BMPs restrict the position of premuscle masses in the limb buds by influencing Tcf4 expression

Developmental Biology 299 (2006) 330 – 344 www.elsevier.com/locate/ydbio

BMPs restrict the position of premuscle masses in the limb buds by influencing Tcf4 expression Alexander Bonafede 1 , Thomas Köhler 1 , Marc Rodriguez-Niedenführ 2 , Beate Brand-Saberi ⁎ Institute of Anatomy and Cell Biology II, University of Freiburg, PO Box 111, D-79001 Freiburg, Germany Received for publication 17 January 2005; revised 22 December 2005; accepted 6 February 2006 Available online 15 June 2006

Abstract Previous studies have shown the distally retreating source of Scatter factor/Hepatocyte growth factor (SF/HGF) can account for the distal migration of myogenic precursor cells in the limb bud mesenchyme. However, the normal expression pattern of Sf/Hgf alone does not explain the distribution of muscle precursor cells. Hence, the position of the dorsal and ventral premuscle masses suggests the presence of additional patterning factors. We present evidence that BMP2 and 4 can act as such factors by inhibiting the expression of Tcf4, a downstream element of the canonical Wnt pathway. The normal position of muscle cells depends on the correct distribution of BMP and SF/HGF throughout the limb bud mesenchyme. Removal or inhibition of the BMP signals within the limb margins leads to a shift in position resulting in the fusion of the dorsal and ventral premuscle masses towards the manipulated areas. In the absence of BMPs, mispositioning requires the presence of SF/HGF. Consequently, ectopic application of exogenous SF/HGF in the presence of BMP signals does not change muscle positioning. We conclude that correct positioning of the premuscle masses in the limb buds is controlled by the combined influence of SF/HGF signals – guiding cells mainly in the proximo-distal axis – and BMP signals that restrict the positioning to the dorsal and ventral central portions of the limb buds. © 2006 Elsevier Inc. All rights reserved. Keywords: BMP; TCF4; SF/HGF; Cell migration; Limb muscle

Introduction Limb muscle precursor cells, a subpopulation of hypaxial muscles, detach from the lateral dermomyotomal lips at fore and hindlimb levels and actively migrate towards the limb buds (Christ et al., 1974, 1977; Chevallier et al., 1977; Jacob et al., 1979). It has been shown that delamination of the muscle precursor cells results from interaction of SF/HGF secreted by the mesenchymal cells of the lateral plate at limb bud level and its receptor cMet on the premyogenic cells in the dermomyotomes (Bladt et al., 1995; Brand-Saberi et al., 1996a,b; Heymann et al., 1996). cMet is a tyrosine kinase receptor specific for SF/ ⁎ Corresponding author. Institut für Anatomie und Zellbiologie, Lehrstuhl II, Albertstrasse 17, 79104 Freiburg, Germany. Fax: +49 761 203 53 87. E-mail address: [email protected] (B. Brand-Saberi). 1 These authors have contributed equally to this study. 2 Present address: Department of Craniofacial Development, King's College, London SE1 9RT, UK. 0012-1606/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2006.02.054

HGF (Sonnenberg et al., 1993; Thery et al., 1995; Heymann et al., 1996; Matsumoto and Nakamura, 1996; Yang et al., 1996). Transcription factors expressed in the muscle precursor cells, especially Lbx1, Mox-2 and Pax3 are crucial for correct limb muscle distribution (Goulding et al., 1994; Daston et al., 1996; Epstein et al., 1996; Mankoo et al., 1999; Schäfer and Braun, 1999; Brohmann et al., 2000; Gross et al., 2000; Uchiyama et al., 2000; Mennerich and Braun, 2001). Among them, the homeobox-containing gene, Lbx1 is the only one to be expressed exclusively in hypaxial migratory muscle precursors and to mark the entire migratory subpopulation of somite-derived muscle precursor cells (Jagla et al., 1995; Dietrich et al., 1998, 1999; Mennerich et al., 1998). Once myogenic precursors cells reach the limb buds, they segregate into a clearly separated dorsal and ventral stream of cells that will eventually form the dorsal and ventral premuscle masses (Schramm and Solursh, 1990). During further limb bud development, myogenic cells continue to migrate distally and SF/HGF has been shown to have a dual role in this process: it

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

enables further migration and at the same time inhibits the differentiation of the migrating cells by inhibiting MyoD expression (Dietrich et al., 1999; Scaal et al., 1999). This argues for a close link between myogenic precursor cell migration, positioning of premuscle masses and muscle differentiation (Kardon, 1998; Christ and Brand-Saberi, 2002). The muscle pattern in the limbs is controlled by the stationary mesenchyme (Birchmeier and Brohmann, 2000; Brand-Saberi et al., 1996b). There has been some progress concerning molecular players involved in this process. Apart from SF/HGF, cadherin-mediated interactions (Brand-Saberi et al., 1996a,b), BMP (Amthor et al., 1998) and Wnt signaling (Kardon et al., 2003) have been described to play a role. Recently, Tcf4 has been found to prefigure muscle position in the limb buds (Kardon et al., 2003). TCF/LEF proteins are transcription factors that are activated as a result of canonical Wnt signaling. They are characterized by a DNA binding domain of the high mobility group (HMG) and an N-terminal beta-catenin binding domain (Hurlstone and Clevers, 2002). However, it is still unclear how these molecules interact to pattern the muscles in the context of the limb bud mesenchyme. Here, we address the first morphological manifestation of such a pattern, the future antagonist muscle groups – represented by a dorsal and a ventral premuscle mass – located at a distance from the limb bud margins. This means that, remarkably, the most anterior, posterior and distal regions of the outgrowing limb are devoid of muscle precursor cells, despite of strong Sf/Hgf expression anteriorly and distally (Brand et al., 1985; BrandSaberi and Christ, 1992; Myokai et al., 1995; Scaal et al., 1999). This discrepancy led us to postulate the existence of additional signaling mechanisms residing in the limb margins which counteract the migration sustaining function of SF/HGF. To test whether such signals emanate from the limb bud margins, we removed the margins individually and subsequently analyzed the position and distribution of the premuscle masses. Here, we demonstrate that removal of limb margins (anterior, posterior or distal) results in displacement and fusion of the premuscle masses across the resected margin. We present experimental evidence for the specific involvement of BMPs in this process. Our data suggest that BMPs control the correct positioning of migrating muscle precursors by inhibition of Tcf4 expression. Because of the implication of Tcf4 in muscle positioning (Kardon et al., 2003), it is a likely candidate to explain BMP-dependent changes in limb muscle pattern. Additional evidence for the involvement of BMPs is provided by the fact that exogenous SF/ HGF is able to produce a fusion of muscle masses exclusively in regions devoid of BMPs, e.g. the central limb, but not in the anterior or posterior margins where BMPs are expressed. We thus show that factors in the limb bud mesenchyme exert a dual function: firstly they act on myogenic precursor cells directly by inhibiting differentiation, and secondly they act on the lateral plate-derived limb bud mesenchyme by controlling Tcf4 expression. We therefore suggest a model in which muscle positioning in the limbs depends on the balance between attracting SF/HGF signals and repressing BMP signals both residing in the limb bud mesenchyme. While SF/HGF acts on the migrating myogenic precursor cells, BMPs act primarily on the stationary mesenchyme.

