- Email: [email protected]
Pax-3 is necessary but not sufficient for lbx1 expression in myogenic precursor cells of the limb
Mechanisms of Development 73 (1998) 147–158
Pax-3 is necessary but not sufficient for lbx1 expression in myogenic precursor cells of the limb Detlev Mennerich1, Konstanze Scha¨fer1, Thomas Braun* Department of Cell and Molecular Biology, University of Braunschweig, Spielmannstrasse 7, 38106 Braunschweig, Germany Received 8 January 1998; revised version received 2 March 1998; accepted 2 March 1998
Abstract In vertebrates all skeletal muscles of trunk and limbs are derived from condensations of the paraxial mesoderm, the somites. Limb muscle precursor cells migrate during embryogenesis from somites to limb buds where migration stops and differentiation occurs. We have characterized lbx1 homeobox genes in chicken and mice and found them to be expressed in migrating limb muscle precursor cells in both species. Analysis of splotch mutant mice showed that lbx1 and c-met are differently affected by the lack of Pax-3. Limb buds of splotch (Pax-3 mutant) mice were devoid of lbx1 transcripts, while expression of c-met was still detectable at a low level. The presence of c-met-positive cells in splotch mice entering the limbs indicates that migration of cells from somites to limbs is not entirely dependent on Pax-3. We show that induction of epithelial to mesenchymal transition of Pax-3-positive cells by SF/HGF was not sufficient to induce ectopic lbx-1 expression at the inter-limb level, while ectopic limb formation was able to activate lbx1 expression. We postulate that Pax-3 is necessary for lbx1 expression in the lateral tips of somites but additional, yet unknown signals derived from limb buds are needed to initiate lbx1 expression. The role of limb bud-derived signals involved in targeted muscle precursor cell migration, and lbx1 activation was further confirmed by analysis of explanted somite/limb bud co-cultures in collagen gels. 1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: Migration; Epithelial-mesenchymal transition; lbx1; Pax-3; c-met
1. Introduction Somites form after gastrulation by condensation of paraxial mesoderm in cranial to caudal direction. In vertebrates they are the sole source for all skeletal muscle cells of the trunk and limbs (Chevallier et al., 1977; Christ et al., 1978). During development cell populations emerge in the somite, which acquire different fates and are dependent on signals from neighboring structures (for reviews see Brand-Saberi et al., 1996a; Lassar and Mu¨nsterberg, 1996). Based on transplantation experiments, two different myogenic cell lineages were identified which are distinguished by their prospective migration pattern and their dependence on different neighboring structures (Ordahl and LeDouarin, 1992). Cells located in lateral parts of somites migrate toward the limb fields, while cells in the medial part remain * Corresponding author. Tel.: +49 531 3915722; fax: +49 531 3918178; e-mail: [email protected] 1 D.M. and K.S. contributed equally to this study.
0925-4773/98/$19.00 1998 Elsevier Science Ireland Ltd. All rights reserved PII S0925-4773 (98 )0 0046-X
in place and form the autochthonous axial musculature and myotomes. Migration seems incompatible with muscle cell differentiation, since only resident cells express myogenic bHLH proteins which are characteristic for muscle cell differentiation (Braun et al., 1992). Migrating limb muscle precursor cells initially do not depend on myogenic factors and express myogenic factors only after they have reached their target (Sassoon et al., 1989; Braun et al., 1994; Tajbakhsh and Buckingham, 1994). Ablation experiments in chicken have demonstrated that cells in the medial half of the dermomyotome from which the epitaxial musculature is derived depend on signals from the neural tube and/or the notochord, while cells from the lateral half which later form the hypaxial musculature do not show such a dependency (Rong et al., 1992; Bober et al., 1994a). In contrast, signals from the forming limb bud appear to be required for the epithelio-mesenchymal transition of dermomyotomal cells in the lateral part of somites, a step absolutely essential for migration of limb muscle precursors into limb buds (Hayashi and Ozawa, 1995). Until
148
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
recently the molecular nature of these signals was enigmatic. During the last years, however, a number of different molecules was shown to play important roles in different processes and interactions required for the formation of the migrating muscle cell precursors (Sonnenberg et al., 1993; Bladt et al., 1995; Hayashi and Ozawa, 1995; Brand-Saberi et al., 1996c; Heymann et al., 1996). The epistatic relationship between different genes has been suggested based on their expression patterns (Daston et al., 1996; Epstein et al., 1996; Yang et al., 1996). However, only relatively few functional studies were performed to reveal the operating mechanisms and networks controlling migration of limb muscle precursors. An initial step appears to be controlled by the Pax-3 gene which is expressed very early during formation of the paraxial mesoderm and later becomes confined to the dermomyotomal layer. Mutation of the Pax-3 gene leads to changes of the morphology of the dermomyotome and to abrogation of muscle precursor cell migration (Bober et al., 1994b; Goulding et al., 1994; Williams and Ordahl, 1994). A second step is defined by effects mediated by the HGF/c-met ligand/receptor system. Knock-out experiments in mice (Bladt et al., 1995) and in ovo manipulations using HGF-soaked beads (Brand-Saberi et al., 1996c; Heymann et al., 1996) suggested that HGF expressed by the developing limb bud initiates the epithelial to mesenchymal transition of the dermomyotomal epithelium required for limb muscle precursor cell migration. Consequently, mutation of either gene results in an inhibition of migration at an early stage. Analysis of splotch mice suggested that c-met might be regulated by Pax-3, since mutation of the Pax-3 gene leads to a downregulation of c-met in somites and limb buds (Daston et al., 1996; Yang et al., 1996). However, reduced levels of c-met were still detectable by RT-PCR in limb buds of mutant mice (Yang et al., 1996). It remains controversial whether expression of c-met and HGF in migrating cells and limb buds is necessary to maintain the mobility of migrating cells or serves as a chemotactic agent to direct targeted migration. Other signals controlling limb muscle precursor cell migration into the limb field and their homing into prospective muscle forming areas are elusive. It appears likely that surface molecules or cell adhesion molecules either alone or together with secreted signal molecules are involved in this process (Lassar and Mu¨nsterberg, 1996). It has been suggested, for example, that N-Cadherin plays a critical role in myoblast migration (Brand-Saberi et al., 1996b). However, a N-Cadherin mutation in Drosophila does not result in changes in muscle cell migration (Masatoshi Takeichi, Kyoto University, pers. commun.) and the knock-out analysis in mice has not been informative since homozygous mutant embryos die before migration starts (Radice et al., 1997). It is clear, however, that somitic cells reach the limbs by active locomotion and that a passive matrix-driven locomotion plays only a minor role if any (Sze et al., 1995). Homeobox-containing genes have been shown to control
body axis and pattern formation within a number of different tissues. Although it is still unclear how homeobox genes exert their role as decisive components in morphogenetic processes, evidence has been accumulated that homeobox genes may control expression of cell surface and cell adhesion molecules and thereby control pattern formation (Edelman and Jones, 1995; Newman, 1996). In addition, homeobox genes have been shown to be involved in determination/differentiation of various cell types, a process which is of paramount importance for the formation of muscle and other organs (Anderson et al., 1997). In the present study we have investigated the role of the lbx1 homeobox-containing gene in the regulation of events controlling limb muscle precursor cell migration in chicken and mice. We show that expression of lbx1 in limb muscle precursor cells is strictly dependent on Pax-3 expression, while expression of c-met in Pax-3 deficient mice is not completely abrogated. Pax-3 appears to be necessary but not sufficient for lbx1 expression, since induction of epithelial to mesenchymal transition of Pax-3-positive cells by HGF does not lead to ectopic lbx-1 expression. Finally, migration of cells from somites to limb buds was studied in an explant culture system in vitro.
2. Materials and methods 2.1. Molecular cloning of mouse and chicken lbx1 A mouse genomic library was screened under reduced stringency using a PflMI restriction fragment of the S59 Drosophila cDNA as a probe (kind gift of M. Frasch, Mount Sinai Medical Center, New York) which contains the homeobox. Hybridizing clones were purified and analyzed using standard procedures (Braun et al., 1989). To generate probes for in situ hybridization appropriate restriction fragments were cloned in pKSII (Stratagene), linearized, and used to produce run-off cRNA transcripts. The chicken lbx1 cDNA clone was isolated by screening 2 × 106 pfu of a chicken limb bud library (kindly provided by William Upholt, University of Conneticut Health Cente) using a 700-bp PCR fragment comprising the mouse homeobox domain. The nucleotide sequence from the longest cDNA clone was determined on both strands by primer walking using an ABI automatic sequencer. All DNA work was performed with standard procedures as outlined in Sambrook et al. (1989). 2.2. Mouse and chicken embryos Splotch1H mice used in this study were obtained from Prof. Thomas Franz (University of Bonn, Germany). Splotch1H mice were backcrossed to C57/Bl6 mice for propagation of the strain (obtained from Harlan Winkelmann, Paderborn, Germany). Homozygous splotch mice were identified by the presence of exencephaly and spina bifida.
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Wild-type embryos (C57/BL6) and Splotch embryos were obtained from timed matings. The morning of the appearance of the vaginal plug was designated E0.5. Fertilized chicken eggs were purchased from Lohmann Bruteier, Cuxhaven, Germany, incubated at 38°C, and staged according to Hamburger and Hamilton (1951). 2.3. Chicken embryo manipulation and application of FGF2 and SF/HGF beads All manipulations were performed as described by Heymann et al. (1996). After windowing of eggs at different developmental stages (Hamburger and Hamilton, 1951) the vitelline membrane overlaying the prospective implantation areas was torn away with a sharpened tungsten needle. The ectoderm covering the lateral plate mesoderm was cut at the appropriate somite level and a bead was carefully positioned inside the slit. Heparin acrylic beads (H5263, Sigma) were washed several times in Ringer solution, selected for size and incubated for 2 h at room temperature in a glass capillary filled with 500 mg/ml FGF-2 (R&D Systems, Minneapolis, MN) or 200 mg/ml SF/HGF (R&D Systems) before use. 2.4. In situ hybridization Non-radioactive whole-mount hybridizations were performed with digoxigenin-labeled cRNA probes as described by Wilkinson (1992). For double-labeling experiments individual templates were transcribed in the presence of digoxigenin-UTP and fluorescin-UTP, respectively. cRNA probes were pooled and used for hybridization following the standard protocol. Embryos were first reacted with alkaline phosphatase coupled to anti-fluorescin antibodies (Boehringer Mannheim, Germany) and stained with fast red (Boehringer Mannheim). Bound antibodies were removed by incubation in glycine and embryos were subjected to a second antibody reaction using alkaline phosphatase coupled anti-digoxigenin antibodies. After staining with NBT and X-phosphate (Boehringer Mannheim), embryos were fixed in 4% paraformaldehyde and stored in phosphate-buffered saline. For sectioning, embryos were embedded in a gelatin/ albumen mixture and cut with a vibratome at 20–100 mm as described previously (Bober et al., 1994b). 2.5. Explant cultures Co-cultivation of lateral parts of somites with limb buds and MRC-5 cell pellets was done essentially as described by Guthrie and Lumsden (1994). HH stage 18/19 embryos were washed thoroughly in Ringer salt solution before sagittal cuts were made to divide somites in a lateral and a medial half using a micro-scalpel. Lateral parts were removed with a micropipette and yellow tips and placed on a drop of gelled collagen (10 ml) in a 4-well petri dish. Limb bud fragments and MRC-5 pellets were added and positioned
149
at a distance of approximately 150 mm relative to somite fragments. After removal of as much liquid as possible, 80 ml of collagen were added. DMEM (1ml) with 10% FCS was added to each well after the collagen gel was set and dishes were incubated at 10% CO2 for the indicated length of time. Collagen was prepared from rat tails following exactly the protocol from Guthrie and Lumsden (1994). MRC-5 cells were obtained from the ETCC. Secretion of SF/HGF from MRC-5 cells was verified by testing MRC-5 tissue culture supernatants for scatter factor activity on MDCK cells as described by Montesano et al. (1991). Pellets of MRC-5 cells were obtained by culture of cells in microtiter plates with a conical-shaped bottom and mechanical mobilization of resulting cell aggregates.
