ABSTRACT
The amphibian limb regeneration blastema is used here to examine whether irradiated, non-dividing tissue can participate in the development of new patterns of morphogenesis. Irradiated blastemas were rotated 180° on normal stumps and normal blastemas rotated on irradiated stumps. In both cases supernumerary elements developed from the unirradiated tissue. The supernumeraries were defective but this did not seem to be due to a lack of tissue. Rather it suggested that this could be a realization of compartments in vertebrate development or simply reflect the limited regulative ability of the blastema. The results are also discussed in relation to a recent model of pattern formation.
INTRODUCTION
There have been several theories proposed to explain the inhibitory action of X-rays on limb regeneration. The original suggestion, that irradiation disrupts cell division, resulted from observations of abnormal mitoses (Butler, 1933; Horn, 1942) and is consistent with contemporary radiobiological theory. Recent work has provided ample support for this contention, as it has been shown that irradiated blastemal cells become blocked in the G1 or G2 phases of the cell cycle (Maden, 1979) and those cells that do divide most likely die as a result of loss of genetic material due to induced chromosome breakages (Maden & Wallace, 1976).
A second, but far less tangible view, is that X-rays destroy the ‘morphogenetic field’ of the limb, thereby rendering it incapable of regeneration (Trampusch, 1958; Oberpiller, 1968). In modern terminology we would say that positional signalling has been inhibited. This contrasts with the first situation where we would say that positional signalling occurs as normal, but cells simply cannot divide. The regenerating limb provides an excellent opportunity to clearly distinguish between these two alternatives since limb field potential is expressed in the development of new patterns of morphogenesis following rotation of the regeneration blastema (Bryant & Iten, 1976; Tank, 1978; Maden & Turner, 1978). Thus by irradiating one half of the interacting system we can examine whether the new pattern elements, which normally appear, still form or are abolished.
Results from other developmental systems reveal that irradiated tissue can communicate with normal tissue, i.e. the ‘morphogenetic field’ is not destroyed. For instance, intercalary regeneration between fragments of tissue derived from opposite ends of Drosophila imaginal discs is still stimulated when one of the fragments has been irradiated with up to 100 krad (Adler & Bryant, 1977). Similarly, in the chick wing-bud a heavily irradiated zone of polarizing activity (ZPA) can induce reduplications in the same fashion as a normal graft (Smith, Tickle & Wolpert, 1978). The amphibian limb affords us an additional opportunity to differentiate irradiated and normal tissue by exploiting the natural colour difference between black and white axolotls. In the work described here either the stump or blastema was irradiated and a blastema of opposite genotype to the stump was grafted and rotated through 180°. Thus it was revealed that irradiated, non-dividing tissue is capable of positional signalling which leads to the development of supernumerary elements.
MATERIALS AND METHODS
The experiments were performed on black and white axolotls, Ambystoma mexicanum, 70 –80 mm long, whose forelimbs had been amputated through the mid-humerus. When the blastemas had developed to the early digit stage the forelimbs and shoulders of half of the animals were irradiated. The rest of the body was covered with 3 mm thick lead plates during exposure to 2000 R of filtered X-rays from a Newton and Victor Ltd X-ray machine at 90 kV, 5·5 mA.
Immediately after irradiation, blastemas were exchanged between the irradiated and the unirradiated animals to produce irradiated blastemas on unirradiated stumps and vice versa. All blastemas were placed on the corresponding stump of the host and then rotated through 180°. Both grafts and hosts were marked with a thin red line of carmine particles inserted under the epidermis with a tungsten needle (see Maden & Turner, 1978) in order to observe any displacement of the graft. Black and white animals were used in pairs such that all grafts were of opposite colour to the host and in this way the tissue that contributed to any new growth could be determined (see Wallace & Wallace (1973) for a detailed discussion of the use of this colour marker).
All animals were left for 2 h at 4°C to promote normal healing and sticking of the grafts and then returned to water. The limbs were observed weekly for 5 –6 weeks, then fixed in neutral formalin and stained with Victoria blue to examine the skeletal structure.
RESULTS
Controls for irradiation
Four limbs were irradiated with 2000 R and then amputated to ensure a sufficient dose of X-rays had been administered. None of them regenerated. In addition to these controls, all the animals which supplied irradiated grafts were kept to ensure no regeneration occurred from the stumps. Also, after fixing the limbs at the termination of the experiment, all those animals which had irradiated stumps were maintained for several weeks. This amounted to a total of 34 irradiated stumps, none of which regenerated, confirming that the dose of irradiation was adequate to suppress any further tissue growth.
Series I: White blastemas on irradiated black stumps
This was a control series to ensure that the trauma of transplantation and/or the heterotypic stump tissue would not cause any supernumerary outgrowths. The APDV axes were kept in harmony.
