ABSTRACT
The dynamics of boundary regions have been investigated mainly by axial grafting and with the emphasis on the head end. The time to resist inhibition of head-end formation and the ability to inhibit head-end formation have been assayed under a variety of conditions. The times increase with distance from the head end. The times required by a boundary region to acquire the inhibitory properties of a head end are longer than those required to acquire resistance to inhibition. Determination of a head end is faster at a cut surface and at higher temperatures. The results are discussed in terms of a model involving two gradients. Some anomalous results are reported.
INTRODUCTION
In a previous paper (Hicklin, Hornbruch, Wolpert & Clarke, 1973) we considered the results of axial grafts of hydra in terms of a model based on two gradients, one in a diffusible positional signal S, and the other in a more stable parameter, the positional value P. It was suggested that the results could be interpreted on the basis that a new head end would form if S fell a threshold amount below P. One can think of P as a parameter which tells a cell what state it currently is in, and S as a parameter which can respond to the state of neighbouring cells (Wolpert, Hornbruch & Clarke, 1974). In the model, S is assumed to be synthesized at the head-end boundary and it effectively acts as an inhibitor of head-end formation elsewhere. Evidence has recently been obtained that the transmission of this inhibitory influence is consistent with a mechanism based on diffusion (Wolpert, Clarke & Hornbruch, 1972).
In order to construct a model it is necessary to know how P and S change with time. These cannot, at present, be investigated directly and their dynamics can only be inferred from experiments providing information on the change in properties of regions with time. In this paper, we investigate the dynamics of the boundary region. The boundary region is that region which will become a head end or a foot end. The properties, whose change with time we assay, are resistance to inhibition and the ability to inhibit. Preliminary results were briefly reported in Wolpert, Hicklin & Hornbruch (1971).
The basic approach is that adopted in earlier papers (Webster & Wolpert, 1966; Webster, 1966a, b; reviewed Webster, 1971), and involves determining how long a region takes to become determined as a head end (in the sense that it can no longer be prevented from forming a head end in the particular graft combination used), and how long a region takes to acquire the ability to inhibit head-end formation elsewhere. These parameters can be considered in terms of our current model (Wolpert et al. 1974), which is partly based on the results to be presented here, though it must of course be remembered that other models are possible. The model assumes that the ‘switch’ associated with resistance to inhibition only occurs when P has increased to the value of the head-end boundary. It is assumed that at any time prior to this, a sufficient concentration of S will stop P becoming a boundary region and thus prevent head-end formation. It is also assumed that the acquisition of inhibitory properties is associated with the synthesis of S when P becomes the head-end boundary value. These points are illustrated in Fig. 1. It is very important to realize that the time for determination will be very much affected both by the experimental conditions and the nature of the assay. For example, true determination can only be assayed by placing the piece under consideration in a position where S is at its maximum value. Again, since the time for determination involves both a fall in S and a rise in P, the conditions affecting the rate of fall of S must be taken into account.
Diagram to illustrate the assumed relation between S and P during regeneration. S(a) shows the equilibrium condition. Following the removal of the distal region the concentration of S falls until it reaches a threshold amount below P (b). This initiates synthesis of P and also a small amount of S. P increases now to the boundary value, but if S rises before this value is attained, synthesis of P is inhibited. Thus, in (c) an increase of S could block further synthesis of P. Once P reaches the boundary value as in (d) S is synthesized at the boundary and increases until a linear gradient is re-established (e). Note that no mechanism is proposed here for changes in P away from the boundary.
Diagram to illustrate the assumed relation between S and P during regeneration. S(a) shows the equilibrium condition. Following the removal of the distal region the concentration of S falls until it reaches a threshold amount below P (b). This initiates synthesis of P and also a small amount of S. P increases now to the boundary value, but if S rises before this value is attained, synthesis of P is inhibited. Thus, in (c) an increase of S could block further synthesis of P. Once P reaches the boundary value as in (d) S is synthesized at the boundary and increases until a linear gradient is re-established (e). Note that no mechanism is proposed here for changes in P away from the boundary.
