|Home | About | Journals | Submit | Contact Us | Français|
The predominant source of myogenic cells in vertebrates is the dermomyotome (DM). In teleost fish, recent research has provided a useful but limited picture of how myogenic precursors originate from the DM and how they develop into muscle fibers. Here, we combine detailed morphological analysis with examination of molecular markers in trout to describe the cellular mechanisms by which the lateral fast muscle growth zone is created during second phase myogenesis. Results suggest that this occurs by lateral-to-medial immigration of myogenic cells de-epithelializing from the posterior DM lip. These cells then appear to stop proliferation and migrate anteriorly to finally differentiate into muscle fibres. This seems to be a continuation of the rotational cell movement that creates the teleost DM during early somite development. These findings suggest an evolutionary conserved role of the posterior DM lip in amniotes and fish.
The dermomyotome is a transient structure of the vertebrate somite. Being formed as a dorsolateral epithelial sheet during somite maturation, the dermomyotome has been identified as the most important turntable of mesodermal cell fate choice in the amniote embryo, giving rise to a variety of tissues, most importantly myotomal muscle and dermis (Stockdale et al., 2000; Tajbakhsh, 2003; Scaal and Christ, 2004).
Modern research on somite differentiation using fish progressively indicates that patterns of muscle development in these animals are in many aspects similar to those in the amniotes. This all the more since evidence has been obtained for the existence of a dermomyotome-like epithelium in the fish somite (Devoto et al., 2006). This structure is briefly referred to here as the dermomyotome (DM), although proof for its dermogenic role in fish is still rather incomplete (Hollway et al., 2007; Dumont et al., 2008). Recent work in the zebrafish has demonstrated that the teleost DM layer arises from the anterior compartment of the somite by a process of cell rearrangement (Hollway et al., 2007; Stellabotte et al., 2007). Once established, the teleost DM has been shown to generate myogenic precursors for the expansion of the pre-established primary myotome. It has been shown that during the second phase of myogenesis, myogenic precursors detaching from the DM account for stratified hyperplasia at the dorsal and ventral extremes of the myotomes (Steinbacher et al., 2006, 2007). Cells deriving from the DM have also been shown to contribute to muscle growth at the lateral boundaries of the fast muscle domains (Hollway et al., 2007; Stellabotte et al., 2007), but knowledge as to how this exactly occurs is as yet missing. The cell movements which transfer myogenic precursors from the DM into the myotome have been extensively studied in amniotes. There seems to be agreement that the cells involved derive from all four lips of the DM, although some controversy remains over the relative contribution of these lips and the specific migration paths taken by these cells (e.g., Denetclaw et al., 2001; Kahane et al., 2002; Gros et al., 2004).
In the present work, we investigate the formation of lateral fast muscle cells from the dermomyotome in the brown trout (Salmonidae, Teleostei). This species grows to a large adult size and embryonic development is slow, thus providing optimum definition of cellular patterning. Additionally, the salmonid DM consists of numerous large cuboidal cells that facilitate morphological study. We use light microscopy of semithin sections and transmission electron microscopy (TEM) combined with in situ hybridization (ISH) and double-immunostaining for Pax7 (a marker of DMderived muscle precursors) and markers of muscle cell differentiation (Myogenin, MEF2) to further determine the patterns of myogenic cell detachment from the DM layer during the second phase of myogenesis (stratified hyperplasia). Our results provide an important extension to the existing knowledge of the role of the teleost DM during this growth phase and are discussed in relation to the mechanisms that create the amniote (specifically avian) myotome.
During the period of somitogenesis, somites at their earliest phase of differentiation (e.g., those in the anal area of a 40 somite embryo, i.e., approx. somite X, considering somite I as the most recently formed) are already characterized by the feature that the cells of the adaxial walls are morphologically clearly different from all other cells. They are elongated, dorsoventrally flattened, and mononucleated with only lightly staining nuclear chromatin. All other cells of these somites have more darkly staining nuclei and are rounded to polygonal without major morphological difference (Fig. 1A). By contrast, molecular evidence demonstrates that the wall cells of these somites are heterogeneous. The posterior wall cells lateral to the adaxial cells stain for MEF2, as do the adaxial cells themselves. Anterior wall cells, by contrast, stain for Pax7 (Fig. 1B,C).