3. Results 3.1. Isolation of mouse and chicken ladybird like (lbx1) genes A low stringency screening of a mouse genomic library was performed with a probe derived from the Drosophila homeobox gene S59 which is expressed in muscle and neuronal precursor cells of flies (Dohrmann et al., 1990). Isolated genes were hybridized at increasing stringencies with the S59 probe and grouped according to their hybridization strength. cRNA probes were synthesized from various clones of each group and used to hybridize mouse embryos between E9.5 and E11.5. Among the clones which generated discrete hybridization signals one yielded a particularly interesting expression pattern in somites, limb buds, and the CNS. Sequence analysis revealed that this gene represented the mouse homolog of the Drosophila homeobox gene ladybird which was first isolated and described in mouse and human by Jagla et al. (1995) and named lbx1. We next isolated the chicken homologue of lbx1 to compare expression patterns in both species and to benefit from the chicken experimental system with its possibility for micromanipulation and local (over)expression. The homeodomain of the mouse lbx1 gene was isolated and used to screen a chicken limb bud library at low stringency. Among many other known or novel members of the homeobox gene family, we identified a set of clones which appeared to represent the chicken ortholog of the mouse and human genes. The longest of these clones contained an insert of 1254 bp comprising most of the coding region of the mRNA and the 3′-untranslated region. Comparison of the deduced peptide sequence revealed 100% identity to the mouse and 98% identity to human proteins in the homeobox domain; the N-terminal domain is 82% identical to the mouse and 81% identical to the human sequence. In contrast, the Cterminal domains are highly divergent with the exception of isolated amino acid stretches immediately adjacent to the homeobox domain and at the very end of the protein (Fig. 1). Interestingly, in the N-terminal region the highest degree
150
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Fig. 1. Sequence comparison of chicken, mouse and human lbx1 proteins. Deduced amino acid sequences were compared using the ClustalV program. Amino acids conserved between chicken, mouse and human are boxed.The homeodomain is indicated by a blue frame. Chicken and mouse homeoboxes are identical. The human homeobox shows only one conservative exchange. The N-terminal domain of all three proteins has a relatively high degree of identity of >80% while in the C-terminal domain only scattered blocks of amino acids are conserved between chicken and mouse or human. In contrast, identity of human and mouse sequences to each other is much higher in the C-terminal domain.
of similarity is also found directly adjacent to the homeobox. 3.2. Chicken lbx1 is highly expressed in limb muscle precursor cells originating from lateral tips of somites We first investigated the expression profile of lbx1 in developing chicken embryos (Fig. 2). A prominent lbx1 expression domain was detected in somites at the level of developing limb buds (Fig. 2A–C). Interestingly, the onset of lbx1 expression correlated with the outgrowth of limb buds at HH stage 17/18. Vibratome sections were prepared from stained embryos and revealed that lbx1 expression in somites was confined to the ventro-lateral tip of the dermomyotome. Cells in this area express high levels of Pax-3 mRNA, while another Pax-3 expression domain in dorsomedial tips of the dermomyotome was negative for lbx1. Lbx1 was barely detectable in somites of the inter-limb region of chicken embryos and caudal to the leg buds. In a more rostral position, lbx1 expression was observed cranial to wing buds in the lateral tips of dermomyotomes. We also
noticed a characteristic streak of lbx1-positive cells which originated from cervical somites below the branchial arches. It is interesting to note that cells of this structure which most probably represent tongue muscle precursor cells also express Pax-3 at a high level. At later stages (HH stage 19/20) when migrating muscle precursor cells leave somites, lbx1 expression was clearly visible in the migrating cell population (indicated by arrows in Fig. 2C,F), concomitant with the migration of dermomyotomal cells from somites into limb buds lbx1 expression decreased in the dermomyotomal cell layer and appeared in a proximal-distal gradient within limbs and wings. After completion of the migration process, lbx1-positive cells were found only in the prospective muscle forming areas of limbs (Fig. 2G–I), whereas the somites at limb level had lost all lbx1 staining. In mice a relatively strong expression of lbx1 was seen in the neural tube and in some areas of the brain stem. In chicken these expression domains were only detectable after prolonged staining, indicating a relatively low expression level compared to limb muscle precursor cells. Our
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
151
Fig. 2. Expression of lbx1 in chicken embryos. Stage 18 (A–C), stage 20 (D–F), and stage 22 (G–I) chicken embryos were subjected to whole mount in situ hybridization using a DIG-UTP-labelled lbx1 probe. Stained embryos were sectioned on a vibratome. At stage 18, lbx1 transcripts were detected in the lateral edge of dermomyotomes adjacent to wing and leg buds and at lower amounts in cranial somites (A–C). At HH stage 20, lbx1 expression becomes visible in the wing/leg buds (D–F) and at HH stage 22 all lbx1-positive cells have moved from somites to wing/leg buds (G–I). At this stage a weak expression domain in the neural tube is also visible. Arrows indicate migrating limb muscle precursor cells. dm, dermomyotome; nt, neural tube; lb, leg bud.
analysis of lbx1 expression in mouse embryos (data not shown) did not reveal any differences to results previously reported by Jagla et al. (1995). 3.3. Complete absence of lbx1 but not c-met expression in Pax-3 deficient mice The apparent expression of lbx1 in mouse limb muscle precursor cells prompted us to study lbx1 in Pax-3 mutant mice which have a severe deficit in limb muscle formation. In addition, we analyzed the expression of c-met, another marker for migrating limb muscle precursor cells. We found that expression of lbx1 was absent in somites and limbs of Pax-3 mutant mice between E9.5 and E11.5 (Fig. 3). Despite the complete lack of lbx1 transcripts in somites and cells derived thereof, expression of lbx1 in the CNS of splotch mice was only slightly reduced or was not affected at all. Between E9.5 and E10.5, lbx1 expression appeared to be weaker in neural tubes of splotch mice
(Fig. 3D). At present it is unclear whether the lbx1 decrease here is due to the absence of Pax-3, since reduction of lbx1 appeared not restricted to the expression domain shared with Pax-3. In addition, at E11.5 no reduction of lbx1 expression was detected in the neural tube of splotch mice (Fig. 3F,H). Comparison of sections of wild-type and splotch embryos at E11.5 hybridized with the lbx1 probe did not reveal significant differences in this tissue, while lbx1 expression was completely absent in somites and limb buds of splotch mutant mice (Fig. 3G,H). Therefore, a lower expression level of lbx1 in the spinal cord in splotch mice at early stages compared to wild-type might be due to a general retardation of spinal cord development in mutants rather than a specific regulatory effect of Pax-3 on lbx1 expression. In contrast to the complete lack of lbx1 in somites and limb buds of splotch mice, expression of c-met was not entirely abolished. Although expression levels were reduced and c-met expressing cells in somites appeared disarranged, expression of c-met was clearly detectable in somites and
152
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Fig. 3. Absence of lbx1 expression in somites and limb buds but not in the neural tube of splotch mutant mice. E9.5 (A,B), E 10.5 (C,D), and E11.5 (E–H) wild-type (A,C,E,G) and splotch mutant mice (B,E,F,H) were hybridized with a DIG-UTP labelled lbx1 probe. Either whole-mount preparations (A,B,E,F) or vibratome sections (C,D,G,H) of whole-mount hybridizations are shown. lbx1 expression is absent in somites and limb buds of splotch mutant mice from E9.5 to E11.5. Expression of lbx1 in the neural tube becomes most prominent at E11.5. At this stage lbx1 appears to be expressed normally in neural tubes of splotch mutant mice (F,H) while no lbx1 signals are visible in limbs (indicated by arrows in F,H). The normal expression of lbx1 in the neural tube of splotch mice but the absence in somites and limb buds is particularity evident on the section shown in (H).
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
limb buds of splotch mice at E10.5 (Fig. 4A). At the forelimb level, relatively strong c-met expression was detected (Fig. 4C). This expression domain did not easily compare to the dorso-medial or the ventro-lateral domains normally seen in wild-type mice. The observed differences in c-met distribution might be a result of morphological disturbance of somites in splotch mice where no intact epithelium is seen at the normal tips of the dermomyotome, which results in a shortening of the dermomyotomal cell layer (Daston et al., 1996). This compression of the dermomyotome might account for the aberrant location of c-met expressing cells in the dermomyotome at the forelimb level. At the hindlimb level c-met signals were weaker and more diffuse, while in the tail region virtually normal levels of c-met mRNA were reached (Fig. 4A). It should also be mentioned that c-met signals in somites of splotch mutants decreased when the limbs were forming similar to wild-type mice where this decrease corresponds to the migration of c-met-positive cells into the limbs (data not shown). To evaluate the relationship between cells expressing mutant Pax-3 mRNA and cells expressing c-met in splotch embryos, double-label in situ hybridizations were performed. As shown in Fig. 4D and E, cells which expressed mutant Pax-3 mRNA (dark blue staining in Fig. 4D,E) were immobilized in the dermomyotome. In contrast, c-met-positive cells (red staining in Fig. 4D,E) left the dermomyotome and were found along the way into the limb and at the dorso-medial edge of the somites. An additional expression domain of c-met which was not dependent on Pax-3 was visible in mesonephric tubules (Fig. 4E; Woolf et al., 1995). To investigate whether cmet-positive/Pax-3-negative cells in the limb can give rise to differentiated muscle cells, limb buds were excised and cultured in vitro. This procedure allows an extended development of limbs which is not possible in utero due to the early lethality of splotch mice. Despite the presence of c-met-positive cells in limbs of splotch mice, no differentiated myogenic cells were found in these limbs even after extended culture. MyHC-positive cells were present only in the most proximal parts of the limb, most probably representing shoulder muscles which are not affected by the Pax-3 mutation (data not shown). Therefore, it appears that Pax-3 has a Fig. 4. Expression of c-met is reduced and distorted but not absent in splotch mutant mice. E10.5 splotch mutant (A,C–E) and wild-type mice (B) were subjected to whole-mount in situ hybridization with a c-met probe (A–C) and for double label in situ hybridizations with a c-met and a Pax-3 probe (D,E). Localization of Pax-3 transcripts in double label hybridizations (D,E) is indicated by a dark blue precipitate, presence of c-met transcripts is indicated by a red precipitate. Whole-mount preparations (A) and vibratome sections (B–E) are shown. The presence of c-metpositive cells in somites and fore-limb is clearly visible in (A). In (B,C) a comparison of c-met expression in thoracic somites of wild-type (B) and splotch mutant mice (C) is shown. Note that in splotch mutant mice (D,E), Pax-3-expressing cells do not leave somites (dark blue staining indicated by red arrowheads) while c-met-positive cells (red staining indicated by black arrows) are found between the dermomyotome and the limb and in a distal position within the limb. LB, limb bud; NT, neural tube; DM, dermomyotome; oNT, open neural tube in splotch mice.
