Leg regeneration in the cockroach,Blattella germanica

Wilhehn Roux's Archives 179, 57--76 (1976) Roux's Archives of Developmental Biology �9 by Springer-Verlag 1976

Leg Regeneration in the Cockroach, Blattella germanica I. Regene ra t i on f rom a Congruen t T ib ia l G r a f t / H o s t J u n c t i o n

Vernon French

Developmental Biology Group, University of Sussex

Received June 26, 1975

Accepted in revised form October 22, 1975

Summary. The interactions occuring between graft and host leg epidermis at a congruent junction (non-rotated, homopleurM combination of components cut perpendicular to the proximal-distal axis) were studied at the tibia level in the cockroach, Blattella germanica. Grafts were made between dark (B1) and light (br) cuticle eolour mutants.

1) Precise boundaries could not usually be drawn between Bl and br tissue over areas of bare cuticle but spines, hairs and claws could be identified, providing a good indication of the graft or host origin of regenerated structures.

2) When the graft tarsus remained intact, segmented structures were not regenerated from the junction. Host distal tibia and graft proximal tibia interacted to produce a reversed orientation intercalary regenerate, usually composed mainly of host cells which had become more proximal than their level of origin.

3) When the graft tarsus was amputated (or broken off), nearly 50% of congruent junctions regenerated segmented distal structures, which were classified as "autonomous" or "lateral". Amputation of the graft tarsus acted, not through removal of any inhibition, but by hindering healing of the junction because of the apolysis of graft tibial epidermis.

4) Distal structures regenerated autonomously by host and graft components of the junction were either complete or partial (fused at a common level in the tarsus).

5) Lateral regenerates were of joint origin and usually distally incomplete. They were stable and, when amputated, were regenerated to approximately the same level, in the presence or absence of the graft tarsus.

6) I t is concluded that autonomous regeneration occurred from junctions which had totally failed to heal, and that lateral regeneration occurred from an unhealed sector of a junction. Laterals were therefore regenerated from a bilaterally symmetrical, partial circumference. They are compared to other incomplete regenerates found in analogous situations. The relationship between transverse organization and distal incompleteness is obscure.

7) Segmented structures are thus regenerated only in situations where host and graft do not heal and interact (at least initially) over all or part of the circumference of the junction: interaction results in the formation of an unsegmented intercalary regenerate comprising the levels normally lying between host and graft on the proximal-distal axis.

An impor t an t and elusive aspect of any developmenta l process, embryonic or postembryonic , is the development of the reliable spatial pa t t e rns in which the differentiated elements are formed (Wolpert, 1969, 1971; Cooke, 1975; Bui- lt ,re, 1971b). Despite the success of biochemical techniques for invest igat ing the intracel lular processes of differentiat ion and cellular response to known extracellular signals (e.g. hormones), the most fruitful technique for exploring the intercel lular communica t ion involved in spatial pa t t e rn format ion remains the classical one of spatial ly d is turbing the system (e.g. by graft ing and extirpation) and observing the new pa t te rns formed. I t is commonly found tha t the embryonic development of a pa t t e rn involves an early regulative phase when the

58 V. French

final pattern depends on communication between the cells in different areas of the tissue, followed by determination and mosaic behaviour, and ultimately the pattern is expressed in cellular differentiation (Sander, 1971).

A popular system for the study of pattern formation has been the embryonic and postembryonic development of the insect epidermis, a single-layered sheet of cells secreting a characteristically patterned cuticle (see Lawrence, 1970 and Bryant, 1974 for review). In hcmimetabolous insects such as the cockroach, the epidermis secretes a new cuticle before each larval moult. Although the pattern is already expressed in the differentiation of a patterned cuticle, the postembryonic epidermis does not behave in a mosaic way; grafting and extirpation can cause cells or their progeny 4o form different parts of the pattern. In par- ticular, cockroach legs have a great ability to regenerate amputated distal parts, and to regenerate new structures at graft/host junctions.

The cockroach, Blattella germanica has been used to examine the relationship between regeneration and the moult cycle (O'Farrel and Stock, 1953; O'Far- tel, Stock, Rae and Morgan, 1960), and the gross histology of regeneration (O'Far- rell and Stock, 1958). Blattella has not been used to investigate the interactions between different regions of the leg epidermis, confronted at a graft/host june- tion. Many such experiments have been performed on the larger cockroaches; Leucophaea maderae (Bohn, 1965a, b, 1967, 1970a, b, 1971, 1972a, b), Grompha- dorhina portentosa (Bohn, 1971, 1972a, b), Blabera cranii/er (Bulli6re, 1970a, b; French and Bulli6re, 1975a, b), and Periplaneta americana (Penzlin, 1965) and on the stick insect, Carausius morosus (Burr, 1971a, b, 1972). In this and sub- sequent papers (French 1976a, b), many graft combinations will be examined, using cuticle relent mutants of Blattella germanica, in order to compare this insect with other species, to extend the range of graft/host interactions studied, and to determine the origin of the new structures produced. The results of a range of graft/host interactions can suggest the type of information which can be communicated between cells of the epidermis.

I t has been shown in other species of cockroach that the epidermal cells of a leg segment have their position within the segment specified along the proximal- distal axis. Association of different proximal-distal levels of the tibia result in intercalary regeneration of the intermediate levels (Bohn, 1967, 1970a; Bulli6re 1971b). In Blattella there is no way of distinguishing morphologically between the original tissues and the normally orientated intercalary regenerate devel- oping between a host proximal level and graft distal level. Only the combination of host distal level and graft proximal level will be considered here since it gives a morphologically distinct intercalary regenerate, with spines and hairs of reversed orientation.

