction. Each point with its vertical bar represents the mean ± SD of five or six samples.

Effects of Inhibition of HGF Expression on Tongue Development

Figure 4a shows the relative levels of HGF and desmin mRNAs in the mandibular explants from E10 embryos cultured for 10 days without ODN (control), with HGF A-ODN, and with HGF control ODN (C-ODN) as analyzed by competitive RT-PCR. HGF A-ODN reduced the levels of both HGF and desmin mRNAs by 40% that of control (P < 0.0001), suggesting that decreased expression of intrinsic HGF caused by HGF A-ODN leads to decrease in the number of myogenic cells in mandibular explants. In contrast, no significant difference in HGF or desmin mRNA level was detected between the control and HGF C-ODN–treated explants, confirming that the toxicity of ODN itself is negligible.

Figure 4.

Graph (a) and fluorescent photomicrograph (b) showing the effects of hepatocyte growth factor (HGF) antisense oligodeoxyribonucleotide (A-ODN) supplemented in the culture media. a: Relative amounts of mRNA for HGF (filled bar) and desmin (open bar) in the embryonic mouse mandibular explants cultured for 10 days in the absence (control) or presence of HGF A-ODN or HGF control oligonucleotide (C-ODN). In HGF A-ODN–treated mandibles, relative content of HGF and desmin mRNAs are significantly (P < 0.01) less than in control and HGF C-ODN-treated mandibles (asterisk) b: Fluorescein isothiocyanate labeling representing the incorporation of ODN is recognized in the entire explant and the cells (inset). Scale bar = 500 μm in b, 200 μm in inset.

Figure 4b demonstrates the cross-section of E10 mandibular explants cultured for 4 days in the presence of fluorescein isothiocyanate (FITC) -labeled ODN in the medium. All cells in the mandibular explants, including those of the tongue, were positive for FITC fluorescence. This finding clearly indicated that the present ODN was efficiently incorporated by the cells of mandibular explants.

Under a dissection microscope, the effect of HGF A-ODN was quite distinct in the explants. The swelling peculiar to the tongue primordium was recognized in the mandibular explants cultured for 10 days without ODN (Fig. 5a). In hematoxylin and eosin (H&E) -stained sections, the tongue was identified by the presence of thicker stratified epithelium and bilateral lingual sulci (Fig. 5b). When the mandibular explants were cultured with HGF A-ODN, the dorsal surface of the explants appeared smooth and no tongue had been formed at any portions (Fig. 5c). In sections, no swelling or sulci was found in the mid-portion of dorsal surface. The central area of mesenchymal mass where the tongue could have developed was occupied with the loose mesenchymal tissue under the tongue-specific thick epithelium (Fig. 5d). In contrast, the explants cultured with HGF C-ODN exhibited normally developed tongue as in the control (no ODN) explants (Fig. 5e). Moreover, the defect of tongue development caused by HGF A-ODN was completely rescued by adding recombinant HGF together with HGF A-ODN in the culture medium (Fig. 5g). Histologic analysis of the tongues in C-ODN–treated and rescued explants also showed the normal structure (Fig. 5f,h). These results strongly suggest that deficiency of endogenous HGF caused by HGF A-ODN gives rise to tongue malformation.

Figure 5.

Whole appearance (a,c,e,g), hematoxylin and eosin–stained frontal sections (b,d,f,h), and desmin-immunohistochemistry (k–n) of mandibular explants cultured in the absence (a,b,k,m), or presence of hepatocyte growth factor (HGF) antisense oligodeoxyribonucleotide (A-ODN) (c,d,l,n), HGF control oligodeoxyribonucleotide (C-ODN) (e,f), or HGF A-ODN plus recombinant HGF (g,h), and schematic drawings of 5a (i) and 5c (j). Note that HGF A-ODN–treated mandibles show aglossia characterized by the lack of tongue swelling and lingual sulci (c,d), whereas the mandibles in other groups are normally developed (asterisk for tongues, arrowheads for sulci, and square bracket for the presumptive tongue area). In the schemas, gray areas indicate the distributions of desmin-positive cells, and lines k-n in i and j indicate the planes for frontal sections k-n, respectively. In the control explants, desmin-immunoreactive cells are abundantly present in the tongue (i,k) but are absent in the posterior end of the explants (m). In HGF A-ODN–treated explants, no immunoreactive cells are observed in the presumptive tongue portion (j,l), whereas numerous immunoreactive cells are present at the posterior end of the explants (n). Scale bars = 1 mm in a (applies to a,c,e,g), 500 μm in b (applies to b,d,f,h), 500 μm in k (applies to k–n).