Somites at a slightly more developed state (e.g., from midtrunk of a 40 somite embryo, i.e., approximately somite XX) are almost evenly divided into a medial and a lateral compartment. The medial compartment, representing the nascent myotome, consists of light-colored elongated cells with light-colored nuclei. These cells most likely constitute a mixture of more advanced slow fibers which already contain small arrays of myofilaments and less differentiated fast fibers (cf. Steinbacher et al., 2007). Preferentially in the posterior half of these myotomal compartments, the differentiating fibers usually alternate with cells containing more condensed cytoplasm and dark nuclei. Such cells are presumably fast fiber precursors that are undifferentiated in the sense that they still show instances of mitotic division and have not entered myofibrillogenesis (Fig. 2A). The lateral compartments of these somites most probably includes the DM as defined by Devoto et al. (2006). They consist of an epithelium of more darkly staining, mitotically active cells that are generally devoid of myofibrils and unreactive to markers of myogenic differentiation (cf. Steinbacher et al., 2007). Following the curved somite contours, this epithelium constitutes not only the lateral surface of the somite but also part of its anterior and posterior surfaces. Anteriorly, the DM epithelium projects far medially toward the notochord and has a distinct border toward the myotome. The posterior DM, however, frequently appears less clearly separated due to the occurrence of darkly staining (DM-type) cells in the underlying posterior part of the myotome (see above; Fig. 2A).
As the somites further differentiate (anal area of 50 somite embryo), the described two-part architecture of the somite is at first maintained, though with a clear shift in volume ratio toward the myotome, particularly due to the establishment of a multilayered fast muscle domain. Mononucleated slow fibers now form a single layer at the lateral surface of the myotome, covering a bulk of less differentiated fast muscle cells. Toward the epidermis, the slow fibers are tightly overlaid by a continuous cuboidal DM epithelium (Fig. 2B). This epithelium is less curved than in the previous stage while the posterior halves are multilayered and appear “turned back” toward the underlying myotome. Additionally, some cells of these halves now seem to have de-epithelialized and “dripped down” from the DM to shift toward the centre of the fast muscle while still being able to undergo mitotic division. Such cells are occasionally located directly at the slow fiber level, suggesting that they pass underneath between these fibers (Fig. 2B). By contrast, the anterior halves of the DM are monolayered and show no indication of cell detachment. The anteroposterior transition from monolayer to multilayer shape within the DM is also clearly evident on transverse sections, as are features of probable DM cell detachment toward the myotome (Fig. 3A; for positional information see Fig. 3B). Immunolabeling indicates that almost all DM cell nuclei are Pax7+. Only a few nuclei of the posterior DM also stain for markers of muscle differentiation such as Myogenin (Mgn) and MEF2 (Fig. 4A,B). Horizontal sections particularly demonstrate that these Pax7+/Mgn+ cells are located at positions that correspond to the inwardcurved part of the posterior DM lip (cf. Figs. Figs.2B,2B, ,4A).4A). The additional presence of many Pax7−/Mgn+ cells in the muscle area beneath the DM indicates the formation of the lateral fast muscle growth zone that has been shown to occur in the embryos of these fish (cf. Steinbacher et al., 2006, 2007).