I t will be shown that Blattella often produces segmented regenerates from a congruent tibia1 junction when the graft tarsus is removed. This relationship will be explored.

Material and Methods Laboratory colonies were kept at 26~ in plastic sandwich boxes and fed a diet of dog

biscuits, dried meat, oats, peanuts and yeast powder. They were provided with moist filter paper. Separate colonies were maintained of wild type BlatteIla @rmanica, a semi-dominant

Leg Regeneration in the Cockroach Blattella Germanica 59

autosomal dark cuticle colour mu tan t (Bl) (Ross and Coehran, 1967) kindly provided by Dr. Mary Ross, and a recessive autosoma[ light colour m u t a n t (br) which arose spontaneously in the wild type colony.

Mature oothecae were removed from the females for experimental use, surface sterilised with 70 % ethanol and hatched on damp filter paper. The subcultures were maintained synchronous from hatching by a regime of food for 4 days al ternating with no food unti l the animals moulted (Kunkel, 1966).

In all experiments some grafting operations were performed on wild type animals (usually between different legs of the same animal) and some were done between the B1 and br animals. 3rd or 4 th (rarely 5th) instar nymphs were used, one or two days after moulting (the normal intermoult period is 7 to 10 days). Operations were performed after anaesthetising with C02, under a binocular microscope, using fine forceps, small spring scissors and knives made from fragments of razor blade. Graft and host cut surfaces were exposed to the air as briefly as possible, and were chosen to be of slightly different diameters so t ha t the graft could be tele- scoped slightly into the end of the host stump. The graft was held in position by dried haemo- lymph. Experimental animals were kept unti l the cuticle had tanned fully after the 1st or 2nd post-operative moult , and then the operated leg was removed, mounted in Gum Chloral (Faur6's or Berlbse) and examined.

, 1 r a m ,

A + t 3 / E t 4

t5

P laws

b . a ~

Fig. l a and b. Left metathoracic leg of Blattella germanica: (a) Camera lueida drawing of anterior view; (b) schematic cross-section, distal view. Tn t roehant in ; C coxa; Tr t rochanter ; F femur; T t i b i a ; t l - t 5 tarsal segments, ap autotomy plane; 8 tibial spine; cot distal coronet of tibial spines; con tarsal condyle art iculating with the distal tibia. A anterior; P posterior;

I internal; E external; dashed line: plane of bilateral symmetry of the animal

60 V. French

. . . . ? [ = .

4.1 , ' r '

:::::..'.~

H

a ,

A

b.

;.':-: 5 ~.. �9 ) : . . . . . .

i i [ ,6 ,

i l , ' < 6

i

10

B

a. b.

Fig. 2A and B. Congruent tibial grafts made between B1 and br mutants, leaving the graft tarsus intact: (a) graft combination: (b) result after 2 moults. The phenotype of the host is shown by stippling and shaded spines, (A) Graft at mid tibia level. (B) Combination of graft proximal tibia and host distal tibia, forming a reversed orientation intercalary regenerate.

1-10 abitrary notation for the proximal-distal levels within the tibia

Fig. 1 shows the metathoracie leg of Blattella, and the labelling convention which will be used. The three pairs of legs have the same basic euticular pattern although the relative sizes of the segments vary and the hair and spine patterns differ slightly between legs. I t should be noted that any leg regenerated from a level proximal to the 4th tarsal segment has a 4-segmented instead of a 5-segmented tarsus (O'Farrel, Stock, Rae and Morgan, 1960) but is otherwise normally segmented.

A congruent junction results from the unrotated homopleural combination of host and graft (both cut perpendicular to the proximal-distal axis). Both transverse axes of the graft and host are in alignment, so every position on the circumference of the host cut surface will be adjacent to the homologous position on the graft.

A) Cell Autonomy o/the Mutations Grafts were made between wild type and the 2 mutants, and between the mutants.

The left pro- or mesothoracic donor leg was removed at the mid tibia level and grafted to the host left metathoracic tibia (Fig. 2A).

B) Intercalary Regeneration The left pro- or mesothoracie donor leg was removed at the proximal tibia level and grafted

onto the distal level of the host left metathoracic leg (Fig. 2 B).

C) Segmented Regeneration The controls for this experiment are the combined series of congruent grafts made between

mid tibiae (Fig. 2A) or between distal host and proximal graft levels (Fig. 2B), and retaining the graft tarsus intact. The experimental grafts were made between mid tibiae or between proximal graft and distul host levels, and the graft was amputated in the proximal tarsus (Fig. 3). Secondary experiments will be introduced in the Results section.

Leg Regeneration in the Cockroach BlatteUa Germanica 61

~.~ ci .

A

b o. b2. r

l

i

b s.

