Myogenesis in the trunk and leg during development of the tadpole of Xenopus laevis (Daudin 1802)

Although myogenesis has been investigated in many vertebrates in vivo and in vitro, few studies have correlated it with the development of motility or behaviour in the embryo. Studies in vitro (Shimada, Fischman & Moscona, 1967 ; Fischman, 1970) can cover only limited aspects ; in vivo studies (Hay, 1961, 1963) have often been based on regenerating muscle rather than on developing embryonic muscle, and therefore on older animals and in the presence of nerves and complex chemical factors. As a result most current work on myogenesis has not been related to the functional state of the animal or its nervous system, but purely to the differentiation of the muscle fibre from the myoblast. Coghill in a series of papers (1914 – 18) and in his book (1929) attempted to correlate the differentiation of the nervous system with the development of behaviour in Ambystoma tigrinum (Green 1825) (punctatum but barely mentioned the structure of the muscles. This paper attempts to correlate myogenesis with the development of the nervous system and the stage of motility, based on a study of the myotome muscles and their innervation during early development and of the hind limb muscles during metamorphosis of Xenopus laevis (Daudin 1802). In the myotomes the muscle cells develop in the absence of nerves, whereas nerves are present during the development of the limb muscle cells. The earliest behaviour in Xenopus tadpoles consists of a contraction of anterior myotomes resulting in a flexing of the neck. By the time the tadpole hatches (48 h after laying) completely co-ordinated swimming is observed, indicating that a considerable degree of structural and physiological development has taken place. After development of the myotome muscles and co-ordinated swimming, metamorphosis begins (10 days after laying) and the hind limb-bud starts to grow. As the limb-bud gradually assumes the structure of a fully functional limb the development of nerves and muscles can be followed and correlated with the degree of motility shown. It is these stages in the myotome and limb muscle which are analysed in this paper.

Adult toads were induced to lay eggs by injection of chorionic gonadotrophin (Pregnyl, Organon Laboratories Limited). The eggs were kept in ‘pond’ water (tap water which had been standing for 24 h) at 20 – 22 °C, and developmental stages identified using the Normal Tables for Xenopus laevis of Nieuwkoop & Faber (1967). Starting at stage 20 when the neural tube had almost closed, batches of embryos were removed from their membranes, had their stage identified using the Normal Tables, and were tested to establish the responses to various stimuli.

The responses were tested by the method described by Youngstrom (1938) using an 8 mm hair mounted on a glass handle to prod the larva. The position, type and effect of the stimulus were accurately controlled and observed, and responses recorded on cine film so that their position and extent could be correlated with histological findings. The sensory response was tested by stroking the skin gently so that no visible indentation was caused. Direct mechanical stimulation of the muscle was caused by prodding with a hair so that the skin and underlying tissues were distorted, but not pierced. Some simple d.c. electrical stimulation (0 · 2 – 0 · 8 mA from a 24 V battery using silver chloride electrodes 5 mm apart on either side of the larva) was also applied. Time-lapse photography was used to record any spontaneous movements made by the developing embryos. Stages in the development of the response in the hind limbs are based on the work of Hughes & Prestige (1967).

Specimens were fixed in Nonidez fixative (Nonidez, 1939) or in-p-toluene sulphonic acid (Malm, 1962) and then sectioned prior to silver impregnation using Protargol, following J. D. Boyd’s modification of Bodian’s technique (Bodian, 1937). Older specimens with less yolk and more differentiated tissue were fixed in Bouin’s fluid (Pantin, 1959) and impregnated with silver by the Holmes (1947) or Palmgren (1948) methods using silver nitrate. A parallel series of all stages was stained with haematoxalin and eosin after fixation in Bouin’s fluid.

The developing embryo passes through five distinct behavioural stages between the end of neurulation at stage 20 and the accomplishment of free swimming at stage 33. These five behavioural stages are (1) non-motile, (II) pre-motile, (III) early flexure, (IV) early swimming or S-flexure, and (V) free swimming. They were described in detail by Muntz (1964) and are summarized in Table 1.

The non-motile stage is defined as the period between stage 20 when the neural tube has just closed and stage 22 when the eyes are beginning to protrude. During this period no form of stimulation produced any visible response, so that the myotome cells appear to be incapable of contraction.