331

Materials and methods Embryos Fertilized eggs of Gallus gallus (White Leghorn) were incubated at 38°C and 80% relative humidity. The embryos were staged according to Hamburger and Hamilton (1951).

Bead preparation For application of rhBMP2, rhBMP4 and SF/HGF, heparin-coated acrylic beads of approximately 80 μm in diameter (Sigma) were rinsed 3× in PBS and individually transferred into the different protein solutions; BMP2, BMP4 (R&D Systems): 100, 10 and 1 μg/ml. SF/HGF was produced in Sf9 insect cells using the baculovirus expression system followed by one-step purification on heparin sepharose (Weidner et al., 1993). Beads were soaked in factor (0.5 μg/μl) at 4°C at least overnight. Noggin (R&D Systems) was soaked on affigel blue beads using a concentration of 0.5 μg/μl at 4°C overnight or longer.

Microsurgical procedures and bead application After the eggs were windowed, the vitelline membrane and the amnion were slit open in the area of operation. Limb margin removal was performed at HHstages 18–24 using electrolytically sharpened tungsten needles to avoid crushing of limb bud mesenchyme. Posterior margin removal defects were covered by gold foil or shell membrane. To implant a bead into an embryo, a small slit was cut through the ectoderm into the underlying tissue using tungsten needles and the bead was pushed through the slit into the desired position with a blunt needle. Embryos were operated at HH-stages 17–24. The operated eggs were sealed with medical tape and reincubated for the time required. Embryos were inspected, sacrificed, fixed in 4% paraformaldehyde in PBS overnight at 4°C, dehydrated in a graded methanol series and stored at − 20°C until further procedures.

Noggin cell beads preparation CHO B3 cells expressing noggin protein and DHFR control CHO cells were kindly provided by Dr. Richard Harland, University of California at Berkeley. Cells were grown on microcarrier beads (Hillex 38–72 m; SoloHill) (Aulehla et al., 2003) under conditions described elsewhere (Lamb et al., 1993).

In ovo electroporation and BMP2-application Eggs were incubated until stage 17 HH as described above. For electroporation, the upper sides of the eggs were windowed to visualize the embryos, the shell membrane was partially removed. The eGFP-plasmid DNA (7 μg/μl) was dissolved in an adjunct solution (high viscosity carboxymethylcellulose 0.33% (Sigma), Fast Green 1% (Sigma), MgCl2 1 mM, phosphatebuffered saline (PBS; 1×, in water)) in a ratio of 2:1. The plasmid DNA was microinjected into the epithelial somites at hindlimb level, and electroporated as previously described (Scaal et al., 2004). The electrodes were placed onto the extraembryonic membranes at each side of the microinjected embryo, and five square pulses of 60 V, 20 ms width, and 200 ms space were applied for each embryo. Upon passing current, the plasmid DNA with negative charges was sent to the tissue of the lateral somites adjacent to the anode side (Fig. 4A). One hour after electroporation, acrylamide beads soaked with BMP2 or BSA in PBS were implanted into the hindlimb bud as described above. After 17–20 h reincubation, the injected embryos were fixed in PBS, 4% PFA and the eGFP expression was visualized under fluorescence microscopy and photographed.

In situ hybridization For hybridization, embryos were rehydrated in a graded methanol series and processed as described by Nieto et al. (1996). Visualization of the hybridization product was achieved by using the digoxigenin RNA labeling and detection kit by Boehringer (Germany) according to the recommendations of the supplier.

332

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

The following probes were used: Sf/Hgf (Thery et al., 1995), qPax3, MyoD and Myf5, Lbx1, Bmp2 and Bmp4, Tcf4. The probes were used for hybridization at a concentration of 1 μg/ml.

Nile blue staining for apoptosis For Nile blue staining, extra-embryonic membranes were removed from the embryos in PBS and specimens then stained in toto for 30 min at 37°C in a Nile blue sulfate solution (1:20000 w/v). After incubation, specimens were rinsed twice with cold PBS, put on ice and immediately photographed under a microscope (van den Eijnde et al., 1997). Apoptotic areas showed up intensely blue due to uptake of the dye. The specimens were then fixed in 4% PFA for subsequent in situ hybridization.

Sections Whole mount embryos were embedded in 4% agar and sectioned with a vibratome at 50–100 μm.

Results Removal of the limb margins causes shifting and fusion of dorsal and ventral premuscle masses During normal development, the dorsal and ventral premuscle masses are clearly separated and located in the dorsal and ventral myogenic zone, respectively (Figs. 1A–C, H–J). To analyze the effect of the marginal limb bud mesenchyme on the separation of the premuscle masses, we removed the anterior, posterior, or distal limb margins at HH-stages 18–19 and examined the distribution of muscle masses after 24 h of reincubation using Lbx1 as a marker for the migrating muscle cells. Equivalent results could be observed using Pax3, Myf5, or MyoD instead of Lbx1 as markers (data not shown). After removal of the distal limb margin (Fig. 1D), the ventral and dorsal premuscle masses (Lbx1 positive) extended further distally and fused at the distal tip of the limb (n = 19/23, Figs. 1E–G). Notably, both premuscle masses were shifted in total towards the distal tip and were consequently situated at an increased distance from the base of the limb bud (Comparison of Figs. 1B and F). Removal of the posterior limb bud margin (Fig. 1K) resulted in a posterior shift of the dorsal and ventral premuscle masses and their fusion across the posterior limb mesenchyme (n = 17/21, Figs. 1L–N). The muscle precursor cells of both ventral and dorsal premuscle masses thus formed a U-like shape, viewed from anteriorly as well as in cross-sections of the limb bud (Fig. 1M). Similarly to distal mesenchyme removal, the manipulation resulted in a shift of the premuscle masses towards the resected margin. In this situation, the distance of the premuscle masses from the anterior margin was found to be increased, indicating a displacement towards the posterior margin (Fig. 1M). This also occurred when the defect after removal was subsequently covered by gold foil or shell membrane (data not shown) indicating that neither exposure at the wounding edge nor ectoderm regeneration play a role in the observed phenotype. Likewise, removal of the anterior limb margin (Fig. 1O) resulted in an anterior displacement of the premuscle masses and their fusion across the anterior mesenchyme (n = 15/20,