In embryos from the end of somite formation until hatching, the DM cells in pre-anal somites become increasingly flattened but continue to divide mitotically. Only the posterior lips usually still comprise two or more stacked cells, also with occasional mitotic figures. At these sites, cells that are structurally similar to the DM cells are found inserted between the slow fibers or immediately underneath (Figs. 2C–E, ,3C).3C). Sagittal sections running precisely through the plane of the slow fiber myofibrils further confirm that almost all inserted cells are located next to the posterior end of the myotome (Fig. 3D). Only rarely, such cells are found at the anterior myotome end. This suggests that cells further de-epithelialize from the posterior DM lip, now becoming directly delivered to the posterior part of the lateral fast muscle growth zone (Fig. 2C–E). The function of the posterior DM lip as the main source of myogenic cells destined to be fed into this growth zone appears to be present along the entire dorsoventral extension of the somite (Fig. 2F). As in the previous stages, the anterior DM edge shows no sign of such cell detachment. Pax7+ cells are detected within the DM but also in the myotome, where they are preferentially located next to the posterior end of the lateral fast muscle growth zone (Figs. 4C,D, ,5B;5B; anteroposterior localization of labeled cells within the myotome explained in Figs. Figs.3B,3B, ,5A).5A). Comparison of in situ hybridization and immunostaining results further suggest that only cells at this site are still transcribing the pax7 gene (Fig. 4E). Occasionally, Pax7+ nuclei are also inserted directly between two fibers of the slow muscle layer, even at positions next to the anterior DM lip (Fig. 4F). Double-immunolabeling for Pax7 and the muscle differentiation marker Mgn further shows that most of the Pax7+ cells at the posterior end of the lateral fast muscle are Mgn−, while more anteriorly located Pax7+ cells are mainly Mgn+. Similarly, Pax7−/Mgn+ cells appear to occur preferentially in the anterior and central areas of the lateral fast muscle (Figs. 4C,D, ,5B).5B). These observations are confirmed by quantitative data from brown trout embryos at 4 weeks before hatching (Fig. 5A-C) revealing that numbers of Pax7+/Mgn− cells in the posterior quarter (zone 4) of the lateral fast muscle are indeed significantly higher than those in the all other quarters (zones 1–3; p < 0.01). Zone 3 still contains more Pax7+/Mgn− cells than the more anterior zones 1 (p = 0.06) and 2 (p = 0.02). Similarly, Pax7+/ Mgn+ cells are clearly more abundant in the posterior zones 3 and 4. By contrast, numbers of Pax7−/Mgn+ cells are almost equal for zones 1 to 3, but significantly lower in the posterior zone 4 (p < 0.01). Labeling for proliferating cell nuclear antigen (PCNA) reveals that only the posterior lateral fast muscle comprises large numbers of proliferative cells (Fig. 4G).
The present work provides evidence that de-epithelialization of myogenic cells from the posterior lip of the DM is the main mechanism to promote stratified fast muscle growth during the second phase of teleost myogenesis. This is particularly obvious in trout which for a long time maintain a cuboidal DM epithelium. Horizontal sections clearly suggest that within each trout somite, a “flow” of undifferentiated (mitotically active) cells is released from the medial part of the multilayered posterior DM lip (Fig. 2A–E). Detached cells appear to pass directly underneath between the superficial slow fibers (Fig. 2C,D; see also sagittal section in Fig. 3D) to invade the fast muscle domain without entering the intersegmental gaps. Accordingly, cells that resemble those assumed to detach from the DM are scattered over the lateral part of the fast muscle (Fig. 2B–E) where previous studies have shown the formation of an important zone of stratified hyperplasia (Rowlerson and Veggetti, 2001; Steinbacher et al., 2006, 2007). The specific accumulation of DM-like cells in the posterior lateral fast muscle is also clearly evident on transverse sections, as is the occasional insertion of such cells between the overlying slow fibers (Fig. 3A,C; for positional information see Fig. 3B). From the present observations it appears plausible, that similar cell detachment also occurs at the anterior DM lip, although only rarely.