153
dual role both in the initiation of cell migration as well as in differentiation of migrating limb muscle precursor cells. 3.4. Delamination of Pax-3-positive cells from the dermomyotome in the inter-limb region is not sufficient to up-regulate lbx1 expression The absence of lbx1 expression in somites of splotch mice suggests that Pax-3 is a critical component needed to
154
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
activate lbx1 transcription. The lack of lbx1 expression in inter-limb somites which also express high amounts of Pax3 in the dermomyotome, however, indicates that Pax-3 alone is not sufficient for lbx1 expression. Since lbx1 is expressed predominantly in somites adjacent to limb buds, it is reasonable to assume that signals from the limb might affect lbx1 expression. SF/HGF is so far the only known secreted signal from the limb mesenchyme which controls decisions of the dermomyotomal epithelium, such as epithelial to mesenchymal transition (Heymann et al., 1996). We therefore investigated whether initiation of migration caused by exogenous SF/HGF together with the presence of Pax-3 would be sufficient to instruct lbx1 expression or whether additional factors are necessary. To address this question, we implanted heparin beads soaked either in SF/ HGF or in FGF-2 in ovo into the flank region of HH stage 15–17 embryos (n = 20). Manipulated embryos were returned to the incubator and allowed to develop. After 24 and 48 h, embryos were isolated and subjected to in situ hybridization with a lbx1 probe. As shown in Fig. 5A–C, no lbx1 expression was detected at the inter-limb level close to implanted SF/HGF beads, while strong expression was found in somites adjacent to limb buds. To ensure that implantation of SF/HGF beads initiated an epithelial to mesenchymal transition of dermomyotomal cells at the inter-limb level, embryos from the same series of experiments were hybridized with the Pax-3 probe. All embryos carrying a SF/HGF bead showed Pax-3-positive cells which delaminated from the dermomyotome and started to migrate (Fig. 5D,E).
The lack of lbx1 up-regulation in delaminating Pax-3positive dermomyotomal cells at the inter-limb level indicated that newly forming limb buds may contribute activities in addition to SF/HGF to induce lbx1 expression in Pax-3-positive cells and activate targeted migration. Therefore, we implanted FGF-2 coated heparin beads into the flank region of HH stage 15–17 chicken embryos (Cohn et al., 1995) and analyzed the expression of lbx1 in ectopically forming limbs (Fig. 5F). As expected, lbx1 expressing migrating cells were clearly detectable in newly formed limbs. As shown previously for Pax-3 expression, these cells were derived from somites directly adjacent to newly formed limbs (Heymann et al., 1996). It appears likely that FGF-mediated activation of lbx1 expression is an indirect process, since lbx1 expression did not occur in the immediate vicinity of FGF coated beads (Fig. 5F). 3.5. Limb buds but not HGF secreting MRC-5 cells are able to stimulate cell migration from lateral somite fragments in explant cultures To further study the relationship between targeted migration and SF/HGF-mediated delamination of dermomyotomal cells, lateral parts of somites of HH stage 17–19 chicken embryos were isolated and co-cultured with limb bud fragments of HH stage 18/19 embryos or with aggregates of SF/HGF expressing MRC-5 lung fibroblasts in collagen matrix gels at a distance between 150 and 350 mm. As expected, no cells were found between both tissue fragments derived from lateral somite halves and limb buds,
Fig. 5. Expression of lbx1 in chicken embryos after implantation of FGF- and HGF-coated heparin beads. Beads soaked in 200 mg/ml SF/HGF (A–E) or 500 mg/ml FGF-2 (F) were implanted into flanks of HH stage 16 chicken embryos and hybridized with a lbx-1 probe (A–C,F) and a Pax-3 probe (D,E) after 24 h of incubation. Despite induction of epithelial to mesenchymal transition of Pax-3-positive dermomyotomal cells by SF/HGF and initiation of migration (indicated by arrowheads in D,E), no induction of lbx1 expression is seen in the inter-limb region (A–C). In contrast, induction of an ectopic limb in the interlimb region by FGF-2 induces lbx1 expression and migration of lbx1-positive cells into the additional limb bud (indicated by arrowheads in F).
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
155
Fig. 6. Explant co-cultures of lateral somite halves with limb bud fragments and with SF/HGF-expressing MRC-5 cell aggregates in collagen gels. Lateral halves of somites were co-cultured with limb bud fragments (A,B) and with MRC-5 cell aggregates for 6 h (A,C) or for 3 days (B,D) in collagen gels. No cells were found between tissue fragments derived from lateral somite halves and limb buds (A) or MRC-5 cell aggregates (C) after 6 h. After 3 days of incubation, cells were found only between limb bud and lateral somite fragments (B) but not between MRC-5 cell aggregates and lateral somite halves (D). lb, limb bud; so, lateral somite half; M-5, MRC-5 cell aggregates.
or MRC-5 cell aggregates shortly after explantation (Fig. 6A,C). However, upon further incubation a cell bridge formed which spanned the gap between somitic and limb bud fragments, while no cells filling the gap between somitic fragment and aggregates of HGF expressing cells were detected (Fig. 6B,D). After extended culture periods in collagen gels we did observe a radial cell outgrowth from explants which partially obscured the picture after 8 days in culture. But clearly, the directional outgrowth from the somitic fragment was not seen with MRC-5 cells. Immunohistochemical staining of collagen cultures with the monoclonal antibody MF-20 against MyHC confirmed that, concomitant with ongoing migration, the number of MyHC-positive cells within limb bud explants increased (data not shown). Additional control experiments in which tissue fragments from the head or the heart of chicken embryos were cultured together with lateral halves of somites did not result in a directed outgrowth or cell migration. Only radial outgrowth from tissue fragments was observed. Reduction of the distance below 100 mm between somitic and limb bud fragments or aggregates of MRC-5 cells did not increase the amount of cells found between explants (data not shown). In contrast, when the distance between explants was extended beyond 600 mm, no directed movement of cells was observed. These results indicate that the range of limb-derived signals is restricted in distance. Interestingly, the limited range of limb-derived signals seems to match the restricted expression of lbx1 in the 4–5 somites located directly adjacent to limb buds.
4. Discussion Development of limb muscles can be divided in several steps including commitment of somitic cells, epithelial to mesenchymal transition of the dermomyotomal epithelium, migration of limb muscle precursor cells into limb buds and determination of myoblasts. Only recently some of these steps have been addressed at the molecular level. In the present study we have characterized the homeobox gene lbx1 in chicken and mice and analyzed cell migration into limb buds by various means. We found that lbx1 is expressed at all stages of limb muscle precursor cell development, which suggests that lbx1 may be an important part of the machinery controlling limb muscle precursor cell development and migration. Expression of lbx1 appeared to be dependent on several distinct mechanisms. (a) Pax-3 expression: In the lateral parts of somites and in migrating limb muscle precursor cells the expression of lbx1 depends on Pax-3. Lack of Pax-3 in splotch mutants completely abolished lbx1 expression in these structure. However, Pax-3 is not sufficient to drive lbx-1 expression since lbx-1 is not expressed in the inter-limb region of chicken embryos where Pax-3 is present at the same concentration as in the limb regions. (b) Neural system specific factors: At E10.5 of mouse development, expression of lbx1 is reduced but not absent in the neural tubes of splotch mice. At E11.5, virtually normal levels of lbx1 RNA were found in the neural tubes of mutants. These results clearly demonstrate that additional factors are present in the neural tube which can activate lbx1 expression.
156
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Whether Pax-3 is important for high-level lbx1 expression at early stages or whether a reduced lbx1 expression is due to a general retardation of neural tube development in splotch mice is hard to distinguish and may require analysis of additional marker molecules which specifically reflect the stage of neural tube development. (c) Limb bud-derived factors: The formation of an ectopic limb in the inter-limb region leads to initiation of lbx-1 expression, indicating that limb buds contain signals required for lbx-1 activation. This activity is different from SF/HGF, since SF/HGF-coated beads did not induce lbx1 when implanted into the interlimb region despite the delamination of Pax-3-positive dermomyotomal cells in this area. From our experiments it is clear that during limb bud formation, a process is initiated which acts back on somites to induce lbx1 expression. This may contribute to the commitment of muscle precursor cells and to their migration pattern. It is also clear that this activity is distinct from SF/HGF, although it cannot be excluded that SF/HGF is a part of this signal. The existence of an additional signaling pathway, necessary for lbx1 expression and targeted cell migration, was further supported by our explantation experiments. Only in explant cultures where lateral halves of somites were co-cultivated with limb bud fragments within a distance of 150–350 mm was a targeted migration of cells apparent. The limited range of putative signals in explant cultures may correspond to the restriction of lbx1 expression to somites which are located adjacent to limb buds. According to this model, the lack of lbx1 expression in the inter-limb region and caudal to the hind-limb may result from the absence or limited range of inducing substances supplied by developing limb buds. In contrast to the complete absence of lbx1 in somites of splotch mice, expression of c-met was only reduced and the distribution of c-met transcripts was disarranged. We also noted differences in the expression profile between cranial and caudal somites and between forelimbs and hindlimbs. A more localized expression was found in forelimbs, while in hindlimbs c-met-expressing cells appeared to have lost their orientation and were scattered in the body wall and the hindlimb bud. Cranial to caudal differences in the expression of c-met in splotch mice as well as the disseminated distribution of c-met-positive cells, a general reduction of the expression level, and analysis of only some developmental stages may have contributed to the failure to detect c-met expression in splotch mice by whole-mount hybridization (Daston et al., 1996; Yang et al., 1996). However, using RT-PCR Yang et al. (1996) were able to detect c-met mRNA in limbs of splotch mice which is in line with our findings. Using radioactive in situ hybridization on sections, Epstein et al. (1996) reported a deficiency of c-met expression in the lateral dermomyotome of splotch mice at E11.5 but a relatively strong expression in the dorsal forelimb which was attributed to the distal ectoderm. In contrast, we have observed a disarranged expression of c-met in the dermomyotome of splotch1H mice at E10.5 which starts to
disappear as cells migrate into the limb bud. Although we agree with Epstein et al. (1996) that c-met is expressed in limbs of splotch mutants, analysis of our single and double whole-mount in situ hybridizations does not suggest an expression of c-met in the ectoderm of limbs. It is hard to pinpoint the nature of c-met-positive cells in the limb of Pax-3 mutant mice, since these cells do not differentiate into muscle cells as indicated by lack of MyHC expression, nor do they express other known differentiation markers. A number of different explanations can be given to account for the presence of the c-met-positive cell population. First, the expression of c-met in the limb of splotch mice may be not related to any migrating cell population, in particular myogenic precursor cells, but may occur de novo in the limb mesenchyme. Second, c-met may be expressed in a population of cells which only partially overlaps with Pax-3-positive cells and which is not myogenic. Angioblasts, for example, migrate from the dorso-lateral quadrant of the somite into the wing field and the ventro-lateral body wall and may represent such cells (Wilting et al., 1995). Third, c-met expressing cells may represent a sub-population of myogenic precursor cells which are not dependent on Pax-3 for migration but for differentiation. Although none of these hypotheses can be completely excluded, the first explanation appears most unlikely. Bladt et al. (1995) have not observed cells which express a mutant c-met transcript in limbs of c-met knock-out mice despite the preservation of other c-met expression domains in these mice. This finding clearly indicates that c-met expression in limbs is derived from migratory cells. In addition, we have observed a continuous layer of c-met expressing cells located between the ventro-lateral edge of the somite and the limb bud of splotch mutant mice. This staining pattern is characteristic for migrating cells and suggests that c-met cells in the limb of splotch mice originate from somites. While the migratory character of c-met-positive cells located between somites and the limb bud and within the limb buds of splotch mice appears very likely, it is much harder to decide whether these cells represent non-myogenic cells, such as angioblasts, or myogenic cells which cannot longer differentiate into muscle because of the absence of Pax-3. The lack of any differentiated muscle cell in c-met containing limbs of splotch mice may favor the view that these cells represent non-myogenic, migratory cells. However, it has been shown recently that Pax-3 can activate the myogenic program via MyoD and Myf-5 under certain conditions (Maroto et al., 1997; Tajbakhsh et al., 1997). Therefore, it cannot be excluded that c-met-positive/Pax3-negative cells in the limbs of splotch mice represent myogenic precursor cells which can no longer differentiate due to the absence of Pax-3. It appears likely that differentiation of most limb muscle cells is independent of Pax-3, since the majority of cells transplanted from lateral halves of somites of splotch mice into limb buds of chicken embryos differentiate into muscle (Daston et al., 1996). Nevertheless, it is
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
possible that Pax-3 is necessary for the differentiation of only a subset of cells. Failure of a subpopulation of cells to differentiate may be easily obscured by other differentiating cells and prevent the detection of cell-autonomous muscle defects in Pax-3 mutant mice. The presence of lbx1 in Pax-3-positive cells makes it also difficult to distinguish, without a mutation of lbx1 being available, whether both molecules are necessary for limb muscle differentiation or whether Pax-3 alone is sufficient for muscle cell determination. Although it is possible that c-met + /Pax-3 − cells are not myogenic but rather represent another cell type, this does not necessarily mean that these cells are not involved in the migration process or the pattern formation of muscle cells. For example, c-met + /Pax-3 − cells may serve as pioneer cells during migration as indicated by their location proximal to Pax-3-positive cells. Such a hypothetical function does not require that c-met + /Pax-3 − cells become muscle. Future research, in particular the careful analysis of the c-met + /Pax-3 − cells, will probably disclose their identity and their role in the migration of cells from somites to limbs. In addition, the generation of lbx1 null mutants will reveal whether lbx1 is only a useful marker molecule for migrating limb muscle precursor cells or whether it fulfils a critical role in the specification of the migrating cell population and in the pathway, leading to an arrangement of muscle precursor cells within their future locations in the limb. Acknowledgements The expert technical assistance by Stefanie Willenzon is gratefully acknowledged. We would like to thank Peter Gruss for chicken Pax-3 cDNA, Manfred Frasch for the Drosophila S59 probe, Carmen Birchmeier for the mouse c-met probe and William Upholt for the chicken limb bud cDNA library. We are indebted to Hans-Henning Arnold for continuous support, encouragement and improvements to the manuscript, Eva Bober and Markus Ko¨ster for helpful discussions and critically reading the manuscript. This work was supported by the SFB 271: ‘Molekulare Mechanismen morphoregulatorischer Prozesse’ project B2 to T.B. and by the ‘Fonds der Chemischen Industrie’.