MATERIALS AND METHODS
All experiments were performed using Hydra littoralis as described in Hicklin et al. (1973) and Hicklin & Wolpert (1973) and were incubated at 26 °C. When grafts were made onto regenerating animals, the healed end was carefully prised open with pins. The regions of hydra are referred to by the H1234B56F terminology. The results are scored in terms of structures formed at the junctions as described earlier. Very briefly, N signifies a normal animal with no structures at the junction; H is head at the junction; HF is a head and a foot at junction; F is foot at the junction.
EXPERIMENTAL RESULTS
Dynamics of region 1
Using lateral grafting, Webster & Wolpert (1966) showed that when the H is removed, a new H is determined at the top of region 1 after about 4-6 h. Since the combination 12/12…F nearly always gave structures at the junction whereas H12/12…F was never observed to do this (Hicklin et al. 1973) one may ask how long a regenerating 12…F could be left before an Hl2 graft could no longer inhibit head formation at the junction. After removing the head from prospective host animals, they were allowed to regenerate for up to 6 h, and a freshly excised H12 region was grafted on each (Table 1). After 1–3 h of regeneration before grafting, a very few animals formed distal structures at the junction; those that did formed two or three tentacles only. However, none of the animals which had been regenerating for 5 h could be inhibited when combined with an H12 graft. In every case, distal structures, usually comprising an entire head,formed at the junction and a few animals formed a foot end as well. Feet were not observed to form at the junction in animals which had been regenerating for 6 h before grafting. It is interesting to note that an estimate of the time required for half of the animals to escape from inhibition by a H12 graft gives a figure of about 4–5 h. This is similar to the time for determination of a region 1 using lateral grafting (Webster & Wolpert, 1966) but in view of the different geometries in the two assays this correspondence may be misleading.
The next experiment was designed to determine when the inhibitory properties of a H12 were acquired by a 12; how long must 12…F regenerate before the 12 region will behave like H12 when it is combined with 12…F? Animals were cut at the top of region 1 to remove the head and allowed to regenerate for up to 18 h, by which time tentacle buds had formed at the distal end. At various times after cutting, the 12 region was excised and combined with a freshly cut 12…F host. It was found (Table 2) that about 15 h of regeneration were required prior to transplantation before a substantial number of combinations did not form structures at the junction. Even after 18 h of regeneration, three out of twenty-one animals formed tentacles at the junction. It can be seen that the time of 15 h obtained in the present experiment is considerably longer than the 4–6 h period of regeneration necessary for head-end determination at the top of region 1, which suggests that it takes some time for full inhibitory properties to be established at the distal end. This is a somewhat different conclusion from that which Webster (1966a) arrived at concerning the re-establishment of inhibition along the axis during regeneration; he found that a lateral graft of tissue from region 1 was normally absorbed when implanted into a 12…F animal which had been regenerating for 4–6 h, whereas at earlier times the graft led to the formation of a new head end. The probable explanation for the difference is that the relative position of the tissues in his and our experiments is rather different.
In obtaining information concerning the dynamics of the gradients of positional value and that of the signalling substance, an important question is whether during regeneration it makes a difference whether the regenerating boundary is at a free cut end or not (Hicklin et al. 1973). Combinations consisting of F…21/12…F always form heads (usually two) at the junction (Hicklin et al. 1973) and the time for head-end determination for region 1 from such combinations was obtained by lateral grafting into region 2 of intact animals. This time was compared with the time required by region 1 of regenerating 12…F. The results (Table 3) show that very few grafts taken from F…21/12…F which had been regenerating for less than 6 h resulted in the formation of a new head end in the host, whereas after the same incubation period, the distal tip taken from the controls, regenerating 12…F, induced a head in most instances. After 8 h of regeneration, however, by which time grafts from the region 1 of regenerating 12…F all induced a head end upon transplantation, the majority of grafts taken from regenerating F…21/12…F also appeared to be determined by this time. Whereas just under 5 h of regeneration appeared to be sufficient for half of the animals of the control series to become determined, for the experimental series a little more than 7 h of regeneration seem to be required. Head determination appears to be faster at a free cut surface. It was interesting that grafts taken from F…21/12…F often formed supernumerary tentacles and occasionally two heads were formed following transplantation to a host animal. This was never observed in the control series, where the inductions were always of a single head, and suggests autonomy of the two 1 regions.
