BMP-4 and Noggin signaling modulate dorsal fin and somite development in the axolotl trunk

Authors


Abstract

BMP-4, a member of the TGF-beta superfamily of growth factors, is involved in various developmental processes. We investigated the effects of BMP-4 and its antagonist Noggin on axolotl trunk development. Implantation of BMP-4-coated microbeads caused inhibition of muscle and dorsal fin formation in the vicinity of the microbeads. At some distance, myotomes developed with reduced height but increased width, which was accompanied by increased cell proliferation. These effects could be modulated by co-implanting Noggin-coated beads. Immunostaining of Pax7 further revealed that although the dermomyotome was absent in the vicinity of BMP-4-coated beads, at some distance from them, it was thicker than in controls, indicating that moderate amounts of BMP-4 stimulate this layer of undifferentiated cells. In contrast, Noggin generally inhibited the dermomyotome, possibly indicating premature differentiation of dermomyotome cells. We conclude that BMP-4 and Noggin are involved in the regulation of cell proliferation and differentiation during somite development. Developmental Dynamics 236:2464–2474, 2007. © 2007 Wiley-Liss, Inc.

INTRODUCTION

Bone morphogenetic protein 4 (BMP-4), a member of the TGF-β superfamily of growth factors, plays diverse roles during vertebrate development. It is involved in embryonic axis and germ layer induction and regulates the formation of various tissues and organs such as brain, neural crest, muscle, bone, or cartilage (Hogan,1996; Mehler et al.,1997; Whitman,1998; Dale and Jones,1999; Massague et al.,2000; Shi and Massague,2003). BMP-signaling is negatively regulated at various levels by several inhibitors, affecting either BMP-receptor interactions or intracellular signal propagation (Massague et al.,2000; Miyazono et al.,2000; Shi and Massague,2003). The secreted proteins Chordin, Noggin, and Follistatin, for example, bind BMP-4 and thereby downregulate BMP-activity during frog gastrulation (Piccolo et al.,1996; Zimmerman et al.,1996; Fainsod et al.,1997). Noggin-BMP-interactions also play a role later in development, e.g., during somite formation and differentiation in chick and mouse embryos (Marcelle et al.,1997; Reshef et al.,1998; McMahon et al.,1998; Linker et al.,2003).

We analysed the influence of BMP-4 and its antagonist, Noggin, during trunk development of the Mexican axolotl (Ambystoma mexicanum). Recent experiments in the system of wild type (dark) and white mutant axolotl embryos showed that pigment cell migration on the lateral (subepidermal) neural crest migratory pathway is regulated by balanced signaling between BMP-4 and Noggin (Hess et al., in press). Here, we focus on the influence of both proteins on the development of somites and dorsal fin mesenchyme.

The mesenchyme of the dorsal fin in lower vertebrates has classically been regarded as derived from the neural crest (Raven,1931; DuShane,1935). However, recent experiments in axolotl and Xenopus demonstrated a contribution from the somites (Sobkow et al.,2006; Garriock and Krieg,2007). Most of the current knowledge about somite development comes from chick and mouse (Christ and Ordahl,1995; Scaal and Christ,2004; Christ et al.,2004; Pourquié,2004; Kalcheim and Ben-Yair,2005). Somites are mesodermal structures that condense from paraxial somitogenic mesenchyme on either side of the neural tube into pairs of epithelial spheres, a process that reflects vertebrate segmentation. During further development, the somites become subdivided into a medial sclerotome, which gives rise to axial cartilage and bone, and a lateral dermomyotome. The dermomyotome is a source of precursor cells for dermis and muscle, in both amniotes (Scaal and Christ,2004; Ben-Yair and Kalcheim,2005) and anamniotes (Devoto et al.,2006; Scaal and Wiegreffe,2006) and gives rise to all skeletal muscles of the trunk and limb as well as to dermis and vascular endothelia (Scaal and Christ,2004). In axolotl, differentiated somites consist of central myotomes bordered by narrow medial sclerotomes (Epperlein et al.,2007) and lateral dermomyotomes (termed dermatome in Sobkow et al.,2006). To date, both the dermomyotome and sclerotome are poorly characterized in axolotl embryos.

Here we show that the development of somites and dorsal fin mesenchyme in axolotl can be influenced by BMP-4 and Noggin. Moderate doses of BMP-4 stimulate cell proliferation, affect somite morphology, and induce dermomyotome growth. Strong BMP-4 signaling interferes with the development of the myotome, dermomyotome and dorsal fin. Noggin alone stimulates dorsal fin formation, and interferes with dermomyotome formation but does not affect early myotome development. Our results indicate that interactions between BMP-4 and its inhibitor Noggin in axolotl embryos may be crucial to balance differentiation and cell proliferation in somites and their descendants.

