ABSTRACT
The time required for foot end formation increases with distance from the foot end. Using lateral grafting it was shown that both the peduncle and basal disc could induce a proximal axis when grafted into the gastric region. The time for foot end determination was shown to be about 4 h at the proximal end of the gastric region and to increase towards the head end. In general the determination of the foot end is similar to that of the head end.
INTRODUCTION
Much less attention has been given to the development of the foot region than the head region during pattern regulation in hydra (see Webster, 1971). Our previous papers (Wolpert, Hicklin & Hornbruch, 1971; Hicklin, Hornbruch, Wolpert & Clarke, 1973) have also concentrated on the head end boundary, and the model using gradients in positional signal and positional value was primarily developed to account for the formation of the head end. Boundary regions are of key importance in considering positional fields (Wolpert, 1971) and it is essential to know whether the foot end boundary has properties similar to the head end and what additional features must be added to the model to account for formation of the foot end. The model for head formation suggests that a new head will form when the concentration of a diffusible substance S falls a threshold amount below its original concentration which is represented by a non-diffusible parameter P. Since S is made at the head end, and is constantly being destroyed, removal of the head will lead to S falling and this leads to the formation of a new head. It would be very attractive if the interactions between S and P could also account for foot end formation: for example, might it not be possible that foot end formation was initiated when the concentration of S rose a threshold amount above P? Unfortunately this simple symmetrical model cannot easily be sustained since there are a variety of grafts in which just the contrary is observed. For example, in grafts of H12/12…F (the hydra is represented by H1234B56F as in previous papers) S will rise at the proximal end of the distal 2 region but no foot forms, whereas in 12/12…F, when S is lower there, a foot often forms. From a variety of such experiments we have concluded that a signal from the head end can inhibit foot end formation.
There is good evidence that the presence of a foot can prevent the formation of another foot in adjacent regions (MacWilliams & Kafatos, 1968). This ability of the foot end to inhibit foot formation is of course analogous to the properties of the head end and this analogy is further strengthened by our results (Hicklin et al. 1973) which showed the symmetrical behaviour of H12/12…F and H12…56/56F type grafts. It would seem that the simplest model to accommodate the results would be to postulate another gradient in a diffusible substance S ′ which would be produced at the foot end. Foot formation would then depend on S ′ falling to within a threshold amount of A–P, where A is a constant greater than the maximum value of P. In addition S would tend to inhibit head formation. It will be important to know if S ′ can also influence P. Is, for example, the formation of regions 56, the peduncle, controlled by a positional signal from the foot end? The answer to these questions is not clear. If there are two positional signals, what might be the relationship between them? There is some evidence that the presence of a head inhibits foot formation (Hicklin et al. 1973) and thus that a signal from the head may be dominant. In this paper we investigate the factors controlling the formation of the foot using the lateral grafting techniques employed by Webster (Webster & Wolpert, 1966; Webster, 1966) for the head end. In this, the dynamics of foot end formation are investigated by finding the time of foot determination as well as the time at which foot end structures appear. When investigating determination of the head end, grafts were taken from the tips of regenerating animals at various times – if these resulted in the formation of a head end when grafted laterally into host gastric regions (rather than being absorbed), a new head end was considered to have been determined in the regenerating animal. The time of determination of the head end was found to vary according to axial position of the regenerating distal end. Here we use a similar technique for the foot end.
MATERIALS AND METHODS
All experiments were performed using Hydra littoralis as described in Hicklin et al. (1973) and carried out at 26°C.
Lateral grafts were performed as described by Webster (Webster & Wolpert, 1966). Pieces of tissue were obtained from donor animals by removing transverse rings of tissue from selected body regions and cutting these into small pieces. If the donor animal was regenerating, a small disc of tissue was taken from the regenerating end, bisected, and one of the two pieces was used as the graft. The graft was implanted in the body wall of the host in a small incision at the selected site. Great care was taken to ensure that the endoderm of host and graft were in contact for correct healing, that the graft did not project beyond the edge of the wound and that the size of the grafts was constant throughout all these experiments.