Mention must be made of two puzzling observations. First, that if 12…F is allowed to regenerate for 18 h, by which time tentacle buds have appeared at the distal end and then combined with a freshly excised 12…F, as in F…21/12…F, head formation is not inhibited in the second part of the combination and two rings of tentacles form. Secondly, if the 12…F is allowed to regenerate for 18 h, and then a 1 region is grafted onto the distal end with the same polarity, head formation is inhibited at the top of the graft region 1 and a ring of 12-15 tentacles is formed at the junction. A foot does not form at the distal end of the graft which remains for at least 8 days as a lump situated above the circle of tentacles.
Dynamics of region 3
As for region 1, the time for region 3 to become determined as a head end was assayed by finding out how long 34…F can be left before distal structures will form at the junction when it is combined with H12. At various times, a fresh Hl2 was grafted to the regenerating end of the 34…F (Table 4). About 8-10 h of regeneration are required for a regenerating 34…F to escape from inhibition by an Hl2 graft. This time should be compared with the 4–5 h required by region 1. The result is consistent with the results of Webster & Wolpert (1966) that the time for head-end determination increases with distance from the head end.
It was of interest to compare the time of 10 h for the determination of region 3 obtained by axial grafting with that required using lateral grafting, since Webster & Wolpert (1966) did not consider this region in their study. The procedure employed was similar to theirs, except that half of the regenerating tip was grafted at the end of the incubation period instead of the entire tip (Table 5). Grafts taken from the regenerating tip of 34…F did not result in the formation of any inductions when transplanted to the mid-gastric region of intact hosts when taken from animals which had been regenerating for 6 h only. After 10 h of regeneration, however, a number of the grafts elicited the formation of secondary axes consisting of hypostome and tentacles, although about 15 h of regeneration were necessary before the majority of the 34…F had become determined. The time for head-end determination may be estimated to be 11–12 h, which is similar to the value obtained by axial grafting. On one occasion, the graft from the regenerating distal tip of region 3 resulted in the formation of an axis consisting of peduncle and basal disc; this suggests that determined bud material may have been included as the graft. If so, it is possible that the results may on other occasions have been affected by the proximity of the budding region to the site of distal regeneration.
Grafts of 12/34…F do not form structures at the junction (Hicklin et al. 1973). It is possible, therefore, to investigate how long regenerating 34…F may be left before it can ‘escape’ from inhibition by a graft 12 region. 34…F regions were excised from donor animals, incubated for a certain time and then a 12 region was grafted on each. It was found (Table 6) that after as little as 1 h of regeneration, distal structures formed at the junction in well over half of the animals when they were combined with a 12 region. This result is also of interest because it shows that changes can occur quite rapidly. Moreover, if starvation delays wound healing, it might provide an explanation as to why 12/34…F which have been starved give heads at the junction (Hicklin et al. 1973).
The next experiment was undertaken to find out how long a period of regeneration 34…F requires before a graft of the region 3 behaved like a Hl2 region in that it would inhibit the formation of structures at the junction when combined with a 12…F host. It was found (Table 7) that between 12 and 16 h of regeneration are required for about 50% inhibition to be obtained. The time for determination seems to be about 14 h. It should be noted that in none of the 0–h control series was the polarity of the graft reversed (compare table 8, Hicklin et al. 1973).
In contrast to 12/34…F which gives no structures at the junction, 34/34…F usually gives structures at the junction (Hicklin et al. 1973). How long does it take for a 3 to become like a 12 such that when it is grafted to a 34…F no structures form at the junction? This is similar to the previous experiment except that acquisition of inhibition is being assayed against region 3 rather than region 1 (Table 8). The results show that it requires about 8–10 h before a region 3 can prevent head formation from 34…F. This is, as might be expected, shorter than that required to inhibit 12…F as found in the previous experiment.
Changes at the proximal end
Investigations using axial grafting of the time required for changes in the positional value of a region destined to become a foot-end boundary were limited to one experimental situation only: more extensive study involving lateral grafts is reported elsewhere (Hicklin & Wolpert, 1973). An investigation was carried out to determine how long regenerating Hl2 required to be left until a foot was formed at the junction when it was combined as a graft with 12…F (Table 9). The time required is between 3 and 5 h. This can be compared with the 12 h required for foot-end determination by region 2, when assayed by lateral grafting (Hicklin & Wolpert, 1973). Since the foot end exerts an inhibitory influence on foot-end determination, this difference in times may be due to the difference in the distance between the region being assayed and the foot end; in the axial graft it is greater.