Immunohistochemical analysis further clarified the effect of HGF A-ODN on tongue development. Desmin-immunopositive myogenic cells were abundantly observed in the tongue of cultured control explants (Fig. 5i), indicating that normal myogenesis with appropriate movement and proliferation of myogenic precursors in tongue had taken place. In explants cultured with HGF A-ODN, desmin-positive myogenic cells were hardly detected in the middle portion of the explants (Fig. 5j) but were accumulated at the area adjacent to the posterior margin of explants, irrespective of the extent of tongue defect (Fig. 5k). This finding indicated that HGF A-ODN had inhibited the migration of myoblasts into the tongue-forming region in mandibular explants.

The defect of tongue development and the accumulation of myogenic cells at posterior margins of mandibular explants caused by HGF A-ODN are consistent with the hypothesis that endogenous HGF produced in developing tongue positively regulates the migration of myoblasts along the posterior to anterior axis of mandible.

Effects of Inhibition of c-Met Expression in Mandibular Explants

Mandibular explants cultured with c-Met A-ODN showed results similar to those with HGF A-ODN, namely, aplasia or severe hypoplasia of tongue without apparent effect on the growth of whole mandibular explant (Fig. 6a). In contrast, c-Met C-ODN did not affect tongue development (Fig. 6b). These results further confirm that endogenous HGF, through its signal transduction, is responsible for the development of tongue and especially for the migration of myogenic cells into the tongue.

Figure 6.

Whole mandibular explants of mandibles cultured with c-Met antisense oligodeoxyribonucleotide (A-ODN) (a) or c-Met control (sense) oligonucleotide (C-ODN) (b). c-Met A-ODN induces aglossia similar to that induced by hepatocyte growth factor (HGF) A-ODN, whereas c-Met C-ODN does not affect tongue development (asterisk). Scale bar = 1 mm in a (applies to a,b).

Effects of Inhibition of HGF/c-Met Expression for Different Time Periods

In the mandibular explants cultured in the presence of HGF A-ODN only for the initial 4 days, aglothia or microglothia was induced at 10 days (Fig. 7a). In contrast, when HGF A-ODN was present only for the last 4 days, the mandibular explants showed normal tongue appearance (Fig. 7b). Similar results were obtained when HGF A-ODN was replaced with c-Met A-ODN (not shown). These results suggest that the role of HGF in tongue myogenesis is restricted to the earlier periods, when migration, proliferation, or both, of myogenic cells are the major phenomena rather than differentiation of myofibers.

Figure 7.

Whole mandibular explants cultured with hepatocyte growth factor (HGF) antisense oligodeoxyribonucleotide (A-ODN) for the first 4 days (a) or for the last 4 days (b). Aglossia is induced in the explants treated for the first 4 days, whereas those treated for the last 4 days show normal tongue development (asterisk). Scale bar = 1 mm in a (applies to a,b).