The anterior myotomes are composed of plump elongated cells, 120 μ m long, occupying the full length of the myotome and having a single large oval nucleus (6x18 μ m) centrally placed in the cell (Fig. 1). The cytoplasm is packed with yolk granules, as are most of the other cells of the embryo at this stage. Viewed with the light microscope there is no sign of striated myofibrils.

No motor nerves leave the cord at this stage, although a few motor cells bodies can be distinguished in the ventral part of the cord opposite the first somite. In the dorso-lateral part of the cord Rohon-Beard sensory cells may be recognized by their large size and the possession of a short thick process protruding from the cord at the inter-myotomal position. There is no anatomical sign of any link between the sensory and motor systems within the cord.

The earliest contractions are shown in response to either a hard mechanical prod or a direct current stimulus. There does not appear to be any myogenic contraction before this. Contractions occur on the side stimulated (homolateral), are slow and involve only the most anterior myotomes, giving rise to a neck flexure. This stage, when the stimulation of sensory nerves does not produce any response, has been called the pre-motile stage (Muntz, 1964) and demonstrates the ability of the myotomes to contract; it occurs at stages 22 – 24 when there are 9 – 15 somites similar in appearance to those of the non-motile stage. However, viewed with the light microscope at the pre-motile stage (Fig. 2) the anterior myotomes contain a small number of striated myofibrils which are about 1 μ m thick and occur singly in the muscle cell just within the cell membrane. The myotomes are not apparently innervated by motor nerves at this stage, as no nerves can be seen leaving the cord in preparations though longitudinal fibres can be seen within it. In the dorsal part of the cord, fibres of Rohon-Beard cells leave the cord and run between the myotomes. There is still no apparent connexion between the sensory and motor systems.

A few hours later (stage 25/26) the myotomes become innervated by motor nerves which can be seen in longitudinal horizontal sections leaving the cord. At this stage the embryos first respond to sensory stimulation from which it may be concluded that there are now functional connexions between the skin, sensory and motor systems. Myofibrils occupy a larger volume of the muscle cell, still uninucleate (Fig. 3).

When the muscle cells first contract they are still undergoing development and are structurally unlike differentiated muscle fibres, or even myotomal muscle in older tadpoles. The dimensions of the earliest striated fibrils observed suggest that they may be single myofibrils. As development proceeds the strands get thicker but remain transversely aligned in bundles presumably in relation to the T-system, each bundle occurring singly in the cell (Figs. 3, 4). As the bundle of myofibrils thickens the nucleus is pushed slightly to one side (Fig. 5).

When the tadpole has hatched and is free swimming (stage 33 – 46) myofibrils fill the cell pushing the nucleus against the cell membrane (Figs. 6, 7). The cells are probably still uninucleate but have increased in length and diameter and become cylindrical rather than spindle-shaped, and are innervated by both motor and sensory nerves. In sections (Fig. 6) small nuclei (4 × 10 μm) can be seen at the ends of the muscle cells, but it is difficult to tell whether these are inside or outside the cell membrane.

During the period between the onset of swimming (stage 33) and the end of pre-metamorphosis (Etkin, 1964) (stage 48) the tadpole increases in overall length from 5 to 15 mm (Nieuwkoop & Faber, 1967). The large nuclei are still present in the cells during pro-metamorphosis (stage 50), but the small nuclei are more evenly distributed, and seem in sections to be peripherally placed within the muscle cells.

At a later stage in metamorphosis (stage 50 – 55) the muscle cells take on the appearance of adult skeletal muscle fibres (Figs. 8, 9), numerous small nuclei now occupying peripheral patches of cytoplasm. At this stage the average size of the nuclei is 4 × 11 μ m, the large nuclei having disappeared and the length of the muscle fibre and myotome block is 300 μ m.