Fig. 3a--b ~ 6 Graft, at mid-tibia level, between B1 and br mutants, with removal of the graft tarsus: (a) graft combination; (b~ 6) classes of results after one moult. (b ~ :No segmented regenerate; (b 1 and b 2) incomplete lateral; (b ~) complete lateral; (b 4) completely autonomous regeneration; (b 5 and b 6) partially autonomous regeneration. The phenotype of the

host is shown by stippling and shaded spines and claws

Results and Discussion

A) Cell Autonomy o/the Mutations Graft and host mid tibiae usually healed together, forming no new structures

(Fig. 2A). The eolour difference between wild type and the br m u t a n t was in- sufficient to permit analysis, but the clear difference between wild type and B1 and (especially) between B1 and br cuticle was mainta ined over 2 or more moults (Fig. 4 A and B). The colour boundary was sometimes not distinct over areas of bare cuticle, but hairs and spines could be easily classified as host or graft. I t was no t possible to draw a precise boundary at the cellular level, I t cannot be concluded from these results t h a t the mutan t s are absolutely eeil-a~tonomous (as indistinct boundaries could easily have resulted from local cell interactions affecting cuticle colonr) but the mu tan t phenotypes are clearly very stable and provide a good indication of the host or graft origin of the cuticular structures.

B) Intercalary Regeneration The results of combining graft proximal tibia and host distal tibia (Fig. 2]3)

are given in Table 1. Almost all legs which survived to the 2nd post-operat ive

62 V. French

Fig. 4A--E. Congruent tibial graft/host junctions healed without forming segmented structures; results after 2 moults of grafts between B1 and br mutants. (A and B) Graft at mid tibia level (Fig. 2 A). (C--E) Combination of graft proximal tibia and host distal tibia (Fig. 2 B), forming reversed orientation intercalary regenerate (ir) between host (h) and graft (.q) levels


Table 1. Results of combining graft proximal tibia and host distal tibia graft tarsus not amputated

a) Incidence of intercalary regeneration

Successful Structures formed at the graft/host junction operations

reversed orientation segmented structures no structures intercalary regenerate

126 115 10 1

b) Composition of intercalary regenerates resulting from grafts between B1 and br mutants

Total Composition n o .

all host nearly host 2/3 host 1/2 nearly not clear all host graft 1/3 graft 1/2 all graft

71 3 43 i0 6 1 8

moult and bore no segmented structures (see section C) showed intercalary regeneration, clearly distinguishable by the reversed orientation of hairs and spines and the increase in length of the operated tibia (Fig. 2B, Fig. 4C, D, E). In most legs resulting from grafts between mutants, an approximate boundary could be drawn between host-derived and graft-derived portions of the new structure. The intercalary regenerate was usually derived largely from the host component, regardless of whether it was B1 (Fig. 4C) or br (Fig. 4D), but had a graft contribution (even if it consisted only of a very small area adjacent to the graft and carried only a few hairs). Rarely were host and graft contributions approximately equal (Fig. 4E).

The results show tha t intercalary regeneration occurs in the tibia of Blattella as in other cockroaches. Proximal-distal level is continuously specified down the tibia, and confrontation of different levels provokes growth and the formation of the levels which normally separate host and graft in the order. This ordering of the cells has been described as a "gradient" by Bohn (1967, 1970a) and a "sequence" by Bulli~re (1971a, b), and has been shown to be repeated in the coxa, femur, tibia and tarsus (Bohn, 1970a, b; Bulligre, 1971b). I t has recently been demonstrated that position around the circumference of a segment is speei- fled independently of proximal-distal level, but by an analogous ordering of the cells (French and Bulli~re 1975a, b).

Concerning the origin of the intercalary regenerate, these results complement those of previous workers. Bulligre (1971b) studied the metameric nature of the reversed orientation intercalary regenerate produced by grafting between pro- and metathoraeic tibiae and found it was derived from both host and graft components. Grafting between the differently pigmented tibiae of Leucophaea and Gromphadorhina, Bohn (1971) determined the contributions of host and graft to the normally orientated intercalary regenerate. He found that it was usually derived mainly from the graft (i.e. the more distal of the confronted levels).

64 V. French

Thus it seems that both graft and host surfaces undergo a localised prolif- eration to provide the tissue of the intercalary regenerate, with the major con- tr ibution tending to come from the more distal of the levels. Hence much of the intercalary regenerate is composed ol ceils which occupy a level more 1)fox- imal than t h a t of their ancestors. Regeneration from a cut surface produces only the structures" distal to tha t level, even if the cut surface is proximal-facing (Bohn, 1965b; Bulligre, 1970a). I t has been suggested that the cells of regenerating systems such as the insect or amphibian limb can proliferate and form structures distal but not proximal to their level of origin (Rose, 1970; Wolpert, 1971). This "Rule of Distal Transformation" is clearly contravened during intercalary regeneration by cells derived from the more distal level.

C) Segmented Regeneration Table 2 shows that congruent tibial grafts retaining the graft tarsus intact

characteristically heal smoothly (with intercalary regeneration where appropriate) and form no segmented structures at the junction (Figs. 2 and 4). Of the 9 legs producing segmented structures, 3 bore incomplete laterals (see below) and 6 (all resulting from combinations of proximal graft and distal host levels) formed an articulation between the graft and host tibiae (Fig. 5).

Table2. Results of MI congruent tibial grafts with the graft ta.rsus not amputated

Successful Graft tarsus retained ir~t~et; structures Graft tarsus lost operations formed at the graft/host junction and regenerated a

no segmented structure segmented structure

278 223 9 46 a

a Animals incorporated into Table 3

When the graft tarsus was amputa ted (or accidentally broken off) almost 50% of the graft/host junctions produced segmented structures at the 1st post- operative moult (Fig. 3, Table 3). The terminal graft tarsus almost always regenerated.