References Anderson, S.A., Qiu, M., Bulfone, A., Eisenstat, D.D., Meneses, J., Pedersen, R., Rubenstein, J.L., 1997. Mutations of the homeobox genes Dlx-1 and Dlx-2 disrupt the striatal subventricular zone and differentiation of late born striatal neurons. Neuron 19, 27–37. Bladt, F., Rietmacher, D., Isenmann, S., Aguzzi, A., Birchmeier, C., 1995. Essential role for the c-met receptor in the migration of myogenic precursor cells into the limb bud. Nature 376, 768–771. Bober, E., Brand-Saberi, B., Ebensperger, C., Wilting, J., Balling, R., Paterson, B.M., Arnold, H.H., Christ, B., 1994a. Initial steps of myogenesis in somites are independent of influence from axial structures. Development 120, 3073–3082.
157
Bober, E., Franz, T., Arnold, H.-H., Gruss, P., Tremblay, P., 1994b. Pax-3 is required for the development of limb muscles: a possible role for the migration of dermomyotomal muscle progenitor cells. Development 120, 603–612. Brand-Saberi, B., Wilting, J., Ebensperger, C., Christ, B., 1996. The formation of somite compartments in the avian embryo. Int. J. Dev. Biol. 40, 411–420. Brand-Saberi, B., Gamel, A.J., Krenn, V., Muller, T.S., Wilting, J., Christ, B., 1996. N-Cadherin is involved in myoblast migration and muscle differentiation in the avian limb bud. Dev. Biol. 178, 160–173. Brand-Saberi, B., Muller, T.S., Wilting, J., Christ, B., Birchmeier, C., 1996. Scatter factor/hepatocyte growth factor (SF/HGF) induces emigration of myogenic cells at interlimb level in vivo. Dev. Biol. 179, 303–308. Braun, T., Buschhausen, D.G., Bober, E., Tannich, E., Arnold, H.-H., 1989. A novel human muscle factor related to but distinct from MyoD1 induces myogenic conversion in 10T1/2 fibroblasts. EMBO J. 8, 701–709. Braun, T., Rudnicki, M.A., Arnold, H.-H., Jaenisch, R., 1992. Targeted inactivation of the muscle regulatory gene Myf-5 results in abnormal rib development and perinatal death. Cell 71, 369–382. Braun, T., Bober, E., Rudnicki, M.A., Jaenisch, R., Arnold, H.H., 1994. MyoD expression marks the onset of skeletal myogenesis in Myf-5 mutant mice. Development 120, 3083–3092. Chevallier, A., Kieny, M., Mauger, A., 1977. Limb-somite relationship: origin of the limb musculature. J. Embryol. Exp. Morphol. 41, 245– 258. Christ, B., Jacob, H., Jacob, B., 1978. On the formation of the myotomes in avian embryos. An experimental and scanning electron microscope study. Experientia 34, 514–516. Cohn, M.J., Izpisua-Belmonte, J.C., Abud, H., Heath, J.K., Tickle, C., 1995. Fibroblast growth factors induce additional limb development from the flank of chick embryos. Cell 80, 739–746. Daston, G., Lamar, E., Olivier, M., Goulding, M., 1996. Pax-3 is necessary for migration but not differentiation of limb muscle precursors in the mouse. Development 122, 1017–1027. Dohrmann, C., Azpiazu, N., Frasch, M., 1990. A new Drosophila homeobox gene is expressed in mesodermal precursor cells of distinct muscles during embryogenesis. Genes Dev. 4, 2098–2111. Edelman, G.M., Jones, F.S., 1995. Developmental control of N-CAM expression by Hox and Pax gene products. Phil. Trans. R. Soc. Lond. B Biol. Sci. 29, 305–312. Epstein, J.A., Shapiro, D.N., Cheng, J., Lam, P.Y.P., Maas, R.L., 1996. Pax3 modulates expression of the c-met receptor during limb muscle development. Proc. Natl. Acad. Sci. USA 93, 4213–4218. Goulding, M., Lumsden, A., Paquette, A.J., 1994. Regulation of Pax-3 expression in the dermomyotome and its role in muscle development. Development 120, 957–971. Guthrie, S., Lumsden, A., 1994. Collagen gel coculture of neural tissue. Neuroprotocols 4, 116–120. Hamburger, V., Hamilton, H., 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49–92. Hayashi, K., Ozawa, E., 1995. Myogenic cell migration from somites is induced by tissue contact with medial region of the presumptive limb mesoderm in chicken. Development 121, 661–669. Heymann, S., Koudrova, M., Arnold, H.-H., Koster, M., Braun, T., 1996. Regulation and function of SF/HGF during migration of limb muscle precursor cells in chicken. Dev. Biol. 180, 566–578. Jagla, K., Dolle, P., Mattei, M.G., Jagla, T., Schuhbaur, B., Dretzen, G., Bellard, F., Bellard, M., 1995. Mouse Lbx1 and human LBX1 define a novel mammalian homeobox gene family related to the Drosophila lady bird genes. Mech. Dev. 53, 345–356. Newman, S.A., 1996. Sticky fingers: Hox genes and cell adhesion in vertebrate limb development. Bioessays 18, 171–174. Lassar, A.B., Mu¨nsterberg, A.E., 1996. The role of positive and negative signals in somite patterning. Curr. Opin. Neurobiol. 6, 57–63. Maroto, M., Reshef, R., Munsterberg, A.E., Koester, S., Goulding,
158
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
M., Lassar, A.B., 1997. Ectopic Pax-3 activates MyoD and Myf-5 expression in embryonic mesoderm and neural tissue. Cell 89, 139– 148. Montesano, R., Matsumoto, K., Nakamura, T., Orci, L., 1991. Identification of fibroblast-derived epithelial morphogen as hepatocyte growth factor. Cell 67, 901–908. Ordahl, C.P., LeDouarin, N., 1992. Two myogenic cell lineages within the developing somite. Development 114, 339–353. Radice, G.L., Rayburn, H., Matsunami, H., Knudsen, K.A., Takeichi, M., Hynes, R.O., 1997. Developmental defects in mouse embryos lacking N-cadherin. Dev. Biol. 181, 64–78. Rong, P.M., Teillet, M.A., Ziller, C., Le Douarin, N.M., 1992. The neural tube/notochord complex is necessary for vertebral but not limb and body wall striated muscle differentiation. Development 115, 657–672. Sassoon, D., Lyons, G., Wright, W.E., Lin, V., Lassar, A., Weintraub, H., Buckingham, M., 1989. Expression of two myogenic regulatory factors myogenin and MyoD1 during mouse embryogenesis. Nature 341, 303– 307. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1659 pp. Sonnenberg, E., Meyer, D., Weidner, K.M., Birchmeier, C., 1993. Scatter factor/Hepatic growth factor and its receptor, the c-met tyrosine kinase, can mediate a signal exchange between mesenchyme and epithelia during mouse development. J. Cell Biol. 123, 223–235.
Sze, L.-Y., Lee, K.K.H., Webb, S.E., Li, Z., Pauline, D., 1995. Migration of myogenic cells from the somites to the fore-limb buds of developing mouse embryos. Dev. Dyn. 203, 324–336. Tajbakhsh, S., Buckingham, M., 1994. Mouse limb muscle is determined in the absence of the earliest myogenic factor myf-5. Proc. Natl. Acad. Sci. USA 91, 747–751. Tajbakhsh, S., Rocancourt, D., Cossu, G., Buckingham, M., 1997. Redefining the genetic hierarchies controlling skeletal myogenesis: Pax-3 and Myf-5 act upstream of MyoD. Cell 89, 127–138. Yang, X.-M., Vogan, K., Gros, P., Park, M., 1996. Expression of the met receptor tyrosine kinase in muscle progenitor cells in somites and limbs is absent in Splotch mice. Development 122, 2163–2171. Wilting, J., Brand-Saberi, B., Huang, R., Zhi, Q., Ko¨ntges, G., Ordahl, C.P., Christ, B., 1995. Angiogenic potential of the avian somite. Dev. Dyn. 202, 165–171. Wilkinson, D.G., 1992. Whole mount in situ hybridization of vertebrate embryos, in: D.G. Wilkinson (Ed.), In Situ Hybridization: A Practical Approach, IRL Press, Oxford, pp. 75–83. Williams, B.A., Ordahl, C.P., 1994. Pax-3 expression in segmental mesoderm marks the early stages in myogenic cell specification. Development 120, 785–796. Woolf, A.S., Kolatsi-Joannou, M., Hardman, P., Andermarcher, E., Moorby, C., Fine, L.G., Jat, P.S., Noble, M.D., Gherardi, E., 1995. Roles of hepatocyte growth factor and the met receptor in the early development of the metanephros. J. Cell Biol. 128, 171–184.