At stage 48 (7 · 5 days) towards the end of pre-metamorphosis, when the myotome muscle cells are uninucleate, the hind limb-bud is hemispherical. A week later (15 days) it has begun to elongate so that the length of the bud is greater than its breadth, and is beginning to constrict at the base. By stage 52 (21 days) the foot is starting to be marked off from the rest of the limb, which is still immobile. Finally, by stage 63 (51 days) the legs are mobile and are the sole method of propulsion. Hughes & Prestige (1967) described the development of motility in the hind limb of Xenopus and showed that it passes through stages which are similar to those of the myotome system (Muntz, 1964). Four stages in limb development can be distinguished: (i) non-motile stage, (ii) pre-motile stage, (iii) motile stage, (iv) fully functional stage.

At stage 52 (21 days) the hind limb-bud is insensitive to any form of stimulation (Hughes & Prestige, 1967). The skin, a distally placed terminal blood vessel and a branching nerve can be distinguished from the main bulk of the mesenchyme which is undifferentiated (Fig. 10), except for denser areas of cells representing the future position of the cartilage.

During stage 53 (24 days) the limb trembles, at stage 54 (26 days) the movement described as a flare by Hughes & Prestige (1967) is seen. At this stage the limb is directly excitable by electrical stimulation of the 8th spinal nerve, but is still insensitive to touch. The mesenchymal region shows differentiation into cartilage cells and areas of elongated cells with nuclei 3 × 6 μ m destined to become muscle (Fig. 11). With light microscopy some striated myofibrils have been seen in these areas (Fig. 12) and they have been observed in electron microscopic preparations.

The limb first shows sensitivity to touch at stage 55 (32 days), giving what Hughes & Prestige (1967) described as an elaborate flare response, where the hip and knee joint are flexed and extended in sequence. Later at stage 56 (38 days) the limb performs stepping movements. In section (Fig. 13) the limb appears to have well-differentiated skin, blood vessels, nerves and cartilage. Distinct blocks of muscle with striated myofibrils can be identified clearly both with the light microscope (Fig. 14) and the electron microscope. Numerous nuclei still 3 × 6 μ m lie within fibres which appear to be only 4 – 5 μ m wide.

During stage 60 – 61 (47 days) legs and tail are used for swimming but by stage 63 (51 days) the legs are the only form of propulsion (Hughes & Prestige, 1967), though the tail is still as long as the body (Nieuwkoop & Faber, 1967). At this stage the limb appears fully differentiated (Fig. 15), with normal adult skeletal muscle now similar to the myotome muscle except that its fibres are 20 – 25 μ m wide with numerous 4 × 10 μ m nuclei (Fig. 16), whereas the myotome muscle fibres are 30 – 40 μ m wide with relatively fewer 4 × 11 μ m nuclei (Fig. 8).

The two main topics arising from this work are, first, the development of motility in a vertebrate embryo and, secondly, the difference in histogenesis between myotome muscle and limb muscle in the same embryo. The development of motility can be studied in its simplest form in vivo in the myotomes of an anamniote because there are no antagonistic muscle systems, and the muscle is not innervated initially. In the muscles of the limb different systems whose actions may be antagonistic develop at the same time, and differentiation proceeds in the presence of nerves, making the systems more difficult to analyse. The limb muscle has therefore a different environment in vivo compared with the myotome muscle, and correlated with this a different pattern of development.

In studies on the development of motility in vertebrate embryos, teleosts (Whiting, 1954) and elasmobranchs (Harris & Whiting, 1954; Harris, 1962) were found to have myogenic contractions as their first form of motility. These were rhythmic in elasmobranchs and non-rhythmic in teleosts, and initiated entirely within the muscle cell independent of any innervation as in elasmobranchs they were unaffected by complete removal of the nerve cord or apparently any form of external stimulation. In amniote skeletal muscle it seems that the myogenic phase does not exist, and that the earliest activity is spontaneous and neurogenic (Hooker, 1952; Hamburger, 1963, 1964). In Amphibia there may be an intermediate phase. In Xenopus the earliest spontaneous motility is seen when the muscles are striated and innervated both by sensory and motor nerves (early flexure stage, Muntz, 1964), but a pre-motile stage precedes this, when the myotomes can be made to contract by mechanical or electrical stimulation. At this stage striated myofibrils are present, whilst at the non-motile stage none is visible. In the absence of motor fibres leaving the cord the contractions initiated in the pre-motile stage must be assumed to be accomplished by direct stimulation of the muscle cell, either by deformation or externally initiated depolarization of the membrane. This is not strictly a myogenic contraction as it is not initiated within the muscle cell itself.