RESULTS

BMP-4 Inhibits Somite and Dorsal Fin Formation and Modulates Pigment Cell Distribution

Microbeads soaked in BMP-4 protein solution were grafted into the trunk of axolotl embryos in order to study the influence of BMP-4 on trunk development. Each embryo (wild type, premigratory neural crest, stages 24–26) received one microbead in the left dorsolateral trunk. For implantation, the epidermis over the implantation area was lifted and the dorsal portion of the 7th or 8th somite was removed to provide space for the microbead. Uncoated beads were implanted as controls to rule out non-specific effects of the procedure. Manipulated embryos were allowed to develop for up to 5 days following the operation (stage 38).

Control beads had no apparent effect on somite development, dorsal fin formation, or pigment cell distribution (Fig. 1A,B). The gap in the somites visible in Figure 1B (asterisk) was generated prior to bead implantation to provide space for the bead. Initially, we injected 10 nl of a 50-μg/ml BMP-4 protein solution into the dorsal somite region of wild-type embryos at stage 25. This caused a reduction in the number of melanophores at the injection site and a slight decrease in dorsal fin height above the injection site (Fig. 1C). Dorsal fin development was, however, more severely inhibited following the implantation of microbeads soaked in 50 and 100 μg/ml BMP-4 solutions (Fig. 1D,E). Because beads soaked in 100 μg/ml BMP-4 caused the most pronounced effects on pigment cell distribution and dorsal fin development, this concentration was used to coat beads during further experiments. In addition to a strong reduction of dorsal fin height, we observed an indentation of the trunk around the implanted bead on the ipsilateral but a thickening on the contralateral side (Fig. 1E,F) opposite the implanted BMP-4-coated bead. Horizontal sections through the trunk at the level of the microbead that were stained with a muscle marker (12/101; Kintner and Brockes,1984), revealed a loss of myotome tissue on the ipsilateral side (Fig. 1F) but an apparent increase in myotome width on the contralateral side (Fig. 1B,F). Interestingly, the myotome also appeared to be wider on the ipsilateral side at some distance from the microbead (Fig. 1F), possibly indicating concentration-dependent effects of BMP-4 on myotome tissue, a loss of tissue at high concentrations and an expansion of tissue at lower concentrations. Horizontal and transverse sections through control larvae showed myotomes with equal width on either side of the trunk (Fig. 1B).

Figure 1.

Effects of BMP-4 and Noggin on pigment cell migration, dorsal fin formation, and somite development. A: Lateral aspect of 7-day (d) control larva (stage 38) carrying an uncoated white microbead dorsal to one midtrunk somite (arrow). The bead affected neither pigment cell distribution nor dorsal fin development. B: Horizontal vibratome section through a 7-d larva at the level of a control bead, stained with a muscle-specific antibody (12/101) and DAPI. The microbead (asterisk) does not affect the morphology of adjacent or more distant somites. CE: Lateral aspects of stage-38 larvae injected with BMP-4 protein (10 nl of a 50-μg/ml solution; C) or after implantation of BMP-4-coated microbeads (arrow) soaked in 50 μg/ml (D) and 100 μg/ml (E) solutions, respectively. At some distance from the bead, a concentric belt developed free of melanophores and the lateral border of pigmentation was shifted more dorsally. The dorsal fin was slightly reduced in height after protein injection (C) and became progressively reduced after implantation of microbeads loaded with increasing amounts of BMP-4 (D,E). F: Horizontal section through the larva shown in E. The myotomes are missing in the vicinity of the bead (asterisk) but appear to be more prominent at some distance, both on the ipsi- and the contralateral side. G: Lateral aspect of a stage-38 larva carrying a blue Noggin-coated microbead (arrow). Noggin causes an increase of the dorsal fin above the implantation site. H: Horizontal section at the level of the neural tube through the larva in G. Noggin has no influence on the size of myotomes. I: Lateral aspect of a stage-38 larva carrying one white microbead loaded with BMP-4 in the position of the 7th dorsal somite (white arrow) and one blue Noggin coated microbead in the position of the 8th somite (blue arrow). Note the reduction of pigment-free area compared to C. K: Horizontal section through the larva in I. A moderate increase in myotome width was observed on either side of the trunk. The dotted lines in A,C–E,G,I indicate dorsal fin margins, and the dashed lines in A,E,G,I the sectioning planes of the horizontal sections shown in B,F,H,K. df, dorsal fin; my, myotome.

Noggin Antagonizes BMP-4 Induced Effects on Dorsal Fin Formation, Pigment Cell Distribution, and Somite Development

BMP-4 signals can be antagonized by a number of secreted proteins, including Noggin, Follistatin, and Chordin (Piccolo et al.,1996; Zimmerman et al.,1996; Fainsod et al.,1997). To demonstrate the specificity of BMP-4 activity during pigment cell migration and somite development, we used Noggin as a specific BMP-4 inhibitor. We first tested the influence of Noggin on the normal morphogenesis of the axolotl trunk. Agarose beads coated with Noggin were implanted at the same location in the dorsolateral trunk of dark embryos (stage 24–27) as the BMP-4 beads in the previous experiments. Approximately five days after bead implantation, a local increase in the dorsal fin height was observed above the implantation site (Fig. 1G). However, Noggin beads had no influence on pigment cell migration or the development of paraxial mesoderm. The mesoderm developed into myotomes on either side of the trunk (Fig. 1H) equivalent to controls (Fig. 1B).

