Positional signals: evidence for a possible role in muscle fibre-type patterning of the embryonic avian limb

The ability of peripheral motoneurones to determine muscle fibre-type characteristics, as first demonstrated by the elegant cross-reinnervation experiments of Buller et al. (1960), is well documented (for review, Jolesz & Sreter, 1981). Although these experiments were performed during the postnatal period, the influence of motoneurones on fibre-type determination had been extended to include early myogenesis (see Sanes, 1987). The latter assumption, however, has been challenged recently by the aneuro-genic chick embryo model designed specifically to investigate the influence of peripheral nerves on the initial expression of muscle fibre types (Butler et al. 1982). Aneurogenic brachial muscles, prepared by deletion of the brachial neural tube at day 2 in ovo, exhibit fibre-type profiles equivalent to those characteristic of innervated brachial muscles. This observation favours the position that primary generation myogenic cells are intrinsically committed to distinct fibre-type cell lineages. Current evidence in support of this hypothesis is derived from further analyses of aneurogenic avian muscles in vivo (Phillips & Bennett, 1984; Bloom et al. 1985; Phillips et al. 1986; Sohal & Sickles, 1986) as well as embryonic in vitro clonal analyses (Miller & Stockdale, 1986, 1987), functional denervation studies (Crow & Stockdale, 1986) and embryonic cross-innervation experiments (Laing & Lamb, 1983; Butler et al. 1986; Vogel & Landmesser, 1987), using either myosin-ATPase profiles or monoclonal antibodies as markers to distinguish fibre types. Although the experiments cited eliminate neuronal influences and imply an intrinsic commitment of myogenic cells to distinct cell lineages, the possibility exists that a non-neurogenic, extramyogenic factor(s) may be associated with early embryonic fibre-type diversification. The present study, designed to explore this possibility, assessed the fibre-type composition of brachial muscles derived from myogenic precursor cells of the thoracic region of a donor embryo.

Our rationale for selecting this experimental model was based on previous analyses of intraspecific (chick/chick) and interspecific (chick/quail) chimaeras formed by transplanting unsegmented thoracic somitic mesoderm to replace extirpated brachial somites at 48 –52 h in ovo. Recently reported motility analyses of intraspecific chimaeras demonstrate that the donor tissue is viable and capable of forming a functional wing musculature (Cauwenbergs et al. 1986a). These findings correlate with previous histological analyses of interspecific chimaeras. Historically, the latter studies were the first to determine that ‘the site of implantation… controls the destiny of somitic presumptive muscle cells’ (Chevallier et al. 1977; Chevallier, 1979) and emphasize that the ‘myogenic potentials of the somitic mesoderm are not regionalized’ (Chevallier et al. 1978). Thus, both functional and structural data attest to the highly plastic nature of somitic premyogenic cells and suggest that their ultimate ability to form specific individual muscles is determined by positional signals of the host embryo. Consequently, we now query whether or not positional information also influences the differentiation of specific muscle fibre types.

To answer this question, the brachial musculature of intraspecific chick/chick chimaeras previously analysed for functional activity (Cauwenbergs et al. 1986a) has now been examined histochemically to determine the fibre-type composition of muscles derived from heterotopically transplanted unsegmented thoracic somitic mesoderm. This model was highly suitable for achieving our objective since (1) unsegmented somitic mesoderm represents the earliest source of cells destined to form muscle fibres in vivo (Wachtler et al. 1982); and (2) the fibre-type profiles expressed by muscles derived from in situ thoracic somites differ from those exhibited by in situ brachial muscles (Butler & Cosmos, 1981b; Butler et al. 1986).

Comparative histochemical analyses of experimental and control embryos from stage 34 (day 8E) through stage 45 (day 19 –20E) demonstrated that an appropriate brachial musculature formed, on schedule, when unsegmented thoracic somitic mesoderm was grafted to replace extirpated brachial somites at day 2 in ovo. Moreover, each experimental brachial muscle, derived from transplanted thoracic somitic mesoderm, expressed a fibre-type profile equivalent to that characteristic of its control counterpart. Since the site of transplantation influenced the fate of grafted myogenic cells from the perspectives of muscle formation and fibre-type expression, we propose that non-neuronal environmental cues are associated with embryonic fibre-type determination. A preliminary report of these experiments has been presented (Butler & Cosmos, 1987a).

