ABSTRACT
The normal development of the brachial ventral horn of the frog Xenopus laevis and the response of the brachial ventral horn to complete forelimb extirpation at five developmental stages were assessed histologically. Differentiation of brachial ventral horn neurons occurred in pre-metamorphic tadpoles between stages 52/53 and 57. Mean cell number in the brachial ventral horn reached a peak of 2576 (S.E.M. = ± 269, n = 2) per side of the spinal cord at stage 55 and decreased to 1070 (S.E.M. = ± 35, n = 7) by the end of metamorphosis. Cell degeneration was presumed to be the mode of cell loss since it was most prevalent during the period of rapid decrease in cell numbers. The response of the ventral horn to forelimb removal varied with the stage of the animal at amputation. Following amputation at stage 52/53 or 54 the ipsilateral ventral horn neurons appeared less differentiated than those on the control side and a rapid cell loss of about 80% occurred on the operated side. These effects occurred more rapidly after ablation at stage 54 than at stage 52/53. Amputation at stage 58, 61, or 66 caused chromatolysis in the ventral horn, a period of relative cell excess on the operated side, and a delayed neuronal loss of 32–66%. It was concluded that excess cell degeneration accounted for cell loss and that suppression of normal neuronal degeneration caused the relative cell excess on the operated side. The data indicate that the brachial ventral horn was indifferent to the periphery before stage 54, was quickly affected by limb removal between stages 54 and 58, and by stage 58 had entered a phase in which a delay preceded cell death. No forelimb regeneration occurred.
INTRODUCTION
Both autonomous and dependent differentiation play important roles in the development of the vertebrate central nervous system. Motor and sensory neurons have been particular targets of research efforts to distinguish between peripheral and central influences on the developing spinal cord. The present study investigated the relative importance of autonomous and dependent differentiation in the development of the brachial ventral horn neurons of the frog Xenopus laevis.
The ventral horns, or lateral motor columns, are groups of nerve cell nuclei that are located in the spinal cord at brachial and lumbar levels and provide motor innervation to the limb and girdle muscles. In Xenopus laevis the ventral horn is formed in larvae at stage 50–51 (see Table 1 for staging guide) by migration of undifferentiated neuroblasts of the ventricular layer (Prestige, 1973) and neuronal differentiation is complete by stage 57–58 (Kollros, 1956; Hughes & Tschumi, 1958). During this period of growth and differentiation cell degenerations result in a decrease in total cell number in the ventral horn (Hughes, 1961; Prestige, 1967). This developmental pattern is similar to that exhibited in other vertebrate species (Hamburger, 1958; Hughes, 1968 a, b; Kollros, 1968).
Abbreviated staging guide for Xenopus laevis (taken from Nieuwkoop & Faber, 1956)

Previous studies on several vertebrate classes suggested that although the initial migration and differentiation of motor neuroblasts may be independent of the periphery, the presence of the limb is necessary for the maintenance of the motor neurons that innervate it (Hamburger, 1958; Hughes, 1968 a, b;Kollros, 1968). Although nerve cell death is the final effect of limb removal, Prestige (1967) found that the time of nerve cell body degeneration in Xenopus lumbar ventral horns depended on the stage at which limb amputation was performed. Prestige has suggested that lumbar motor horn neurons pass through three temporally successive phases. Phase I neurons are insensitive to the periphery; limb removal has no effect. When neurons enter phase II they come under peripheral control and quickly degenerate if the limb is amputated. Phase III neurons are not immediately affected by limb removal, but die after a period of time. The older the animal at the time of amputation, the longer the
Forelimb amputation during development in Xenopus 455 phase III neurons survive. To account for this variation of response with time, Prestige suggested that neurons receive and accumulate a ‘maintenance factor’ from the limb and that they die only when their store of the factor is exhausted.
In the present experiments forelimbs of Xenopus tadpoles or juveniles were extirpated at five developmental stages and the response of the brachial ventral horn was assessed histologically at various times after amputation. Since Hughes (1968 d) reported that Xenopus forelimbs have extensive powers of regeneration, it was expected that regeneration would follow forelimb removal and that a study of ventral horn nuclei reacting simultaneously to extirpation and regeneration would further elucidate the mechanism of peripheral control of ventral horn size.