Where it was important to minimize variation between time of cutting and time of grafting, only small groups of animals (5–10) were used; the totals in the tables therefore represent the pooled results of several experiments.
Experiments involving the transplantation of small amounts of stained hypostome and basal disc (Browne, 1909; Yao, 1945) have shown that where a secondary axis is developed as a result of lateral grafting, the graft persists at its distal end. It is thought that the presence of the graft influences cell movement and cell orientation in the host (Webster, 1971). In the present experiments it was noted that the first change which indicated positive induction was a localized swelling of the body wall, and became evident within 2 or 3 h after grafting. The development of the outgrowth and its differentiation occupied no more than 48 h. The secondary axes could be recognized as partial distal or partial proximal. A secondary distal axis, the characteristic result of grafting hypostomal tissue, usually consisted of a short outgrowth the diameter of which was approximately equal to that of the distal gastric region of the host; it terminated in a head region (Fig. 1a). Three other results have been described by Browne (1909), Yao (1945) and Webster (Webster & Wolpert, 1966) in similar experiments and some of these are illustrated in Fig. 1 (b,c). These were occasionally observed in the course of the experiments described here. The only type which was not classed as a positive induction was Fig. 1(c), where just one or two tentacles formed. This was rare and with time the tentacles were almost invariably absorbed. All the other types were characteristically stable and were not absorbed.
Examples of induced secondary distal axes following grafting (arrow). (A) Short axis with hypostome and tentacles. (B) Short axis with long single tentacle. (C) Single tentacle.
A secondary proximal axis consisted of a peduncle and a basal disc, and it was formed when pieces of tissue from either of these regions were grafted to the gastric region of a host animal (Fig. 2). Occasionally, tissue grafted from proximal regions apparently differentiated as a small basal disc in the side of the animal. Such basal discs sometimes later disappeared and were not counted as positive inductions.
It was noted that induced proximal axes did not appear to increase in size after the first 48 h; the same appeared to be true for induced distal axes unless the animals were fed. Fed animals with two heads became Y-shaped and then V-shaped before splitting in two. On a very few occasions it was observed that when the branch point of the two distal axes reached the budding zone, an outgrowth developed on one of the distal axes close to the fork, and this outgrowth differentiated into a peduncle and basal disc. In these instances, the separation occurred by cleavage of the isthmus of tissue which connected each set of distal and proximal structures.
One interesting phenomenon observed was the speed at which an induced proximal axis would migrate down the gastric region, traversing perhaps its entire length in a matter of a few days. However, once a proximal axis had reached the lower limit of the budding zone it moved only very slowly, gradually fusing with the host’s peduncle over the period of a week or so. Distal inductions remained at the distal end of the animal and appeared to compete with the host head for control of the axis.
RESULTS AND DISCUSSION
Time required for different regions to form basal disc
A simple cutting experiment was first of all performed to investigate the time taken for foot regeneration in animals cut at different regions. The formation of a basal disc at the proximal end was gauged by testing for the known adhesive properties of the basal disc region: the dish containing the animals was agitated for a few seconds and if an animal remained attached to the bottom it was judged to possess a basal disc. Meticulous care was taken to keep the dishes free from any debris which might prevent it from attaching to the dish. Any detached buds were removed.
The results (Table 1) show differences in the time required to form a basal disc. In the gastric region these times appear to be graded up the axis while basal disc formation occurs much faster from the 5 region than from the 2 region. H1 regions formed basal discs very slowly, and most had not done so even after several weeks. It is interesting that H1234B animals which had been cut just under the youngest bud took as long to regenerate the basal disc as did H1 regions. Results from H1234 animals were similar to those from H1234B5 animals: by 24–48 h they had all regenerated a basal disc.