The unilateral replacement of brachial (Br) somites by unsegmented somitic mesoderm of the thoracic (Thor) region was accomplished using the following protocol. Fertile eggs homozygous for the dystrophic (am⁺⁺) gene, obtained from the University of Connecticut (Storrs), and normal White Leghorn fertile embryos, obtained from either Storrs or a local hatchery (Martindale Farms), were incubated at 37 ·4°C and 56% humidity until potential donor embryos reached stage 14 (50 –53 h) and host embryos were at stage 13 (48 –52 h). Embryonic age was determined by the criteria of Hamburger & Hamilton (1951).

Donor embryos at the appropriate stage were removed from their eggs, placed in a Petri dish containing sterile Hank’s Balanced Salt Solution (HBSS) and the thoracic region extending from somite 22 to presumptive somite 26 inclusive was excised. The excised tissue was then placed in a sterile solution of 0 ·25% trypsin/Ca²⁺- and Mg²⁺-free HBSS for 10 –15min, while the host embryo was being prepared. Following trypsinization, the thoracic somitic mesoderm excised from the right side of the donor embryo was dissected free from surrounding tissues (ectoderm, endoderm, neural tube, mesonephric duct and lateral plate mesoderm), transferred by pipette to trypsin inhibitor (5 % fetal bovine serum/HBSS) for approximately 10s, and then transplanted to the host embryo. Host embryos at the appropriate stage were prepared by extirpating the brachial somites (somite 16 to presumptive somite 21 inclusive) from the right side of the embryo. Removal of host brachial somites was achieved by using an electrophoretically sharpened tungsten needle. After penetrating the vitelline membrane, a longitudinal incision extending the length of the removal site was made through the surface ectoderm dorsal to the right lateral edge of the brachial neural tube. Transverse incisions through the ectodermal layer were then made to join this initial cut, first at the cranial end of the removal site (between somites 15 and 16) and then at the caudal end (between presumptive somites 21 and 22). The flap of ectoderm overlying the brachial somites was then folded laterally to expose the somites below. Two deeper longitudinal incisions were then made medial and lateral to the brachial somites, being careful not to damage the posterior cardinal vein, underlying aorta or mesonephric duct. Transverse incisions were then made at the cranial and caudal limits of the removal site and the brachial somites were gently lifted away with fine forceps. After removal of the brachial somites, the operative site was suctioned with a finely drawn glass pipette to remove any possible fragments of somitic mesoderm; the removal site was inspected for injuries to either the brachial neural tube or lateral plate mesoderm. Damaged embryos were discarded.

Following removal of the host brachial somites, the donor thoracic somitic mesoderm was manipulated into the removal site of the host embryo (Fig. 1) using a fine wire probe. By cutting the caudal end of the donor tissue at a slight angle at the time of removal from the donor embryo, it was possible to maintain proper craniocaudal and dorsoventral orientation of the graft. After the donor somitic mesoderm was positioned correctly into the host removal site, some of the fluid overlying the embryo was suctioned away using a finely drawn glass pipette. This procedure prevented subsequent shifting of the grafted tissue within the operative site. The ectodermal flap produced at the start of surgery was then folded back over the transplanted somitic mesoderm using a wire probe. Following surgery, the eggs were closed with a glass coverslip sealed with soft dental wax and returned to the incubator to develop to the desired postoperative stage. A total of 177 embryos was subjected to the experimental (Thor –Br) manipulation, using either genotypically normal (n = 34) or genotypically dystrophic (n = 143) embryos as donors; in all cases, the hosts were of normal genotype.

A series of unoperated embryos of normal (n = 123) and dystrophic (n = 90) genotype as well as embryos surviving the homotopic somite transplantation procedure (n = 5) served as controls. Homotopic (Br –Br) somite transplants were performed by replacing extirpated brachial somites of a normal host embryo with brachial somites from either a normal or dystrophic donor embryo at stage 13 (48 –52 h). In addition, the right (operated) brachial musculature of all surgically manipulated embryos was compared to that of the left (unoperated) side.