MATERIALS AND METHODS
Experimental animals were the progeny of three matings of animals of the Cornell Xenopus colony and were maintained at 22–24 °C in aged tap water. Left forelimbs of animals at stages 52/53, 54, 58,61, and 66 were amputated as close to the body wall as possible. Operations on stage-66 animals were performed on the day metamorphosis was completed. A small tear was made in the operculum in order to gain access to the limb-bud at stages 52/53 and 54. This wound healed within a few days.
Animals were anesthetized with M.S. 222 (tricaine methane sulfonate, 1:1000) and operations were performed in full strength Steinberg saline (Hamburger, 1960) to facilitate blood clotting. Animals were maintained in water containing sulfadiazine sodium for several days after amputation. The loss of the forelimb did not cause the animals any difficulties in swimming or eating and all developed normally. Operated animals were observed closely for forelimb regeneration.
Animals were fixed in half-strength Bouin’s at various times after limb extirpation. Specimens were decalcified with Decal solution (Scientific Products) for 3–12 h prior to dehydration and clearing. Serial sections (10 μm) of the brachial region were cut and stained with Erlich’s hematoxylin and eosin.
Motor neurons of the brachial ventral horn were counted in every third section on both the operated and control sides of the spinal cord and the appearance of the cells was noted. The brachial region of the ventral horn is located in the area of the second and third spinal ganglia (Kollros, 1956) and is fairly well demarcated since organization of motor neurons into a definite column projecting from the mantle layer occurs only at hind and forelimb levels. Ventral horn neurons were identified by both position in the motor column and appearance of the cells. Neurons were counted only if they met the following criteria: nuclei paler and larger than those of mantle layer neuroblasts, presence of cytoplasm around the nucleus, and presence of a nucleolus. Since at stages 52–54 neurons are still in the initial stages of differentiation, the cytological criteria could not be applied as rigorously at these stages. Presence of a nucleolus and position in the ventral horn were the principal counting criteria at these earlier stages. The nucleolus is small in comparison to section thickness (Prestige, 1967) and Jones (1937) concludes that if sections are approximately 10 μm thick, no correction need be made for split nucleoli if only definite and distinct nucleoli are counted. Neuron counts included chromatolytic cells. Chromatolysis is a neuronal reaction to axotomy that is characterized by disappearance of the Nissl substance and is followed by a recovery phase in which the nucleus is surrounded by a dense ring of basophilia (Prestige, 1967). Chromatolytic cells were identified by the basophilic perinuclear ring.
The error introduced by the sampling procedure was estimated by complete cell counts of three ventral horns which had 941, 457, and 51 neurons. The average percentage error due to sampling for these counts was ± 1·8 %, ± 9·0 %, and ± 15·7 %, respectively. The corresponding coefficients of variation are 0·02, 0·10, and 0·18. Sample counts ranged from 2844 to 675 on the control side and from 2557 to 46 on the operated side. On the control side the sampling error was usually less than ± 1·8% since most counts were greater than 941. On the operated side the sampling error was less than ± 1·8% for the largest counts and ± 15·7 % for the smallest count. Although the sampling error is large for the smallest counts, the size of the experimental change induced in these cases was much greater than the sampling error.
Degenerating neurons in the ventral horn were not classed as living cells, but were counted separately. No sampling procedure was employed for these counts; every section was examined for degenerating neurons.
RESULTS
Forelimb amputation and regeneration
There was no evidence of regeneration following the 128 forelimb amputations performed in the course of these experiments. Animals whose forelimbs had been amputated at stage 58, 61, or 66 usually developed small bony stumps at the site of amputation. In contrast, ablation at stage 52/53 or 54 resulted in a slight indentation at the site of amputation.