Following transplantation of 12 early digit regenerates they regressed back to the notch stage, a consistent observation throughout this work. After a delay of about a week development of the grafts resumed and in all cases a complete limb resulted. Nine limbs were perfectly normal, one produced one extra carpal, two had one extra digit (d.4) and two or three extra carpals (Table 1). It is not unusual for one extra digit to be produced following control blastemal trans-plantation.
The plane of the graft was readily observable in cleared limbs by the abrupt black/white boundary halfway down the humerus (Fig. 1), thus demonstrating the usefulness of this natural colour marker. Despite irradiation having prevented intercalary regeneration from the stump, any slight deletions in the humerus which might be expected due to loss of proximal tissue at transplantation were not observed.
The result of grafting a white proximal blastema onto a black proximal limb stump keeping the axes in harmony. The boundary between graft and host is clearly marked (arrow) by the absence of melanophores in the soft tissues of the white graft. White tissue prevents the migration of melanophores and hence their penetration from adjacent black tissue. Victoria blue staining.
The result of grafting a white proximal blastema onto a black proximal limb stump keeping the axes in harmony. The boundary between graft and host is clearly marked (arrow) by the absence of melanophores in the soft tissues of the white graft. White tissue prevents the migration of melanophores and hence their penetration from adjacent black tissue. Victoria blue staining.
Series II: White blastemas rotated 180° on irradiated black stumps
From 12 rotations which developed (three regressed), ten blastemas produced supernumerary outgrowths (Table 1). In contrast to the previous control series (Series I), all these rotated blastemas lost proximal tissue and the resulting regenerate had at least the distal humerus and proximal radius and ulna missing (Fig. 3). Thus the effect of rotation must have caused this loss of proximal tissue from the blastema. Eight of the ten limbs produced single supernumeraries, the other two were double. All supernumeraries were composed of white tissue (Fig. 3) and they arose mostly at the zeugopodial level. Thus irradiation of the stump did not decrease the high incidence of supernumerary induction after 180° rotation (Maden & Turner, 1978).
However, three digits was the maximum development obtained in any supernumerary (Fig. 3). Furthermore, in those cases where their polarity could be determined the single supernumeraries were all found to be composed of posterior digits. In the two limbs that produced double supernumeraries, one was a posterior part of a limb and the other possibly anterior (Fig. 3). This situation contrasts markedly with the ease with which complete supernumeraries are produced after 180° rotation in the absence of irradiation (e.g. Fig. 2). A comparison of these two figures is particularly revealing for we might have expected irradiation of the stump to result in only one supernumerary being produced rather than two. Instead both supernumeraries seem to have been attempted, but neither completed. During normal supernumerary induction the extra limbs are composed of both stump and blastemal tissue (Maden & Turner, 1978). Thus it seemed that after stump irradiation the rotated blastema only produced what it would have done anyway (with some bias towards posterior tissue) without compensating for the missing stump contribution. The implications of this phenomenon are considered in the Discussion.
Two supernumerary limbs (S1 and S2) produced after rotating a normal blastema 180° onto a normal hindlimb stump. Compare with Fig. 3.
The result of grafting a white blastema 180° onto a black irradiated stump (left forelimb). Deletions in the proximo-distal axis occur and supernumerary elements arise at the carpal level. Note that all the supernumerary elements (S1 and S2) are composed of white (no melanophores) tissue. The arrow marks the stump/ graft junction. S1 is a three-digit posterior half of a left limb. S2 has one digit and four carpals and is perhaps the anterior part (d.l) of another supernumerary.
The result of grafting a white blastema 180° onto a black irradiated stump (left forelimb). Deletions in the proximo-distal axis occur and supernumerary elements arise at the carpal level. Note that all the supernumerary elements (S1 and S2) are composed of white (no melanophores) tissue. The arrow marks the stump/ graft junction. S1 is a three-digit posterior half of a left limb. S2 has one digit and four carpals and is perhaps the anterior part (d.l) of another supernumerary.
Of the 15 blastemas rotated in this series, four derotated, demonstrating that blastemas can derotate on irradiated stumps.
Clearly then we can conclude that irradiated and normal tissue can interact to produce supernumerary outgrowths even though the former does not contribute cells to the extra limbs.
Series 111: Irradiated white blastemas rotated 180° on normal black stumps
In all ten cases in this series intercalary regeneration from the stump was extensive. The irradiated blastemas contributed at most to the digits of the limb and in several cases had completely disappeared (Fig. 4). Nevertheless five out of ten limbs produced supernumerary elements (Table 1). In all cases these were composed of black tissues, from the stump, and the splitting occurred at the level of the zeugopodium as in Series II. Again, as in Series II, the supernumeraries were deficient comprising one (sometimes both) zeugopodial element and at most three digits (Fig. 4).