Form of the Regenerated Structures

The structures regenerated from the junction are classified in Table 3b. Lateral structures developed at the level of the graft/host junction and the

more complete laterals usually developed from the internal face. In almost all cases, only one lateral structure developed from a junction. When distally complete, the laterals consisted of tibial apex and a 4-segmented tarsus bearing a single central claw (Fig. 6A) or, rarely, a set of 2 daws (Fig. 6B and 3ba). I-Iowever, they were usually distally incomplete, lacking claws (Fig. 6C) or the distal segments of the tarsus (Fig. 3b2). The least-developed laterals consisted of a tibia apex and no tarsus (Fig. 3 bl). A single tarsal condyle could be found on some well-developed laterals, and many laterals may have been bilaterally symmetrical (Fig. 6C).


Fig. 5. Formation of an articulation between host and graft tibiae: results after 2 moults of combining graft proximal tibia and host distal tibia (Fig. 213) h host (br); ~ graft (B1);

r regenerate comprising distM tibia, articulation and proximal tibia

Table 3. l~usults of congruent tibial grafts with the graft tarsus amputated (or broken off after the operation)

a) Incidence of regeneration of segmented structures

Structures formed at graft/hest junction Type of operation Successful operations

host level graft level segmented regenerate no segmented structures

Distal tibia proximM tibia 200 79 121 Hid tibia mid tibia 152 93 59

Total 352 172 180

b) Classification of segmented structures formed from the graft/host junction

Type of op~r~tion Structure formed at graft/host junction

host level graft level lateral regenerate autonomous other regenerate

apex tibia apex tibia apex eom- of @in- + complete plete partial t ibia complete tarsus

tarsus

Distal tibia proximM tibia 5 30 12 2 24 6 Mid tibia mid tibia 9 52 9 2 15 6

Total 14 82 21 4 39 12

5 Wilhehn tloux's Archives

66 V. French

Fig. 6 A - - D

Leg t~egeneration in the Cockroach Blattella Germanica 67

Fig. 6A--F, Formation of segmented regenerates from a congruent tibial junction following removal of the graft tarsus (Fig. 3): results after 1 moult, h host; g graft; rt tarsus regenerated from the distal end of the graft; It lateral tarsal structure, hs spine of host origin; gs spine of graft origin; hc claw of host origin; gc claw of graft origin; hh hair of host origin; gh hair of graft origin, hr autonomous regenerate from host; gr autonomous regenerate from the graft component of the junction. (A--C) Lateral regenerates in situ; (12)) isolated lateral tarsus; (E) completely autonomous regeneration (BI host and br graft); (F) partially autonomous

regeneration (br host and B1 graft)

S~ructures termed "completely au tonomous" (Table 3b) consisted of a mirror- image sequence of 2 sets of tibial apex plus complete tarsus separating the original graft and host levels and joined by their distal tips (Figs. 6E and 3b4). Structures were te rmed "part ial ly au tonomous" when only 2 partial sets of distal structures separated graft and host levels. Fusion could occur at a proximal tarsal level or much more distally (Fig. 6F) but always occurred between the same levels on the 2 components. F rom the point of fusion there was either no lateral structure (Fig. 3bS), an incomplete lateral (Fig. 3b 5) or a complete lateral with a single (or partial) set of el~ws (Fig. 6F).

Animals not fit t ing into this classification had either 2 separate incomplete lateral structures, or formed an articulation between host and graft tibiae.

Origin of the Regenerated Structures

Many of the grafts were made between B1 and br mutan ts and this allows some conclusions to be made about the tissue origin of the regenerated structures (although t h e y were often poorly pigmented).

68 V. French

Completely autonomous structures were, as implied, produced one from the host and one from the graft component (Fig. 6E), as were the partially autonomous structures up to the point of fusion (Fig. 6F).

I t was almost always evident tha t both host and graft contributed to the ring of spines of the lateral tibia apex, even when the lateral tarsus seemed to be of host origin (Fig. 6A). In many cases the lateral tarsus was clearly of joint origin; Fig. 6B shows a lateral with one host-derived and one graft-derived claw, and BI and br hairs are evident on the incomplete lateral tarsus shown in Fig. 6 D. Thus both host and graft can participate in the formation of the single lateral structures.

Why is the Graft /Host Junct ion Liable to Regenerate ?

Removal of the tarsus from an otherwise nnoperated leg never results in structures being regenerated from the mid tibia region, so why is a tibial graft/ host junction prone to morphogenesis ? There seem to be two possible reasons:

i) Accidental slight misalignment of the "congruent " graft might make the junction liable to regenerate. 180 ~ rotation of a homoplenral graft does increase the frequency of regeneration from the junction (French, 1974) but in many cases graft and host just heal together with a partial re-rotation of the graft into alignment with the host (a phenomenon already described by Bohn, 1965b and BulliSre, 1970b).

ii) The wounding and healing following the grafting operation will disrupt communication between cells, and this may make the junction area liable to regenerate, whereas the corresponding area of an intact leg is not. This inter- pretation is supported by the results of experiments in which the side of the tibia was wounded by cutting a deep notch: a partial or complete lateral regenerate sometimes developed from the wounded area, but only if the tarsus was removed (French, 1974).