Experimental (Thor –Br) and control (Br –Br, unoperated) embryos were removed from the egg, decapitated, eviscerated and frozen for cryostat sectioning. The exact developmental stage of each embryo was assessed according to the morphological criteria of Hamburger & Hamilton (1951). Serial longitudinal sections of frozen embryos were analysed by the myosin-ATPase reaction following alkali (pH 10 · 0) or acid (pH 4 ·35) preincubation to monitor differentiation of muscle fibre types, as described (Butler & Cosmos, 1981a,b). In addition, innervation patterns and end-plate formation were monitored by using our modification of the silver-cholinesterase reaction of Toop (Butler & Cosmos, 1981a) and the acetylcholinesterase reactions of Karnovsky & Roots (1964) or Koelle, as modified by Silver (1963).

Verification that thoracic somitic mesoderm is viable following heterotopic transplantation to a host embryo was derived from (1) functional and (2) anatomical criteria. (1) Using the distinctive wing motility patterns characteristic of dystrophic and normal embryos as our marker, we recently reported that transplants performed between donor and host chick embryos form a functionally viable wing musculature (Cauwenbergs et al. 1986a). To illustrate, comparative wing motility analyses performed after transplanting either normal or dystrophic thoracic somitic mesoderm to a normal host are presented in Fig. 2. Similar to the trends exhibited by unoperated normal and dystrophic wings (Cauwenbergs et al. 1986 a,b), experimental muscles of dystrophic origin exhibit a significantly lower frequency of wing movements from day 10E to day 15E when compared to similarly aged experimental muscles of normal genotype. The observation that experimental wing muscles function according to the genotype of the donor embryo attests to the viability of the transplanted tissue. The present study represents histochemical analyses of nineteen experimental embryos from the series previously assessed for functional activity (see Fig. 2); without exception, experimental embryos that received genotypically dystrophic donor tissue ex-pressed the dystrophic phenotype of impaired wing motility whereas those that received genotypically normal donor tissue exhibited a normal motility pattern. (2) As shown previously (Kieny et al. 1972), the formation of supernumerary ribs serves as an anatomical marker for the viability of transplanted thoracic somitic mesoderm. Serial reconstruction of all chick/chick chimaeras analysed for the present study revealed that transplantation of thoracic somitic mesoderm to the brachial region resulted in the formation of supernumerary ribs, located anterior to the in situ rib cage. Fig. 3 compares the unoperated (left) and operated (right) sides of the brachial region of a stage-35 (day 8 – 9E) Thor-Br embryo. Note the absence of ribs opposite the left coracoid bone and the presence of supernumerary ribs (arrowheads) opposite the right coracoid bone. Serial reconstruc-tion of control (Br-Br) embryos revealed the absence of supernumerary ribs on the operated side.

Once the viability of the donor tissue was con-firmed using functional (Fig. 2) and anatomical (Fig. 3) criteria, the formation and fibre-type compo-sition of experimental and control muscles were compared from stage-34 (day 8E) to stage-45 (day 19 – 20E). The latter analyses were restricted to proxi-mal muscles derived from cleavage of the premuscle masses of the wing bud; distal (forewing) muscles or myotomally derived brachia! muscles were not exam-ined.

