|Home | About | Journals | Submit | Contact Us | Français|
Muscle development in teleost embryos has been shown to depend on myogenic cell recruitment from the dermomyotome (DM). However, little is known as to the cellular mechanisms that account for myotome growth after the dissociation of the DM. Here we combine immunolabeling for cell-specific markers with quantitative analysis to determine the sources and patterns of activation of myogenic cells in pearlfish larvae. Results demonstrate that appearance of mitotically active myogenic precursors inside the myotome coincides with the dissociation of the DM. Such cells are preferentially aggregated within the posterior lateral fast muscle. We therefore propose a growth model in which a pool of proliferative DM-derived precursors transferred to the posterior lateral fast muscle functions as an important source of myogenic cell spread to carry forward stratified fast muscle hyperplasia. This indicates that postembryonic teleost muscle growth includes a cellular mechanism that has no direct equivalent in the amniotes.
Muscle development and growth in vertebrates is a multiphase process that lasts long into fetal/larval life and depends on the recruitment of different lineages of myogenic cells. The predominant source of myogenic cells in the embryo is the dermomyotome (DM), a transient epithelial structure of the somite (Buckingham, 2006; Stellabotte and Devoto, 2007; Bryson-Richardson and Currie, 2008). This is well established for amniotes, and has recently been also demonstrated for fish (Devoto et al., 2006). In the teleost embryo, the DM arises from the anterior compartment of the somite by a process of cell rearrangement and covers the lateral surface of the primary myotome that is created during the first phase of myogenesis (Hollway et al., 2007; Stellabotte et al., 2007). It constitutes the main (and possibly exclusive) source of the myogenic precursors that subsequently account for the expansion of the primary myotome (Steinbacher et al., 2006, 2007; Hollway et al., 2007; Stellabotte et al., 2007; Stellabotte and Devoto, 2007). In this context, we have recently shown that the posterior lip of the DM is the main source of myogenic cells that promote second phase stratified fast muscle growth in the trout embryo (Steinbacher et al., 2008). The lateral-to-medial translocation of these cells appears to be a continuation of the above-described rotatory movement of DM cells from the anterior to the lateral part of the somite.
Less is known about myogenic cell recruitment in the time when the DM is no longer capable to ensure continuation of myotome growth after its disappearance at around the beginning of the larval period. It is long established that teleost muscle growth during the larval and juvenile periods depends upon the proliferation and differentiation of precursor cells that are located inside the myotome (Rowlerson and Veggetti, 2001). These cells are either recruited during continued (second phase) stratified hyperplastic growth or account for third phase mosaic hyperplastic growth. Cells for stratified lateral fast muscle growth have been reported to arise from a “germinal layer” or proliferation zone located just under the superficial slow muscle layer (Rowlerson and Veggetti, 2001). The lateral fast muscle area along its entire dorsoventral extent is certainly the main growth engine for early larval myotome expansion before mosaic hyperplasia reaches relevant levels. However, the circumstances by which this pool of proliferative myogenic precursor cells emerges have remained enigmatic. A similar level of uncertainty is attached to the possible relationships of these lateral fast muscle precursors with the DM and the later-emerging precursors of mosaic hyperplasia.
In the present work, we use embryos and larvae of the pearlfish Rutilus meidingeri to examine myogenic cell proliferation and differentiation in the time during and after the dissociation of the DM. Pearlfish fry grow much slower but to a larger size than those of the species' widely used cyprinid congener, the zebrafish. This facilitates morphological study and offers optimum definition of cellular patterning. We combine light microscopy of semithin sections and transmission electron microscopy (TEM) with double-immunostaining for Pax7 (a marker of DM-derived muscle precursors) and Myogenin (Mgn, a marker of muscle cell differentiation) or Phospho-Histone H3 (H3P, a marker of mitotically dividing cells; Hammond et al., 2007) to screen for myogenic cells that promote fast muscle stratified hyperplasia. Results provide evidence for a mechanism of DMderived myogenic precursor translocation and intermittent storage that has no equivalent in amniotes.