Pax-3 is necessary but not sufficient for lbx1 expression in myogenic precursor cells of the limb Detlev Mennerich1, Konstanze Scha¨fer1, Thomas Braun* Department of Cell and Molecular Biology, University of Braunschweig, Spielmannstrasse 7, 38106 Braunschweig, Germany Received 8 January 1998; revised version received 2 March 1998; accepted 2 March 1998
Abstract In vertebrates all skeletal muscles of trunk and limbs are derived from condensations of the paraxial mesoderm, the somites. Limb muscle precursor cells migrate during embryogenesis from somites to limb buds where migration stops and differentiation occurs. We have characterized lbx1 homeobox genes in chicken and mice and found them to be expressed in migrating limb muscle precursor cells in both species. Analysis of splotch mutant mice showed that lbx1 and c-met are differently affected by the lack of Pax-3. Limb buds of splotch (Pax-3 mutant) mice were devoid of lbx1 transcripts, while expression of c-met was still detectable at a low level. The presence of c-met-positive cells in splotch mice entering the limbs indicates that migration of cells from somites to limbs is not entirely dependent on Pax-3. We show that induction of epithelial to mesenchymal transition of Pax-3-positive cells by SF/HGF was not sufficient to induce ectopic lbx-1 expression at the inter-limb level, while ectopic limb formation was able to activate lbx1 expression. We postulate that Pax-3 is necessary for lbx1 expression in the lateral tips of somites but additional, yet unknown signals derived from limb buds are needed to initiate lbx1 expression. The role of limb bud-derived signals involved in targeted muscle precursor cell migration, and lbx1 activation was further confirmed by analysis of explanted somite/limb bud co-cultures in collagen gels. 1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: Migration; Epithelial-mesenchymal transition; lbx1; Pax-3; c-met
1. Introduction Somites form after gastrulation by condensation of paraxial mesoderm in cranial to caudal direction. In vertebrates they are the sole source for all skeletal muscle cells of the trunk and limbs (Chevallier et al., 1977; Christ et al., 1978). During development cell populations emerge in the somite, which acquire different fates and are dependent on signals from neighboring structures (for reviews see Brand-Saberi et al., 1996a; Lassar and Mu¨nsterberg, 1996). Based on transplantation experiments, two different myogenic cell lineages were identified which are distinguished by their prospective migration pattern and their dependence on different neighboring structures (Ordahl and LeDouarin, 1992). Cells located in lateral parts of somites migrate toward the limb fields, while cells in the medial part remain * Corresponding author. Tel.: +49 531 3915722; fax: +49 531 3918178; e-mail: [email protected] 1 D.M. and K.S. contributed equally to this study.
0925-4773/98/$19.00 1998 Elsevier Science Ireland Ltd. All rights reserved PII S0925-4773 (98 )0 0046-X
in place and form the autochthonous axial musculature and myotomes. Migration seems incompatible with muscle cell differentiation, since only resident cells express myogenic bHLH proteins which are characteristic for muscle cell differentiation (Braun et al., 1992). Migrating limb muscle precursor cells initially do not depend on myogenic factors and express myogenic factors only after they have reached their target (Sassoon et al., 1989; Braun et al., 1994; Tajbakhsh and Buckingham, 1994). Ablation experiments in chicken have demonstrated that cells in the medial half of the dermomyotome from which the epitaxial musculature is derived depend on signals from the neural tube and/or the notochord, while cells from the lateral half which later form the hypaxial musculature do not show such a dependency (Rong et al., 1992; Bober et al., 1994a). In contrast, signals from the forming limb bud appear to be required for the epithelio-mesenchymal transition of dermomyotomal cells in the lateral part of somites, a step absolutely essential for migration of limb muscle precursors into limb buds (Hayashi and Ozawa, 1995). Until
148
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
recently the molecular nature of these signals was enigmatic. During the last years, however, a number of different molecules was shown to play important roles in different processes and interactions required for the formation of the migrating muscle cell precursors (Sonnenberg et al., 1993; Bladt et al., 1995; Hayashi and Ozawa, 1995; Brand-Saberi et al., 1996c; Heymann et al., 1996). The epistatic relationship between different genes has been suggested based on their expression patterns (Daston et al., 1996; Epstein et al., 1996; Yang et al., 1996). However, only relatively few functional studies were performed to reveal the operating mechanisms and networks controlling migration of limb muscle precursors. An initial step appears to be controlled by the Pax-3 gene which is expressed very early during formation of the paraxial mesoderm and later becomes confined to the dermomyotomal layer. Mutation of the Pax-3 gene leads to changes of the morphology of the dermomyotome and to abrogation of muscle precursor cell migration (Bober et al., 1994b; Goulding et al., 1994; Williams and Ordahl, 1994). A second step is defined by effects mediated by the HGF/c-met ligand/receptor system. Knock-out experiments in mice (Bladt et al., 1995) and in ovo manipulations using HGF-soaked beads (Brand-Saberi et al., 1996c; Heymann et al., 1996) suggested that HGF expressed by the developing limb bud initiates the epithelial to mesenchymal transition of the dermomyotomal epithelium required for limb muscle precursor cell migration. Consequently, mutation of either gene results in an inhibition of migration at an early stage. Analysis of splotch mice suggested that c-met might be regulated by Pax-3, since mutation of the Pax-3 gene leads to a downregulation of c-met in somites and limb buds (Daston et al., 1996; Yang et al., 1996). However, reduced levels of c-met were still detectable by RT-PCR in limb buds of mutant mice (Yang et al., 1996). It remains controversial whether expression of c-met and HGF in migrating cells and limb buds is necessary to maintain the mobility of migrating cells or serves as a chemotactic agent to direct targeted migration. Other signals controlling limb muscle precursor cell migration into the limb field and their homing into prospective muscle forming areas are elusive. It appears likely that surface molecules or cell adhesion molecules either alone or together with secreted signal molecules are involved in this process (Lassar and Mu¨nsterberg, 1996). It has been suggested, for example, that N-Cadherin plays a critical role in myoblast migration (Brand-Saberi et al., 1996b). However, a N-Cadherin mutation in Drosophila does not result in changes in muscle cell migration (Masatoshi Takeichi, Kyoto University, pers. commun.) and the knock-out analysis in mice has not been informative since homozygous mutant embryos die before migration starts (Radice et al., 1997). It is clear, however, that somitic cells reach the limbs by active locomotion and that a passive matrix-driven locomotion plays only a minor role if any (Sze et al., 1995). Homeobox-containing genes have been shown to control
body axis and pattern formation within a number of different tissues. Although it is still unclear how homeobox genes exert their role as decisive components in morphogenetic processes, evidence has been accumulated that homeobox genes may control expression of cell surface and cell adhesion molecules and thereby control pattern formation (Edelman and Jones, 1995; Newman, 1996). In addition, homeobox genes have been shown to be involved in determination/differentiation of various cell types, a process which is of paramount importance for the formation of muscle and other organs (Anderson et al., 1997). In the present study we have investigated the role of the lbx1 homeobox-containing gene in the regulation of events controlling limb muscle precursor cell migration in chicken and mice. We show that expression of lbx1 in limb muscle precursor cells is strictly dependent on Pax-3 expression, while expression of c-met in Pax-3 deficient mice is not completely abrogated. Pax-3 appears to be necessary but not sufficient for lbx1 expression, since induction of epithelial to mesenchymal transition of Pax-3-positive cells by HGF does not lead to ectopic lbx-1 expression. Finally, migration of cells from somites to limb buds was studied in an explant culture system in vitro.
2. Materials and methods 2.1. Molecular cloning of mouse and chicken lbx1 A mouse genomic library was screened under reduced stringency using a PflMI restriction fragment of the S59 Drosophila cDNA as a probe (kind gift of M. Frasch, Mount Sinai Medical Center, New York) which contains the homeobox. Hybridizing clones were purified and analyzed using standard procedures (Braun et al., 1989). To generate probes for in situ hybridization appropriate restriction fragments were cloned in pKSII (Stratagene), linearized, and used to produce run-off cRNA transcripts. The chicken lbx1 cDNA clone was isolated by screening 2 × 106 pfu of a chicken limb bud library (kindly provided by William Upholt, University of Conneticut Health Cente) using a 700-bp PCR fragment comprising the mouse homeobox domain. The nucleotide sequence from the longest cDNA clone was determined on both strands by primer walking using an ABI automatic sequencer. All DNA work was performed with standard procedures as outlined in Sambrook et al. (1989). 2.2. Mouse and chicken embryos Splotch1H mice used in this study were obtained from Prof. Thomas Franz (University of Bonn, Germany). Splotch1H mice were backcrossed to C57/Bl6 mice for propagation of the strain (obtained from Harlan Winkelmann, Paderborn, Germany). Homozygous splotch mice were identified by the presence of exencephaly and spina bifida.
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Wild-type embryos (C57/BL6) and Splotch embryos were obtained from timed matings. The morning of the appearance of the vaginal plug was designated E0.5. Fertilized chicken eggs were purchased from Lohmann Bruteier, Cuxhaven, Germany, incubated at 38°C, and staged according to Hamburger and Hamilton (1951). 2.3. Chicken embryo manipulation and application of FGF2 and SF/HGF beads All manipulations were performed as described by Heymann et al. (1996). After windowing of eggs at different developmental stages (Hamburger and Hamilton, 1951) the vitelline membrane overlaying the prospective implantation areas was torn away with a sharpened tungsten needle. The ectoderm covering the lateral plate mesoderm was cut at the appropriate somite level and a bead was carefully positioned inside the slit. Heparin acrylic beads (H5263, Sigma) were washed several times in Ringer solution, selected for size and incubated for 2 h at room temperature in a glass capillary filled with 500 mg/ml FGF-2 (R&D Systems, Minneapolis, MN) or 200 mg/ml SF/HGF (R&D Systems) before use. 2.4. In situ hybridization Non-radioactive whole-mount hybridizations were performed with digoxigenin-labeled cRNA probes as described by Wilkinson (1992). For double-labeling experiments individual templates were transcribed in the presence of digoxigenin-UTP and fluorescin-UTP, respectively. cRNA probes were pooled and used for hybridization following the standard protocol. Embryos were first reacted with alkaline phosphatase coupled to anti-fluorescin antibodies (Boehringer Mannheim, Germany) and stained with fast red (Boehringer Mannheim). Bound antibodies were removed by incubation in glycine and embryos were subjected to a second antibody reaction using alkaline phosphatase coupled anti-digoxigenin antibodies. After staining with NBT and X-phosphate (Boehringer Mannheim), embryos were fixed in 4% paraformaldehyde and stored in phosphate-buffered saline. For sectioning, embryos were embedded in a gelatin/ albumen mixture and cut with a vibratome at 20–100 mm as described previously (Bober et al., 1994b). 2.5. Explant cultures Co-cultivation of lateral parts of somites with limb buds and MRC-5 cell pellets was done essentially as described by Guthrie and Lumsden (1994). HH stage 18/19 embryos were washed thoroughly in Ringer salt solution before sagittal cuts were made to divide somites in a lateral and a medial half using a micro-scalpel. Lateral parts were removed with a micropipette and yellow tips and placed on a drop of gelled collagen (10 ml) in a 4-well petri dish. Limb bud fragments and MRC-5 pellets were added and positioned
149
at a distance of approximately 150 mm relative to somite fragments. After removal of as much liquid as possible, 80 ml of collagen were added. DMEM (1ml) with 10% FCS was added to each well after the collagen gel was set and dishes were incubated at 10% CO2 for the indicated length of time. Collagen was prepared from rat tails following exactly the protocol from Guthrie and Lumsden (1994). MRC-5 cells were obtained from the ETCC. Secretion of SF/HGF from MRC-5 cells was verified by testing MRC-5 tissue culture supernatants for scatter factor activity on MDCK cells as described by Montesano et al. (1991). Pellets of MRC-5 cells were obtained by culture of cells in microtiter plates with a conical-shaped bottom and mechanical mobilization of resulting cell aggregates.