Previous homotopic somite transplants, using the chick/quail chimaeric model (Christ et al. 1977), demonstrate that all brachial muscles are somitic in origin and that premyogenic cells migrate from the somites to form the primary dorsal and ventral muscle masses of the wing bud. Subsequently, the latter undergo a series of cleavages, according to a precise chronological schedule, to form the brachial musculature, i.e. intrinsic muscles of the wing as well as muscles associated with the pectoral girdle (Sullivan, 1962). To determine if transplanted thoracic somitic mesoderm was capable of forming, on schedule, an appropriate brachial musculature, experimental (Thor – Br) and control (Br – Br) embryos at stages 34 (day 8E), 35 (day 8 – 9E) and 36 (day 10E) were sectioned completely and compared to similarly staged unoperated embryos of both normal and dystrophic genotype. These specific stages were selected since the individualization of all muscles derived from the primary muscle masses of the wing bud is virtually complete during this period (Sullivan, 1962; Butler et al. 1982). Furthermore, since the process of cleavage is not impaired in the dystrophic mutant (Butler & Cosmos, 1987b), the transplantation of donor tissue from a dystrophic embryo was not a factor to be considered in this phase of our study. By stages 34 (day 8E) – 36 (day 10E), the dorsal muscle mass of both unoperated normal and dystrophic wing buds has cleaved to form twelve individual wing muscles, whereas the ventral muscle mass has formed six distinct wing muscles (Table 1). A similar pattern was observed for all Thor-Br and Br – Br embryos examined during these stages. To illustrate, the stage-36 (day 10E) Thor – Br embryo presented in Fig. 4A shows the formation of distinct deltoideus (D), scapulohumeralis anterior (SHA) and posterior (SHP), subscapularis et subcoracoideus (SS+SC), anterior latissimus dorsi (ALD) and triceps brachii pars scapularis (Ts) muscles. This particular section of the right (operated) brachial region was selected because it includes muscles derived from each of the four divisions of the dorsal muscle mass, as shown in Table 1. More ventral sections from the Thor-Br embryo presented in Fig. 4A showed that all wing muscles originating from the ventral mass were also present and located in anatomically correct positions. Analyses of all Thor – Br embryos beyond stage 36 (day 10E) confirmed that transplanted thoracic somitic mesoderm was capable of forming an appropriate brachial musculature. The only anomaly observed was the absence of the right PLD muscle in one Thor-Br embryo examined at stage 42 (day 16E).

To determine if the brachial musculature derived from thoracic somitic mesoderm duplicated the fibretype patterning characteristic of similar muscles derived from brachial somitic mesoderm, the fibre-type composition of brachial muscles was compared among Thor – Br, Br – Br and unoperated embryos from stage 34 (day 8E) onward. Although the majority of brachial muscles of unoperated embryos are composed predominantly of fibres that exhibit alkali-stable myosin-ATPase activity only, specific muscles are composed entirely of fibres that exhibit dual alkali and acid stability (marked ‡ in Table 1; Butler et al. 1982). Experimental Thor-Br muscles demonstrated a similar pattern. The profiles of representative stage-42 (day 16E) experimental brachial muscles derived from grafted thoracic somitic mesoderm are presented in Fig. 5A – D. Experimental ALD, posterior latissimus dorsi (PLD), dorsocutaneous latissimus dorsi (DLD), metapatagial latissimus dorsi (MLD) and SHP muscles all exhibit alkali stability. Only the ALD, DLD and MLD, however, are also acid stable; the PLD and SHP are essentially acid labile (Fig. 5A,B). Thus, similar to unoperated embryos (see Table 1), three of the four muscles derived from the latissimus dorsi division expressed dual stability in all fibres. Sections from a more ventral region of the same Thor – Br embryo shown in Fig. 5A,B demonstrate that, typical of the unoperated SHP (Butler et al. 1982), acid stability is noted in fibres located on the ventral surface of the SHP, adjacent to the scapula (Fig. 5C,D). A similar restriction of acid-stable fibres to specific regions within the predominantly alkali stable D muscle is also observed. At stage 42 (day 16E) all experimental muscles were well-innervated and exhibited endplates.

Examination of younger Thor – Br embryos between stage 34 (day 8E) and stage 39 (day 13E) confirmed that brachial fibre-type profiles were expressed by muscles derived from thoracic somitic mesoderm. To illustrate, at stage 36 (day 10E), the experimental ALD and SHA exhibit dual stability whereas the SHP, Ts and D muscles are alkali stable only (Fig. 4A,B). At this stage, unoperated muscles express similar profiles. Analyses of Thor-Br embryos during earlier developmental stages was important since a previous study (Butler et al. 1982) determined that fibre-type differentiation proceeds independent of neuronal influences during this phase of development. Therefore, the possibility that innervation from brachial nerves influenced the fibre-type characteristics expressed by Thor-Br muscles was eliminated. Thus, throughout the developmental period analysed, individual brachial muscles derived from transplanted thoracic somitic mesoderm exhibited fibre-type profiles equivalent to those characteristic of their unoperated counterparts. Likewise, the profiles expressed by Br-Br muscles resembled those typical of unoperated muscles. Furthermore, these observations applied to donor tissue of either normal or dystrophic gentotype, confirming a previous observation that differentiation of appropriate myosin-ATPase profiles is not impaired in dystrophic embryos (Butler & Cosmos, 1987b).