In embryos shortly before hatching (e.g., at 10 days postfertilization [dpf], 3 days before hatching), the myotomes in preanal somites are covered by a continuous DM epithelium comprised of flattened cells (Fig. 1A). Immunolabeling indicates that almost all nuclei of the DM are positive for Pax7 (Fig. 1B). Only rarely, Pax7+ cells are present inside the myotome. In such cases, the inside Pax7+ cells are all located next to the posterior end of the lateral fast muscle and stain also for Mgn, a marker of myogenic differentiation (Fig. 1C, for positional information see Fig. 2D,E). At this developmental stage, no H3P+ (i.e., mitotically active) cells were found inside the myotome.
Fish at hatching (13 dpf) do no longer exhibit a continuous DM epithelium. In the central areas of the somites, only a few scattered squamous DM cell-like cells have remained at the former DM position (not shown). More frequently, such cells are found at the former DM lips, particularly at the posterior lip (Fig. 2A). These cells are all morphologically undifferentiated in the sense that they exhibit a high nucleus/cytoplasm ratio, condensed nuclear chromatin and lack of myofibrils. Even at the dorsal and ventral boundaries of the somites, such undifferentiated cells are preferentially clustered at the posterior ends, indicating that their nonuniform distribution appears to be present along the entire dorsoventral extension of the somite (Fig. 2B). Specifically horizontal sections demonstrate that cells of such kind also bend inward toward the lateral fast muscle at the posterior myotome edges (Fig. 2A), thus coming into immediate contact with the lateral fast muscle growth zone. This zone is known to strongly foster myotomal muscle increase in the larval period (Rowlerson and Veggetti, 2001; Steinbacher et al., 2006). Sagittal sections running precisely through the plane of the slow fiber myofibrils provide further indication about how these small undifferentiated cells could enter the myotome. Thorough examination of such sections from 5 individuals (on average analyzing 12 myotomes per individual) revealed that all of such cells detected (n = 55) were exclusively inserted between the posterior ends of the slow fibers, but absent from the myoseptal gaps (Fig. 2C). This provides good morphological support for the idea that DM cell entry to the myotome by direct passage between the posterior slow fiber ends, as already described for the intact DM that exists during the embryonic period (Steinbacher et al., 2008), continues to occur until the terminal dissociation of the DM in the posthatching period.
Morphological analysis is strongly supported by molecular evidence from immunostaining using the DM-marker Pax7. Pax7+ nuclei are found at all positions where the above undifferentiated cells reside (Fig. 2F,G) indicating that most of these cells are indeed residual DM cells. In contrast to the prehatching stage, Pax7+ cells are now also occasionally found inside the myotomes, where they preferentially occur in the posterior quarter (zone 4) of the lateral fast muscle (p < 0.01, Fig. 2F,H). Labeling for mitotic activity with H3P reveals that mitotically active cells have appeared at the same position (Fig. 2G). Almost all of these H3P+ cells are also Pax7+. Quantitative analysis of the presence of H3P and Mgn in the Pax7+ cells of zone 4 (at approximately 32% and 50%, respectively) demonstrates that these intramyotomal Pax7+ cells constitute a heterogeneous population that either already differentiate or are still mitotically active (Fig. 2H). Consistent with the distribution of undifferentiated cells determined by morphological analysis (see above), numbers of such Pax7+/Mgn+ and Pax7+/H3P+ cells are significantly higher in the posterior zone 4 (p < 0.01) than in all other zones (1–3) of the lateral fast muscle (Fig. 2H). Pax7−/Mgn+ cells are more abundant than Pax7+/Mgn+ cells and, in contrast to the nonuniform distribution of the latter, almost evenly distributed over all four lateral fast muscle zones (Fig. 2H). Together, these results show that after the disappearance of a continuous DM layer at the surface of the myotome, a pool of mitotically active Pax7+ muscle precursors has become established within the lateral part of the myotome itself. Our morphological and quantitative data suggest that these muscle precursors are directly delivered from the residual posterior DM lip to the underlying part of the lateral fast muscle.