Normal development of the brachial ventral horn
Normal developmental changes in size, appearance, and number of brachial ventral horn neurons were determined by observations of control animals and the contralateral sides of experimental animals. At stage 52/53 the ventral horn cells were still in the neuroblast stage. The neuroblasts were grouped in a distinct rounded mass at the lateral border of the ventral half of the mantle layer (Fig. 1). No cytoplasm was visible around the closely packed, oval neuroblast nuclei. The nuclei had a long axis of about 11 μm, while the diameter of mantle cell nuclei was about 7 μm. Nucleoli were slightly more distinct in motor column nuclei than in mantle layer cells and were approximately 1·5 μm in diameter.
A cross-section through the brachial spinal cord of a stage-52/53 tadpole. The arrows indicate the brachial ventral horns, × 96.
The neuroblasts gradually developed into mature neurons between stage 52/53 and stage 57. The nuclei became round rather than oval, attaining a final diameter of 11–14 μm, while the diameter of the mantle layer cells remained at about 7 μm. The motor cell nucleoli gradually enlarged to a diameter of 3–4 μm, while the surrounding nucleoplasm became increasingly pale. Cytoplasm was first observed around the nuclei at stage 55 and increased in volume as the neurons matured. At stage 52/53 the motor column was a rounded mass projecting from the mantle layer at an angle of about 80° from the dorso-ventral vector of the central canal of the spinal cord (Fig. 1). By stage 58 the neurons were arranged in a column projecting from the mantle layer at an angle of about 45° from the dorso-ventral vector of the central canal (Fig. 2).
A cross-section through the brachial spinal cord of a stage-58 tadpole. The arrows indicate the brachial ventral horns, × 60.
Changes in the total number of neurons accompanied these changes in size, appearance, and position of individual neurons. Table 2 and Fig. 6 A show the average number of neurons in the brachial ventral horn at most stages from 52/53 to 66. While the length of the brachial ventral horn region increased from 680 to about 1000 μm as the forelimb developed, the number of ventral horn neurons dropped gradually from a mean of 2576 (S.E.M. = ± 269, n = 2) at stage 55 to 1070 (S.E.M. = ± 35, n = 7) at the completion of metamorphosis. Counts continued to decrease slightly after stage 66 was reached and stabilized at around 900 per side, although there was much individual variation. Thus, in noting and interpreting changes in neuron number caused by limb extirpation, it must be remembered that the operated side is being compared to a progressively smaller control number.
Numbers of living (—) and degenerating (‐ ‐ ‐ ‐ ‐) neurons in normal brachial ventral horns (A) and numbers of living (—) and degenerating (‐ ‐ ‐ ‐ ‐) neurons in brachial ventral horns on the side of forelimb amputation compared to control ventral horns from the same animals (B–F). The arrows indicate the time of forelimb amputation. Most points represent one animal; some points represent the mean of 2–5 animals (see Tables 2–7).
Numbers of living (—) and degenerating (‐ ‐ ‐ ‐ ‐) neurons in normal brachial ventral horns (A) and numbers of living (—) and degenerating (‐ ‐ ‐ ‐ ‐) neurons in brachial ventral horns on the side of forelimb amputation compared to control ventral horns from the same animals (B–F). The arrows indicate the time of forelimb amputation. Most points represent one animal; some points represent the mean of 2–5 animals (see Tables 2–7).
The number of degenerating neurons observed in control animals was initially very low (1·0 ± 0·7). Degenerations reached a maximum at stages 56–58 (38 ± 14·0–29 ± 11·6) and then declined as metamorphosis approached (Table 2). Only a few cell degenerations were observed in normal ventral horns after metamorphosis (Fig. 6 A).
Amputation at stage 52/53
The first quantitative difference in living cells between operated and control sides occurred 5 days after limb amputation at stage 52/53 and by the end of the first week the number of neurons on the operated side was 62% that of the control side (Table 3; Fig. 6B). By the end of the second week only 34% of the normal number of neurons was observed on the operated side. The changes induced by ablation seemed to be complete by one month after the operation when only 17 % of the normal number of cells remained. During the first week after amputation excess cell degenerations occurred on the operated side as compared to the control side (Table 3; Fig. 6B). After the first week there were fewer degenerating cells on the operated side as compared to the control.