Next, we analysed the antagonizing effect of Noggin on BMP-4-induced changes in pigment cell and myotome development. In these experiments, one BMP-4 bead (white) was implanted into the position of the left 7th dorsal somite (stage 24–27) and one Noggin bead (blue) beside it in the position of the left 8th dorsal somite, i.e., the beads were implanted at the same location of the dorsolateral trunk as the single BMP-4 or Noggin beads in the previous experiments. After approximately 3 days, melanophores disappeared in the implantation area and the dorsal fin failed to develop above the bead. These effects became more obvious at later larval stages (Fig. 1I). However, the melanophore-free area that developed when only one BMP-4 bead was implanted did not appear. Other effects caused by a BMP-4 bead alone such as a dorsal shift in the left lateral pigmentation border or the indentation of left trunk myotomes were less pronounced. Furthermore, the BMP-4-induced loss of myotome tissue on the ipsilateral side was compensated for and the width of myotomes appeared slightly increased (Fig. 1K). Finally, contralateral myotomes were thicker but not as pronounced as in embryos carrying a single BMP-4 bead (compare Fig. 1K with 1F).

Together, these results reveal specific responses of the developing axolotl trunk to increased BMP-4 levels, and suggest that both BMP-4 and Noggin may be involved in pigment cell migration, dorsal fin formation, and myotome development.

Quantitative Evaluation of the Influence of BMP-4 and Noggin on Myotome Development

As shown above, ectopic BMP-4 causes striking changes in myotome development, presumably in a concentration-dependent manner, and these effects can be modified or rescued by Noggin (Fig. 1I,K). The apparent increase in muscle tissue at some distance from the BMP-4 source, i.e., both on the ipsi- and contralateral side of the trunk with respect to the implanted bead, appears peculiar. Analysis of transverse sections through the trunk at the level of the implanted bead, however, revealed a correlation between an increase in myotome width and a decrease in myotome height (Fig. 2A–D; see also Fig. 3). These observations prompted us to further investigate whether BMP-4 can cause a bona fide increase in myotome mass or simply an unusual rearrangement of prospective muscle tissue. To distinguish between these alternatives, we compared myotome areas on transverse plastic sections through stage-38 larvae. At least 3 larvae in each group were analysed carrying either one BMP-4 bead (Fig. 2A), one BMP-4 and one Noggin bead (Fig. 2B), one Noggin bead (Fig. 2C), or one uncoated bead (Fig. 2D). On transverse sections through these different embryos, we determined the somite area and the mean cell number/myotome area (Fig. 2E). We focused on the right side of the trunk, i.e., the contralateral side relative to bead implantation, where the “growth” effect was clearly apparent with a BMP-4 bead alone as well as with the BMP-4 and Noggin combination.

Figure 2.

Influence of BMP-4 and Noggin on myotome development. A–D: Transverse semithin plastic sections through the trunk of axolotl larvae (stage 38) carrying different microbeads in the dorsal midtrunk. A: One white BMP-4-coated microbead (location on the left, outside the image): The left neural tube has a thinner wall, the left notochord is malformed, and most of the muscle tissue on the left side of the trunk is missing. Conversely, the muscle mass on the right side of the trunk seems to be increased. B: One white BMP-4-coated and one blue Noggin-coated microbead: Similar effects on trunk structures as in A but generally less pronounced. C,D: One blue Noggin-loaded microbead (asterisk, C) and one uncoated control bead (asterisk, D). No effects are observable on trunk structures. The red line in the embryo sketch (top) indicates the sectioning plane. E: Quantitative evaluation of BMP-4- and Noggin-induced effects on myotome area (top) and cell number/myotome area at the vicinity of the microbead (bottom). The differences in total myotome area are not significant, but the cell numbers/myotome area of BMP-4- and BMP-4+Noggin-treated embryos are significantly higher than in controls (asterisks in E, Student's t-test; P < 0.05). At least 3 animals were analysed for each group.

Figure 3.

Cell division in midtrunk myotomes of untreated wild type axolotl embryos (AE) and after stimulation with BMP-4 (FK). Microbeads soaked in BMP-4 (100 μg/ml) were implanted at stage 23/24. Transverse vibratome sections through control and experimental embryos were stained after 4 hr (A,F), 12 h (B,G), 24 (C,H), 2 d (D,I), and 4 d (E,K) with antibodies against mitotic cells (phospho-Histone H3, red fluorescence) and prospective muscle cells (12/101, green fluorescence). In control embryos, cell proliferation becomes progressively reduced, whereas in BMP4-treated embryos the cell division rate remains at high levels (compare E and K). In addition, muscle development appears to be normal in BMP4-treated embryos until 1–2 days after bead implantation. Between 2 and 4 days after bead implantation, the myotome disappears on the ipsilateral side and is malformed on the contralateral side. not, notochord; nt, neural tube; som, somite. Asterisks in F,G indicate the implanted microbead.