Regeneration of the peduncle and basal disc in animals cut across the gastric region (i.e. H1, H12, H1234) was marked by a series of visible changes whose duration appeared to vary according to the axial position of the regenerating end. There was first a noticeable narrowing of the most proximal end, which in the H1234 animals was evident as early as 4 or 5 h after isolation. This narrowed zone then began to acquire the pale translucent appearance characteristic of peduncle tissue. At a later stage the basal disc could be distinguished at the proximal end, though it was not until the basal disc had visibly differentiated that the peduncle assumed quite its normal form and proportions, having until then appeared somewhat shorter and broader than normal.
Although H1234B regions also appeared to follow this sequence of morphological changes, in most instances the animals remained in the first stage for several days without regenerating a basal disc. After 7 days all had regenerated. Neither bud initiation nor bud detachment appeared to be affected during regeneration in these animals.
Lateral grafts of peduncle and basal disc
No reports are known in which it is shown that region 56, the peduncle, is capable of giving rise to a new foot end following lateral transplantation into another animal. However, it is well known that the outcome of any grafting operation depends on (a) the source of the graft and (b) the site to which the graft is transplanted in the host. For example, Webster (1966) found that a graft taken from region 1 was absorbed when transplanted into the midgastric region of an intact animal, but resulted in the formation of a head end when transplanted to its foot.
The following experiment investigated the behaviour of peduncle and basal disc grafts transplanted to different positions on the host axis (Table 2).
Pieces of tissue of approximately the same size were isolated from three positions on the linear axis of donor animals: from the foot (about one quarter of the basal disc) and from the top of regions 5 and 6. Grafts were then transplanted laterally into hosts. Grafting basal discs was particularly difficult because this region is so sticky; also the thick ectoderm caused some grafts to roll up and enclose the endoderm, the cells of which are relatively sparse in this part of the animal. It is suspected that this rolling-up may have affected the success of the grafting operations in a few instances.
Grafts of basal disc were always found to be absorbed when transplanted to a host’s head region; occasionally the graft passed into an elongating tentacle where it was not absorbed, but was lost from the tip within a few days. Basal disc grafts to the gastric region resulted in the formation of a proximal axis (i.e. consisting of a peduncle and basal disc) in 20 out of 27 animals. From its earliest stages the induced axis had the morphological appearance of peduncle tissue and within 24 h a basal disc of normal size was evident at its tip. When the peduncle was chosen as host site, the foot graft did not induce an axis but migrated to the basal disc as a small sticky patch. Once this migration had been accomplished, the graft disappeared, having fused with the glandular surface of the disc.
Grafts from regions 5 and 6 to the midgastric region of hosts resulted in a high proportion of proximal axes forming. Peduncle tissue grafted into the gastric region resulted in the formation of an outgrowth which retained the dense texture and pink colour of gastric region tissue for the first 24 h or so after grafting. Between 24 and 48 h after grafting this outgrowth became pale and translucent and a basal disc formed at its tip. When the graft was taken from region 6, the basal disc seemed to form a little sooner than when the graft was from region 5. Grafts of peduncle (region 5) into the peduncle of hosts were invariably absorbed.
In one animal (not in Table 2) a graft into the head region healed into the subhypostomal region of the host, resulting in the formation of a tiny axis consisting of a peduncle and basal disc – it was not clear whether this was an induction or whether part of the graft had reorganized into peduncle.
The most important point to emerge from these results is that both peduncle and basal disc appear to be able to bring about the formation of a secondary proximal axis when transplanted to the midgastric region of a host animal. The finding that peduncle tissue can induce seems to conflict with previous reports concerning the behaviour of grafts taken from this region. It may represent a species difference. The very high proportion of proximal inductions which resulted from grafting basal disc into the midgastric region suggest that this is a standard result. The technical difficulty of grafting fragments of basal disc could well account for the finding that seven of the grafts were absorbed.