The Incomplete Nature of the l~egenerated Structures

Regeneration from the graft/host junction was unlike most insect leg regeneration phenomena in tha t it only occurred in some cases, and the regenerated structures were usually distally incomplete. Terminal regeneration of the amputated graft tarsus almost always occurred, however, and was complete. I t is known (O'Farrel and Stock, 1953; O'Farrel, Stock, Rae and Morgan, 1960; Bulli6re, 1968) tha t leg regeneration is dependent on the stage of the moult cycle. If amputat ion occurs before a critical point in the instar it is followed by complete regeneration; if it occurs after the critical point no regeneration occurs in that instar, but the leg regenerates completely between the 1st and 2nd postoperative moults. If lateral regeneration is not governed in this way, there may be insuffi- cient t ime for its completion before the cuticle synthesis phase of the instar. If so, it would be expected that incomplete laterals would be completed by the 2nd post-operative moult, and possibly that legs bearing no laterals at the 1st moult might subsequently show some regeneration. Some animals were therefore kept until the 2nd moult.

Legs initially bearing no laterals did not produce laterals by the 2nd moult (29 cases). Incomplete laterals regenerated further in only 1 case (Fig. 7A) and


P

Fig. 7A--C. Evolution of incomplete laterals; (a) structure after 1st postoperative moult; (b) structure after 2nd post-operative moult

remained virtually unaltered in 10 cases (Fig. 7B and C). The condition of having no lateral or an incomplete lateral therefore remains virtually stable, so incomplete regenerates do not result from a normal regeneration process being halted prematurely by the moult cycle. They must result from some other factor present at their initiation or intervening during their development.

Relationship between Amputat ion of the Graft Tarsus and Regeneration from the Graft /Host Junction

I t has been shown that the tibial graft/host junction regenerated a segmented structure in only 9/232 animals with an intact graft tarsus, but in 172/352 eases when the graft tarsus was removed at the t ime of operation (or lost after the operation). This effect of tarsus removal also occurs in relation to lateral regeneration from a tibial notch (see above). The amputat ion has an effect on a site at some distance proximal to it. There seem to be three possible explanations of this relationship.

i) Absence o/the Tarsus ItseI] In Hydra the concept ot inhibition by the existing hypostome of hypostome

formation elsewhere has been used to explain many experimental results (see Webster 1971 for a review). A broadly analogous model for the insect leg would postulate tha t the presence of an intact tarsus inhibits distal transformation at a tibial graft/host junction. After removal of the graft tarsus, the junction is released from inhibition until the tarsus is regenerated. This model is compatible with the stability of the incomplete laterals when left to moult again, and

70 V. French

k ? ':

D E

Fig. 8 A - E . Evolution of lateral regenerates following secondary operations; (a) leg after the 1st post-operative moult (A) or lateral structure removed after 1st post-operative moult (B, C, D, E) ; (b) leg one moult after the secondary operation. (A) Stability of incomplete lateral after re-amputation of the graft tarsus. (13) and (C) Lateral regeneration following amputation of the lateral (shown in a). (D) and (E) l~egeneration of lateral and terminal tarsi

after removal of both

can be tested in 2 ways by amputating parts of the leg after the ist post-operative moult.

i. It is predicted by the "tarsal inhibitor" model that incomplete laterals might regenerate further, possibly to completion, if the terminal tarsus were re- amputated and the inhibition thus removed. When this experiment was performed, legs in i t i a l ly bear ing no la te ra ls d id no t p roduce la te ra ls a t the 2nd moul t (45 cases) and incomple te la te ra ls r ema ined v i r tua l ly unchanged (Fig. 8A) in 6 cases. There was only one ease of fu r the r regenera t ion dur ing the absence of t he graf t ta rsus : a 4 -segmented l a t e ra l ta rsus deve loped a single claw.


2. I t is further predicted by the "tarsal inhibitor" model that, if the terminal tarsus were left intact and the lateral structure amputated, there would be no lateral regeneration. This experiment was performed by pulling the lateral tarsus out of the lateral tibia] apex. Of the 20 complete or incomplete laterals amputated, lateral regeneration failed to occur in only 3 cases. The majori ty of the remaining 17 animals regenerated approximately the same lateral structure as before (Fig. 8B, C). In addition, 2 animals had the lateral structure amputated and the graft tarsus removed; both animals regenerated both the graft tarsns and the lateral structure (Fig. 8D, E).

These results are definitely not compatible with the "tarsal inhibitor" model. Graft tarsus removal is only effective at the time of the grafting operation. "Tarsal inhibition" could only be rescued by the ad hoc postulate that it is disto- proximally polarised and hence cannot act proximo-distally down a lateral once it is initiated.

In Hydra, the transmission of hypostome inhibition is not disto-proximally polarised ; a hypostome grafted to the proximal end of an axis inhibits hypostome formation at the distal end before altering the polarity of the axis (Wilby and Webster, 1970).

ii) "Wound E//ect"

Creating an open wound in the proximal tarsus may stimulate morphogenesis of the tibiM junction by means of some transmissible "wound effect" such as that postulated by Bart (t970) to explain morphogenesis of the trochanter of the stick insect. This seems unlikely as the maximum effect of any wound would surely be on the cells nearby, and the major wound area (the graft and host cut surfaces) is present regardless of whether or not a comparatively small wound surface is created elsewhere by tarsal removal.

As it is impossible to wound the tarsus without removing it, this hypothesis co-ald only be tested by cutting a deep notch in the host proximal tibia well above the graft/host junction, leaving the graft tarsus intact. The experiment had a low success rate but, of the 24 successful operations, only 1 bore a lateral structure: a frequency of regeneration from the junction comparable to that resulting from the operation done without the wound in the tibia (9/232, see Fig. 2A, Table 2).