Figs. 1P–R). Again, not only fusion was observed but also an increased distance between the posterior limb margin and the premuscle masses (Fig. 1Q). Finally, we excised both anterior and posterior limb margins (Fig. 1S). Consistent with our prior observations, this led to a circular fusion of muscle cells at the anterior and posterior margins of the limbs (n = 9/12, Figs. 1T–V). Sf/Hgf, Bmp2 and Bmp4 expressions are altered after removal of the distal, posterior and anterior limb margins The observed mispositioning of the premuscle masses after removal of the limb margins indicated that factors expressed in, or secreted by these areas are involved either directly or indirectly in the correct distribution of muscle precursor cells. Therefore, we investigated the expression pattern of several factors known to be present in these areas. Sf/Hgf expression shows anterio-posterior differences in normal development. It retreats to the anterio-distal portion of the limb bud from HH-stage 21 on (Fig. 2A). After AER removal, Sf/Hgf was downregulated (n = 11/14, Fig. 2B), whereas posterior margin removal led to upregulation of Sf/Hgf posteriorly (n = 13/15, Fig. 2C). Extirpation of anterior mesenchyme caused a loss of Sf/Hgf positive tissue, but expression was maintained at lower levels in the adjacent mesenchyme (n = 11/11, Fig. 2D). Therefore, the presence or absence of Sf/Hgf alone cannot account for the muscle phenotype described. Among the secreted factors present in the marginal zones are members of the Bmp family (Bmp2 and Bmp4). Bmp2 is expressed in the posterior margin as well as in the AER (Fig. 2E). We wondered if Bmp expression would be induced ectopically or extinguished after marginal mesenchyme removal. We found that removal of the distal limb margin resulted in ablation of the Bmp2 domain in this area (n = 13/15), while expression in the posterior limb bud mesenchyme was maintained at normal levels (Fig. 2F). After posterior margin removal, Bmp2 expressing mesenchyme was almost totally ablated in the posterior limb bud mesenchyme, except for a slight stripe of expression at the more proximal margin (n = 11/13, Fig. 2G). Removal of the anterior limb margin did not affect Bmp2 expression at the posterior margin (n = 12/12, Fig. 2H). In contrast to Bmp2, Bmp4 is strongly expressed in the anterior mesenchyme, exhibiting a weaker expression in the AER and posterior mesenchyme (Fig. 2I). Distal limb margin removal resulted in a complete ablation of Bmp4 expressing tissue, while posterior and anterior expression remained at normal levels (n = 14/15, Fig. 2J). Following posterior margin excision, the posterior Bmp4 expression domain was no longer detectable (n = 11/11, Fig. 2K). Removal of the anterior limb mesenchyme resulted in the disappearance of anterior Bmp4 expression (n = 10/10, Fig. 2L). Taken together, Sf/Hgf was upregulated after posterior margin excision, whereas anterior and distal margin removals decreased its expression. Bmp2 and Bmp4 expression domains were ablated following manipulations that resulted in fusion of the premuscle masses, suggesting that BMPs prevent a fusion in the normal situation.

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

BMPs rescue the normal phenotype after anterior or posterior mesenchyme removal To further address the question, whether BMPs are responsible for the maintenance of muscle-free spaces at the limb bud margins, we combined the anterior and posterior excision experiments with implantation of BMP2/4 soaked beads (Figs. 3A, E, D, H). The first bead was implanted simultaneously to excision and after 12 h of reincubation, a second bead was implanted more distally. BMP implantation resulted in a rescue of the normal phenotype (anterior: n = 8/11, posterior: n = 10/12, Figs. 3C, G) inhibiting the fusion of premuscle masses observable after the excision experiments (Figs. 3B, F) whereas application of a PBS-soaked bead was not able to inhibit the fusion of the premuscle masses at the limb bud margins (not shown). These findings demonstrate that BMPs were able to maintain the separation of dorsal and ventral premuscle masses and to rescue a normal phenotype after manipulations that led to fusion of premuscle masses indicating a crucial role in the positioning of muscles during normal limb development. BMP inhibits migration of undifferentiated precursor cells The observed results pointed to a possible role of BMPs in the process of correct muscle positioning. To challenge this hypothesis, we tested the effect of BMP on the migration of cells delaminating from the lateral dermomyotomes. We labeled the lateral dermomyotome cells at hindlimb bud level with the help of electroporation of eGFP (Fig. 4A) and implanted a BMP2 soaked bead into the hindlimb bud. Analysis under a fluorescence microscope showed hindlimb buds devoid of labeled cells in the neighborhood and distal to the bead (n = 8/10, Fig. 4B). Controls with BSA beads showed that this effect is BMP specific and not due to a shielding effect (Fig. 4C). In support of this BMP effect on migration of cells into the limb bud, the BMP-binding protein Noggin, which serves as a potent and specific inhibitor of BMP function in the embryo (Zimmerman et al., 1996) should have the opposite effect. After placing several beads soaked in Noggin or Noggin-expressing cells close to the anterior and posterior margin of the limb buds of HH-stage 18–22 embryos (Fig. 5A), analysis of the premuscle masses showed a distortion and extension towards the limb margin (Figs. 5B–D). Viewed from cranially, dorsal and ventral premuscle masses were found to have fused across the anterior limb margin (Fig. 5D) corresponding to the effects caused by limb margin removals. Thus, BMP signals are indeed specifically involved in the process of premuscle mass positioning. This always occurred in the case of anterior bead application (n = 12/12), but never posteriorly (n = 0/ 12), presumably due to the absence of SF/HGF during later stages of limb bud development (Scaal et al., 1999). Thus, muscle can form only in the absence of BMPs if SF/HGF is present.

333

analyzed the expression pattern of Tcf4, a gene that had previously been reported to prefigure the position in the stationary mesenchyme where premuscle masses and subsequently limb muscles will develop (Kardon et al., 2003). In all types of experiments that had previously led to a shift in premuscle masses, a corresponding change in the expression domains of Tcf4 was observed (Fig. 6). This means that Tcf4 was strongly shifted (Fig. 6B) and even expressed across the new margins (Fig. 6E) in the posterior removal experiments (n = 32/35); BMP application caused a reduction in Tcf4 expression (n= 15/19, Fig. 6H). Noggin application lead to extensions of theTcf4 expression domains towards the Noggin source (n = 18/25, Fig. 6K). Thus, there was a high correlation of Tcf4 expression and experimental muscle position arguing for an involvement of canonical Wnt signaling in this process. SF/HGF alone is not able to produce fusion of premuscle masses at the limb margins As reported previously, Sf/Hgf is expressed in the mesenchyme of early limb buds. From HH-stage 23 on, it retreats to the anterio-distal portion of the limb bud (Scaal et al., 1999). Thus, the dynamic expression of Sf/Hgf during normal development of the limb prefigures the route of myogenic cell migration with respect to the proximo-distal axis. Application of exogenous SF/HGF ectopically in the proximal region of the limb bud (Figs. 7A, D) resulted in ectopic localization of myogenic cells near the source of SF/HGF (Scaal et al., 1999) as well as migration through the central core (chondrogenic region; n= 5/8, Figs. 7B, C) leading to a close approximation of the dorsal and the ventral premuscle masses. As posterior limb margin removal is known to ectopically upregulate Sf/Hgf expression in the posterior limb bud mesenchyme (Scaal et al., 1999, this paper), we wanted to test whether this Sf/Hgf upregulation alone could be responsible for the fusion of premuscle masses in the posterior mesenchyme. SF/HGF beads were inserted into the posterior limb mesenchyme (Fig. 7D). However, in this case, the position of the premuscle masses remained unaffected after a reincubation period of 24 h (n = 9/9, Fig. 7E). The equivalent result was observed for the anterior limb margin, where additional exogenous SF/HGF was also unable to cause fusion of the dorsal and ventral premuscle masses (data not shown). Therefore, upregulation of Sf/Hgf in the limb bud margins alone – in the presence of BMPs – is not sufficient to produce the fusion effects. As was shown for anterior and posterior limb margin removals, ectopic muscle was always exclusively found in areas where SF/HGF was present and BMPs were absent. This supports the hypothesis of a repulsive BMP influence in the mesenchyme of the limb margins, whereas the central limb region lacks those signals allowing cell redistribution towards the source of ectopic SF/HGF.