On the basis of these results alone, it is not possible to differentiate between the behaviour of grafts taken from the peduncle and those from the basal disc. There was, however, a difference when grafts of peduncle and of foot were transplanted to the mid-peduncle of host animals. No inductions were formed as a result of grafting either of these regions, but peduncle grafts were completely absorbed while foot grafts appeared to be stable until lost by attrition. This confirms the finding of MacWilliams & Kafatos (1968) that a grafted basal disc will inhibit the formation of another disc in its neighbourhood (see also MacWilliams, Kafatos & Bossert, 1970).
Time of determination of proximal regions
The above experiment seems to indicate that both peduncle and basal disc can elicit the formation of a secondary proximal axis consisting of peduncle and basal disc when transplanted into the midgastric region of an intact host animal. These results do not, however, resolve the question whether peduncle specifically can induce a proximal axis without first having been determined as a foot end. The test also provides a means of investigating the time at which a regenerating proximal end is first capable of inducing proximal structures when grafted laterally into a host animal, which may be taken to be the time at which the new foot end is determined. We have investigated the time of determination of the proximal regions in regenerating H1, H12, H123, and H1234 tested by lateral grafting into the midgastric region (regions 2 and 3) of intact host animals.
Donor animals were cut to provide H1, H12, H123, and H1234 pieces and at various times after cutting the proximal tip of the regenerating animal was removed, split into two pieces and one of these pieces was then transplanted laterally into the midgastric region of an intact host animal. All grafts were made before the animals had regenerated a basal disc.
As might be expected the results (Table 3) show differences in the time at which proximal structures are determined in the four groups of regenerating animals. From the results, estimates were made of the approximate time after which 50% of the grafts belonging to each series induced proximal axes. This will be referred to as the T50 for foot determination. For the H1234 this was about 4 h, for the H123 about 8 h, and for H12 about 12 h. These results indicate an axial gradient for proximal end determination, which may be analogous to that for head end determination (Webster & Wolpert, 1966).
Occasionally, distal axes and sometimes mixed distal and proximal axes were obtained when the proximal end of a regenerating H1234 region was transplanted, but the frequency of such inductions diminished with longer regeneration periods. This result is probably due to the inclusion of determined bud material in the 4 region (Sanyal, 1966). Where mixed inductions were obtained they have been included as proximal inductions in Table 3. In a small number of animals belonging to each series the graft only appeared to differentiate into a basal disc. These have been recorded in Table 3 as having been absorbed. Although these could be classed separately, because of their rareness this was not done.
Determination of the head end from the distal end of region 5 (the peduncle)
Webster & Wolpert (1966) showed by lateral grafting that a new head end became determined at the top of region 5 in regenerating 56F about 18 h after cutting. The results in Table 2 indicate that a graft taken from the distal end of region 5 can induce a peduncle and basal disc when immediately grafted into the gastric region of a host animal. There will therefore presumably be a period when, if tested by lateral grafting, regenerating 5 tissue can be said to be undergoing a change in properties. This may be shown either by a period when the inductions are predominantly both distal and proximal or a period when foot end inductions have ceased and head end inductions have not yet started. An experiment was performed to investigate these possibilities.
Animals were cut at the top of region 5 and at various times after cutting the regenerating top was removed, bisected and half was then implanted into the gastric region of an intact host animal. It was found (Table 4) that grafts taken from the distal tip of 56F which had been regenerating for less than 6 h formed a high proportion of proximal axes consisting of peduncle and basal disc. After 12 h of regeneration, fewer proximal axes resulted but in a number of instances the graft differentiated either into a basal disc or into one or two tentacles which were eventually absorbed. After 18 h of regeneration, distal axes predominated in grafting tests, a finding which accords with Webster & Wolpert’s (1966) results.
These results suggest that the switchover point from foot-like to head-like transplantation properties occurs at the distal end of an isolated 56F region after about 12 h of regeneration.