The quantitative analysis showed no significant differences of the mean myotome area among experimental groups (summary in Fig. 2E; mean somite areas: controls 47,000 ± 9,780 μm2, Noggin 56,000 ± 6,570 μm2, BMP-4 45,500 ± 14,400 μm2, BMP-4+Noggin 35,100 ± 3,540 μm2). Therefore, there appears to be no net growth of muscle tissue in response to BMP-4. The mean number of cells/myotome area in BMP-4- and BMP-4/Noggin-treated embryos, however, was significantly higher than in controls (Fig. 2E; mean cell number/myotome area (mm2): controls 4,850 ± 130, BMP-4 6580 ± 1340, BMP-4+Noggin 6270 ± 950; P < 0.05). In Noggin-treated animals, this parameter was not significantly different from that of controls (Fig. 2E; mean cell number/myotome area (mm2): controls 4,850 ± 130, Noggin 5,070 ± 760). These data show that the apparent BMP-4-induced increase in myotome width on the contralateral side is due to a shape change in myotome tissue rather than to tissue growth; therefore, the tissue mass is similar in both cases. This change in muscle shape correlates with an increase in cell number/myotome area, however, suggesting increased cell proliferation.

BMP-4-Coated Microbeads Stimulate Cell Division of Somite Cells

Above we describe two different effects of BMP-4 on myotome development: a loss of muscle tissue in the vicinity of the implanted BMP-4 bead and a change in myotome morphogenesis at some distance to the bead accompanied by an increase in the proliferation of muscle cells. The loss of muscle tissue on the ipsilateral side of embryos carrying the BMP-4 bead could either be due to apoptosis of differentiated cells or to dedifferentiation of prospective muscle cells that subsequently become incorporated into other tissues. Despite the use of several assays for the detection of apoptotic cells (DAPI staining, TUNEL-assay, immunostaining of caspase 3), we were not able to detect a significant stimulation of apoptosis 1, 2, or 3 days after BMP-4 bead implantation (data not shown). Therefore, we focused on the role of BMP-4 in regulating cell proliferation and cell differentiation during somite development. Normally, cells undergoing final differentiation exit the cell cycle (see Denetclaw et al.,1997; Vernon and Philpott,2003; Vernon et al.,2003). We wondered whether muscle progenitor cells in BMP-4-treated embryos also follow this rule and compared the development of muscle cells and their proliferation in embryos carrying one BMP-4 bead with that of untreated controls. Experimental and embryos were fixed at different time points between 4 hr and 4 days after bead implantation. Transverse vibratome sections through the midtrunk were stained with an antibody against the phosphorylated form of Histone H3 (mitosis marker, Hendzel et al.,1997) and the muscle-specific antibody 12/101.

As expected, differentiation of tissues including the myotomes in control embryos correlated with a decrease in the mitotic rate (Fig. 3A–E). In contrast, BMP-4-treated embryos retained higher rates of cell proliferation in somites and adjacent tissues (Fig. 3F–K). Higher proliferation was not restricted to tissues adjacent to the microbead on the ipsilateral side, but extended to areas on the contralateral side. This result indicates long-ranging effects of locally applied BMP-4. In addition, myotome development, as revealed by 12/101 immunostaining, appeared normal in embryos carrying BMP-4 beads until approximately 2 days following bead implantation. Anti-12/101 staining was clearly present on the ipsilateral side, but the myotome appeared to be thicker. Somite morphology appeared to be normal on the contralateral side (Fig. 3D,I). Between 2 and 4 days post-implantation, however, prospective muscle tissue was dramatically reduced on the ipsilateral side whereas the myotomes were increased in width and reduced in height on the contralateral side as described above (Figs. 1,2). These data demonstrate that the morphological changes induced by BMP-4 are correlated with changes in the regulation of cell proliferation.