The experimentally induced shift in muscle position correlates with alterations in the Tcf4 expression domains

SF/HGF expression is independent of BMP

In additional series of marginal mesenchyme removal experiments, BMP applications and Noggin applications, we

In the anterior limb bud mesenchyme, Sf/Hgf and Bmp4exhibit overlapping expression domains during normal

334

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

Fig. 1. Abnormal positions of premuscle masses in chick embryos at HH-stage 24 after AER, ZPA and anterior mesenchyme removal, visualized by Lbx1 in situ hybridization, in comparison to the control situation: (A) control limb viewed from cranially. (B) Longitudinal section of panel A. Note the proximal and distal margins of the dorsal and ventral premuscle masses (arrowheads). The dotted bracket indicates the distance of the proximal margin of the dorsal premuscule mass to the limb base in the normal situation. (C) Schematic drawing of longitudinal section through a control limb with normal positioning of premuscle masses. (D) Schematic drawing of AER removal. (E) Anterior view of whole mount embryo after AER removal. Note distally fused premuscle masses (arrowhead). (F) Longitudinal section of (E) reveals distally shifted (bracket indicates normal situation) and fused premuscle masses at the distal limb tip (arrowheads). (G) Schematic drawing showing the results after AER removal, remember the absence of distal Bmp expression in comparison to panel C. Distal view of a control limb as a whole-mount embryo (H), in a transverse section (I) and schematic drawing (J). Note the clearly separated ventral and dorsal premuscle masses (arrowheads) and their distance to the anterior and posterior edges of the limb (brackets). (K) Schematic drawing of posterior mesenchyme (roughly equaling ZPA) removal. (L) Whole-mount embryo and (M) transverse section of (L) revealing a fusion of the premuscle masses at the posterior limb margin (arrowhead). Note the posteriorly shifted margin of the dorsal premuscle mass (arrowhead and bracket). (N) Schematic drawing showing the results after ZPA removal. (O) Schematic drawing of anterior limb mesenchyme removal. (P) Distal view of a whole mount embryo showing a fusion of the premuscle masses at the anterior limb margin (arrowhead). (Q) Transverse section of panel P with anteriorly shifted ventral premuscle mass (arrowhead and bracket). (R) Schematic drawing showing the results after anterior limb mesenchyme removal. (S) Schematic drawing of anterior and posterior mesenchyme removal. (T) Distal view of a whole mount embryo showing a circular fusion of the premuscle masses at the anterior and posterior margins (arrowheads). (U) Transverse section of panel T with circular fused premuscle masses (arrowheads). (V) Schematic drawing showing the results after anterior and posterior limb mesenchyme removal.

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

development. In contrast, Bmp2 and Sf/Hgf are mutually exclusive. To test for the possible influence of BMPs on Sf/Hgf expression, we inserted BMP2 and BMP4 soaked beads (1, 10, 100 μg/ml) into the anterior, posterior and central limb bud mesenchyme (Figs. 8B–D). Since BMPs are known to be able to induce apoptosis (Yokouchi et al., 1996), we performed Nile blue staining to examine the apoptotic activity after bead implantation. Both BMP2 and BMP4 caused additional apoptosis in the anterior (n = 17/20) and posterior (n = 15/18) mesenchyme around the bead (Figs. 8F, H), but not when grafted into the center of the limb bud (n = 20/20, Fig. 8G). This effect was more evident after a reincubation period of 12 h than after 24 h, probably due to a decrease in the release or to decay of BMP protein. Subsequently, the specimens were then stained for Sf/Hgf. In situ hybridization revealed a loss of Sf/Hgf expression at the

335

anterior limb margin only in an area that correlated to the apoptotic area, but not further distally (anterior: n = 7/8, posterior: n = 7/9, Figs. 8B, F). In the central limb region where no apoptosis was observed, expression of Sf/Hgf remained unaffected by the BMP applications (n = 13/13, Figs. 8C, G). Taken together, expression of Sf/Hgf in the limb mesenchyme did not seem to be specifically regulated or suppressed by BMPs, although its expression is upregulated in the posterior mesenchyme following posterior limb margin removal. Discussion The limb buds develop from two main sources of origin: the framework is formed by lateral plate-derived mesoderm, which is subsequently invaded by a number of migrating cell

Fig. 1 (continued).

336

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

Fig. 2. Alterations in the expression pattern of signaling and cell adhesion molecules in the limb mesenchyme after AER, posterior and anterior mesenchyme removal. (A) Normal Sf/Hgf expression pattern at HH-stage 24. (B) Extremely reduced Sf/Hgf expression after AER removal (arrowheads). (C) Intensified Sf/Hgf expression in the distal margin and ectopic expression in the posterior mesenchyme (arrowheads) following removal of the posterior marginal mesenchyme. (D) Reduced Sf/Hgf expression in the anterior limb mesenchyme (arrowheads), but still present expression in the distal mesenchyme. (E) Normal Bmp2 in situ hybridization at HHstage 24. (F) Maintained Bmp2 expression in the posterior mesenchyme after AER removal. Arrowheads indicate the lack of Bmp2 expression after AER removal. (G) Almost complete loss of Bmp2 expression after ZPA removal in the posterior mesenchyme (arrowheads) except for a stripe of expression proximally. (H) Unaffected Bmp2 expression after anterior mesenchyme removal. Arrowheads indicate the position of the manipulation. (I) Normal Bmp4 in situ hybridization showing expression in the anterior and posterior limb mesenchyme and in the AER (arrowheads) at HH-stage 23–24. (J) Bmp4 expression is maintained only in the proximal anterior mesenchyme after AER removal (arrowheads). (K) Loss of Bmp4 expression in the posterior mesenchyme after posterior margin removal (compare with panel L). (L) Almost totally absent Bmp4 expression (arrowheads) in the anterior mesenchyme after anterior margin removal (compare with panel K).

populations, the largest of which are the precursor cells for the skeletal muscle derived from the somitic dermomyotomes. While many classical and recent investigations have focused on the morphogenesis of the stationary (lateral plate-derived) limb bud mesenchyme yielding the cartilaginous skeleton and the connective tissue, our understanding of the limb muscle pattern is still refractory.