Effect of the presence of the head on the formation of proximal regions
A question of considerable theoretical significance is whether the presence of the head affects the regeneration of proximal regions. Previous attempts to investigate this have in general been unsatisfactory, relying as they have done solely on the reappearance of the basal disc at the cut end (Mookerjee & Bhattarcharjee, 1967). We (Hicklin et al. 1973) have suggested that the head end may inhibit foot formation. We have further investigated this by comparing the behaviour of grafts placed in intact hosts with those placed in hosts from which the head had been removed.
Pieces of tissue were isolated from regions 3 and regions 5. In one series, region 3 grafts were transplanted to the midgastric region of host animals, and in another to the foot; region 5 grafts were transplanted to the midgastric region of hosts. Immediately after each grafting operation the host’s head region was removed by cutting at the top of region 1. Control experiments using intact animals were run for each series (Table 5).
The results of grafting region 5 give some indication that the presence of a head region may have an inhibitory effect on the development of a foot end. Region 5 grafts transplanted into decapitated animals resulted in a significantly greater number of proximal axes forming than in the intact control group (P less than 0·05 x2 using test). No such difference appeared when grafts were taken from region 3 and transplanted to the midgastric region of host animals; instances of induction were rare anyway. Grafts of region 3 to the foot sometimes resulted in the formation of a tiny secondary distal axis, and though the proportion of animals giving this result was slightly higher in the decapitated group than in the control series the difference is not significant. The development of these secondary axes was peculiar in that, unlike grafts of head region to the gastric region, an outgrowth formed which only later developed tentacles at its distal end. In this respect it resembled bud development, except that the induced distal axis did not separate from the host, at least for several days.
This formation of a secondary axis in the peduncle as a result of grafting gastric region into it extends Webster’s (1966) finding. In this, a graft of region 1 when transplanted to the foot was found to result in a secondary axis with hypostome and tentacles whether or not the host was regenerating a head at its distal end.
DISCUSSION
The interesting point to emerge from these experiments is that determination and development of the foot end appears to be symmetrical with that at the head end. Like grafts from the head end, lateral grafts from the foot end are capable of ‘inducing’ or organizing a secondary axis. This property is also shown by peduncle tissue grafted to the gastric region though not to the peduncle. This is a new finding.
The times for foot regeneration and determination are also similar to those for the head end (see Webster & Wolpert, 1966). Firstly there is a gradient in the time required for the development of the foot after cutting, depending on the distance of the regenerating end from the foot. A foot will form only very slowly from the cut proximal end of H1. This is comparable to the slow formation of the head in 6F animals (Webster & Wolpert, 1966). There is also a gradient in the time for foot end determination as assayed by lateral grafting. The times are short and determination rapid. For peduncle tissue it is not even possible to measure a time for determination since it will induce immediately following lateral grafting, suggesting that determination takes place within 1 h. For region 4 the determination time is only 3 – 4 h. For head determination, the time required by the 3 region is about 10 h (Hicklin & Wolpert, in preparation) and a similar time is required for foot determination. In view of the ability of region 5 to ‘induce’ a proximal axis when grafted into the gastric region it was interesting to observe the behaviour of a 5 region when it was regenerating the head: there was a switch from proximal to distal inductions between 12 and 18 h. It is perhaps surprising that the ability to induce proximal axes had not been lost even after 12 h of regeneration.
We have obtained some further evidence that the presence of the head tends to inhibit foot formation which supports our previous conclusion (Hicklin et al. 1973). A higher percentage of proximal inductions was obtained with grafts from region 5 when the host’s head was removed than when it was present.
These similarities of behaviour exhibited by the foot and head confirm our view that they are the boundary regions of a bipolar positional field. These boundary regions appear capable of organizing a new axis, presumably by establishing a new boundary to the positional field. The head end may be dominant in so far as it may inhibit foot end formation. One may account for foot end formation in a similar way to head end formation: their symmetrical behaviour with respect to axial grafting has already been established (Hicklin et al. 1973). What is far from clear is whether they contribute to the assignment of positional information to the intermediate regions and if so what their relative contribution may be.
Acknowledgement
This work is supported by The Nuffield Foundation.