BMP-4 Affects Dermomyotome Development in a Concentration-Dependent Manner

To strengthen the conclusion that BMP-4 mediates changes in somite differentiation, we analysed additional markers for the dermomyotome, which develops into muscle and dermis. The dermomyotome in vertebrates can be specifically identified via the expression of two transcription factors, Pax3 or Pax7 (Ben-Yair and Kalcheim,2005; Devoto et al.,2006). In axolotl adults, Pax7 is expressed in the dorsal aspect of the spinal cord (Schnapp and Tanaka,2005; Schnapp et al.,2005), and is also present in a thin cell layer between the epidermis and the myotome of stage 39 larvae. There it was termed “dermatome” (Sobkow et al.,2006), but probably represents the amphibian dermomyotome layer (Grimaldi et al.,2004; Devoto et al.,2006). To further characterize the putative axolotl dermomyotome, we analysed the expression of Pax7 in the somites and myotomes of untreated embryos and larvae (stages 25–43). We were interested in determining whether Pax7 is present in progenitor cells of muscle and dermis. At stage 25, Pax7 was expressed throughout early somites (Fig. 4A). Later, the staining became confined to outer somite regions, the presumed dermomyotome (stage 30–35, Fig. 4B,C). In stage-43 larvae, Pax7-positive cells were randomly distributed under the epidermis and scattered throughout the skeletal muscles of the trunk (Fig. 4D). In addition, the dorsal aspect of the neural tube was stained (Fig. 4D; see also Fig. 5A,B). In the tail muscles of a juvenile, Pax7-positive cells were observed adjacent to myofibres, a position characteristic for satellite cells (Fig. 4E-G). These observations confirm that Pax7 may function as a marker for embryonic dermomyotome and adult satellite cells.

Figure 4.

Distribution of Pax7-positive cells during normal dermomyotome development. Horizontal paraffin sections (A–C, notochord level; D, neural tube level) through embryos at stage 25 (A) and 30 (B), larvae at stage 35 (C) and 43 (D), and the tail of a 7-cm juvenile animal (E) were stained with antibodies against Pax7 (brown color). Pax-7-expressing cells are initially found throughout the somites (A), later become restricted to the lateral aspect of the somites (B,C), and finally form a thin dermomyotome layer (dm) between epidermis (epi) and myotome (my) (D). Single Pax7-positive cells are also found within the myotome (arrows in D). F,G: Close-ups of the insets indicated in E display putative satellite cells (asterisks). bm, basement membrane; epi, epidermis; mu, muscle fibre; my, myotome; not, notochord; nt, neural tube; som, somite.

Figure 5.

Influence of BMP-4 on dermomyotome development. Immunolocalisation of Pax7 (brown color) on cross-sections through untreated control larvae (A–C) and larvae after implantation of a BMP-4- (D–G) or a Noggin-loaded microbead (H–K) at stage 38. A: Untreated control, overview. Pax7 is detected dorsally in the neural tube (nt) and in the dermomyotome (dm). B: Close-up displaying Pax7 in the dorsal aspect of the neural tube. C: Enlarged view of the dermomyotome. Nuclei of Pax7-expressing cells are flattened except in the dorsomedial lip (DML) where they appear thicker. Note the Pax7-positive cells (arrows) in the myotome (my). D: Larva with a BMP-4 microbead (*), overview. Note the lack of Pax7-positive cells on the ipsilateral side and the thickening of prospective dermomyotome (dm) on the contralateral side. E: Close-up of Pax7 staining in the neural tube (nt). On the ipsilateral side, Pax7 staining is weaker and ventrally shifted, on the contralateral side Pax7 is expressed normally. F: The dermomyotome layer is thicker and the cell nuclei are not as flat as in controls (compare to C). G: Cross-section through another BMP-4-treated larva. In the vicinity of the BMP-4-coated microbead (*), myotome tissue appears to be replaced by connective tissue (arrows. H: Larva with a Noggin microbead (*), overview. Pax7 staining in the dermomyotome and the neural tube is reduced in response to ectopic Noggin. I: Close-up of the neural tube displays weak Pax7 signal in the dorsal half. K: Somitic Pax7 is confined to the ventral-most part of the dermomyotome (dm) and to single cells within the myotome (arrows). df, dorsal fin; en, endoderm; epi, epidermis; my, myotome; not, notochord.

Next, the distribution of Pax7 in transverse paraffin sections through the trunk of embryos carrying one BMP-4 or one Noggin bead was compared with that of sections through the trunk of controls (Fig. 5). In stage-38 controls, Pax7 was expressed in the dorsal neural tube (Fig. 5A,B) and in the dermomyotome (Fig. 5A,C). At the lateral aspect of the somite, the dermomyotome cells were flattened and formed a thin layer. However, this layer appeared thicker at the dorsomedial lip because the cells were more rounded (Fig. 5C). In addition, some Pax7-expressing cells were scattered in the myotome and at its ventral margin (Fig. 5C). On the ipsilateral side of BMP-4-treated larvae, Pax7 immunostaining was reduced significantly in both the dermomyotome and the neural tube (Fig. 5D,E). The myotome tissue appeared to be replaced by connective tissue (Fig. 5G) and neural tube staining was shifted ventrally (Fig. 5E). On the contralateral side, dermomyotome staining was enhanced and the cells were less flattened than in controls (Fig. 5F). Pax7 immunostaining was reduced both in the neural tube and dermomyotome of larvae containing one Noggin-coated bead that displayed the phenotype with an elevated dorsal fin (Fig. 5H–K). A few positive cells were also observed at the ventral margin of the dermomyotome and scattered within the myotome (Fig. 5K). In larvae displaying a weaker dorsal fin phenotype, the neural tube staining of Pax7 is normal but the dermomyotome is still reduced (data not shown), indicating that the dermomyotome is more sensitive to Noggin. The strongly reduced anti-Pax7 staining in the dermomyotome of Noggin-treated larvae may be explained by premature differentiation and emigration of dermomyotome cells to the myotome and to the dorsal fin.