Effect of Exogenous HGF on Myogenic Cell Proliferation in Mandibular Explants

In our previous study, the bone and cartilage formation in cultured mandibles was enhanced by exogenous HGF without apparent effects on the growth and appearance of whole explants (Amano et al., 1999). In the present study, further analyses focusing on the tongue revealed the effect of exogenous HGF on the proliferative activity of myogenic cells. The explants cultured with recombinant HGF for 10 days showed no apparent change in the size and appearance of tongue from the control explants, and desmin-immunoreactive myogenic cells occupied most of the cells in tongue in both explants (not shown). However, [3H]thymidine radioautography revealed that the number of labeled cells in tongue at 10 days is larger in HGF-treated explants than in the control explants (Fig. 8a,b). The quantitative analysis of labeling index confirmed a significant increase (P < 0.01) in the cell proliferation in tongue by HGF (Fig. 8c). Furthermore, c-Met A-ODN administered together with HGF in the culture medium reduced the effect of HGF significantly (P < 0.01), whereas c-Met C-ODN had no effect, suggesting that the promotion of myogenic cell proliferation of is specifically to exogenous HGF.

Figure 8.

Effect of exogenous hepatocyte growth factor (HGF) supplemented for 10 days on the number of [3H]thymidine-incorporating cells in the tongues of cultured mandibles. a,b: Autoradiography showing the labeled cells in the absence (a) or presence (b) of 100 ng/ml HGF. c: The labeling index of the cells in tongue in the absence or presence of exogenous HGF, with or without c-Met antisense oligodeoxyribonucleotide (A-ODN) or control (sense) oligonucleotide (C-ODN). * P < 0.01 compared with the control (n = 3); ** P < 0.01 compared with HGF plus c-Met C-ODN (n = 3). Scale bars = 200 μm in a,b.

Effects of Exogenous HGF Present for Different Time Periods on Myogenic Cell Proliferation

The mandibular explants were cultured in the presence and absence of recombinant HGF for either the initial 4 days or the last 4 days of the 10-day culture period. In both cases, the explants were labeled with [3H]thymidine during the last 24 hr of HGF treatment and were analyzed by radioautography at the end of culture (Fig. 9a–c). Relative labeling index in HGF-treated tongue to that in the control tongue was 2.5 times when labeled at 4 days, which was significantly higher than 1.5 times when labeled at 10 days (P < 0.005). In the control explants, labeling index of the 4th day were significantly higher (P < 0.001) than that of the 10th day.

Figure 9.

Effect of exogenous hepatocyte growth factor (HGF) supplemented for different time periods on the number of [3H]thymidine-incorporating tongue cells in the cultured mandible. a,b: Autoradiography showing the labeled cells in the presence of HGF for the first 4 days (a) or the last 4 days (b). c: The labeling index of the cells in tongue in the absence or presence of exogenous HGF supplemented for different periods. * P < 0.005 compared with control/last 4 days (n = 3); ** P < 0.01 compared with control/first 4 days (n = 3); *** P < 0.01 compared with HGF/first 4 days (n = 3).

Higher mitotic activity of tongue cells in earlier period of mandibular development was confirmed in the embryonic mouse tongue in vivo. The labeling index in tongues of E12 mice, equivalent to the 4th day of cultured mandibles, were than 1.5 times that of E15 mice, equivalent to the 10th day of cultured mandibles (P < 0.005) (Fig. 10). These results suggest that HGF functions in promoting the proliferation of myogenic cells during the earlier period rather than the later period of tongue development.

Figure 10.

Graph showing the labeling index of [3H]thymidine-incorporating tongue cells in the mouse embryos at embryonic day (E) 12 and E15. * P < 0.005 compared with E15 (n = 3).

DISCUSSION

Although knockout experiments such as null-mutation of specific genes have advanced a great deal our knowledge of the growth factor functions, they are often inappropriate for analyzing the roles of such multifunctional growth factors as HGF, because when the organogenesis is arrested at one of the earlier critical steps, it is difficult to determine whether the growth factor is also responsible for any of the subsequent critical steps. In this context, the mandibular organ culture system, which has been confirmed to simulate almost perfectly the embryonic development of oral tissues, including tongue (Slavkin et al., 1989, 1990; Mayo et al., 1992; Shum et al., 1993, Chai et al., 1994; Yamane et al., 1998a,b; Amano et al., 1999), has great advantages. By using this system, we have revealed that HGF can positively regulate the tongue development in at least two critical aspects, i.e., migration and proliferation of myogenic cells.