In fish of the developmental period from hatching to about midlarval stage (53 dpf), TEM and immunolabeling for Pax7 both continue to show that elongated squamous cells (most probably representing residual DM cells) are still present at the lateral surface of the slow muscle layer, again predominantly at positions next to the previous posterior DM lip (Fig. 3E, TEM not shown, but see Fig. 3C). Small undifferentiated cells (as above characterized by a more condensed nuclear chromatin and lack of myofibrils) are also present at the lateral fast muscle surface, especially in the posterior part (Fig. 3A) but also at more anterior positions (Fig. 3B). Other small cells at these sites contain initial clusters of myofilaments indicating onset of myogenic differentiation (not shown). Quantitative analysis in the larvae at 53 dpf reveals that relevant numbers of Pax7+ cells are present in all four zones of the lateral fast muscle, but still with highest amounts being found in the posterior zone 4 (Fig. 3D–F). In contrast to the situation at hatching, many of these Pax7+ cells are now Mgn−. Numbers of both Pax7+/Mgn− and Pax7+/Mgn+ cells are significantly higher in the posterior zone 4 than those in the anterior zones 1-3 (p < 0.01; Fig. 3F). Additionally, the anterior zone 1 contains significantly more Pax7+/Mgn− and Pax7+/Mgn+ cells than each of the two central zones 2 and 3 (p = 0.02 each). Pax7−/Mgn+ cells are approximately equally distributed over all four zones, similar to the data at hatching. As in the previous stage, almost all H3P+ cells in the lateral fast muscle proved Pax7+. Quantitative analysis of the distribution of Pax7+/H3P+ cells reveals that most mitotically active muscle precursor cells are found in zone 4 (p < 0.01 vs. zone 2; p = 0.01 vs. zones 1 and 3; Fig. 3F). However, in contrast to the situation at hatching, the vast majority of the Pax7+ cells is now H3P− (88% in zone 4; 95% on average for zones 1–3).
More advanced larvae (53–102 dpf) show a notable increase in myotomal muscle mass. Both TEM analysis and immunolabeling for Pax7 again confirm that DM cell-like cells are still present at the lateral surface of the slow fibers of these fish, particularly at the posterior ends (Fig. 3C). As in the younger larvae, undifferentiated cells together with cells at initial stages of myofibrillogenesis are also present at the lateral fast muscle surface (not shown but see Fig. 3A,B). Quantitative analysis of double immunolabeling in fish at the end of the larval period (102 dpf) also partly mirrors the result of midlarval stage, in that the number of Pax7+/Mgn− cells in zone 4, although decreased compared to 53 dpf, is still significantly higher than in all other zones (Fig. 3G). Also, numbers of Pax7−/Mgn+ cells are almost the same for all four zones. Furthermore, Pax7+/H3P− cells are significantly more frequent in the posterior zone 4 than in the zones 1–3 (Fig. 3G), and Pax7+/H3P+ cells preferentially occur in the posterior zone 4 (p = 0.05 vs. zones 1 and 2; no significance between zone 3 and 4). However, in contrast to 15 dpf and 53 dpf, Pax7+/Mgn+ cells no longer exhibit statistically significant differences in their anteroposterior position, although a trend to accumulate posteriorly is maintained.
In the present work, we provide first morphological evidence aided by quantitative analysis indicating that the proliferative myogenic precursor cells accounting for posthatching stratified fast muscle growth in teleost larvae derive from the DM. We suggest that these cells migrate underneath from the posterior DM lip to the underlying posterior fast muscle surface. At this site, they likely continue to replicate thus generating daughter cells which subsequently spread out in anterior direction before they eventually differentiate into muscle fibers. This interpretation is particularly supported by quantitative immunocytochemical analysis demonstrating that mitotically active Pax7+ cells almost exclusively occur in the posterior lateral fast muscle (Figs. 2F–H, 3D–G). The Pax7+/H3P+ cells clustered at the posterior lateral fast muscle surface may thus be regarded as the functional successors of the myogenic precursors within the DM epithelium of the teleost embryo.
There are several reasons to suppose that these successor cells indeed originate from the DM: (1) They appear simultaneously with the dissociation of the continuous DM epithelium; (2) Similar to embryonic DM cells, they are Pax7+ and mitotically competent; and (3) They arise precisely beneath the posterior DM lip (Fig. 2A–C,F–H), which is likely to be the site of centripetal myogenic cell supply to the myotome in the teleost embryo (Fig. 1C; for evidence in trout, see Steinbacher et al., 2008).