3. Results 3.1. Isolation of mouse and chicken ladybird like (lbx1) genes A low stringency screening of a mouse genomic library was performed with a probe derived from the Drosophila homeobox gene S59 which is expressed in muscle and neuronal precursor cells of flies (Dohrmann et al., 1990). Isolated genes were hybridized at increasing stringencies with the S59 probe and grouped according to their hybridization strength. cRNA probes were synthesized from various clones of each group and used to hybridize mouse embryos between E9.5 and E11.5. Among the clones which generated discrete hybridization signals one yielded a particularly interesting expression pattern in somites, limb buds, and the CNS. Sequence analysis revealed that this gene represented the mouse homolog of the Drosophila homeobox gene ladybird which was first isolated and described in mouse and human by Jagla et al. (1995) and named lbx1. We next isolated the chicken homologue of lbx1 to compare expression patterns in both species and to benefit from the chicken experimental system with its possibility for micromanipulation and local (over)expression. The homeodomain of the mouse lbx1 gene was isolated and used to screen a chicken limb bud library at low stringency. Among many other known or novel members of the homeobox gene family, we identified a set of clones which appeared to represent the chicken ortholog of the mouse and human genes. The longest of these clones contained an insert of 1254 bp comprising most of the coding region of the mRNA and the 3′-untranslated region. Comparison of the deduced peptide sequence revealed 100% identity to the mouse and 98% identity to human proteins in the homeobox domain; the N-terminal domain is 82% identical to the mouse and 81% identical to the human sequence. In contrast, the Cterminal domains are highly divergent with the exception of isolated amino acid stretches immediately adjacent to the homeobox domain and at the very end of the protein (Fig. 1). Interestingly, in the N-terminal region the highest degree
150
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Fig. 1. Sequence comparison of chicken, mouse and human lbx1 proteins. Deduced amino acid sequences were compared using the ClustalV program. Amino acids conserved between chicken, mouse and human are boxed.The homeodomain is indicated by a blue frame. Chicken and mouse homeoboxes are identical. The human homeobox shows only one conservative exchange. The N-terminal domain of all three proteins has a relatively high degree of identity of >80% while in the C-terminal domain only scattered blocks of amino acids are conserved between chicken and mouse or human. In contrast, identity of human and mouse sequences to each other is much higher in the C-terminal domain.
of similarity is also found directly adjacent to the homeobox. 3.2. Chicken lbx1 is highly expressed in limb muscle precursor cells originating from lateral tips of somites We first investigated the expression profile of lbx1 in developing chicken embryos (Fig. 2). A prominent lbx1 expression domain was detected in somites at the level of developing limb buds (Fig. 2A–C). Interestingly, the onset of lbx1 expression correlated with the outgrowth of limb buds at HH stage 17/18. Vibratome sections were prepared from stained embryos and revealed that lbx1 expression in somites was confined to the ventro-lateral tip of the dermomyotome. Cells in this area express high levels of Pax-3 mRNA, while another Pax-3 expression domain in dorsomedial tips of the dermomyotome was negative for lbx1. Lbx1 was barely detectable in somites of the inter-limb region of chicken embryos and caudal to the leg buds. In a more rostral position, lbx1 expression was observed cranial to wing buds in the lateral tips of dermomyotomes. We also
noticed a characteristic streak of lbx1-positive cells which originated from cervical somites below the branchial arches. It is interesting to note that cells of this structure which most probably represent tongue muscle precursor cells also express Pax-3 at a high level. At later stages (HH stage 19/20) when migrating muscle precursor cells leave somites, lbx1 expression was clearly visible in the migrating cell population (indicated by arrows in Fig. 2C,F), concomitant with the migration of dermomyotomal cells from somites into limb buds lbx1 expression decreased in the dermomyotomal cell layer and appeared in a proximal-distal gradient within limbs and wings. After completion of the migration process, lbx1-positive cells were found only in the prospective muscle forming areas of limbs (Fig. 2G–I), whereas the somites at limb level had lost all lbx1 staining. In mice a relatively strong expression of lbx1 was seen in the neural tube and in some areas of the brain stem. In chicken these expression domains were only detectable after prolonged staining, indicating a relatively low expression level compared to limb muscle precursor cells. Our
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
151
Fig. 2. Expression of lbx1 in chicken embryos. Stage 18 (A–C), stage 20 (D–F), and stage 22 (G–I) chicken embryos were subjected to whole mount in situ hybridization using a DIG-UTP-labelled lbx1 probe. Stained embryos were sectioned on a vibratome. At stage 18, lbx1 transcripts were detected in the lateral edge of dermomyotomes adjacent to wing and leg buds and at lower amounts in cranial somites (A–C). At HH stage 20, lbx1 expression becomes visible in the wing/leg buds (D–F) and at HH stage 22 all lbx1-positive cells have moved from somites to wing/leg buds (G–I). At this stage a weak expression domain in the neural tube is also visible. Arrows indicate migrating limb muscle precursor cells. dm, dermomyotome; nt, neural tube; lb, leg bud.
analysis of lbx1 expression in mouse embryos (data not shown) did not reveal any differences to results previously reported by Jagla et al. (1995). 3.3. Complete absence of lbx1 but not c-met expression in Pax-3 deficient mice The apparent expression of lbx1 in mouse limb muscle precursor cells prompted us to study lbx1 in Pax-3 mutant mice which have a severe deficit in limb muscle formation. In addition, we analyzed the expression of c-met, another marker for migrating limb muscle precursor cells. We found that expression of lbx1 was absent in somites and limbs of Pax-3 mutant mice between E9.5 and E11.5 (Fig. 3). Despite the complete lack of lbx1 transcripts in somites and cells derived thereof, expression of lbx1 in the CNS of splotch mice was only slightly reduced or was not affected at all. Between E9.5 and E10.5, lbx1 expression appeared to be weaker in neural tubes of splotch mice
(Fig. 3D). At present it is unclear whether the lbx1 decrease here is due to the absence of Pax-3, since reduction of lbx1 appeared not restricted to the expression domain shared with Pax-3. In addition, at E11.5 no reduction of lbx1 expression was detected in the neural tube of splotch mice (Fig. 3F,H). Comparison of sections of wild-type and splotch embryos at E11.5 hybridized with the lbx1 probe did not reveal significant differences in this tissue, while lbx1 expression was completely absent in somites and limb buds of splotch mutant mice (Fig. 3G,H). Therefore, a lower expression level of lbx1 in the spinal cord in splotch mice at early stages compared to wild-type might be due to a general retardation of spinal cord development in mutants rather than a specific regulatory effect of Pax-3 on lbx1 expression. In contrast to the complete lack of lbx1 in somites and limb buds of splotch mice, expression of c-met was not entirely abolished. Although expression levels were reduced and c-met expressing cells in somites appeared disarranged, expression of c-met was clearly detectable in somites and
152
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Fig. 3. Absence of lbx1 expression in somites and limb buds but not in the neural tube of splotch mutant mice. E9.5 (A,B), E 10.5 (C,D), and E11.5 (E–H) wild-type (A,C,E,G) and splotch mutant mice (B,E,F,H) were hybridized with a DIG-UTP labelled lbx1 probe. Either whole-mount preparations (A,B,E,F) or vibratome sections (C,D,G,H) of whole-mount hybridizations are shown. lbx1 expression is absent in somites and limb buds of splotch mutant mice from E9.5 to E11.5. Expression of lbx1 in the neural tube becomes most prominent at E11.5. At this stage lbx1 appears to be expressed normally in neural tubes of splotch mutant mice (F,H) while no lbx1 signals are visible in limbs (indicated by arrows in F,H). The normal expression of lbx1 in the neural tube of splotch mice but the absence in somites and limb buds is particularity evident on the section shown in (H).
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
limb buds of splotch mice at E10.5 (Fig. 4A). At the forelimb level, relatively strong c-met expression was detected (Fig. 4C). This expression domain did not easily compare to the dorso-medial or the ventro-lateral domains normally seen in wild-type mice. The observed differences in c-met distribution might be a result of morphological disturbance of somites in splotch mice where no intact epithelium is seen at the normal tips of the dermomyotome, which results in a shortening of the dermomyotomal cell layer (Daston et al., 1996). This compression of the dermomyotome might account for the aberrant location of c-met expressing cells in the dermomyotome at the forelimb level. At the hindlimb level c-met signals were weaker and more diffuse, while in the tail region virtually normal levels of c-met mRNA were reached (Fig. 4A). It should also be mentioned that c-met signals in somites of splotch mutants decreased when the limbs were forming similar to wild-type mice where this decrease corresponds to the migration of c-met-positive cells into the limbs (data not shown). To evaluate the relationship between cells expressing mutant Pax-3 mRNA and cells expressing c-met in splotch embryos, double-label in situ hybridizations were performed. As shown in Fig. 4D and E, cells which expressed mutant Pax-3 mRNA (dark blue staining in Fig. 4D,E) were immobilized in the dermomyotome. In contrast, c-met-positive cells (red staining in Fig. 4D,E) left the dermomyotome and were found along the way into the limb and at the dorso-medial edge of the somites. An additional expression domain of c-met which was not dependent on Pax-3 was visible in mesonephric tubules (Fig. 4E; Woolf et al., 1995). To investigate whether cmet-positive/Pax-3-negative cells in the limb can give rise to differentiated muscle cells, limb buds were excised and cultured in vitro. This procedure allows an extended development of limbs which is not possible in utero due to the early lethality of splotch mice. Despite the presence of c-met-positive cells in limbs of splotch mice, no differentiated myogenic cells were found in these limbs even after extended culture. MyHC-positive cells were present only in the most proximal parts of the limb, most probably representing shoulder muscles which are not affected by the Pax-3 mutation (data not shown). Therefore, it appears that Pax-3 has a Fig. 4. Expression of c-met is reduced and distorted but not absent in splotch mutant mice. E10.5 splotch mutant (A,C–E) and wild-type mice (B) were subjected to whole-mount in situ hybridization with a c-met probe (A–C) and for double label in situ hybridizations with a c-met and a Pax-3 probe (D,E). Localization of Pax-3 transcripts in double label hybridizations (D,E) is indicated by a dark blue precipitate, presence of c-met transcripts is indicated by a red precipitate. Whole-mount preparations (A) and vibratome sections (B–E) are shown. The presence of c-metpositive cells in somites and fore-limb is clearly visible in (A). In (B,C) a comparison of c-met expression in thoracic somites of wild-type (B) and splotch mutant mice (C) is shown. Note that in splotch mutant mice (D,E), Pax-3-expressing cells do not leave somites (dark blue staining indicated by red arrowheads) while c-met-positive cells (red staining indicated by black arrows) are found between the dermomyotome and the limb and in a distal position within the limb. LB, limb bud; NT, neural tube; DM, dermomyotome; oNT, open neural tube in splotch mice.