Numbers of ventral horn neurons following forelimb amputation at stage 52/53 (O = operated side, C = control side)

Neurons on the operated side appeared less differentiated as compared to those of the control side from the fifth day after amputation. By the eleventh day there was no motor column as such on the operated side and this was true of all animals killed after that time (Fig. 3). The spinal cord on the operated side was slightly reduced in cross-sectional area by the ninth day; the size difference between the two sides increased with the age of the animals.
A cross-section through the brachial spinal cord of an animal fixed 2 months after forelimb amputation at stage 52/53. Note the absence of a ventral horn on the operated side (right side of photograph), × 60.
Amputation at stage 54
Following limb amputation at stage 54, the first difference in numbers of living cells was noted 3 days later when the operated side had 21 % fewer neurons than the control ventral horn (Table 4; Fig. 6C). The relative difference was 40 % by the end of the first week. The decrease in relative numbers was completed within 2 weeks, when the operated side had only 18% as many neurons as the contralateral side. For the first 5 days after amputation the operated side exhibited a higher rate of cell degeneration than the control side; but starting with day 7, cell degeneration was lower than on the control side (Table 4; Fig. 6C).
Numbers of ventral horn neurons following forelimb amputation at stage 54 (O = operated side, C = control side)

Most neurons on the operated side were less mature by one or two stages than those on the control side during the first week after limb extirpation. After about 9 days there was no real motor column. The few cells that remained on the side of amputation were as mature as those on the control side, but there were not enough of them to form a column. By the twelfth day the operated side of the spinal cord was slightly smaller in cross-sectional area in the brachial region than the unoperated side and this difference was also seen in all animals killed after the twelfth day.
Amputation at stage 58
The number of neurons on the operated and control sides remained approximately equal for the first 5 days following forelimb extirpation at stage 58 (Table 5; Fig. 6D). From the seventh through the thirteenth day, the number of neurons on the amputated side was consistently higher than on the control side, with a difference between the two sides of as much as 18%. By day 16 the operated side had 84% as many neurons as the control side. A gradual decrease followed and after 8 months the number of motor cells on the operated side was 36% that of the control side. From the third through the seventh day cell degeneration on the operated side was lower than on the contralateral side; but from day 10 through day 16 excess degeneration occurred on the operated side (Table 5; Fig. 6D).
Numbers of ventral horn neurons following forelimb amputation at stage 58 (O = operated side, C = control side)

Some neurons on the operated side, particularly in the more posterior regions of the brachial ventral horn, appeared to be in a state of mild chromatolysis between day 3 and day 16. By 25 days after amputation chromatolysis was no longer evident, but there were holes or gaps in the spinal cord in the area of the motor column (Figs. 4, 5). Such vacant spaces were also seen in animals killed after 25 days and were, for the most part, confined to the posterior half of the ventral horn. The spinal cord on the operated side was smaller in cross-sectional area in all animals killed from day 25 on.
A cross-section through the brachial spinal cord of an animal fixed 3 months after forelimb amputation at stage 58. The arrow indicates the ventral horn region on the operated side; note the holes in the tissue in this area, × 60.
Amputation at stage 61
Cell numbers were maintained at a normal level ipsilaterally during the first week after forelimb ablation at stage 61 (Table 6; Fig. 6E). By the end of the first week there was a small relative excess on the operated side. This suppression of cell loss on the operated side produced an excess of 22% by day 18. However, in the following 2 weeks cell numbers decreased sharply on the operated side. The loss continued at a slower rate in the succeeding weeks, yielding a final cell deficit of about 60 %. Numbers of degenerating cells were similar on the two sides except on day 3, when the operated side had a much lower rate, and on day 18 when there were excess degenerating cells on the operated side (Table 6; Fig. 6E).