Muscle precursor cells enter the limb buds after detaching from the lateral dermomyotomal lips, divide into a clearly separated dorsal and ventral stream of cells, and migrate towards the distal tip along two clearly separate routes, the dorsal and ventral myogenic zones (Christ et al., 1977). During normal development, no exchange between dorsal and ventral muscle precursor cells can be observed (Schweizer et al., 2004).

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

337

Fig. 3. BMP beads rescue the fusion phenotype after anterior and posterior margin removal. In situ hybridization for Lbx1. (A, E) Operation schemes. (B) Anterior margin removal causes intense anterior fusion of the premuscle masses (arrowhead). (C) Insertion of BMP soaked beads to the anterior mesenchyme after anterior margin removal prevents fusion and reconstitutes the normal situation of separated ventral and dorsal premuscle masses (arrowheads). (D) Schematic drawing of the situation after anterior margin removal and anterior BMP bead application. As shown in Fig. 1L (arrowhead), removal of posterior mesenchyme results in fusion of the premuscle masses across the posterior margin. (G) BMP-soaked beads grafted into the posterior mesenchyme after posterior margin removal inhibits fusion and maintain the position of a separated dorsal and ventral premuscle mass. (H) Schematic drawing of the situation after posterior margin removal and bead insertion. (F) Schematic orientation of the displayed limb buds in panels B, C and G, note that the limb buds in panels B and C are shown from anteriorly, while the limb bud in panel G is shown from posteriorly.

The dorsal and the ventral premuscle mass does normally not extend to the anterior, posterior and distal margins of the limb bud. Although they move distally into the limb bud mesenchyme, they always remain located at a distance of 200 to 600 μm from the AER (Newman et al., 1981; Brand et al., 1985; Brand-Saberi and Christ, 1992). Muscle precursor cell migration, proliferation and differentiation are intensely studied processes (see reviews BrandSaberi and Christ, 1999; Dietrich, 1999; Arnold and Braun, 2000; Birchmeier and Brohmann, 2000; Buckingham, 2001; Christ and Brand-Saberi, 2002; Church and Francis-West, 2002; Buckingham et al., 2003; Francis-West et al., 2003). However, the mechanisms that control the positioning and separation into a dorsal and ventral premuscle mass within the limb bud remain unknown. It has been suggested recently that Tcf4, a downstream constituent of the canonical Wnt pathway of signal transduction, is involved in patterning the arrangement of skeletal muscle in the limbs (Kardon et al., 2003. Also, Ephreceptors and their ligands have been implicated in this process (Swartz et al., 2001). There is ample evidence for the fact that the chondrogenic zone located centrally between the dorsal and

ventral myogenic zones does not support long-term invasion by myogenic precursor cells in normal development due to the quality of the extracellular matrix and the distribution of adhesion molecules (Schramm and Solursh, 1990; Brand-Saberi et al., 1996b; Schweizer et al., 2004). However, the question of dorsal and ventral separation of the developing muscle blastemas throughout the limb bud – an important functional prerequisite – remained unanswered. Limb bud morphogenesis and muscle precursor cell invasion are tightly linked processes. In this connection, SF/HGF was described as a mediator of limb bud morphogenesis and muscle development, because it is controlled by FGFs from the AER and is capable of maintaining the myogenic cells in an undifferentiated state and of redirecting their migration under experimental conditions (Scaal et al., 1999). FGF2 and FGF4 can repress MyoD expression in myogenic cell lines in vitro and are considered to account for the absence of differentiated muscle in the distal limb region (Robson and Hughes, 1999). However, these signals do not explain the absence of migrating myoblasts in the unmanipulated distal limb bud, and in particular not in the anterior and posterior limb

338

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

Fig. 4. (A) Schematic drawing of electroporation. (B) Hindlimb bud with eGFP labeled cells delaminating from the lateral dermomyotomes and migrating into the proximal limb bud only up to the BMP2 soaked bead (circle), not further distally. (C) Control shows eGFP labeled cells populating the entire hindlimb bud. The bead is not visible because of the strong fluorescence of the myogenic cells.

bud regions. Here, we present evidence that BMPs located in the marginal limb bud mesenchyme inhibit the expression of Tcf4 thereby restricting the position of the premuscle masses and keeping them apart. The spatio-temporal sequence of Tcf4 expression closely parallels the distribution of myogenic cells on the one hand and is mutually exclusive with that of BMPs. Thus, our experimental results are strongly supported by descriptive data of normal development (this study; Kardon et al., 2003; Francis et al., 1994). The anterior, posterior and distal limb margins are required for correct muscle positioning It is known that myogenic precursors are guided into their correct position along their routes by a SF/HGF-cMet-mediated interaction (Dietrich et al., 1999) and that these precursor cells migrate towards ectopic sources of SF/HGF (Scaal et al., 1999; this study). Considering the migration-supporting function of SF/HGF, the myogenic cells should be located preferably near the anterior and distal borders of the limb bud where Sf/Hgf is strongly expressed. However, this is not the case as the ventral

and the dorsal premuscle masses are located at some distance from the anterior and posterior limb bud margins. If muscle cells are positioned by local cues from the lateral plate mesoderm, two different models could be conceivable to explain why the premuscle masses remain centrally in the ventral and dorsal mesenchyme despite of the SF/HGF influence located at the anterior side. In one model, additional properties have to be postulated which keep the cells exclusively in the central limb regions on their distally directed path. Alternatively, a second model is conceivable suggesting that non-permissive signals restrict the expansion or migration of the muscle masses towards the limb margins. Additionally, it has to be taken into account that a balance between proliferation of the stationary and the invading cell population also contributes to the correct pattern of the limb tissues. In this connection, however, no marked increase in proliferation could be detected in the limb buds after BMP-application (Köhler et al., 2004). The first possibility can be ruled out by the fact that in our experiments the shift of the premuscle masses always occurred towards the removed margin arguing against an attractive signal at the opposite side. This would suggest the existence of

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

339

Fig. 5. Noggin-induced anterior fusion of dorsal and ventral muscle masses. Myf5 in situ hybridization. (A) Operation scheme of bead implantation. (B) Dorsal view of operated right forelimb containing Noggin beads anteriorly and posteriorly in comparison to the unmanipulated left limb. Arrowhead points to anterior extension of dorsal premuscle mass (broken line). Note the absence of the extension in the left limb bud (arrow) and in the posterior region of the operated right limb bud. (C) Cranial view of operated right hindlimb bud containing Noggin beads anteriorly. Arrowhead pointing to anterior extension across the anterior margin from the dorsal premuscle mass. Inserted picture shows the contralateral side with no extension. (D) Same operation with extensions from dorsal and ventral premuscle mass, leading to fusion (arrowhead).