From these observations, we conclude that in the axolotl embryo local application of BMP-4 disturbs and/or inhibits the development of the dermomyotome on the ipsilateral side and modulates the balance of myotome and dermomyotome development on the contralateral side. Assuming long-range diffusion of BMP-4 signals from the local bead source, these differences suggest that somite cells respond to different concentrations of BMP-4.

DISCUSSION

In this investigation, we analyzed the effects of ectopically administered BMP-4 and its antagonist, Noggin, on trunk development in the Mexican axolotl. Local overexpression of BMP-4 in the dorsal trunk following implantation of a protein-loaded microbead resulted in a decrease of dorsal fin formation, a change in pigment cell distribution, a lack of myotome and dermomyotome in the vicinity of the implanted bead, altered myotome morphology, and a thickened dermomyotome layer at some distance from the bead, on the contralateral side. Ectopic Noggin increased dorsal fin height and inhibited the dermomyotome layer. Here we focus on the effects of BMP-4 and Noggin on dorsal fin and somite development. The effects on pigment cell migration are analyzed in a separate report (Hess et al., in press).

BMP-4 and Dorsal Fin Outgrowth

We observed that ectopic BMP-4 inhibited dorsal fin outgrowth in a dose-dependent manner (Fig. 1C–E). In contrast, ectopic administration of the BMP-4 inhibitor Noggin stimulated dorsal fin outgrowth locally (Fig. 1G). These observations suggest that specific levels of BMP-4 are necessary for proper dorsal fin development.

Recent experiments in axolotl and Xenopus embryos revealed that the dorsal fin mesenchyme recruits from two tissues, neural crest and somites (Sobkow et al.,2006; Garriock and Krieg,2007). BMPs affect the development of both sources.

First, BMP-4 is known to play a role during premigratory stages of the neural crest and during neural crest delamination in chick embryos (Selleck et al.,1998; Sela-Donenfeld and Kalcheim,1999,2000; for a recent review see Knecht and Bronner-Fraser,2002). However, when applied later, ectopic BMP-4 inhibits neural crest delamination (Sela-Donenfeld and Kalcheim,1999). Because the dorsal fin epidermis is induced by the neural crest (Bodenstein,1952), it is possible that implanted of BMP-4-coated microbeads interferes with neural crest delamination, induction of dorsal fin epidermis and the formation and migration of neural crest mesenchyme into the dorsal fin.

Second, it has been reported that downregulation of BMP-4 is important for the normal development of the dorsomedial somite (Reshef et al.,1998; Linker et al.,2003), and that BMP-4 is secreted by the adjacent epidermis and the dorsal neural tube. Active repression of BMP-signaling is, therefore, needed to initiate muscle differentiation in the dorsomedial part of the somite, e.g., the BMP-antagonists Noggin and Follistatin that are expressed in this region (Amthor et al.,1996; Reshef et al.,1998). Because the somite is the second source of dorsal fin mesenchyme, a local downregulation of BMP-signaling may be important for the determination of dorsal fin precursor cells and their presumed migration from the somite to the fin. Taken together, these data indicate that BMP-4 signaling and its regulation by inhibitors such as Noggin may be required for the migration of dorsal fin mesenchyme precursor cells from neural crest and somites into the dorsal fin.

Pax7 and the Axolotl Dermomyotome

The dermomyotome in higher vertebrates can be specifically identified through the expression of Pax3 and/or Pax7, two paired-box transcription factors that are both required for normal myogenesis in amniote embryos (Pownall et al.,2002). In the chick, for example, Pax7 is a pan-somitic marker at epithelial stages that later becomes restricted to the dermomyotome (Ben-Yair and Kalcheim,2005), and Pax3- and Pax7-positive cells contribute to muscle and dermis (Ben-Yair and Kalcheim,2005). In the mouse, cells expressing Pax3 and Pax7 have been identified as muscle progenitor cells during development. If these cells fail to express Pax3 or Pax7, they die or assume a nonmyogenic fate (Relaix et al.,2005).

We used Pax7 as a marker to monitor dermomyotome development in axolotl embryos and larvae (Figs. 4 and 5). In stage-25 embryos, Pax7 is expressed in the lateral halves of the somites. Later it becomes restricted to a thin layer of cells between the outer periphery of the myotomes and the epidermis that most likely represents the axolotl dermomyotome.