The result of grafting a white irradiated proximal blastema onto a black stump. The resulting supernumerary elements S (radius, four carpals and three digits) is entirely composed of black tissue-note presence of melanophores throughout the limb. Thus the white grafted blastema had disappeared but managed to remain long enough for supernumerary elements to develop.
The result of grafting a white irradiated proximal blastema onto a black stump. The resulting supernumerary elements S (radius, four carpals and three digits) is entirely composed of black tissue-note presence of melanophores throughout the limb. Thus the white grafted blastema had disappeared but managed to remain long enough for supernumerary elements to develop.
Thus despite the fact that the irradiated, rotated blastema virtually disappears, half of the limbs in this series produced supernumerary elements at the zeugopodial level, confirming that irradiated and normal tissue can interact.
None of the blastemas in this series derotated by more than 60°. In order to confirm that this was not due to the suppression of derotation by irradiation of the blastema, an additional experiment was performed in which both the stump and blastema of four limbs were irradiated and then rotated. Two of these four blastemas completely derotated back to 0° even though no further development occurred. Clearly then, derotation is not a product of directional growth.
DISCUSSION
The aim of the work reported here was to determine whether irradiated tissue, which is prevented from further cell division, can interact with adjacent unirradiated tissue to stimulate new patterns of morphogenesis. When normal blastemas are rotated 180°, in a high percentage of cases supernumerary limbs develop. When either the stump (Series II) or blastema (Series III) was irradiated supernumerary limbs appeared with a similar frequency. Derotation was also observed and hence not inhibited by irradiation. Clearly then, despite the inhibition of division, positional signalling still occurs in X-rayed tissue and the ‘morphogenetic field’ of the limb is not destroyed (Trampusch, 1958; Oberpiller, 1968). Thus the behaviour of the amphibian limb in this regard is similar to the chick limb-bud and Drosophila imaginal discs (see Introduction).
Black and white axolotls were used to provide a colour difference between the rotated blastema and stump. In this way it was revealed that the supernumerary limbs were always derived from unirradiated tissue, thereby attesting to the value of this natural genetic marker. It is important to emphasize that these supernumeraries were never fully formed-two or three digits was the maximum development obtained and they were almost all posterior digits. In two limbs two partial supernumeraries were produced, one (possibly an anterior partial limb) being more poorly developed than the other (posterior partial limb). It is likely that this is due to a specific interaction rather than a simple insufficiency of tissue for several reasons. Firstly, during normal supernumerary induction both stump and blastema contribute to the extra limbs (Maden & Turner, 1978) and it was noted here that the unirradiated graft seemed to produce approximately what it would have done if the host had not been irradiated. Secondly, this phenomenon is consistent with the nonequivalence in developmental capacity revealed in double half-limb studies (Stocum, 1978) where the mean number of toes formed by double posterior regenerates was twice that from double anterior regenerates. My own unpublished observations of the regeneration of half irradiated limbs (as in Goss, 1957) also reveal an inequality in that the posterior half produces more digits than the anterior half. These observations suggest that there is more to this anterior-posterior inequality than the position of the major nerves in the limb (Singer, Ray & Peadon, 1964): possibly it is a manifestation of an unequal anterior-posterior fate map of the limb-bud. But whether the inability of the blastema to compensate for the missing stump contribution is the first realization of the presence of compartments in vertebrate development or simply related to the limited regulative ability of the blastema, awaits further experimentation.
Finally, regarding the formalism known as the clockface model (French, Bryant & Bryant, 1976) it is clear that these partial supernumeraries invalidate the complete circle rule for distal transformation. This rule predicts that all regenerated limbs, whether supernumeraries or normal regenerates, will either be complete or be incapable of any distal growth. The above results, therefore, add to the growing body of evidence from the regeneration of double half-limbs (Stocum, 1978; Slack & Savage, 1978), the regeneration of half irradiated limbs (Goss, 1957; Maden, unpublished) and the production of supernumeraries at angles other than 180° (Maden & Turner, 1978; Wallace, 1978) which suggest that this formalism is an unsatisfactory description of the events of limb regeneration. Perhaps it would be better therefore to dispense with the hypothesis altogether, in favour of one which has no complete circle rule, permits supernumerary limbs to appear after passing a threshold of rotation in the APDV axis and takes account of unequal anterior-posterior contributions to the limb. However, the rich diversity of behaviour of the amphibian limb field which is currently being revealed must surely present severe difficulties for any new, all-encompassing theoretical expositions.