This hypothesis, like the previous one, can only be saved by special pleading of a disto-proximal polarity.

iii) E]/ect o] the Terminal Regeneration Process

I t was argued above that the junction is liable to regenerate because it is a region of disrupted cell contact, and any additional operation which impedes rapid healing (e.g. 180 ~ rotation of the graft) will increase the possibility of regeneration. Amputat ion in the proximal tarsus of the cockroach leg is followed by a separation of the epidermis from the cuticle, not only adjacent to the cut surface, but extending for some distance proximally into the tibia (O'Farrel, Stock, Rae and Morgan, 1960; Bulli~re, 1972; Bulli~re, Bulli~re and Sengel, 1969).

72 V. French

When the graft tarsus is removed, this change in the graft epidermal cells may extend back as far as the junction in the tibia, and impede healing with the host cut surface. I t can be seen from Table 3 that regeneration of segmented structures occurred more frequently from a junction involving the mid tibia of the graft than one involving the proximal tibia. This is consistent with an effect on the junction of an apolysis extending for a variable distance back from the terminal amputat ion site.

When the tarsus is amputa ted in the distal segment, apolysis does not involve the tibia (O'Farrel, Stock, Rae and Morgan, 1960) and therefore should not affect a tibial graft/host junction. Congruent grafts were made between host distal tibia and graft proximal tibia, and the graft tarsus was amputated in the distal segment. Only animals regenerating a 5-segmented tarsus were considered (l-segmented tarsi were assumed to have resulted from elimination of part of the tarsus after the distal amputat ion, and regeneration from a more proximal level). Only 1 of the 16 successful legs bore a regenerate from the junction; a frequency comparable to tha t resulting from the operation done without the very distal amputat ion (9/232. see Fig. 2A, Table 2).

Thus it seems tha t amputat ion in the proximal graft tarsus greatly increases the frequency of regeneration of segmented structures from the tibial graft/host junction because the apolysis associated with the terminal regeneration hinders the healing together and hence the interaction, of the graft and host surfaces. When healing and interaction do not occur at all, graft and host cut surfaces each regenerate all distal structures, as they would in isolation, giving the complete autonomous structure (Fig. 3ba). When healing is delayed until some inde- pendent development has occurred, a partially autonomous structure is formed (Fig. 3b6). If healing occurs only over part of the circumference of the original junction or the junction between the autonomous elements, distal structures produced by regeneration from the unhealed sector will protrude laterally (Fig. 3bl,2,8,5).

The Formation of Stable Incomplete Regenerates

I t was shown above tha t amputat ion of an incomplete lateral usually resulted in regeneration, from the lateral coronet, of a similar incomplete lateral, in the presence or absence of the graft tarsus. This implies that an abnormal transverse organisation at the level of the coronet affects the distal level to which the lateral will regenerate.

Bohn (1965b) found distally incomplete regenerates when regeneration occurred from an incomplete or abnormal cross-section. When two external or internal longitudinal tibia halves were combined, a tarsus regenerating from the distal end was bilaterally symmetrical and almost always distally incomplete. I t is now established tha t cells are not irreversibly determined with respect to their circumferential position (Bohn, 1972 b). If normally non-adjacent circumferential positions are confronted, the intervening positions are formed between them (French and Bulligre, 1975a and b), but two homologous sectors of the circumference joined in bilateral symmetry will not form any of the missing positions as there are no discontinuities between one position and its neighbour.

Leg t~egeneration in the Cockroach Blattella Germanica 73

Evidently this abnormal transverse organisation of the s tump is preserved in a regenerate and limits its distal extent.

When a V-shaped notch was cut in one side of the tibia of Leucophaea, a distally incomplete and bilaterally symmetrical regenerate could develop, and it contained only structures corresponding to the wounded face (Bohn, 1965b). The two edges of the notch, being homologous sectors of the circumference, would form a bilaterally symmetrical partial cross-section for the regeneration of the lateral.

In the experiments reported here, if healing occurred over only par t of the circumference of the congruent junction, homologous sectors of the graft and host cut surfaces would be left separated from each other. This would form a partial, bilaterally symmetrical cross-section (similar to tha t resulting from a V-notch) from which a lateral could form, derived jointly from graft and host. The lateral would be bilaterally symmetrical and usually distally incomplete. In both the twin half-tibiae and V-notch experiments, Bohn obtained tarsal structures which were distally more complete from internal rather than external partial cross-sections; the more complete Blattella laterals also tended to arise from the internal region of the junction.

The lateral and autonomous categories of regenerates each form a series of increasing distal completeness. This suggests that lateral regeneration proceeds in a proximo-distal direction but does not strictly prove it, as the various partial structures cannot be shown to be stages of a single process since, in general, they do not regenerate further. The existence of distally incomplete regenerates creates great problems for any theory of regeneration in which an initial event is the formation of the most distal region (Chandebois, 1973), e.g. as a response to confrontation of opposite faces of the limb (Bart, 1969) or lack of appropriate neighbours (Bulli6re, 1971a).