Previous studies concerning the roles of HGF/c-Met system on tongue development has focused on the periods before myogenic cells arrive at the tongue area. Dietrich et al. (1999) showed that, in mice lacking c-Met, the migratory precursors, including those of tongue muscles, fail to take up long-range migration from dermomyotome. The present study has demonstrated that the level of HGF expression in tongue is highest at E11 and decreases gradually thereafter, and a short-term supplementation of A-ODN for HGF or c-Met in mandibular culture explants starting from E10 is sufficient to prevent myoblasts from migrating into the tongue area. These results strongly suggest that endogenous HGF produced locally in the mandibles is essential for the migration of myogenic cells into tongue.

The source of HGF for skeletal muscle development, i.e., myoblast itself or nonmyogenic mesenchymal cells surrounding myoblasts, has been controversial (Seale and Rudnicke, 2000). In general, HGF has been considered essentially a paracrine factor secreted by mesenchymal cells and effective on epithelial cells, except in some autocrine examples such as several types of tumor cells (Bellusci et al., 1994; Maier et al., 1996; Kurimoto et al., 1998). Our previous study revealed the cellular localization of HGF and its receptors in the embryonic mouse mandible, which supported both paracrine and autocrine functions of HGF in bone and cartilage development (Amano et al., 1999). In the present study, although the receptor c-Met is continuously detected only in the myogenic lineage throughout tongue development, the ligand HGF is initially detected in nonmyogenic mesenchymal cells in mandibular process at E10 before the arrival of myogenic cells and shifts to myoblasts at around E12. This finding indicates that the mode of HGF function changes from paracrine to autocrine during the earlier stages of tongue myogenesis. The local expression of HGF and absence of c-Met–containing myogenic cells in the tongue presumptive area observed at E10 suggests the chemotactic action of HGF to remote targets. In support of this notion, Takayama et al. (1996) demonstrated an inappropriate formation of skeletal muscle in the central nervous system induced by an ectopic expression of HGF in the adjacent neural tube, suggesting that, in normal development, HGF serves as a chemotactic agent to guide migratory c-Met–containing myogenic precursor cells. The present study has shown that both the deficiency of endogenous HGF caused by HGF A-ODN and the interruption of HGF signal transduction caused by c-Met A-ODN lead to the failure of tongue myogenesis accompanied by the retention of myogenic cells at the posterior end of cultured explants. This finding clearly indicates that HGF plays an essential role in the migration of myogenic cells during tongue morphogenesis and that this action of HGF is mediated by its specific receptor c-Met.

A previous study by Mayo et al. (1992) demonstrated that the explants of E10 mandible contain only a small number of desmin-positive myogenic cells in the tongue area, and that they increase in number dramatically toward the end of the 9-day culture period. There is also evidence that the level of desmin mRNA expression in the mandible is elevated between E12 and E15 (Yamane et al., 2000). Because the migration of myoblasts from the occipital somites into tongue primordium is considered to be accomplished around E11, subsequent increase in the number of myoblasts must be primarily caused by the proliferation of myoblasts in the developing tongue. Indeed, [3H]thymidine uptake demonstrates extensive proliferation of tongue myoblasts during the culture period (Mayo et al., 1992). The present result showing elevation of the level of c-Met mRNA in tongue between E11 and E15 may also represent the proliferation of myoblasts. The present study has further demonstrated that exogenous HGF added to the culture medium causes a significant increase in the number of proliferating cells in tongue, which is blocked by A-ODN for c-Met, suggesting that HGF is capable of specifically promoting not only migration but also proliferation of tongue myoblasts. As both HGF and c-Met are expressed in myoblasts after E12, the action of intrinsic HGF on myoblast proliferation is considered to be autocrine. This result is consistent with a previous in vitro study showing that C2, a skeletal muscle satellite cell line, expresses high levels of HGF and c-Met when they are actively proliferating as myoblasts in the medium containing 20% fetal calf serum (Anastasi et al., 1997). However, considering that HGF expression is continuously down-regulated in the developing tongue after E11 (Fig. 3) and that HGF A-ODN prevents tongue development when it is present in the earlier 4 days but not when it is present in the later 4 days of culture period, the contribution of intrinsic HGF on the myoblast proliferation in vivo is likely to be restricted to the earlier periods of tongue development.