In Xenopus no contractions take place before striated myofibrils can be observed. Lewis (1915) and Howarth & Dourmashin (1958) found that in the chick in vivo contractions were observed only after the appearance of organized myofibrils. On the other hand, Renyi & Hogue (1934) using chick in vivo and Harris & Whiting (1954) using dogfish found that contractions occurred before the appearance of visible myofibrils.

In evaluating and comparing these reports there are three variables besides differences in the species used: first, the method used for determining whether striated myofibrils are present; secondly, the method used for determining whether a muscle can contract spontaneously, as a result of stimulation, or not at all; and thirdly, the work may have been carried out in vivo or in vitro, on developing or regenerating muscle, or on different anatomically distinct muscles. In anamniotes the earliest movements, whether spontaneous or stimulated, are observed in the axial myotome muscle, which is the first muscle to develop in a fish or fish-like embryo. In both, the muscle is able to contract before it is innervated and before hatching. In these embryos the myotome muscles are the only organs of locomotion available when the larva first hatches and is able to swim, but is still at a relatively early stage in development. It might be expected therefore that these muscles would show a more simple structure and type of contraction than muscles differentiating later in embryonic life, or muscles having longer to differentiate before becoming functional. In amniotes the embryo is at a much more advanced stage of development when it hatches or is born, and uses leg muscles for locomotion, so that the functional development of the myotome muscle in these cases is quite different from the anamniote type. During this period of development to the more advanced stage reached by an amniote before it is free living, the relative rates of development of muscle and nervous control and co-ordination may be different, so that as in the amphibian limb muscle, nerves are present before the muscle begins to differentiate. In the amniote embryo the earliest movements observed are spontaneous neurogenic or reflex, and occur first in the neck and trunk, then generally in the body and limbs (Hamburger, 1964). The embryo of the tree frog Eleuthero-dactylus martinicensis (Tschudi 1839), described by Hughes (1965,1966) provides an interesting exception to the normal anamniote in that its development is wholly embryonic and it hatches as a small frog capable of jumping. The myotome muscle has quite a different function in that it never passes through a stage of providing a simple primary locomotory system as in the free swimming larva of other anamniotes, but just causes an axial wriggling of the embryo within the egg capsule. Hughes (1966) found that the first signs of movement were spontaneous and reflex behaviour, which was very similar, except for a much shorter duration, to that found by Hamburger (1963) in the chick.

In the range of types of earliest motility in vertebrate embryos and larvae, the amphibians may occupy an intermediate position between the fish with true myogenic contraction and the amniotes with solely neurogenic contractions. When investigating the earliest form of motility, larval or embryonic material must necessarily be used and this may be exceptionally specialized or simplified depending upon what role it performs in the life of the adult.

For studying the variation in differentiation pattern of different muscles of the same animal, the amphibian tadpole is convenient in possessing both myotome muscle, which has to develop and function at an early stage, and limb muscle which develops later during metamorphosis. The myotome muscle cells of Xenopus are unusual in that they develop to a fully functional state and large size whilst remaining uninucleate. During the early part of metamorphosis (late pre-metamorphosis, Etkin, 1964) the myotome cells change from this uninucleate state to normal multinucleate vertebrate skeletal muscle ; and at this time the limb-buds begin to grow and differentiate.

The first stage in the development of multinucleation appears to be accumulation of smaller nuclei at the ends of the myotome cell, rather like the satellite cells described by Mauro (1961). These probably fuse with the myotome cell whose original large nucleus becomes smaller, as in the multinucleate myotome cell all the nuclei are equal-sized and small.