restrictive cues in the marginal mesenchyme that had been removed by the operations. However, this non-permissive signal could be acting directly on the myogenic precursor cells or alternatively on the stationary lateral plate mesoderm. Our results show that Wnt signaling is critically involved in the positioning of the premuscle masses, as their position highly correlates with the experimentally induced position of Tcf4, a downstream element of the canonical Wnt pathway which belongs to the Lef1 group of transcription factors that form complexes with beta-catenin and enter the nucleus. Interestingly, such a correlation had previously been described in normal development (Kardon et al., 2003). The expression of Tcf4, in turn, is negatively controlled by BMP-signaling in the limb bud margins. This is paralleled by other developmental systems (Jamora et al., 2003), where BMP inhibition is required for Lef1-type transcription factor expression. Another interesting observation is the fact that the ectopic position of the premuscle masses described here cannot be

interpreted simply as an expansion, but as a shift of premuscle masses. This suggests that myogenic precursor cells must be communicating with one another. At the same time, this finding argues against the view that fusion of the dorsal and ventral premuscle masses occurs as a direct consequence of our manipulations, because tissue is also affected where no manipulation was carried out. Furthermore, fusion also occurred when the defect of the removal was subsequently covered by gold foil or shell membrane, which inhibited ectodermal wound healing or retraction processes that could lead to a passive fusion of the muscle masses. Thus, our results strongly suggest the existence of properties at the anterior, posterior and distal limb margins which either repress the invasion of muscle precursor cells or have an influence on the balance between the expansion of the stationary and the invading cell population. Analyzing the secreted factors expressed in the limb, we focused on members of the BMP family, since Bmp2,

340

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

Fig. 6. Normal expression of Tcf4 in chicken limb bud at stage HH 25 in panels A, D, G and J. Operated chicken limb buds at stage HH 25 in panels B, E, H and K. Schematic orientation of the displayed limb buds in each column in panels C, F, I and L. Removal of posterior mesenchyme leads to a posterior shift of the Tcf4 expression domain (compare brackets in panels A and B, dorsal view), and it leads to a fusion of the dorsal and ventral premuscle masses (white arrowheads in panels D and E, caudal view). BMP beads inserted centrally into the limb bud downregulate Tcf4 (black arrowheads in panels G and H, view from anterior). Beads coated with Noggin-expressing cells placed anteriorly lead to upregulation of Tcf4 and a fusion of the dorsal and ventral Tcf4-positive expression domains (white arrowhead in panel K, anterior view). Inserted pictures in panels J and K are showing sections at the position of the dotted white line, black arrowhead in panel K shows the anterior position of the implanted bead (bead lost during sectioning).

expressed in the posterior limb margin and AER and progress zone, and Bmp4, expressed in the anterior and posterior limb margins (Francis et al., 1994), are present in areas required for normal positioning of the premuscle masses and are strongly reduced in ablation experiments that produced shifting and fusion of premuscle masses. This correlation between the absence of BMPs and ectopic muscle formation is supported by the two findings: (1) exogenous BMP application rescues the normal pattern after marginal mesenchyme removal, and (2) muscle fusion can be induced by Noggin application which blocks BMP binding to their receptors (Zimmerman et al., 1996).

Opposing effects of SF/HGF and BMPs In normal development, the dynamic changes in the Sf/Hgf expression guide myogenic precursor cells on their way (Scaal et al., 1999; Dietrich et al., 1999). It was furthermore shown that SF/ HGF can redirect migratory muscle cells towards an ectopic SF/ HGF source (Scaal et al., 1999, this study). In the present study, we demonstrate that this effect is location-dependent, meaning that muscle precursor cells can only react to an ectopic SF/HGF source when the source is located in regions devoid of BMPs such as the central chondrogenic limb region, but not in the anterior or posterior limb margins. This corresponds to the

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

341

Fig. 7. SF/HGF-induced ectopic migration of myogenic precursor cells. Myf5 in situ hybridization. (A) Schematic drawing of central SF/HGF application, dorsal view and section. (B) Operated hindlimb containing an SF/HGF bead in the central core. Arrowhead points to projection from the ventral premuscle mass towards the bead. (C) Section through the same limb demonstrates centrally located muscle cells (arrowhead), bead broken due to sectioning. (D) Schematic drawing of posterior SF/HGF application, dorsal view. (E) Forelimb bud containing beads posteriorly (viewed from caudally). Arrowheads point to the not fused dorsal and ventral premuscle masses.

situation during normal development where Sf/Hgf and Bmps are expressed in overlapping domains in the anterior limb mesenchyme, but muscle precursors are not localized in this region. We furthermore suppose that the absence of SF/HGF posteriorly is the reason that Noggin-induced fusions occur only anteriorly but

never posteriorly. However, after posterior or anterior margin removal, Sf/Hgf expression is present in the absence of BMPs which results in shift and fusion of the muscle masses towards the margins. In a recent study, Mic and Duester (2003) presented evidence for involvement of RA-signaling in the guidance of

342

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

Fig. 8. Influence of BMPs on Sf/Hgf in its mesenchymal expression domain and on apoptosis. Dorsal views of control limb showing normal Sf/Hgf expression at HH-stage 21 (A). Application of a BMP2 or 4 bead to the anterior mesenchyme resulted in a downregulation of Sf/Hgf (arrowheads in panel B) that was related to apoptosis (arrowheads in panel F). Note that fusion of the premuscle masses (Fig. 4B) only happened distal to the zone of apoptotic activity. Application of BMP2 or 4 to the central limb mesenchyme did not alter Sf/Hgf expression (C), or apoptotic activity (G). Application of BMP2 or 4 to the posterior limb mesenchyme did not alter Sf/Hgf expression, as it was absent as in the normal situation (D), but was linked to an enhanced apoptotic activity (arrowheads in panel H).

cMet expressing migrating muscle precursors through an effect on spatial patterning of Sf/Hgf expression. In limb buds lacking RA, Sf/Hgf expression is abnormally shifted to the anteriorproximal limb margin, correspondingly guiding muscle cells towards the anterior margin, similar to our results after anterior limb margin removal. Analysis of Bmp4 in those limbs revealed expression anterior-distal, mutually exclusive to the Sf/Hgf and cMet muscle domains. This group suggested that the effect of RA signaling might be mediated by BMP4 which negatively regulates Sf/Hgf and thereby causes abnormal muscle positioning anteriorly. In our experiments, we could show that BMPs do not negatively regulate Sf/Hgf expression in the limb buds except for an apoptotic effect in the anterior margin. In RA-deficient limb buds, Sf/Hgf and Bmp4 expressions do not overlap and therefore SF/HGF guides the muscle cells into the abnormal positions at the anterior margin. Indeed, no cMet expressing cells were found in the Bmp4 expressing region. This is also well in line with our finding that Tcf4 expression as a marker of muscle-associated connective tissue is repressed by BMPs.

homologue transcription factor Lbx1 keeps the myogenic cells in a proliferative state especially when they are migrating within the lateral plate-derived mesenchyme (Mennerich and Braun, 2001). Because of the lack of myogenic precursor cells in the hindlimbs and the dorsal premuscle mass of the forelimb buds in Lbx1 mutants, it has been suggested that a proliferation defect may account for this phenotype. Hence it remains unclear as to what extent the effects caused by SF/HGF are due to an increase in cell motility or

Interfering with migration or proliferation In addition to SF/HGF which acts on myogenic precursor cell motility and mitosis via cMet signaling, the ladybird

Fig. 9. Proposed model for muscle positioning (details view text). (A) Lateral view of an HH-stage 22 limb and (B) cranial view of the same limb.