This suggested shift of muscle precursor cell generation from the DM layer itself to a stockpile of DM-derived cells in the posterior part of the larval fast muscle appears to represent the final stage of myogenic cell supply to the teleost myotome along a “rotatory” pathway already pioneered in the time just after somite formation (Hollway et al., 2007; Stellabotte et al., 2007; Steinbacher et al., 2008).
These results serve as a basis for drawing up the following growth model: In the embryonic period, with a continuous DM epithelium in place, myogenic Pax7+ cells enter the myotome by means of the posterior DM lip in a postmitotic, already differentiating (Mgn+) state (Fig. 1C). This situation changes at the onset of the larval period, when the continuous DM epithelium has dissociated but many cells of the posterior DM lip are still in place (Fig. 2A,B). By that time, the myotome receives nondifferentiating (Mgn−) Pax7+ precursors which maintain mitotic activity even after their arrival at the fast muscle surface. This establishes a proliferative (DM-like) myogenic precursor cell pool inside the myotome which, in turn, is likely to account for a continued anteriorward spread of differentiating (Mgn+) muscle cells (Figs. 2F–H, 3D–G). The lasting presence of mitotically active Pax7+/Mgn− cells indicates that a DM-like nature of the precursor cell pool in the posterior lateral fast muscle is maintained throughout larval life (Fig. 3F). Only at the most advanced larval stage investigated, the dynamics of this precursor cell pool appears to slacken in pace while its dominance for myogenic precursor cell generation is still strong (Fig. 3G). Specifically, the almost exclusive presence of mitotically active Pax7+/ H3P+ cells in zone 4 makes it plausible that only the posterior domain of the lateral fast muscle operates as a long-lasting source of myogenic cells to drive ongoing lateral fast muscle hyperplasia. This notwithstanding the fact that the more anterior zones of the lateral fast muscle now also contain relevant numbers of Pax7+/ Mgn− cells (Fig. 3G). However, given the very small number of Pax7+/ H3P+ cells at these sites, such cells may be largely destined for differentiation without further proliferation rather than they would replenish the precursor cell pool.
Remarkably, stockpiling of mitotically active muscle precursor cells appears to occur along the entire dorsoventral extent of the posterior myotome including the dorsal and ventral extremes where such cells are most abundant (Fig. 2B). This indicates that posteriorly stored muscle precursor cells (rather than randomly distributed precursors) are also the main “growth engine” behind the powerful dorsal and ventral growth zones of the teleost myotome thus revealing these growth zones as direct continuations and endpoints of the lateral fast muscle growth zone.
These results lead us to the interpretation that stratified myotome growth in early teleost life consists of two distinct successive phases whereby the precursor cells involved are generated in different cellular, and most likely also signaling, environments. We propose that the first (embryonic) phase is driven by externally derived cells from the DM, which lies between the slow muscle layer and the ectodermal epidermis (Fig. 4A). By contrast, the second (larval) phase would then be powered by internally generated cells that are entirely surrounded by muscle cells without ectodermal neighbors (Fig. 4B).
As to the fact that individual residual Pax7+ DM cell-like cells are still present at the site of the former posterior DM lip throughout larval life (Fig. 3C–E), work in zebrafish has suggested that such cells account for the larval and juvenile muscle growth (Hollway et al., 2007; Stellabotte et al., 2007). However, the low numbers of such cells together with the present finding of a mitotically active population of precursor cells within the posterior lateral fast muscle makes it unlikely that the lip cells alone are responsible for the intense stratified muscle growth during larval life. Rather, these cells may function to continuously replenish the precursor cell pool of the lateral fast muscle and/or give rise to slow fibers or dermis cells.