Numbers of ventral horn neurons following forelimb amputation at stage 61 (O = operated side, C = control side)

Chromatolytic cells were first evident at the end of the first week and persisted through the next several weeks. By the end of the first month chromatolytic cells were no longer present, but there were a number of holes in the ventral horn area of the spinal cord. These gaps were visible in all animals killed after one month. The chromatolytic cells and holes were confined primarily to the posterior half of the spinal cord.
Amputation at stage 66
There was no neuronal reaction to limb extirpation at stage 66 until the end of the first week. At that time and in the next 3 weeks the operated side of the spinal cord contained more neurons (2 % –34 %) than the control side (Table 7 ; Fig. 6F). During the second month cell loss began on the operated side and resulted in a deficit of 32–45 %. Numbers of degenerating cells were small in both amputated and control ventral horns (Table 7; Fig. 6F).
Numbers of ventral horn neurons following forelimb amputation at stage 66 (O = operated side, C = control side)

Chromatolysis began during the first week after amputation and was still evident after one month. Again, the posterior sections contained most of the chromatolytic cells. Seven weeks after the operation chromatolysis had ceased and there were many large holes in the posterior half of the ventral horn region. The spinal cord on the operated side was smaller in the 7-week animal, but no difference in cross-sectional area was evident in the two animals killed thereafter (at 3 and 5 months).
DISCUSSION
In our experiments Xenopus forelimbs never regenerated after total ablation and thus, we cannot confirm Hughes (1968 a) report of the regenerative abilities of Xenopus forelimbs. In the present experiments the indentations that formed at the sites of amputations performed at stage 52/53 or 54 indicate that ablation removed part of the prospective shoulder. Perhaps regeneration did not occur after amputation at these stages because almost all the ‘regeneration territory’ of the arm and shoulder was removed (Guyénot & Ponse, 1930). At later stages (58, 61, and 66) complete limb ablation left the shoulder girdle intact, but by these stages the animals were apparently incapable of regeneration after total limb removal.
The pattern of normal development of the brachial ventral horn was similar to that previously reported for the lumbar ventral horn. As individual cells differentiated, brachial cell number fell from as many as 2576 neurons per side to about 900 (Table 2; Fig. 6 A). Hughes (1961) and Prestige (1967) found that in the lumbar ventral horn a peak number of 3000–6000 neurons per side was reduced to 1200–1600. The smaller number of motor neurons in the brachial as compared to the lumbar ventral horn may be related to the smaller size of the forelimb. The data suggest that reduction in cell number during development of the brachial motor column is caused by cell degeneration. Degenerations were first seen in appreciable numbers just after the peak in cell numbers was reached (Table 2; Fig. 6A). Large numbers of degenerating cells were present only during stages 56–64, the period of rapid cell loss. The size and appearance of brachial motor neurons during development were similar to descriptions previously given by Hughes (1961) and Kollros (1956; 1968,) for the lumbar motor column.
The number and appearance of ventral horn neurons on opposite sides of the spinal cord were always very similar in unoperated animals. The final result of limb extirpation at each of the five stages studied was loss of a considerable number of motor horn neurons on the side of operation. Thus, in Xenopus laevis many brachial ventral horn neurons depend upon the forelimb for maintenance.
The pattern of cell loss and the appearance of brachial motor neurons varied with the stage at which amputation was performed. The older the animal at amputation, the smaller was the final cell deficit (Fig. 6B–F). These differences are probably related to the amount of prospective limb and shoulder tissue removed at each stage, since extirpation was more complete at earlier stages. The timing of cell loss also varied with the stage at amputation. The decrease in cell number following ablation at stage 54 occurred from days 3–18, while operation at stage 52/53 produced cell loss from day 5–30. Thus, although amputation at these two stages resulted in about the same percentage final cell loss, the response was quicker following amputation at stage 54. Amputation at stages 58, 61, and 66 produced a period of relative excess in cell numbers on the operated side. The excess did not result from an increase in cell numbers but reflected a slower rate of cell loss from the operated side as compared to the control side which continued to lose cells at the normal rate. The older the animal at amputation the longer was the period of relative excess : stage 58 – until day 13; stage 61 – until day 18; stage 66 – until day 30 following amputation. In these three experimental series the period of relative excess was followed by a period of cell loss on the operated side. The length of the period of cell loss increased as the age of the animal at operation increased. Cell loss following amputation at stages 58, 61, and 66 was complete by days 25, 31, and 49, respectively (Tables 5, 6, 7; Fig. 6D, E, F).