Regeneration of Segmented Structures from Congruent Tibial Junctions in Other Cockroaches

Congruent grafts between tibiae have been performed by Bohn (1967, 1970a) and Bulli6re (1971b). Graft tarsi were not removed but examination of the published illustrations shows that they were often broken and subsequently regenerated. Bulli6re found no lateral structures, but obtained some "r6g6n6rats inter- ealaires articul6s" which correspond to the partially autonomous structures shown in Fig. 3b 6. The one published photograph shows a regenerated graft tarsus but Bulli6re does not mention this as a common feature, and some of the structures developed in the presence of an intact graft tarsus (Bulli6re,personal communication). He attr ibuted this category of results to a delayed healing of the junction. Bohn (1970 a) obtained many animals with "exzessiven Regenerat," corresponding to incomplete laterals (Bohn, 1970a, Fig. 4e-f), but it seems likely that most of these structures developed in the presence of an intact graft tarsus (Bohn personal communication). Thus there is no evidence from the large cockroaches that graft tarsus removal affects healing at a tibial junction. This may be because the distance between proximal tarsus and mid tibia is of the order of five times as great in the large cockroaches as in Blattella.

74 V. French

General Conclusions

I t can be concluded tha t in te rac t ion between host and graft components at a congruent t ibial junc t ion does no t result in the regenerat ion of segmented structures, bu t only in the in terca lary regenerat ion of tissue normal ly lying between the host and graft levels on the proximal-dis tal axis. Segmented structures are regenerated in s i tuat ions where host and graft do no t heal and interact (at least ini t ial ly) over all or par t of the circumference of the junct ion, forming au tonomous and lateral regenerates respectively. Laterals were usual ly distal ly incomplete, as were regenerates found by Bohn (1965b) in other s i tuat ions where regenerat ion occurred from a partial , bi lateral ly symmetr ica l cross-section. The na tu re of this relat ionship between t ransverse and proximal-distal organisat ion remains to be elucidated (French, 1976a).

This work was supported by a Science Research Council postgraduate studentship. I thank Gerry Webster for years of help and encom-agement, and for advice on the manuscript.

References

Bart, A. : Conditions locales du d4clenchement et du d6veloppement de la r6g4n6ration d'une patte chez l'Insecte Carausius morosus Br. C. R. Acad. Sci. (Paris) 269, 473-476 (1969)

Bart, A. : Blessures et transformations trochant6riennes chez Carausius morosus Br. Ann. Embr. Morph. 3, 379-398 (1970)

Bart, A. : Morphogen~se surnum4raire au niveau de la platte du Phasme Carausius morosus Br. Wilhelm Roux' Archly 166, 331-364 (1971 a)

Bart, A. : Modalit~s de formation et de d4veloppement d'un centre morphog6netique sur- nmn6raire chez Carausius morosus Br. Wilhelm l~oux' Archiv 168, 97-124 (1971b)

Bart, A. : Morphogen~se provoqu6e par une discontinuit6 proximodistale au niveau de la patte de Carausius morosus Br. I. Associations de niveaux diff4renei4s. Arch. Biol. (Liege) 83, 129-166 (1972)

Bohn, H. : Analyse der Regenerationsf~higkeit der Insektextremitgt durch Amputations- and Transplantationsversuche an Larven der afrikanischen Schabe Leucophaea maderae Fabr. (Blattaria). II. Mitt. Regenerationspotenzen. Arch. EntwickI.-Mech. Org. 156, 49 74 (1965a)

Bohn, H. : Analyse der Regenerationsf/~higkeit der Insektenextremitiit durch Amputations- und Transpl~ntationsversuche an Larven der afrikanischen Schabe Leucophaea maderae Fabr. (Blattaria). II. Mitt. Achsendetermin~tion. Arch. EntwickI.-Mech. Org. 156,449-503 (1965b)

Bohn, H. : Transplantationsexpcrimente mit interk~larer Regeneration zum Nachweis eines sieh segmental wiederholenden Gradienten im Bein yon Leucophaea (Blattaria). Verh. dtsch, zool. Ges. Ghttingen (1966), Suppl. 30, 499 508 (1967)

Bohn, H. : Interkalare l~egeneration und segmentale Gradienten bei den Extremit/iten yon Leucophaea-Larven (Blattaria). I. Femur and Tibia. Wilhelm l~oux' Archiv 165, 303-341 (1970a)

Bohn, H. : Interkalare Regeneration und segmentale Gradienten bei den Extremitgten yon Leucophaea-Larven (Blattaria). II. Coxa und Tarsus. Develop. Biol. 23, 355-379 (1970b)

Bohn, H. : Interkalare l~egeneration and segment~le Gradienten bei den Extremiti~ten yon Leucophaea-Larven (Blattaria). III. Die Herkunft des interkalaren Regenerates. Wilhelm Roux' Archiv 167, 209-221 (1971)

Bohn, H. : The origin of the epidermis in the supernumerary regenerates of triple legs in cockroaches (Blattaria). J. Embryol. exp. Morph. 28, 185 208 (1972a)


Bohn, H. : The determination of the symmetry properties in the larval legs of cockroaches (Blattaria). Wilhelm Roux' Archly 170, 354-358 (1972b)

Bryant, P. : Determination and pattern formation in the imaginal discs of Drosophila. Curt. Top. Develop. Biol. 8, 41-80 (1974)

Bulli6re, D. : l~tude de la r6g6n6ration chez un Insecte Blattopteroide, Blabera cranii/er Burm. (Dictyopt6re). II. Influence du moment de l 'amputation dans l ' intermue sur la r6g6n6ra- tion de la patte metathoracique, Bull. See. zool. Fr. 93, 69-82 (1968)

Bullibre, D. : Sur la d6terminisme de la qualit6 r6gionale des r6g@n6rats d'appendices chez la Blatte, Blabera cranii/er. J. Embryol. exp. Morph. 23, 323-335 (1970a)