In conclusion, the present study revealed that intrinsic HGF plays important roles in the earlier stages of tongue morphogenesis in mouse embryos by promoting both the migration and proliferation of myogenic cells through its cognate receptor c-Met in paracrine and autocrine manners.

EXPERIMENTAL PROCEDURES

Animals

Pregnant time-mated ddY mice were purchased from Japan SLC, Ltd. (Hamamatsu, Japan). The day a vaginal plug appeared was designated as embryonic day 0 (E0). At E10, E11, E12, E13, E14, E15, E16, and E17, the pregnant animals were anesthetized by an intraperitoneal injection of sodium pentobarbital (50 mg/kg body weight) and their uteri were taken out and placed in cold Hanks' medium (Nissui, Tokyo, Japan). Under a dissection microscope, embryos were carefully collected from the uteri and were eliminated of their membranes.

Recombinant HGF and Antibodies

Recombinant human HGF and rat HGF were purified from the culture media of Chinese hamster ovary (CHO) cells transfected with the expression vectors containing human and rat HGF cDNAs (Tashiro et al., 1990; Seki et al., 1991). A polyclonal antibody against rat HGF was raised in a rabbit (Montesano et al., 1991), and its specificity was examined in our previous studies (Amano et al., 1994, 1999). A polyclonal rabbit antibody against mouse c-Met was purchased from Santa Cruz Biochemistry (Santa Cruz, CA), and its specificity was confirmed in previous studies by our lab (Amano et al., 1999) and others (Tabata et al., 1996; Kurimoto et al., 1998). A polyclonal antibody against chick desmin was purchased from Sigma (St. Louis, MO), and its specificity and application for mouse embryonic tissues were documented previously (Yamane et al., 1997, 1998a; Dalrymple et al., 1999).

Synthetic Oligodeoxyribonucleotides

Synthetic ODNs purchased from Japan BioService Co. (Asagiri, Japan) and Sawaday Inc. (Osaka, Japan) were used as primers for RT-PCR and as antisense and control ODNs for culture. The ODNs for HGF and c-Met used in RT-PCR were designed according to published mouse HGF and c-Met cDNA sequences (Park et al., 1987; Liu et al., 1993) and further confirmed of their specificity previously by ourselves and others (Santos et al., 1994; Tabata et al., 1996; Ohmichi et al., 1998; Amano et al., 1999). For the internal positive control in RT-PCR, the primers for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were also synthesised according to published GAPDH cDNA sequence (Tso et al., 1985; Zhao et al., 1995). The 15-mer A-ODNs for HGF and c-Met used in the culture experiments were designed to contain the initial translation codons of mouse HGF and c-met cDNAs (Tashiro et al., 1990; Park et al., 1987), respectively, and were phosphorothioated at 3 nucleotides at both end. To confirm penetration of ODN into the explants and cells, FITC-labeled A-ODN for c-Met was also synthesized. The sequences of the ODNs were as follows: HGF A-ODN for culture, GGT CCC CCA CAT CAT; HGF C-ODN for culture, GGA CCC CCT CTA CTA; HGF 5′ primer for RT-PCR, CCA TGA ATT TGA CCT CTA TGA; HGF 3′ primer for RT-PCR, CTG AGG AAT CTC ACA GAC TTC; c-Met A-ODN for culture, GGG GGC CTT CAT TAT; c-Met C-ODN for culture, ATA ATG AAG GCC CCC; c-Met 5′ primer for RT-PCR, AGA TGA ATG TGA ATA TGA AG; c-Met 3′ primer for RT-PCR, CAT ATG AGT TGA TCA TCA TAG; GAPDH 5′ primer for RT-PCR, ACC ACA GTC CAT GCC ATC AC; GAPDH 3′ primer for RT-PCR, TCC ACC ACC CTG TTG CTG TA.