Similar to the situation in teleost fish, postembryonic muscle growth in the amniotes also depends upon a mitotically active population of muscle precursor cells located inside the myotome. Several studies in chick and mouse have recently demonstrated that these muscle precursors are Pax3+/Pax7+ and originate from the central part of the DM (Fig. 4C; Ben-Yair and Kalcheim, 2005; Gros et al., 2005; Kassar-Duchossoy et al., 2005; Relaix et al., 2005). These cells eventually give also rise to a persistent population of muscle stem cells, the satellite cells. The present data now suggest that the source of intramyotomal precursor cells for postembryonic muscle growth is the same for all vertebrates in that all derive from the DM. However, the ways of how these cells enter the myotome are different, as is their primary migratory destination (Fig. 4). In the amniotes, precursor cells migrate directly underneath from the central DM during its terminal dissociation, and contributions of the epithelial DM lips to later muscle growth are negligible (Ben-Yair and Kalcheim, 2005; Gros et al., 2005). By contrast, our results suggest that in fish there is no direct translocation of mitotically active Pax7+ muscle precursors from the dissociating DM centre but instead immigration from the posterior DM lip. A similar difference appears to exist for the immediate destination of these still mitotically active cells after immigration. Hence, these cells instantly infiltrate the entire myotome in the amniotes, but are confined to a restricted area of the myotome (the posterior lateral fast muscle) in fish. It must be added that fish after hatching also possess muscle precursors (the so-called mosaic cells) that are distributed over the entire myotome (Rowlerson and Veggetti, 2001). This situation is morphologically similar to that described for amniotes (see above). The specific ancestry of the mosaic cells is unknown but a DM origin is in debate (Hollway et al., 2007), although a direct (amniotelike) immigration of these cells from the central DM is unlikely in view of the present data. In view of the relative paucity of Pax7+ and/or mitotically active cells as compared to the number of newly generated mosaic fibers (Fig. 3D,E; see also Steinbacher et al., 2006), it is rather plausible that mechanisms different from those in amniotes account for the insertion of the fish mosaic precursors.
Summarizing the above, our findings suggest that fast muscle growth in teleost fish depends upon a population of myogenic precursors that has no equivalent in the amniotes, thus indicating that distinct evolutionary pathways have been followed between vertebrate groups regarding muscle growth from DM-derived precursors.
Investigations were carried out on samples (n ≥ 6 for each technique applied) of pearlfish early life stages from shortly before hatching to the end of the larval period. All fish used in the study were laboratory reared from artificially inseminated eggs. Animals were kept under rising temperature regimens (12°C until hatching, 14°C until onset of exogenous feeding, 16°C during the larval period, ± 0.5°C each). Rates of water flow and recirculation were kept constant throughout the experiment. Larvae were fed a combined diet of live plankton and commercial fish food. All animals sampled were over-anesthetized with MS-222 and, before fixation, larger individuals were cut into smaller pieces to allow sufficient fixative penetration.
For 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. Primary antisera used for immunostaining were murine monoclonal IgG1 against chicken Pax7 (1:20; DSHB), anti–Phospho-Histone H3 (1:100; Upstate), and rabbit anti-Myogenin (1:100; Santa Cruz). Alexa 488-conjugated 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 double-immunolabeling, Pax7 was combined with either Myogenin (Mgn) or Phospho-Histone H3 (H3P). For light microscopy (LM) of semithin sections and for TEM, animals were fixed in PBS-buffered Karnovsky-type fixative, post-fixed 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.
For quantitative assessment of the anteroposterior distribution of labeled cells, lateral fast muscle areas delimited by two successive myosepta were divided into 4 equidistant zones (Fig. 2E). Three developmental stages (newly hatched animals at 13 dpf and larvae at 53 and 102 dpf) were examined. In a first set of specimens (approximately 15 sections and 2 myotomes per section, that is, a total of 30 myotomes from 6 individuals per developmental stage), numbers of Pax7+/Mgn−, Pax7+/Mgn+ and Pax7−/Mgn+ cells were recorded for each zone. A second set of specimens was used to do the same for Pax7 and H3P. Nonparametric Mann-Whitney tests were used for statistical analysis.
The authors thank Hans-Peter Gollmann, Institute of Water Ecology, Fisheries Biology and Lake Research, Scharfling, Austria, for rearing the fish, Synnöve Tholo and Adda Mänhardt, University of Salzburg, Austria, provided excellent technical support.
Grant sponsor: Austrian Science Foundation (FWF); Grant number: P20430.