Considerable support for this morphology-based interpretation is provided by cell labeling for the DMmarker Pax7 and markers of myogenic differentiation (Mgn, MEF2). The emergence of Mgn expression in some Pax7+ cells of the inward-curved part of the posterior DM lip indicates the onset of myogenic differentiation in these cells (Fig. 4A). The common presence of Pax7 in still Mgn− cells within the posterior lateral fast muscle (Figs. 4C,D, ,5B)5B) strongly supports the idea that such cells are directly derived from the medial portion of the posterior DM lip. From the combined immunostaining/in situ hybridization evidence it appears likely that these cells, after lateral-to-medial transfer into the immediately underlying myotome and ongoing proliferation, spread over the entire lateral fast muscle domain. They thereby stop proliferation (Fig. 4G), down-regulate Pax7 and up-regulate Mgn to finally differentiate into fast fibers (Figs. 4A,C–E, ,5B,5B, ,6C).6C). This interpretation is further substantiated by quantitative analysis in prehatching embryos showing that most Pax7+ cells in the posterior part (zone 4) of the lateral fast muscle are Mgn−, while the more anterior parts (zones 1–3) contain more Pax7+/Mgn+ than Pax7+/Mgn− cells, and the vast majority of all Pax7−/Mgn+ cells (Fig. 5C).
These expression patterns are in agreement with the suggested role of Mgn as a down-regulator of Pax7 expression in differentiating murine muscle precursors (Olguin et al., 2007). The proposed translocation of myogenic cells from the posterior lip of the DM into the posterior myotome and the subsequent anterior spread of these cells within the lateral fast muscle may be interpreted as a direct continuation of the rotatory movement of DM precursors from the anterior part of the somite to its lateral surface, as recently described in the zebrafish (Hollway et al., 2007; Stellabotte et al., 2007).
The interpretation that only the medial cells of the posterior fish DM are instantly destined to become muscle fibers would also fit with the assumption that asymmetric cell divisions perpendicular to the orientation of the amniote DM generate medial cells committed to myogenesis and lateral cells for DM self-renewal (Cossu and Tajbakhsh, 2007).
In a wider evolutionary context, these findings allow us to hypothesize that the cellular mechanisms of myotome formation are more similar for teleost fish and amniotes than previously recognized. In each case, myogenic cells from the DM lips contribute essentially to early myotome formation. In the amniotes, this process has been shown to involve all four lips of the DM, i.e. its dorsal, ventral, anterior/cranial and posterior/caudal lips (e.g., Kahane et al., 2002; Gros et al., 2004) already in the time before the myotome receives myogenic cells from the dissociating central DM (Buckingham, 2006). However, the exact immigration routes of the DM cells into the myotome are still a matter of debate. While some authors (Kalcheim and colleagues) suggest that these cells originate from all four lips of the dermomyotome but enter the myotome only from the rostral and caudal lips (Kahane et al., 2002), other authors (Ordahl and colleagues) suggest that such cells originate and immigrate only from the dorsal and ventral lips (Denetclaw et al., 2001).
In teleost fish, evidence to date has been largely confined to myogenic cell supply to the myotome from the dorsal and ventral lips of the DM (Steinbacher et al., 2006, 2007). Just as in the chick (Denetclaw et al., 2001), these lips have been shown to contribute to the incremental growth of the epaxial and hypaxial muscle domains, respectively. Considering the evidence here, a similarly important role may now be attached to the posterior lip of the teleost DM, which is likely the main “growth engine” for lateral fast muscle growth. This finding expands upon the recent interpretation of Pax7 single-labeling data in zebrafish (Stellabotte et al., 2007). These authors have suggested that DM-derived Pax7+ myogenic precursors enter the myotome directly, by moving medially between the slow fibers along their length but not around their ends. The present work basically confirms this interpretation. However, with the results from double-labeling for Pax7 and Mgn taken into account, a more diversified picture comes to the fore showing that most Pax7+ cells next to the posterior end of the lateral fast muscle are at a lesser state of myogenic development than those at more anterior positions (Figs. (Figs.4,4, ,5).5). This together with the morphological observations (Figs. (Figs.11--3)3) leads to the suggestion that most of these myogenic precursors originate from the posterior DM lip, confining their lateral-tomedial movement between the slow fibers to the area next to the posterior termination of these fibers. A similar mechanism may be at work in zebrafish.