Organ Culture of Mandible

Mouse mandibles were cultured in a serumless, chemically defined medium according to the methods established by Slavkin and his coworkers (Slavkin et al., 1989, 1990). E10 ddY mouse embryos at Theiler's stage 18 (Theiler, 1972) were isolated and the mandibular divisions of the FBA were microdissected and explanted. The explants were supported by Millipore type AABP filters with 0.8-μm pore size and 5 mm diameter on steel rafts and were cultured in BGJb medium (Fitton-Jackson's modified BGJ; Gibco BRL, Grand Island, NY) freshly supplemented with 100 μg/ml ascorbic acid and 100 U/ml penicillin-streptomycin (Gibco BRL) and adjusted to pH 7.4, in an atmosphere of 5% CO2 and 95% air with 100% humidity at 37°C. The explants were separated into 17 groups as shown in Figure 11. All experiments were performed with triplicate or more samples for each condition, and repeated three times or more. Culture media supplemented with or without ODNs were changed every other day, and the culture was continued for 10 days.

Figure 11.

Experimental groups of mandibular organ culture prepared in the present study. The indicated materials are supplemented in the culture media for the time periods indicated by arrows. Thick arrows indicate the period supplemented with [3H]thymidine. HGF, hepatocyte growth factor; A-ODN, antisense oligodeoxyribonucleotide; C-ODN, control (sense) oligonucleotide.

To deprive of endogenous HGF in the cultured mandibles, A-ODN or C-ODN corresponding to a portion of HGF cDNA sequence was added at 25 μM to the culture medium from the beginning of the culture (groups 1–5). To confirm the specificity of the effects of HGF A-ODN, 100 ng/ml of human recombinant HGF was added together with 25 μM of HGF A-ODN (group 6). To examine the time-specific effect of HGF A-ODN, two additional groups were prepared. Explants in one group were supplemented with HGF A-ODN for the first 4 days and subsequently with C-ODN for the rest of the 10-day culture period (group 7). Explants of the other group were supplemented with C-ODN for the first 6 days and with HGF A-ODN for the last 4 days (group 8). To inhibit the intracellular signal transduction of HGF through its receptor, A-ODN or C-ODN for c-Met was added to the culture in the same way as that for HGF (groups 9, 10).

Effects of overexpression of HGF were examined by supplementation of 100 ng/ml human recombinant HGF to the culture media (group 12). The optimal concentration of HGF was determined by previous studies (Amano et al., 1999; Ohmichi et al., 1998). To examine the specificity of the effects of exogenous HGF, A-ODM or C-ODN for c-Met was added at 25 μM together with recombinant HGF (groups 13, 14) (Amano et al., 1999). To examine the time-specific effect of exogenous HGF, two additional groups were prepared. Explants of one group were supplemented with recombinant HGF for the first 4 days (group 16), and the explants of the other group were supplemented with HGF for the last 4 days (group 17). To label proliferating cells in the explants, 20 μCi/ml of [methyl-3H]thymidine (740 GBq/mmol, NEN, Boston, MA) was added in the culture media of groups 11–14 and 17 for the last 24 hr of the culture period, and of groups 15 and 16 for 24 hr of the 4th day of culture.

To label proliferating cells in vivo, the tongues of E12 and E15 embryos were separated in cold Hanks' medium and cut into slices. They were cultured in the BGJb medium containing 20 μCi/ml of [methyl-3H]thymidine (740 GBq/mmol) for 24 hr.