The work of Capers (1960), Konigsberg, McElvain, Tootle & Herrman (1960), Maslow (1969), Shimada (1971), Lipton & Konigsberg (1972) and Rash & Fambrough (1973) has shown that multinucleate skeletal muscle fibres form by fusion of cells, not by fission of nuclei as suggested by Kielbowna (1966) and earlier workers. Most of this work has been carried out in vitro and on chick or other amniote muscle, with the result that leg or breast muscles have been used rather than myotome muscle. Muchmore (1962, 1965) working on amphibian myogenesis and Holtfreter (1965) using Ambystoma maculatum (Shaw 1802), found that up to the mid tail-bud stage the myotome cells were isodiametric and uninucleate. Later, at the onset of metamorphosis, large spindle-shaped multinucleate cells appeared throughout the trunk somites. Loeffler (1969,1970) worked on the fusion of myoblasts in amphibians using myotome muscles of Ambystoma tigrinum, A. maculatum, Ranapipiens (Schreber 1782), and Xenopus. He showed clearly either by grafting labelled Ambystoma nuclei homoplastically or grafting R. pipiens or Xenopus tissue xenoplastically into Ambystoma embryos, that the multinucleate fibre was composed of nuclei from both host and donor. The hosts, with multinucleate myotome muscles, were all examined at a stage after the beginning of metamorphosis.

In the amniote in vivo work multinucleate muscle fibres develop early, but in Amphibia it seems that multinucleation is coupled with the onset of metamorphosis. In the tree frog Eleutherodactyius martinicensis (Hughes, 1965, 1966) with an embryonic development and no larval stage, metamorphosis does not occur as distinctly as in other Amphibia. Lynn (1948) working on E. ricordii (Duméril & Bibron 1841), suggested that there may be an early and ‘precocious metamorphosis resulting from an unusually early and intense functioning of the thyroid’. It appears from a preliminary examination of material (Muntz, unpublished) that the myotome muscles of E. martinicensis are multinucleate from an early stage. As early as days, the early limb-bud stage (Hughes, 1962), the myotome muscles appear fully differentiated, though relatively smaller than in Xenopus. During the following days as the limbs grow and become capable of movement, the appearance of the myotome muscles remains the same. When the limb muscle develops in Xenopus its structure is quite different from the myotome muscle and it appears to be multinucleate when the striated myofibrils are developing.

It has been found using chick embryo in vitro (de la Haba, Cooper & Elting, 1966, 1968) that normal development of the myoblast occurs in a culture consisting of basal nutrient medium, only when other substances, such as serum, are added. From this and other work on enzyme requirements (de la Haba et al. 1968) they suggested that a number of factors are involved in the initiation and maintenance of muscle differentiation. Love, Stoddard & Grasso (1969), also using chick embryo in vitro, suggested that during the initial phase of muscle differentiation the main synthetic activity is of nucleic acid, and only after the fusion of myoblasts to form myotubes does it shift to protein synthesis. They suggested that this shift is to some extent under endocrine control. Okazaki & Holtzer (1966) working on chick embryo both in vivo and in vitro and using somite and breast muscle, suggested that the age of the embryo determines whether developing muscle cells fuse before or after synthesis of myosin and actin. Hilfer, Searls & Fonte (1973), looking at the ultrastructural development in the chick limb-bud, reported that myofilaments were identified only in cells that had become multinucleate; whereas in the myotome Dessouky & Hibbs (1965) reported that multinucleation takes place after development of the myofilaments. Hilfer et al. (1973) also pointed out that the processes of differentiation and fusion differ in muscle in vivo and in vitro. They suggested that these processes differ during the first wave of differentiation of an embryo from those at later stages of development. Xenopus follows this pattern in that synthesis of actin and myosin precedes fusion in the myotome muscle developing during the first wave of differentiation, whereas fusion precedes muscle protein synthesis in the limb muscle, developing at a later stage.

The situation in Xenopus is therefore that myotome muscles and limb muscles become multinucleate at more or less the same time in the development of the tadpole. In the former type of fibre this is long after contractility and nervous control have appeared, in the latter it precedes contractility. The establishment of contractility and multinucleation are not, therefore, causally related. It is significant that multinucleation and metamorphosis are temporalily linked and in view of the work detailed in the preceding paragraphs it is suggested that the difference in development of myotome and limb muscle in Xenopus, with respect to the stage at which they become multinucleate, is due to some substance produced just before or during metamorphosis.

I should like to thank Professor Alastair Graham for many helpful discussions, Mrs Mary Tolan for assistance with the histology and Mrs Pat Hawkins for assistance with the photography.