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

proliferation. Most likely, both processes are affected; however, excessive or ectopic SF/HGF does not considerably extend the pool of muscle precursor cells. In contrast to previous suggestions that BMP-dependent muscle positioning depends on a balance between proliferation, differentiation and cell death of myogenic precursor cells as a result of dose-dependent BMP effects (Amthor et al., 1998), our results suggest that mispositioning of premuscle masses as shown here relates to the influence of BMPs on the stationary limb bud mesoderm. Furthermore, BMPs do not cause a detectable upregulation of proliferation within the central limb bud (Köhler et al., 2004). In contrast, apoptosis was found in the stationary marginal limb bud mesoderm after application of BMP. A case for the involvement of cell migration in addition to proliferation is our finding that premuscle masses were not enlarged towards the margins of the limb bud, but entirely shifted with increasing distance to the non-manipulated margin. Conclusions Muscle positioning in the limb is controlled directly and indirectly by signaling molecules produced by the marginal mesenchyme. We have shown that BMPs and SF/HGF are involved in this process. These molecules, instead of directly influencing each others expression, interact with the muscle precursor cells in a way supplementing each other in patterning limb muscle. SF/HGF guides the muscle precursors mainly in the proximo-distal axis as a positive factor and BMP2/4 indirectly control the anterio-posterior positioning of the muscle anlagen acting on the lateral plate mesoderm via the negative control of Tcf4 expression, which in turn prefigures the position of the premuscle masses. Under normal circumstances, the repressing signals for the muscle precursors overrule the supporting ones preventing myogenic precursors from invading the limb margins. This ensures the correct distribution of the cells in a dorsal and a ventral premuscle mass (Fig. 9). Acknowledgments The authors are indebted to Gabrielle Kardon, Philippa FrancisWest and to Charly Ordahl for helpful comments on the manuscript. They furthermore want to thank Ulrike Pein for her excellent technical assistance during in situ hybridization and Anton Gamel for help with the layout of the manuscript. Probes used in this study were kindly provided by Claudio Stern, Gabrielle Kardon, Christophe Marcelle, Bruce M. Paterson, Susanne Dietrich and Paul Brickell. Human recombinant SF/HGF was generously supplied by Carmen Birchmeier and Walter Birchmeier. This study was supported by the Deutsche Forschungsgemeinschaft (Br957/5-2, 5-3). We acknowledge the support of the MYORES project (511978), funded by the EC's Sixth Framework Program. References Amthor, H., Christ, B., Weil, M., Patel, K., 1998. The importance of timing differentiation during limb muscle development. Curr. Biol. 8, 642–652.

343

Arnold, H.H., Braun, T., 2000. Genetics of muscle determination and development. Curr. Top. Dev. Biol. 48, 129–164. Aulehla, A., Wehrle, C., Brand-Saberi, B., Kemler, R., Gossler, A., Kanzler, B., Herrmann, B.G., 2003. Wnt3a plays a major role in the segmentation, clock controlling somitogenesis. Dev. Cell 4, 395–406. Birchmeier, C., Brohmann, H., 2000. Genes that control the development of migrating muscle precursor cells. Curr. Op. Cell. Biol. 12, 725–730. Bladt, F., Riethmacher, D., Isenmann, S., Aguzzi, A., Birchmeier, C., 1995. Essential role for the cMet receptor in the migration of myogenic precursor cells into the limb bud. Nature 376, 768–771. Brand, B., Christ, B., Jacob, H.J., 1985. An experimental analysis of the developmental capacities of distal parts of avian leg buds. Am. J. Anat. 173, 321–340. Brand-Saberi, B., Christ, B., 1992. A comparative study of myogenic cell invasion of the avian wing and leg bud. Eur. J. Morphol. 30, 169–180. Brand-Saberi, B., Christ, B., 1999. Genetic and epigenetic control of muscle development in vertebrates. Cell Tissue Res. 296, 199–212. Brand-Saberi, B., Müller, T.S., Wilting, J., Christ, B., Birchmeier, C., 1996a. Scatter factor/hepatocyte growth factor (SF/HGF) induces emigration of myogenic cells at interlimb level in vivo. Dev. Biol. 179, 303–308. Brand-Saberi, B., Gamel, A.J., Krenn, V., Müller, T.S., Wilting, J., Christ, B., 1996b. N-cadherin is involved in myoblast migration and muscle differentiation in the avian limb bud. Dev. Biol. 178, 160–173. Brohmann, H., Jagla, K., Birchmeier, C., 2000. The role of Lbx1 in migration of muscle precursor cells. Development 127, 437–445. Buckingham, M., 2001. Skeletal muscle formation in vertebrates. Curr. Opin. Genet. Dev. 11, 440–448. Buckingham, M., Bajard, L., Chang, T., Daubas, P., Hadchouel, J., Meilhac, S., Montarras, D., Rocancourt, D., Relaix, F., 2003. The formation of skeletal muscle: from somites to limb. J. Anat. 202, 59–68. Chevallier, A., Kieny, M., Mauger, A., 1977. Limb–somite relationship: origin of the limb musculature. J. Embryol. Exp. Morphol. 41, 245–258. Christ, B., Brand-Saberi, B., 2002. Limb muscle development. Int. J. Dev. Biol. 46, 905–914. Christ, B., Jacob, H.J., Jacob, M., 1974. Über den Ursprung der Flügelmuskulatur. Experientia 30, 1446–1448. Christ, B., Jacob, H.J., Jacob, M., 1977. Experimental analysis of the origin of the wing musculature in avian embryos. Anat. Embryol. 150, 171–186. Church, V.L., Francis-West, P., 2002. Wnt signalling during limb development. Int. J. Dev. Biol. 46, 927–936. Daston, G., Lamar, E., Olivier, M., Goulding, M., 1996. Pax3 is necessary for migration but not differentiation of limb muscle precursors in the mouse. Development 122, 1017–1027. Dietrich, S., 1999. Regulation of hypaxial muscle development. Cell Tissue Res. 296, 175–182. Dietrich, S., Schubert, F.R., Healy, C., Sharpe, P.T., Lumsden, A., 1998. Specification of the hypaxial musculature. Development 125, 2235–2249. Dietrich, S., Abou-Rebyeh, F., Brohmann, H., Bladt, F., SonnenbergRiethmacher, E., Yamaai, T., Lumsden, A., Brand-Saberi, B., Birchmeier, C., 1999. The role of SF/HGF and cMet in the development of skeletal muscle. Development 126, 1621–1629. Epstein, J.A., Shapiro, D.N., Cheng, J., Lam, P.Y.P., Maas, R.L., 1996. Pax3 modulates expression of the cMet receptor during limb muscle development. Proc. Natl. Acad. Sci. U. S. A. 93, 4213–4218. Francis, P.H., Richardson, M.K., Brickell, P.M., Tickle, C., 1994. Bone morphogenetic proteins and a signaling pathway that controls patterning in the developing chick limb. Development 120, 209–218. Francis-West, P.H., Antoni, L., Anakwe, K., 2003. Regulation of myogenic differentiation in the developing limb bud. J. Anat. 202, 69–81. Goulding, M., Lumsden, A., Paquette, A.J., 1994. Regulation of Pax3 expression in the dermomyotome and its role in muscle development. Development 120, 957–971. Gross, M.K., Moran-Rivard, L., Velasquez, T., Nakatsu, M.N., Jagla, K., Goulding, M., 2000. Lbx1 is required for muscle precursor migration along a lateral pathway into the limb. Development 127, 413–424. Hamburger, V., Hamilton, H.L., 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49–92.