The effect of temperature
The effect of temperature on head-end determination was investigated using lateral grafting of region 1 of regenerating 12…F (Table 10) and region 5 of regenerating 56F (Table 11) at 26 °C and 16 °C. For region 1 the time was about 5 h at 26 °C and 11 h at 16 °C. For region 5 the corresponding times were about 16 h at 26 °C and 36 h at 16 °C. The Q10 is thus about 2.
DISCUSSION
The main results are summarized in Table 12. The times for regions at the boundary to become determined as a head end when assayed by grafting a Hl2 on them are essentially the same as those found by lateral grafting. As reported previously (Webster & Wolpert, 1966), the times increase with distance from the head end. The times required by a boundary region to acquire the inhibitory properties of a head end are longer than those required to acquire resistance to inhibition. For example, a regenerating 12…F requires about 4–6 h to resist inhibition, whereas 12–15 h is required for the 12 region to acquire the inhibitory properties of an H12. This is a somewhat different conclusion from that drawn by Webster (1966a) on the basis of lateral grafting: he found that the times for region 1 to resist inhibition and the time for re-establishment of inhibition following head-end removal were about the same. However, the assays used were different.
One can attempt to interpret these results in terms of our current model (Wolpert et al. 1974). This suggests when the positional signal falls a threshold amount below P, P increases until it reaches the head-end boundary value. Until this boundary value is reached, it is assumed that it is possible to inhibit further increase in P if S is increased sufficiently. This has important implications for our assays of head-end determination, for it emphasizes that both P and S must be taken into account. Escape from inhibition need not imply an irreversible determination, but simply that P is sufficiently greater than S, so that it continues towards becoming a head end. This is particularly clear in the case where a 34…F can become resistant to inhibition by a 12 region in a rather short time. In terms of the model, a head end acquiring inhibitory properties occurs only when P has reached the boundary value. The suggestion is that the synthesis of S lags behind that of P, and this could account for the longer times required for the development of inhibitory properties. It is to be expected that, as the results show, less time is required for region 3 to be able to inhibit a 3…F as compared with a 12…F: the concentration of S required is less. The similarity in the times required by region 12 and region 3 to inhibit 12…F is perhaps surprising, but becomes less so if the variability in the results is taken into account, as well as the extra distance that S must diffuse in region 12.
It cannot be too strongly emphasized that other criteria for determination, even in terms of the framework of our type of model, are possible. We cannot, for example, exclude a model in which irreversible determination of the head end occurs when S falls a threshold amount below P. However, the effect of temperature suggests that this is not the case. A Q10 of 2 was found for head-end determination and with H. attenuata a Q10 of 6 was reported (Wolpert et al.1972). This high dependence on temperature might be used to argue against the time for determination being simply due to time for diffusion.
We do not yet have a quantitative model capable of accounting for all the results but we are attempting to construct one. Our model (Wolpert et al. 1974) draws quite heavily on the results obtained here. In fact, the results put severe constraints on any simulation and we do not know if it is possible to construct a simple model which will account for both the instantaneous grafts (Hicklin et al. 1973) and the dynamics given here. It should also be remembered that the quantitative data itself may not be consistent as different batches of hydra do not always give comparable results. For example, our current culture requires a longer time for a 34…F region to escape from a 12 region graft, than those reported here, and 12/56F combinations do not give structures at the junction as reported in Hicklin et al. 1973.
One factor that must be taken into account in the construction of any quantitative model is the effect of a cut surface on the changes in S and P. We have found it difficult to make progress without assuming a significant leakage of S from the cut surface. As pointed out in a previous paper (Hicklin et al. 1973) regenerated heads usually form at surfaces that have been cut. The results on the time for determination of a 1 region in a F…1/1…F graft provide direct confirmation for this since it takes significantly longer than when the 1 is at a free end.
No explanation can be offered for the observation that fully determined head ends, which already have tentacle buds, do not inhibit a region 1 but seem to assimilate such regions into the regenerating head.
In this paper only changes at the boundary have been considered and in a following paper the changes with time away from a boundary will be reported.
ACKNOWLEDGEMENTS
This work was supported by the Nuffield Foundation. We are indebted to Dr D. Summerbell for his constructive comments.