Investigations concerned with the fate of somitic mesoderm following either homotopic or heterotopic transplantation have contributed significantly to an understanding of myogenesis during early embryonic development. With the advent of the quail nuclear marker (Le Douarin & Barq, 1969), interspecific chick/quail chimaeric studies resolved a long-standing controversy concerning the origin of limb musculature. Today, it is generally accepted that skeletal muscle fibres are somitic in origin whereas the tendon and connective tissue components of limb muscles are derived from the somatopleure (for review, Kenny-Mobbs, 1985). Furthermore, although homotopic transplantation of single somites indicates that individual somites contribute to the formation of specific muscles during normal embryogenesis (Beresford, 1983), heterotopic transplants demonstrate that individual somites are not predetermined to form specific muscles exclusively (Chevallier et al. 1977; Chevallier, 1979; Jacob et al. 1978; Cauwenbergs et al. 1986a; present study). Indeed, the heterotopic experiments cited emphasize the plasticity of myogenic somitic cells and suggest that positional cues associated with the site of transplantation control their destiny as far as muscle pattern formation is concerned.

The present study has now focused on the fate of heterotopically grafted somitic mesoderm from the perspective of muscle fibre-type patterning to foster an understanding of factors associated with embryonic fibre-type expression. More specifically, we questioned whether or not positional cues would also influence this developmental process. Therefore, the fibre-type composition of experimental muscles derived from heterotopically transplanted somitic mesoderm was determined.

Intraspecific chick/chick chimaeras were prepared by grafting unsegmented somitic mesoderm from the thoracic region of a donor chick embryo to replace extirpated brachial somites of a host chick embryo (Thor-Br model). Paramount to the interpretation of our experiments is evidence that (1) the donor tissue was viable and (2) the experimental muscles analysed were devoid of host myogenic cells. Evidence concerning the viability of the donor graft is inferred from the interspecific chick/quail study of Chevallier et al. (1977). Utilizing the distinctive quail nucleolus as a marker, they demonstrated that thoracic somitic mesoderm from a quail donor participates in the formation of the wing musculature of the chick host. Myofibres of the wing muscles were exclusively of donor (quail) origin whereas the tendons and connective tissue components were of host (chick) origin. For the present intraspecific chick/chick chimaeric study, viability of the donor tissue was verified directly by functional and structural criteria. The observation that thoracic somitic mesoderm transplanted from a dystrophic donor to a normal host forms muscles that express the dystrophic phenotype of impaired wing motility provides functional verification of graft viability (Cauwenbergs et al. 1986a).

Anatomically, the operated brachial region of Thor-Br embryos exclusively was characterized by supernumerary ribs. As demonstrated previously (Kieny et al. 1972), the presence of these ectopic structures in the brachial region attests to the viability of transplanted thoracic somitic mesoderm. Thus, based on these parameters, the donor tissue was not only viable but capable of differentiating both myogenic and chondrogenic cell lineages.

To ensure that experimental muscles were devoid of host myogenic cells, selection of an appropriately aged host embryo was critical since myogenic cells located in somites opposite limb-forming regions migrate to seed the developing limb musculature during a specific phase of early myogenesis. Based on previous reports that migration within the brachial region does not commence prior to stage 14 (Jacob et al. 1978, 1979; Kenny-Mobbs, 1985), brachial somites were extirpated from host embryos at stage 13, i.e. prior to migration. Indeed, recently Solursh et al. (1987) reported that migration within the brachial region does not occur until stage 15. Using a monoclonal antibody specific to vascular endothelial cells, they demonstrate that mesodermal cells at stage 14, previously described as migrating myogenic cells, are actually endothelial cells belonging to the vascular network. Morphologically identifiable migrating (myogenic) cells within the brachial region of the chick embryo do not appear until stage 15.