153
dual role both in the initiation of cell migration as well as in differentiation of migrating limb muscle precursor cells. 3.4. Delamination of Pax-3-positive cells from the dermomyotome in the inter-limb region is not sufficient to up-regulate lbx1 expression The absence of lbx1 expression in somites of splotch mice suggests that Pax-3 is a critical component needed to
154
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
activate lbx1 transcription. The lack of lbx1 expression in inter-limb somites which also express high amounts of Pax3 in the dermomyotome, however, indicates that Pax-3 alone is not sufficient for lbx1 expression. Since lbx1 is expressed predominantly in somites adjacent to limb buds, it is reasonable to assume that signals from the limb might affect lbx1 expression. SF/HGF is so far the only known secreted signal from the limb mesenchyme which controls decisions of the dermomyotomal epithelium, such as epithelial to mesenchymal transition (Heymann et al., 1996). We therefore investigated whether initiation of migration caused by exogenous SF/HGF together with the presence of Pax-3 would be sufficient to instruct lbx1 expression or whether additional factors are necessary. To address this question, we implanted heparin beads soaked either in SF/ HGF or in FGF-2 in ovo into the flank region of HH stage 15–17 embryos (n = 20). Manipulated embryos were returned to the incubator and allowed to develop. After 24 and 48 h, embryos were isolated and subjected to in situ hybridization with a lbx1 probe. As shown in Fig. 5A–C, no lbx1 expression was detected at the inter-limb level close to implanted SF/HGF beads, while strong expression was found in somites adjacent to limb buds. To ensure that implantation of SF/HGF beads initiated an epithelial to mesenchymal transition of dermomyotomal cells at the inter-limb level, embryos from the same series of experiments were hybridized with the Pax-3 probe. All embryos carrying a SF/HGF bead showed Pax-3-positive cells which delaminated from the dermomyotome and started to migrate (Fig. 5D,E).
The lack of lbx1 up-regulation in delaminating Pax-3positive dermomyotomal cells at the inter-limb level indicated that newly forming limb buds may contribute activities in addition to SF/HGF to induce lbx1 expression in Pax-3-positive cells and activate targeted migration. Therefore, we implanted FGF-2 coated heparin beads into the flank region of HH stage 15–17 chicken embryos (Cohn et al., 1995) and analyzed the expression of lbx1 in ectopically forming limbs (Fig. 5F). As expected, lbx1 expressing migrating cells were clearly detectable in newly formed limbs. As shown previously for Pax-3 expression, these cells were derived from somites directly adjacent to newly formed limbs (Heymann et al., 1996). It appears likely that FGF-mediated activation of lbx1 expression is an indirect process, since lbx1 expression did not occur in the immediate vicinity of FGF coated beads (Fig. 5F). 3.5. Limb buds but not HGF secreting MRC-5 cells are able to stimulate cell migration from lateral somite fragments in explant cultures To further study the relationship between targeted migration and SF/HGF-mediated delamination of dermomyotomal cells, lateral parts of somites of HH stage 17–19 chicken embryos were isolated and co-cultured with limb bud fragments of HH stage 18/19 embryos or with aggregates of SF/HGF expressing MRC-5 lung fibroblasts in collagen matrix gels at a distance between 150 and 350 mm. As expected, no cells were found between both tissue fragments derived from lateral somite halves and limb buds,
Fig. 5. Expression of lbx1 in chicken embryos after implantation of FGF- and HGF-coated heparin beads. Beads soaked in 200 mg/ml SF/HGF (A–E) or 500 mg/ml FGF-2 (F) were implanted into flanks of HH stage 16 chicken embryos and hybridized with a lbx-1 probe (A–C,F) and a Pax-3 probe (D,E) after 24 h of incubation. Despite induction of epithelial to mesenchymal transition of Pax-3-positive dermomyotomal cells by SF/HGF and initiation of migration (indicated by arrowheads in D,E), no induction of lbx1 expression is seen in the inter-limb region (A–C). In contrast, induction of an ectopic limb in the interlimb region by FGF-2 induces lbx1 expression and migration of lbx1-positive cells into the additional limb bud (indicated by arrowheads in F).
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
155
Fig. 6. Explant co-cultures of lateral somite halves with limb bud fragments and with SF/HGF-expressing MRC-5 cell aggregates in collagen gels. Lateral halves of somites were co-cultured with limb bud fragments (A,B) and with MRC-5 cell aggregates for 6 h (A,C) or for 3 days (B,D) in collagen gels. No cells were found between tissue fragments derived from lateral somite halves and limb buds (A) or MRC-5 cell aggregates (C) after 6 h. After 3 days of incubation, cells were found only between limb bud and lateral somite fragments (B) but not between MRC-5 cell aggregates and lateral somite halves (D). lb, limb bud; so, lateral somite half; M-5, MRC-5 cell aggregates.
or MRC-5 cell aggregates shortly after explantation (Fig. 6A,C). However, upon further incubation a cell bridge formed which spanned the gap between somitic and limb bud fragments, while no cells filling the gap between somitic fragment and aggregates of HGF expressing cells were detected (Fig. 6B,D). After extended culture periods in collagen gels we did observe a radial cell outgrowth from explants which partially obscured the picture after 8 days in culture. But clearly, the directional outgrowth from the somitic fragment was not seen with MRC-5 cells. Immunohistochemical staining of collagen cultures with the monoclonal antibody MF-20 against MyHC confirmed that, concomitant with ongoing migration, the number of MyHC-positive cells within limb bud explants increased (data not shown). Additional control experiments in which tissue fragments from the head or the heart of chicken embryos were cultured together with lateral halves of somites did not result in a directed outgrowth or cell migration. Only radial outgrowth from tissue fragments was observed. Reduction of the distance below 100 mm between somitic and limb bud fragments or aggregates of MRC-5 cells did not increase the amount of cells found between explants (data not shown). In contrast, when the distance between explants was extended beyond 600 mm, no directed movement of cells was observed. These results indicate that the range of limb-derived signals is restricted in distance. Interestingly, the limited range of limb-derived signals seems to match the restricted expression of lbx1 in the 4–5 somites located directly adjacent to limb buds.
4. Discussion Development of limb muscles can be divided in several steps including commitment of somitic cells, epithelial to mesenchymal transition of the dermomyotomal epithelium, migration of limb muscle precursor cells into limb buds and determination of myoblasts. Only recently some of these steps have been addressed at the molecular level. In the present study we have characterized the homeobox gene lbx1 in chicken and mice and analyzed cell migration into limb buds by various means. We found that lbx1 is expressed at all stages of limb muscle precursor cell development, which suggests that lbx1 may be an important part of the machinery controlling limb muscle precursor cell development and migration. Expression of lbx1 appeared to be dependent on several distinct mechanisms. (a) Pax-3 expression: In the lateral parts of somites and in migrating limb muscle precursor cells the expression of lbx1 depends on Pax-3. Lack of Pax-3 in splotch mutants completely abolished lbx1 expression in these structure. However, Pax-3 is not sufficient to drive lbx-1 expression since lbx-1 is not expressed in the inter-limb region of chicken embryos where Pax-3 is present at the same concentration as in the limb regions. (b) Neural system specific factors: At E10.5 of mouse development, expression of lbx1 is reduced but not absent in the neural tubes of splotch mice. At E11.5, virtually normal levels of lbx1 RNA were found in the neural tubes of mutants. These results clearly demonstrate that additional factors are present in the neural tube which can activate lbx1 expression.
156
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
Whether Pax-3 is important for high-level lbx1 expression at early stages or whether a reduced lbx1 expression is due to a general retardation of neural tube development in splotch mice is hard to distinguish and may require analysis of additional marker molecules which specifically reflect the stage of neural tube development. (c) Limb bud-derived factors: The formation of an ectopic limb in the inter-limb region leads to initiation of lbx-1 expression, indicating that limb buds contain signals required for lbx-1 activation. This activity is different from SF/HGF, since SF/HGF-coated beads did not induce lbx1 when implanted into the interlimb region despite the delamination of Pax-3-positive dermomyotomal cells in this area. From our experiments it is clear that during limb bud formation, a process is initiated which acts back on somites to induce lbx1 expression. This may contribute to the commitment of muscle precursor cells and to their migration pattern. It is also clear that this activity is distinct from SF/HGF, although it cannot be excluded that SF/HGF is a part of this signal. The existence of an additional signaling pathway, necessary for lbx1 expression and targeted cell migration, was further supported by our explantation experiments. Only in explant cultures where lateral halves of somites were co-cultivated with limb bud fragments within a distance of 150–350 mm was a targeted migration of cells apparent. The limited range of putative signals in explant cultures may correspond to the restriction of lbx1 expression to somites which are located adjacent to limb buds. According to this model, the lack of lbx1 expression in the inter-limb region and caudal to the hind-limb may result from the absence or limited range of inducing substances supplied by developing limb buds. In contrast to the complete absence of lbx1 in somites of splotch mice, expression of c-met was only reduced and the distribution of c-met transcripts was disarranged. We also noted differences in the expression profile between cranial and caudal somites and between forelimbs and hindlimbs. A more localized expression was found in forelimbs, while in hindlimbs c-met-expressing cells appeared to have lost their orientation and were scattered in the body wall and the hindlimb bud. Cranial to caudal differences in the expression of c-met in splotch mice as well as the disseminated distribution of c-met-positive cells, a general reduction of the expression level, and analysis of only some developmental stages may have contributed to the failure to detect c-met expression in splotch mice by whole-mount hybridization (Daston et al., 1996; Yang et al., 1996). However, using RT-PCR Yang et al. (1996) were able to detect c-met mRNA in limbs of splotch mice which is in line with our findings. Using radioactive in situ hybridization on sections, Epstein et al. (1996) reported a deficiency of c-met expression in the lateral dermomyotome of splotch mice at E11.5 but a relatively strong expression in the dorsal forelimb which was attributed to the distal ectoderm. In contrast, we have observed a disarranged expression of c-met in the dermomyotome of splotch1H mice at E10.5 which starts to
disappear as cells migrate into the limb bud. Although we agree with Epstein et al. (1996) that c-met is expressed in limbs of splotch mutants, analysis of our single and double whole-mount in situ hybridizations does not suggest an expression of c-met in the ectoderm of limbs. It is hard to pinpoint the nature of c-met-positive cells in the limb of Pax-3 mutant mice, since these cells do not differentiate into muscle cells as indicated by lack of MyHC expression, nor do they express other known differentiation markers. A number of different explanations can be given to account for the presence of the c-met-positive cell population. First, the expression of c-met in the limb of splotch mice may be not related to any migrating cell population, in particular myogenic precursor cells, but may occur de novo in the limb mesenchyme. Second, c-met may be expressed in a population of cells which only partially overlaps with Pax-3-positive cells and which is not myogenic. Angioblasts, for example, migrate from the dorso-lateral quadrant of the somite into the wing field and the ventro-lateral body wall and may represent such cells (Wilting et al., 1995). Third, c-met expressing cells may represent a sub-population of myogenic precursor cells which are not dependent on Pax-3 for migration but for differentiation. Although none of these hypotheses can be completely excluded, the first explanation appears most unlikely. Bladt et al. (1995) have not observed cells which express a mutant c-met transcript in limbs of c-met knock-out mice despite the preservation of other c-met expression domains in these mice. This finding clearly indicates that c-met expression in limbs is derived from migratory cells. In addition, we have observed a continuous layer of c-met expressing cells located between the ventro-lateral edge of the somite and the limb bud of splotch mutant mice. This staining pattern is characteristic for migrating cells and suggests that c-met cells in the limb of splotch mice originate from somites. While the migratory character of c-met-positive cells located between somites and the limb bud and within the limb buds of splotch mice appears very likely, it is much harder to decide whether these cells represent non-myogenic cells, such as angioblasts, or myogenic cells which cannot longer differentiate into muscle because of the absence of Pax-3. The lack of any differentiated muscle cell in c-met containing limbs of splotch mice may favor the view that these cells represent non-myogenic, migratory cells. However, it has been shown recently that Pax-3 can activate the myogenic program via MyoD and Myf-5 under certain conditions (Maroto et al., 1997; Tajbakhsh et al., 1997). Therefore, it cannot be excluded that c-met-positive/Pax3-negative cells in the limbs of splotch mice represent myogenic precursor cells which can no longer differentiate due to the absence of Pax-3. It appears likely that differentiation of most limb muscle cells is independent of Pax-3, since the majority of cells transplanted from lateral halves of somites of splotch mice into limb buds of chicken embryos differentiate into muscle (Daston et al., 1996). Nevertheless, it is
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
possible that Pax-3 is necessary for the differentiation of only a subset of cells. Failure of a subpopulation of cells to differentiate may be easily obscured by other differentiating cells and prevent the detection of cell-autonomous muscle defects in Pax-3 mutant mice. The presence of lbx1 in Pax-3-positive cells makes it also difficult to distinguish, without a mutation of lbx1 being available, whether both molecules are necessary for limb muscle differentiation or whether Pax-3 alone is sufficient for muscle cell determination. Although it is possible that c-met + /Pax-3 − cells are not myogenic but rather represent another cell type, this does not necessarily mean that these cells are not involved in the migration process or the pattern formation of muscle cells. For example, c-met + /Pax-3 − cells may serve as pioneer cells during migration as indicated by their location proximal to Pax-3-positive cells. Such a hypothetical function does not require that c-met + /Pax-3 − cells become muscle. Future research, in particular the careful analysis of the c-met + /Pax-3 − cells, will probably disclose their identity and their role in the migration of cells from somites to limbs. In addition, the generation of lbx1 null mutants will reveal whether lbx1 is only a useful marker molecule for migrating limb muscle precursor cells or whether it fulfils a critical role in the specification of the migrating cell population and in the pathway, leading to an arrangement of muscle precursor cells within their future locations in the limb. Acknowledgements The expert technical assistance by Stefanie Willenzon is gratefully acknowledged. We would like to thank Peter Gruss for chicken Pax-3 cDNA, Manfred Frasch for the Drosophila S59 probe, Carmen Birchmeier for the mouse c-met probe and William Upholt for the chicken limb bud cDNA library. We are indebted to Hans-Henning Arnold for continuous support, encouragement and improvements to the manuscript, Eva Bober and Markus Ko¨ster for helpful discussions and critically reading the manuscript. This work was supported by the SFB 271: ‘Molekulare Mechanismen morphoregulatorischer Prozesse’ project B2 to T.B. and by the ‘Fonds der Chemischen Industrie’.