The ipsilateral changes in cell number seen after amputation can be accounted for by changes in the number of degenerating cells. Excess cell loss coincided with excess cell degeneration. Following amputation at stages 52/53, 54, and 58 cell degeneration was higher on the operated as compared to the control side during the same times when neuronal numbers were decreasing on the operated side (Tables 3, 4, 5; Fig. 6B, C, D). When amputations were performed at stages 61 or 66, excess cell degenerations occurred on the operated side just prior to excess cell loss, but the number of degenerating cells was too small to be conclusive (Tables 6, 7; Fig. 6E, F). However, the holes or gaps seen in ipsilateral spinal cord tissue just after excess cell loss on that side give indirect evidence that excess cell degenerations were the mode of cell loss (Hughes, 1961).
The relative neuronal excess on the operated side following extirpation at stages 58, 61, and 66 can also be attributed to differential rates of cell degeneration. In the stage-58 and -61 groups there were many fewer degenerating cells on the operated side just prior to the period of relative excess on that side (Tables 5, 6; Fig. 6D, E). In the stage-66 series the numbers of degenerating cells were again too small to be conclusive.
Loss of the forelimb affected the appearance of individual motor cells as well as causing changes in cell number. Motor cell differentiation appeared retarded as compared to the control side after amputation at stages 52/53 and 54. This effect appeared more rapidly after amputation at stage 54 (day 1) than after amputation at stage 52/53 (day 5). Limb ablation at stages 58, 61, and 66 caused many motor cells on the operated side to become chromatolytic during the periods of relative cell excess on that side. The fact that chromatolytic neurons, and also the gaps in spinal cord tissue that were seen after excess cell loss in the stage-58, -61, and -66 series, were confined to the posterior half of the brachial ventral horn suggests that these phenomena are related to excess cell loss since cell loss following amputation at stages 58, 61, and 66 was heaviest in the posterior sections of the brachial motor horn.
The results demonstrate that the response of the brachial ventral horn to limb removal varied with the stage at amputation. At stage 52/53 the neurons seemed to exist independently of the limb since the effects of amputation (cell loss and retardation of development) did not appear until several days after amputation at stage 52/53. In contrast, cell loss and delay of differentiation began almost immediately after forelimb ablation at stage 54. Thus, by stage 54 the brachial motor neurons were under the control of the periphery and depended upon it for survival. At stages 58, 61, and 66 the ventral horn was also under peripheral control, but limb removal at these stages caused an initial chromatolysis and suppression of normal cell loss followed by a later decrease in ventral horn neurons. These results indicate that the cell degenerations that normally occur in the brachial ventral horn during stages 58–66 are under peripheral control. So while the limb is necessary for the differentiation and maintenance of some brachial motor cells, it is necessary for the degeneration of others.
The results of our experiments agree in general with Prestige’s (1967) study of the lumbar ventral horn. His division of motor neuron development into three phases is consistent with the results obtained for the brachial ventral horn. If
Prestige’s phases of neuronal development are applied to the brachial motor column, all brachial motor neurons are in phase I prior to stage 54. Brachial neuroblasts enter phase II between stage 54 and 57. By stage 58 brachial motor cells have become mature phase III neurons. Prestige’s hypothesis that the limb supplies phase III ventral horn neurons with a progressively accumulated maintenance factor can explain delayed cell loss following forelimb amputation at stages 58, 61, and 66, but does not explain why normal cell loss was halted by forelimb removal at these three stages.
The results presented in this paper suggest that the initial migration and differentiation of brachial ventral horn neurons are not under peripheral control. Between stages 54 and 57 the neurons appear to enter a phase of development in which they are dependent on the limb for further differentiation and survival. The results thus support the concept that the development of motor neurons involves both autonomous and dependent differentiation.