344

A. Bonafede et al. / Developmental Biology 299 (2006) 330–344

Heymann, S., Koudrova, M., Arnold, H.H., Köster, M., Braun, T., 1996. Regulation and function of SF/HGF during migration of limb muscle precursor cells in chicken. Dev. Biol. 180, 556–578. Hurlstone, A., Clevers, H., 2002. T-cell factors: turn-ons and turn-offs. EMBO J. 15, 2303–2311. Jacob, M., Christ, B., Jacob, H.J., 1979. The migration of myogenic cells from the somites into the leg region of avian embryos. Anat. Embryol. 157, 291–309. 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. Jamora, C., DasGupta, R., Kocieniewski, P., Fuchs, E., 2003. Links between signal transduction, transcription and adhesion in epithelial bud development. Nature 422, 317–322. Kardon, G., 1998. Muscle and tendon morphogenesis in the avian hind limb. Development 125, 4019–4032. Kardon, G., Harfe, B.D., Tabin, C.J., 2003. A Tcf4-positive mesodermal population provides a prepattern for vertebrate limb muscle patterning. Dev. Cell. 5, 937–944. Köhler, T., Pröls, F., Brand-Saberi, B., 2004. PCNA in situ hybridization: a novel and reliable tool for detection of dynamic changes in proliferative activity. Histochem. Cell Biol. 123, 315–327. Lamb, T.M., Knecht, A.K., Smith, W.C., Stachel, S.E., Economides, A.N., Stahl, N., Yancopolous, G.D., Harland, R.M., 1993. Neural induction by the secreted polypeptide noggin. Science 262, 713–718. Mankoo, B.S., Collins, N.S., Ashby, P., Grigorieva, E., Pevny, L.H., Candia, A., Wright, C.V.E., Rigby, P.W.J., Pachnis, V., 1999. Mox2 is a component of the genetic hierarchy controlling limb muscle development. Nature 400, 69–73. Matsumoto, K., Nakamura, T., 1996. Emerging multipotent aspects of hepatocyte growth factor. J. Biochem. 119, 591–600. Mennerich, D., Braun, T., 2001. Activation of myogenesis by the homeobox gene Lbx1 requires cell proliferation. EMBO J. 20, 7174–7183. Mennerich, D., Schäfer, K., Braun, T., 1998. Pax3 is necessary but not sufficient for lbx1 expression in myogenic precursor cells of the limb. Mech. Dev. 73, 147–158. Mic, F.A., Duester, G., 2003. Patterning of forelimb bud myogenic precursor cells requires retinoic acid signaling initiated by Raldh2. Dev. Biol. 264, 191–201. Myokai, F., Washio, N., Asahara, Y., Yamaai, T., Tanda, N., Ishikawa, T., Aoki, S., Kurihara, H., Murayama, Y., Saito, T., et al., 1995. Expression of the hepatocyte growth factor gene during chick limb development. Dev. Dyn. 202, 80–90. Newman, S.A., Pautou, M.P., Kieny, M., 1981. The distal boundary of myogenic primordia in chimeric avian limb buds and its relation to an accessible population of cartilage progenitor cells. Dev. Biol. 84, 440–448. Nieto, M.A., Patel, K., Wilkinson, D.G., 1996. In situ hybridization analysis of

chick embryos in whole mount and tissue sections. Methods Cell Biol. 51, 219–235. Robson, L.G., Hughes, S.M., 1999. Local signals in the chick limb bud can override myoblast lineage commitment: induction of slow myosin heavy chain in fast myoblasts. Mech. Dev. 85, 59–71. Scaal, M., Bonafede, A., Dathe, V., Sachs, M., Cann, G., Christ, B., BrandSaberi, B., 1999. SF/HGF is a mediator between limb patterning and muscle development. Development 126, 4885–4893. Scaal, M., Gros, J., Lesbros, C., Marcelle, C., 2004. In ovo electroporation of avian somites. Dev. Dyn. 229, 643–650. Schäfer, K., Braun, T., 1999. Early specification of limb muscle precursor cells by the homeobox gene Lbx1h. Nat. Genet. 23, 213–216. Schramm, C., Solursh, M., 1990. The formation of premuscle masses during chick wing bud development. Anat. Embryol. 182, 235–247. Schweizer, H., Johnson, R.L., Brand-Saberi, B., 2004. Characterization of migration behavior of myogenic precursor cells in the limb bud with respect to Lmx1bexpression. Anat. Embryol. 208, 7–18. Sonnenberg, E., Meyer, D., Weidner, K.M., Birchmeier, C., 1993. Scatter factor/ hepatocyte growth factor and its receptor, the cMet tyrosine kinase, can mediate a signal exchange between mesenchyme and epithelia during mouse development. J. Cell Biol. 123, 223–235. Swartz, M.E., Eberhart, J., Pasquale, E.B., Krull, C.E., 2001. EphA4/ephrin-A5 interactions in muscle precursor cell migration in the avian forelimb. Development 128, 4669–4680. Thery, C., Sharpe, M.J., Batley, S.J., Stern, C.D., Gherardi, E., 1995. Expression of HGF/SF, HGF1/MSP, and cMet suggests new functions during early chick development. Dev. Genet. 17, 90–101. Uchiyama, K., Ishikawa, A., Hanaoka, K., 2000. Expression of lbx1 involved in the hypaxial musculature formation of the mouse embryo. J. Exp. Zool. 286, 270–279. van den Eijnde, S.M., Luijsterburg, A., Boshart, L., De Zeeuw, C.I., van Dierendonck, J.H., Reutelingsperger, C., Vermeij-Keers, C., 1997. In situ detection of apoptosis during embryogenesis with Annexin V: from whole mount to ultrastructure. Cytometry 29, 313–320. Weidner, K.M., Sachs, M., Birchmeier, W., 1993. The Met receptor tyrosine kinase transduces motility, proliferation and morphogenic signals of scatter factor/hepatocyte growth factor in epithelial cells. J. Cell Biol. 121, 145–154. 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. Yokouchi, Y., Sakiyama, J., Kameda, T., Iba, H., Suzuki, A., Ueno, N., Kuroiwa, A., 1996. BMP2/4 mediate programmed cell death in chicken limb buds. Development 122, 3725–3734. Zimmerman, L.B., De Jesús-Escobar, J.M., Harland, R.M., 1996. The Spemann organizer signal noggin binds and inactivates bone morphogenetic protein 4. Cell 86, 599–606.