Bulli6re, D.: Interpr6tation des r@g6n6rats multiples ehez les Tnsectes. J. Embryol. exp. Morph. ,03, 337-357 (1970b)

Bulli6re, D. : Les m6ehanismes de la d6termination cellulMre au tours de la morphogen6se, 6tudi@s par des exp6riences de r6g6n6ration ehez un Inseete Dictyopt6re. Thgse de Doc- torat (l~tat). Universit6 de Grenoble (1971 a)

Bulli6re, D. : Utilisation de la r6g6n6ration intercalaire pour l'6tude de la d6termination cellu- laire au cours de la morphogen~se chez Blabera cranii/er (Insecte Dietyopt6re). Develop. Biol. 25, 672-709 (1971b)

Bulli6re, D.: t~tude de la r6g6n6ration d'appendice chez un Insecte: stades de la formation des r6g6n6rats et rapports avec le cycle de mue. Ann. Embr. Morph. 5, 61-74 (1972)

Bulli6re, D., Bulli6re, F., Sengel, P. : R6g6n6ration du tarse chez l 'embryon de Blabera crani- i/er (Insecte Dietyoptere) en culture in vitro. C. 1~. Acid. Sol. (Paris) 269, 355-357 (1969)

Chandebois, t~.: General mechanisms of regeneration as elucidated by experiments on planarians and by a new formulation of the morphogenetie field concept. Acta biotheor. (Leiden) 22, 2-33 (1973)

Cooke, J. : The emergence and regulation of spatial organisation in early animal development. Ann. Rev. Biophys. Bioeng. 4, 185-217 (1975)

French, V. : Pattern formation in cockroach leg regeneration. D. Phil. Thesis, University of Sussex (1974)

French, V. : Leg regeneration in the cockroach, Blattella germanica. IT. Regeneration from a non-congruent tibial graft/host junction (1976a, J. Embryol. exp. Morph., in the press)

French, V. : Leg regeneration in the cockroach, Blattella germanica. ITI. Regeneration from a graft/host junction between different segments (1976b, in preparation)

French, V., Bulli@re, D. : Nouvelles donn6es sur la d6termination de la position des eellules @idermiques sur un appendice de Blatte. C. R. Aead. Sci. (Paris) 289, 53-56 (1975a)

French, V., Bulligre, D.: l~tude de la d6termination de la position des cellules @idermiques: ordonnancement des cellules autour d 'un appendice de Blatte; d6monstration du concept de g6n6ratrice. C. R. Acad. Sci. (Paris) 28(t, 295-298 (1975b)

Kunkel, J. : Development and the availability of food in the German cockroach, Blattella germanica. J. Insect Physiol. 12, 227-235 (1.966)

Lawrence, P. : Polarity and patterns in post-embryonic development of insects. Adv. Insect Physiol. 7, 197-266 (1970)

O'Farrel, A., Stock, A. : Regeneration and the moulting cycle in Blattella germanica. T. Single regeneration during the first instar. Aust. J. biol. Sci. 6, 485-500 (1953)

O'Farrel, A., Stock, A.: Some aspects of regeneration in cockroaches. Prec. 10th Int. Congr. Entomol. 2, 253-259 (1958)

O'Farrcl, A., Stock, A. Rae, C., Morgan, J. : Regeneration and development in the cockroach, Blattella germanica. Acta Soc. Entomol. Cechslov. 57, 317-324 (1960)

Penzlin, H. : Experimentelle Erzeugung yon Dreifachbildungen bei einem heterometabolen Insekt. Naturwissenschaften 52, 217 (1965)

Rose, S. : Regeneration. New York: Appleton-Century-Crofts 1970

Ross, M., Cochran, D.: Genetic variability in the German cockroach. IT. A description of new mutants and linkage tests. J. Here& 58, 274-278 (1967)

76 V. French

Sander, K. : Pattern formation in longitudinal halves of leaf hopper eggs (Homoptera) and some remarks on the definition of "embryonic regulation". Wilhelm Roux' Archiv 167, 336-352 (1971)

Webster, G. : Morphogenesis and pattern formation in hydroids. Biol. Rev. 46, 1-46 (1971) Wilby, 0., Webster, G.: Studies on the transmission of hypostome inhibition in Hydra.

J. Embryol. exp. ~orph. 24, 583-590 (1970) Wolpert, L. : Positional information and the pattern of cellular differentiation. J. theor. Biol.

25, 147 (1969) Wolpert, L.: Positional information and pattern formation. Curt. Top. Develop. Biol. 6,

183-224 (1971)

Dr. V. French Division of Developmental Biology National Institute for Medical Research The Ridgeway Mill Hill London NW7 1AA

Am 31.5. 1975 wurde die ,,fiesellscha~t fiir Entwicklungsbiologie" gegriindet, die sich an Wissensehaft ler wended, deren Tgtigkeitsfeld Entwicklungsphysio]ogie, Entwicklungsgesehichte, Differenzierungsge- ne t ik oder Fortpf lanzungsbiologie ist. Die Gesellschaft wird am 24.4.1976 in KSln ihre erste Tagung u n d Mitgl iederversammlung durchftihren. In teressenten, die noch keine E in ladung erha]ten haben, werden ge- beten, sich an F r a u P r o f . Dr. C. Har te , I n s t i t u t ffir Entwieklungs- physiologie, GyrhofstraBe 17, 5000 K61n 41, zu wenden.

Documents

Leg regeneration in the cockroach,Blattella germanica