References Anderson, S.A., Qiu, M., Bulfone, A., Eisenstat, D.D., Meneses, J., Pedersen, R., Rubenstein, J.L., 1997. Mutations of the homeobox genes Dlx-1 and Dlx-2 disrupt the striatal subventricular zone and differentiation of late born striatal neurons. Neuron 19, 27–37. Bladt, F., Rietmacher, D., Isenmann, S., Aguzzi, A., Birchmeier, C., 1995. Essential role for the c-met receptor in the migration of myogenic precursor cells into the limb bud. Nature 376, 768–771. Bober, E., Brand-Saberi, B., Ebensperger, C., Wilting, J., Balling, R., Paterson, B.M., Arnold, H.H., Christ, B., 1994a. Initial steps of myogenesis in somites are independent of influence from axial structures. Development 120, 3073–3082.
157
Bober, E., Franz, T., Arnold, H.-H., Gruss, P., Tremblay, P., 1994b. Pax-3 is required for the development of limb muscles: a possible role for the migration of dermomyotomal muscle progenitor cells. Development 120, 603–612. Brand-Saberi, B., Wilting, J., Ebensperger, C., Christ, B., 1996. The formation of somite compartments in the avian embryo. Int. J. Dev. Biol. 40, 411–420. Brand-Saberi, B., Gamel, A.J., Krenn, V., Muller, T.S., Wilting, J., Christ, B., 1996. N-Cadherin is involved in myoblast migration and muscle differentiation in the avian limb bud. Dev. Biol. 178, 160–173. Brand-Saberi, B., Muller, T.S., Wilting, J., Christ, B., Birchmeier, C., 1996. Scatter factor/hepatocyte growth factor (SF/HGF) induces emigration of myogenic cells at interlimb level in vivo. Dev. Biol. 179, 303–308. Braun, T., Buschhausen, D.G., Bober, E., Tannich, E., Arnold, H.-H., 1989. A novel human muscle factor related to but distinct from MyoD1 induces myogenic conversion in 10T1/2 fibroblasts. EMBO J. 8, 701–709. Braun, T., Rudnicki, M.A., Arnold, H.-H., Jaenisch, R., 1992. Targeted inactivation of the muscle regulatory gene Myf-5 results in abnormal rib development and perinatal death. Cell 71, 369–382. Braun, T., Bober, E., Rudnicki, M.A., Jaenisch, R., Arnold, H.H., 1994. MyoD expression marks the onset of skeletal myogenesis in Myf-5 mutant mice. Development 120, 3083–3092. Chevallier, A., Kieny, M., Mauger, A., 1977. Limb-somite relationship: origin of the limb musculature. J. Embryol. Exp. Morphol. 41, 245– 258. Christ, B., Jacob, H., Jacob, B., 1978. On the formation of the myotomes in avian embryos. An experimental and scanning electron microscope study. Experientia 34, 514–516. Cohn, M.J., Izpisua-Belmonte, J.C., Abud, H., Heath, J.K., Tickle, C., 1995. Fibroblast growth factors induce additional limb development from the flank of chick embryos. Cell 80, 739–746. Daston, G., Lamar, E., Olivier, M., Goulding, M., 1996. Pax-3 is necessary for migration but not differentiation of limb muscle precursors in the mouse. Development 122, 1017–1027. Dohrmann, C., Azpiazu, N., Frasch, M., 1990. A new Drosophila homeobox gene is expressed in mesodermal precursor cells of distinct muscles during embryogenesis. Genes Dev. 4, 2098–2111. Edelman, G.M., Jones, F.S., 1995. Developmental control of N-CAM expression by Hox and Pax gene products. Phil. Trans. R. Soc. Lond. B Biol. Sci. 29, 305–312. Epstein, J.A., Shapiro, D.N., Cheng, J., Lam, P.Y.P., Maas, R.L., 1996. Pax3 modulates expression of the c-met receptor during limb muscle development. Proc. Natl. Acad. Sci. USA 93, 4213–4218. Goulding, M., Lumsden, A., Paquette, A.J., 1994. Regulation of Pax-3 expression in the dermomyotome and its role in muscle development. Development 120, 957–971. Guthrie, S., Lumsden, A., 1994. Collagen gel coculture of neural tissue. Neuroprotocols 4, 116–120. Hamburger, V., Hamilton, H., 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49–92. Hayashi, K., Ozawa, E., 1995. Myogenic cell migration from somites is induced by tissue contact with medial region of the presumptive limb mesoderm in chicken. Development 121, 661–669. Heymann, S., Koudrova, M., Arnold, H.-H., Koster, M., Braun, T., 1996. Regulation and function of SF/HGF during migration of limb muscle precursor cells in chicken. Dev. Biol. 180, 566–578. Jagla, K., Dolle, P., Mattei, M.G., Jagla, T., Schuhbaur, B., Dretzen, G., Bellard, F., Bellard, M., 1995. Mouse Lbx1 and human LBX1 define a novel mammalian homeobox gene family related to the Drosophila lady bird genes. Mech. Dev. 53, 345–356. Newman, S.A., 1996. Sticky fingers: Hox genes and cell adhesion in vertebrate limb development. Bioessays 18, 171–174. Lassar, A.B., Mu¨nsterberg, A.E., 1996. The role of positive and negative signals in somite patterning. Curr. Opin. Neurobiol. 6, 57–63. Maroto, M., Reshef, R., Munsterberg, A.E., Koester, S., Goulding,
158
D. Mennerich et al. / Mechanisms of Development 73 (1998) 147–158
M., Lassar, A.B., 1997. Ectopic Pax-3 activates MyoD and Myf-5 expression in embryonic mesoderm and neural tissue. Cell 89, 139– 148. Montesano, R., Matsumoto, K., Nakamura, T., Orci, L., 1991. Identification of fibroblast-derived epithelial morphogen as hepatocyte growth factor. Cell 67, 901–908. Ordahl, C.P., LeDouarin, N., 1992. Two myogenic cell lineages within the developing somite. Development 114, 339–353. Radice, G.L., Rayburn, H., Matsunami, H., Knudsen, K.A., Takeichi, M., Hynes, R.O., 1997. Developmental defects in mouse embryos lacking N-cadherin. Dev. Biol. 181, 64–78. Rong, P.M., Teillet, M.A., Ziller, C., Le Douarin, N.M., 1992. The neural tube/notochord complex is necessary for vertebral but not limb and body wall striated muscle differentiation. Development 115, 657–672. Sassoon, D., Lyons, G., Wright, W.E., Lin, V., Lassar, A., Weintraub, H., Buckingham, M., 1989. Expression of two myogenic regulatory factors myogenin and MyoD1 during mouse embryogenesis. Nature 341, 303– 307. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1659 pp. Sonnenberg, E., Meyer, D., Weidner, K.M., Birchmeier, C., 1993. Scatter factor/Hepatic growth factor and its receptor, the c-met tyrosine kinase, can mediate a signal exchange between mesenchyme and epithelia during mouse development. J. Cell Biol. 123, 223–235.
Sze, L.-Y., Lee, K.K.H., Webb, S.E., Li, Z., Pauline, D., 1995. Migration of myogenic cells from the somites to the fore-limb buds of developing mouse embryos. Dev. Dyn. 203, 324–336. Tajbakhsh, S., Buckingham, M., 1994. Mouse limb muscle is determined in the absence of the earliest myogenic factor myf-5. Proc. Natl. Acad. Sci. USA 91, 747–751. Tajbakhsh, S., Rocancourt, D., Cossu, G., Buckingham, M., 1997. Redefining the genetic hierarchies controlling skeletal myogenesis: Pax-3 and Myf-5 act upstream of MyoD. Cell 89, 127–138. Yang, X.-M., Vogan, K., Gros, P., Park, M., 1996. Expression of the met receptor tyrosine kinase in muscle progenitor cells in somites and limbs is absent in Splotch mice. Development 122, 2163–2171. Wilting, J., Brand-Saberi, B., Huang, R., Zhi, Q., Ko¨ntges, G., Ordahl, C.P., Christ, B., 1995. Angiogenic potential of the avian somite. Dev. Dyn. 202, 165–171. Wilkinson, D.G., 1992. Whole mount in situ hybridization of vertebrate embryos, in: D.G. Wilkinson (Ed.), In Situ Hybridization: A Practical Approach, IRL Press, Oxford, pp. 75–83. Williams, B.A., Ordahl, C.P., 1994. Pax-3 expression in segmental mesoderm marks the early stages in myogenic cell specification. Development 120, 785–796. Woolf, A.S., Kolatsi-Joannou, M., Hardman, P., Andermarcher, E., Moorby, C., Fine, L.G., Jat, P.S., Noble, M.D., Gherardi, E., 1995. Roles of hepatocyte growth factor and the met receptor in the early development of the metanephros. J. Cell Biol. 128, 171–184.