The suggested role of the posterior lip of the fish DM as the main source of new lateral fast fibers in second phase myogenesis bears resemblance to cell immigration from the posterior DM lips according to the Kalcheim model of avian myotome formation (Kahane et al., 2001). However, in contrast to this model, the present work provides only little morphological indication that a similar cell detachment also occurs from the anterior DM lip. Nevertheless, the rare occurrence of Pax7+ cells inserted between the slow fibers immediately adjacent to this lip suggests that at least a small subset of muscle precursors enter the fast muscle at this site (Fig. 4F). Whether such cells are in any way different from those that immigrate from the posterior lip is unknown.
Thus, the DM lips of teleosts and amniotes may indeed have similar functions in early myotome formation, although with some variations, e.g., a reduced importance of the anterior DM lip in the teleosts, and cells being added dorsally, ventrally, and medially to the myotome in the chick (e.g., Gros et al., 2004), but dorsally, ventrally and laterally in the teleosts (Fig. 6A–C). A similarly conserved nature may also apply to a 3rd phase of cell emigration from the DM to establish the progenitors of adult satellite cells (Hollway et al., 2007).
Investigations were carried out on embryos of brown trout Salmo trutta lacustris. All fish used in the study were laboratory reared from artificially inseminated eggs. Animals were kept under rising temperature regimes (6.5–8°C ± 0.5°C). Rates of water flow and recirculation were kept constant throughout the experiment. Samples (n ≥ 6) of successive developmental stages were taken within the period from 50% epiboly in the embryo to the hatching stage. All animals were over-anesthetized with MS-222 (Sigma) and, before fixation, larger individuals were cut into smaller pieces to allow sufficient fixative penetration.
For ISH and immunolabeling of cryostat sections, animals were fixed and sectioned as described by Stoiber et al. (2002). All sectioning was confined to positions from mid-trunk to just posterior of the anus. For ISH, plasmids with cDNA encoding for trout pax7 were used to synthesize digoxigeninlabeled probes (protocols in Steinbacher et al., 2006). Primary antisera used for immunostaining were monoclonal mouse anti-chicken Pax7 IgG1 (1:20; DSHB), monoclonal mouse antimouse PCNA IgG2a (1:100; Santa Cruz), polyclonal rabbit anti-rat Myogenin (1:100; Santa Cruz), polyclonal rabbit anti-human MEF2 (1:100; Santa Cruz), monoclonal mouse antipig Desmin IgG1 (1:50; Sigma). Alexa 488-conjugated goat anti-mouse IgG2a (1:800; Molecular Probes), Alexa 488conjugated goat anti-rabbit (1:800; Molecular Probes), and Alexa 546-conjugated goat anti-mouse IgG1 (1:800; Molecular Probes) were applied as secondary antibodies.
For light microscopy (LM) of semithin sections and for TEM, animals were fixed in cacodylate-buffered Karnovsky-type fixative, postfixed in 1% osmium tetroxide, and embedded into epoxy resin. Sections (LM: 1.5 μm, TEM: 80 nm) at transverse, horizontal, and vertical (sagittal) planes were cut on a Reichert Ultracut S microtome and mounted and stained using standard techniques (Stoiber et al., 1998).
Photographs of the results were taken using a Reichert Polyvar microscope, also adapted for fluorescence microscopy, and a Zeiss EM 910 electron microscope.
Quantitative assessment of anteroposterior distribution of Pax7+ and Mgn+ lateral fast muscle cells was undertaken on double-immunolabeled 10 μm anal area cross sections of brown trout embryos at 4 weeks before hatching. Lateral fast muscle areas delimited by two successive myosepta were quartered into 4 zones (Fig. 5A). Numbers of Pax7+/Mgn−, Pax7+/Mgn+, and Pax7−/Mgn+ cells were recorded for each zone. A total of 32 myotomes from 6 individuals was examined. Nonparametric MannWhitney tests were used to analyze the data.
The authors thank Hans-Peter Gollmann, Institute of Water Ecology, Fisheries Biology and Lake Research, Scharfling, Austria, for rearing the fish, and Ben F. Koop, University of Victoria, Canada, for supply of cDNAs. Adda Maänhardt and Synnöve Tholo, University of Salzburg, Austria, provided excellent technical support.
Grant sponsor: Austrian Science Foundation (FWF); Grant number: P20430; Grant number: P16425.