RNA Extraction and Competitive RT-PCR

Total RNA extraction and competitive RT-PCR amplification was performed as previously described (Yamane et al., 1998b, 2000). Briefly, total RNA was extracted according to the specification of the manufacturer (Rapid Total RNA Isolation Kit, 5 Prime→3 Prime, Inc., Boulder, CO). To remove contaminants of genomic DNA, the RNA was treated with 2 U of ribonuclease-free deoxyribonuclease I (Life Technologies, Gaithersburg, MD). The RNA (1.5 μg) was reverse transcribed by 200 U of reverse transcriptase (SuperScript II, Life Technologies). Internal standards (competitors) for the competitive RT-PCR amplification were constructed according to the manufacturer's instructions (PCR MIMIC Construction Kit, Clontech Laboratory, Inc., Palo Alto, CA). Total cDNA (50 ng) was amplified along with the internal standard in the presence of primers specific to the target genes in a thermal cycler (TP3000, Takara Biomedicals, Shiga, Japan). Amplification products were isolated by electrophoresis. Fluorescent intensities in bands were measured by an image analyzer (Argus-100, Hamamatsu Photonics K.K., Hamamatsu, Japan) and ratios of fluorescent intensities in the target gene bands to those in their respective internal standard bands were calculated. The mRNA amount of each target gene for each sample was estimated from the standard curve of the amount of known cDNA concentrations for each target gene, normalized by the amount of GAPDH and expressed as percentage of control value.

Histology and Histochemistry

Mouse embryos at E10, 11, 12, 14, 16, and groups of three or more mandibular explants cultured for 10 days under various conditions were subjected to morphologic analysis. For H&E staining and immunohistochemistry, the specimens were fixed by immersion in 4% paraformaldehyde in 0.1 M phosphate buffer, dehydrated in ethanol series, cleared with xylene, and embedded in paraffin. Consecutive sections of 5 μm thickness were cut from each paraffin block and mounted on silane-coated glass slides (Dako, Glostrup, Denmark). Some paraffin sections were stained with H&E and prepared for histologic observation and morphometric analysis. The rest of the sections were incubated with 3% (v/v) normal porcine serum in PBS for 30 min and then with anti-HGF, anti–c-Met, or anti-desmin antibody, each diluted 1:200 with PBS overnight at room temperature. They were then treated for 1 hr at room temperature with biotinylated anti-rabbit immunoglobulin G (Vector Laboratories, Inc., Burlingame, CA) diluted 1:200 with PBS. The site of immunoreaction was made visible by incubating the sections with horseradish peroxidase-conjugated streptavidin (Dako) diluted 1:300 with PBS for 1 hr, then with 0.01% 3′,3′-diaminobenzidine tetrahydrochloride and 0.002% hydrogen peroxide in 0.05 M Tris-HCl (pH 7.6) at room temperature. For immunohistochemical control, the antibodies preabsorbed overnight at 4°C with recombinant rat HGF (1 μg/ml) or synthetic peptide for c-Met (1 μg/ml) (Santa Cruz) was used as primary antibodies.

Detection of Proliferating Cells

Cell proliferation was analyzed by radioautography as described in our previous studies (Amano et al., 1999; Amano and Iseki, 2001). Briefly, paraffin sections were made from [3H]thymidine-labeled explants and slices as described above, dipped in Kodak NTB2 nuclear track emulsion diluted with 0.33 M ammonium acetate, and then placed in a dark box at 4°C. After 10 days, the sections were developed, stained with H&E, and subjected to observation with the microscope/CCD video camera system (Olympus Co., Tokyo, Japan). At the 480-fold magnification, nine mesodermal fields of tongue tissues were selected randomly from different sections obtained from three explants. The percentage of labeled cells (labeling index) in tongue myogenic cells was calculated in each field and expressed as mean ± SD of nine fields. The statistical significance of the difference in values was examined with Student's t test, and the difference with P < 0.01 was considered significant.

Acknowledgements

The authors thank Mr. S. Yamazaki and Ms. Y. Akabori for their photographic and secretarial assistance. A part of this work was performed in the Tsurumi University High Technology Research Center.

Ancillary