Analyses of muscles derived from the donor tissue demonstrated that the somitic level of origin is associated with neither (1) the formation of specific muscles nor (2) the differentiation of specific fibretype profiles. Unsegmented somitic mesoderm from the thoracic region of a donor embryo transplanted to the brachial region of a host embryo at 48 – 52 h in ovo formed, on schedule, an appropriate brachial musculature in all experimental embryos analysed. Furthermore, without exception, the experimental brachial musculature that formed duplicated the fibre-type patterning characteristic of unoperated and control (Br – Br) brachial muscles at all stages analysed. Therefore, we propose that commitment to distinct fibre-type cell lineages is not endogenously programmed, as previously suggested (for review, see Miller & Stockdale, 1987; Sanes, 1987), but is influenced by non-neuronal cues operative within a limb-forming region.

Although the mechanism whereby positional signals influence fibre-type expression is unknown at present, such cues appear to be operative during a temporally limited period if one correlates previous experiments performed by Laing & Lamb (1983) with the present study. Laing & Lamb observed that a wing bud grafted at stage 18 (65 – 69 h) or 19 (68 – 72 h) to replace an extirpated leg bud develops a wing musculature with appropriate wing-muscle fibre-type profiles. Although designed to examine the influence of foreign innervation on fibre-type expression, the results of their study may be interpreted to indicate that myogenic cells within a transplanted limb bud are already committed to distinct fibre-type cell lineages. Myogenic precursor cells transplanted at the unsegmented somitic mesoderm stage lack such a commitment, as indicated by the present study performed at stage 13 (48 – 52h). Therefore, we propose (1) that commitment to distinct fibre-type cell lineages may be irreversibly expressed during a temporally limited period of approximately 20 h. This specific period represents the interval between the stage when the present unsegmented somitic mesoderm transplants were performed and the stage when the limb bud grafts were done. (2) Alternatively, fibre-type determination may occur at an even later stage. A grafted limb bud is not composed exclusively of myogenic cells. Therefore, a non-neuronal, extramyogenic factors) associated with fibre-type determination, unique to each limb-forming region, may be present in the limb bud graft and, thus, may be operative during the post-transplantation period. The time frame for expression of this putative factor(s), however, would be limited to approximately 24 h posttransplantation since in vivo analyses (Miller & Stockdale, 1986) indicate that distinct classes of myogenic cells exist within the developing limb as early as day 4 in ovo.

In summary, heterotopic transplant studies indicate that somitic mesodermal cells destined for the chondrogenic cell lineage are already regionalized along the cephalocaudal axis at the unsegmented somitic mesoderm stage (Kieny et al. 1972; present study). In contrast, at this stage somitic mesodermal cells destined for the myogenic cell lineage are unspecified as to both muscle pattern formation and fibre-type pattern formation. Instead, these processes appear to be influenced by non-neural, extramyogenic cues associated with the site of transplantation.

Although limb skeletal patterning has been studied extensively and found to be influenced by positional cues derived from the apical ectodermal ridge (Summberbell, 1974) and the zone of polarizing activity (Summerbell, 1979), comparatively few studies have investigated limb muscle pattern formation. These limited studies, however, do indicate that neither peripheral nerves (Butler et al. 1982) nor limb skeletal elements (Lanser & Fallon, 1987) influence initial muscle patterning. Whether or not connective tissue derived from the somatopleure controls this process, as suggested by Kieny et al. (1986), remains to be proven. The elucidation of the mechanism(s) controlling limb muscle pattern formation may provide clues concerning our proposed relationship between environmental signals and incipient limb fibre-type pattern formation.

The authors express their gratitude to E. Paul Allard for his valuable contributions to the present study. This research was supported by grants awarded to E.C. by the Muscular Dystrophy Association of Canada and by the National Science and Engineering Research Council. P.C. was the recipient of a predoctoral fellowship from the Muscular Dystrophy Associaiton of Canada.