In nonamniote vertebrates and amphioxus, a dermomyotome (sensu amniotes) was previously regarded as nonexistent. In amphibian, for example, a single cell layer at the periphery of the somite was referred to as the dermatome until recently (Nieuwkoop and Faber, 1967; Hamilton,1969; Hausen and Riebesell,1991; Keller,2000; Sobkow et al.,2006), and was not regarded as segmented (Hamilton,1969). Recent experiments in Xenopus, however, suggest a closer homology between this cell layer and the segmented dermomyotome of amniotes (Grimaldi et al.,2004). In fact, several studies performed in amphioxus, lamprey, shark, and various bony fishes (from sturgeon to zebrafish) imply the existence of an evolutionarily conserved chordate dermomyotome (Kusakabe and Kuratani,2005; Devoto et al.,2006; Scaal and Wiegreffe,2006).

We also observed single Pax7-positive cells scattered in the myotomes of stage-38 larvae and throughout the tail muscle of a 7-cm-long juvenile. Here, the cells are located at the periphery of the myofibers, a location highly reminiscent of muscle satellite cells, which are a population of small mononuclear cells beneath the basement membrane of adult myofibres (Mauro,1961; Seale and Rudnicki,2000; Seale et al.,2000; Halevy et al.,2004; Chen et al.,2006). These cells are essential for muscle repair and regeneration in adults because they are self-renewable and capable of differentiating into muscle fibres (Gargioli and Slack,2004; Chen et al.,2006; for a recent review, see Wagers and Conboy,2005). In chicken, satellite cells have been shown to be derived from the dermomyotome (Gros et al.,2005). Taken together, Pax7 appears to mark cells with a conserved role during somite development, which are responsible for the maintenance of adult muscle stem cells in vertebrates.

BMP-4 and the Dermomyotome

We observed that implantation of BMP-4-coated microbeads after mesoderm induction and initial mesodermal patterning in axolotl does not influence the early steps of myogenesis (Fig. 3). This is consistent with previous data in the chicken where myogenesis was shown to be a mesoderm-autonomous process (Linker et al.,2003). Later, however, BMPs and BMP-antagonists may influence maintenance and differentiation of dermomyotome-derived muscle precursor cells. Our observations on somite development in axolotl embryos under the influence of ectopically applied BMP-4 support this hypothesis. In the vicinity of implanted e.g., that BMP-4-coated microbeads, muscle development commences normally, but becomes gradually compromised at 2–5 days after bead implantation, and the original muscle tissue may be replaced by other tissues (Figs. 1,3,5). At some distance from the bead, both on the ipsi- and contralateral side, however, we observed enhanced cell proliferation, a thicker Pax7-positive dermomyotome layer, and more differentiated muscle cells per myotome area (Figs. 1–3,5). Finally, implantation of Noggin-coated microbeads caused a strong reduction of the Pax7-positive dermomyotome layer (Fig. 5).

One possible model derived from these data and the current knowledge on amniote myogenesis is the following: Moderate amounts of BMPs induce Pax3 and/or Pax7 expression. Pax3/7-positive cells are then kept in an undifferentiated, proliferative state and serve as a pool of prospective muscle precursor cells (Hirsinger et al.,1997; Marcelle et al.,1997; Reshef et al.,1998; Amthor et al.,1999). BMP-antagonists such as Noggin and Follistatin downregulate BMP-activity, which results in the downregulation of Pax3/7 and the upregulation of muscle differentiation markers such as MyoD and Myf5 (Reshef et al.,1998; Amthor et al.,1996,1999). Eventually the cells differentiate and are withdrawn from the cell cycle (Denetclaw et al.,1997). In this way, new prospective muscle cells are recruited from the dermomyotome and enter the developing myotome in a continuous manner. In contrast, high levels of ectopically applied BMP-4 reduce expression of Pax3/7 and ultimately compromise muscle formation (Pourquié et al.,1995,1996; Reshef et al.,1998). According to this model, the Noggin-induced inhibition of Pax7-staining in the dermomyotome (Fig. 5) may be attributable to premature differentiation of dermomyotome cells, which then migrate to the myotome and dorsal fin. These observations suggest that specific levels of BMPs are necessary for further muscle formation or maintenance beyond the initial development of the myotome that has formed autonomously from myogenic mesoderm.

Is There a Connection Between Dorsal Fin and Muscle Development?

We have shown that BMP-4 can affect such developmental processes as pigment cell migration, dorsal fin formation, and muscle development, and we have demonstrated that differences exist in the response of pigment cell migration vs. dorsal fin/myotome formation (Fig. 1; see Hess et al.,2007). Interestingly, BMP-4 affects the dermomyotome layer (Fig. 5), which is a source of both muscle precursor cells for further myotome growth and of dorsal fin mesenchyme (Scaal and Wiegreffe,2006; Sobkow et al.,2006; Garriock and Krieg,2007). We speculate that the BMP-induced changes in the dermomyotome could at least partially represent the common developmental basis for the fin and muscle phenotypes observed in our study, suggesting a link between these two developmental processes. Further studies are needed to substantiate this hypothesis.

EXPERIMENTAL PROCEDURES

Embryos

Wild-type (dark, D/-) and white mutant (d/d) embryos of the Mexican axolotl (Ambystoma mexicanum) were obtained from the former axolotl colony in Bloomington, IN, or from our own colony in Dresden, Germany. The embryos were kept in tap water at room temperature or at 7–8°C. They were staged according to the normal table of Bordzilovskaya et al. (1989).

Loading and Implantation of Microbeads

Microbeads were handled and coated with proteins as described for operations in the chick (Francis et al.,1994; Ganan et al.,1996; Schneider et al.,1999; Honig et al.,2005). White heparin-acryl microbeads (Sigma) were selected at a diameter of 80–100 μm and blue agarose beads (Affi-Gel R; Biorad) at 60–80 μm. Before loading with protein, beads were washed in distilled water containing antibiotics overnight at room temperature. White beads were soaked in human BMP-4 protein (25, 50, or 100 μg/ml in saline; Genetics, Institute, Cambridge, MA; R&D Systems, Wiesbaden, Germany). Blue beads were soaked in Xenopus Noggin protein (100 μg/ml saline; gift of Richard Harland, Berkeley, CA). Incubation with protein was carried out overnight at 4°C and for 1 hr at room temperature before use. For controls, uncoated beads were implanted.

Control and protein-coated microbeads were implanted into embryos at stages 24–27 (at least 5 embryos in each group received implants). Prior to implantation, the epidermis over the dorsolateral trunk was lifted in the area of the 5th to 10th somite using tungsten needles. The dorsalmost part of the 7th or 8th somite was removed and one white or blue microbead (coated or uncoated) was inserted into the free space. Then the epidermis was folded back and pressed against the somites with a glass bridge for 5 to 10 min to assist healing. In some experiments, a combination of one coated white bead (anteriorly, 7th somite) and one coated blue bead (posteriorly, 8th somite) was implanted. Embryos were allowed to develop for up to 5 days after the operation.

Histology and Immunostaining

For conventional histology, control embryos and embryos with implanted microbeads were fixed in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS) at least overnight, washed in PBS, dehydrated in a graded series of ethanol, and embedded in Technovit 7100 (Heraeus-Kulzer, FRG). Semithin sections (4 μm) were stained with methylene blue.

Vibratome and paraffin sections were used for immunostaining.

Vibratome sections.

Embryos were prefixed in 4% PFA/PBS and postfixed in Dent's fixative (Dent et al.,1989), each overnight. Fixed embryos were stored in methanol at −20°C until use. For recording muscle differentiation during somite development, vibratome sections (100 μm) were stained with a primary muscle specific antibody (12/101; Kintner and Brockes,1984) and a Cy3-conjugated goat-anti-mouse secondary antibody (Dianova). For the analysis of cell proliferation during somite development, vibratome sections were simultaneously stained with 12/101 and an antibody against the phosphorylated form of Histone H3 (Hendzel et al.,1997; mitosis marker, Upstate). For detection, Cy3-conjugated goat-anti-rabbit (Dianova) and Alexa 488-conjugated goat-anti-mouse antibodies (Molecular Probes) were used. Some sections were counterstained with DAPI. One to two sections (comprising the bead area) were analysed for each embryo/larva.

Paraffin sections.

Embryos were fixed in 4 % PFA/ PBS, dehydrated in a graded series of ethanol, and embedded in paraffin. Sections (5 μm) were deparaffinized and rehydrated. Prior to immunostaining, sections were heated in a microwave oven for 2 × 5 min to unmask antibody binding sites. Sections were stained with mouse anti-Pax7 (Kawakami et al.,1997; Developmental Studies Hybridoma Bank), which was raised against chicken Pax7 and cross-reacts with axolotl protein (Schnapp and Tanaka,2005; Schnapp et al.,2005), and with anti-caspase 3 (1:50; New England Biolabs). For the detection of primary antibody, a biotinylated secondary antibody and a tertiary streptavidin-peroxidase complex (Vectastain ABC Elite-Kit) were used. DAB served as a substrate for the peroxidase. Sections were counterstained with Mayer's haematoxylin.

Image Acquisition and Analysis

Brightfield images of whole embryos were photographed with a stereo microscope to which a digital camera was attached. Images of stained plastic sections and Pax-7-immunostained paraffin sections were photographed with a standard microscope to which a Spot RT camera (Visitron) was attached. Fluorescent images of immunostained sections (anti-12/101 staining for muscle cells, anti-phospho-Histone3 staining for mitotic cells, DAPI for cell nuclei) were captured with an epifluorescence microscope and the Spot RT camera. The fluorescent images were merged with Visitron software. Images were processed and combined using Adobe Photoshop (version 7) and Macromedia Freehand (version 8).

Acknowledgements

The authors are grateful to Drs. Richard Harland and Elly Tanaka for the kind donation of reagents, Marc Angermann for evaluating experiments, S. Bramke for excellent technical support, and Sarah Cramton for critically reading the manuscript. The work was supported by a DFG-grant to H.H.E. (EP8/7-1).

Ancillary