Pattern regulation in fragments of Drosophila wing discs which show variable wound healing

Recent models of epimorphic pattern regulation propose that local interactions between cells from formerly disparate positions stimulate intercalation of those intermediary positional values requiring minimal regulation (French, Bryant & Bryant, 1976; Cummings & Prothero, 1978; Russell & Hayes, 1980; Kauffman & Ling, 1981; Lewis, 1981). Evidence in favour of this suggestion is provided by grafting experiments on the larval cockroach limb; the structures regenerated are determined by the normal location of the cells confronted at the host-graft junction (Bohn, 1970a,b; Bullière, 1971; French, 1976, 1978, 1980). Different host-graft combinations stimulate alternative directions of regeneration in both tissues.

When complementary fragments of Drosophila imaginal discs are allowed a period of growth, then generally one fragment will regenerate while the second duplicates (Schubiger, 1971; Bryant, 1971, 1975a, b;van der Meer & Ouweneel, 1974; Bryant & Hsei, 1977; Littlefield & Bryant, 1979a). In these fragments growth is epimorphic, dividing cells being localized at the wound edge (Ulrich, 1971; Postlethwait, Poodry & Schneiderman, 1971; Dale & Bownes, 1980; Abbott, Karpen & Schubiger, 1981). Pattern regulation is preceded by wound healing which apposes cells from disparate positions in the disc (Reinhardt, Hodgkin & Bryant, 1977; Reinhardt & Bryant, 1981).

We have shown that wound healing is required for the initiation of pattern regulation in the imaginal wing disc of Drosophila (Dale & Bownes, 1981). However, it has not been established whether the new structures generated by intercalation are determined by the normal location of apposed cells. If this is the case then variations in the mode of wound healing should lead to variations in the regulative capacity of disc fragments. Pattern regulation in imaginal disc fragments is indeed variable; in many fragments duplication of structures predicted from the fate map is accompanied by regeneration of others (Schubiger, 1971; van der Meer & Ouweneel, 1974; Bryant, 1975a; Duranceau, Glenn & Schneiderman, 1980; Kauffman & Ling, 1981; Karlsson & Smith, 1981; Kirby, Bryant & Schneiderman, 1982) and a number of wing disc fragments are capable of alternatively regenerating or duplicating (Karlsson, 1981; Kauffman & Ling, 1981).

Reinhardt et al. (1977) have described two modes of wound healing which occur in imaginal disc fragments: regular, in which two wound edges heal with each other and irregular, in which two wound edges heal back upon themselves. If the structures produced in pattern regulation depend upon precisely which cells are apposed then these two modes of wound healing should result in the intercalation of different intermediary structures. We have examined 15 wing disc fragments for variations in both wound healing and pattern regulation and we show that regeneration and duplication in these fragments are associated with different modes of wound healing.

Wild-type larvae and adults of the Barton strain of Drosophila melanogaster were used both as donors and hosts. Stocks were maintained at 25°C under sterile conditions on autoclaved David’s medium (Sang, 1978) containing antibiotics (Penicillin, 50μg/ml; streptomycin, 30μg/ml; Tetracyclin 50μg/ml; Kanamycin 50μg/ml; only two antibiotics were added at any time, but they were rotated regularly).

Wing discs were dissected from mature third instar larvae, 120±4h after oviposition, into a Ringer solution (Chan & Gehring, 1971) containing antibiotics (Penicillin, 60μg/ml; streptomycin, 100μg/ml; Gentamycin, 40μg/ml; Kanamycin, 50μg/ml; only two antibiotics were added at any one time). They were then cut into the required fragments (Fig. 2) using tungsten needles. Fragments were either injected directly into mature third instar larvae to induce metamorphosis immediately, or cultured in the abdomen of adult female hosts for 6 days prior to inducing metamorphosis by reimplanting the fragment into a larval host (Usprung, 1967). This allowed the regulative capabilities of these fragments to be established. All fragments (Fig. 2) cultured in adult female hosts were isolated after two days of culture and examined under the dissecting microscope to determine the mode of wound healing. Each disc was then implanted into a second adult female host and cultured for a further 4 days. All operations were performed under sterile conditions. Implants were dissected from newly eclosed flies in 70% alcohol and mounted between coverslips in Gurr’s water mounting medium.

For the 15 types of fragment examined (Fig. 2) three classes of wound healing were identified: (i) regular, the two wound edges healed together (Fig. 3A); (ii) irregular, each wound edge healed back upon itself (fig. 3B); (iii) indistinguishable, the precise mode of wound healing could not be distinguished. Regular and irregular healing modes had previously been suggested by Reinhardt et al. (1977), however to our knowledge this is the first report in which irregular healing has been observed. Furthermore Reinhardt et al. (1977) suggested that irregular wound healing would be transient, the regular healing mode being quickly established. Close -examination of fragment A after successive days of culture and fragments B and C after 6 days of culture shows that in these fragments irregular healing is not transient but permanent.

The frequency of the three classes of wound healing for each of the 15 types of fragment used in this study is presented in Table 1. For three of these fragments (A, B and C; Fig. 2) each wound healing class was subsequently cultured separately. The results of scoring the metamorphosed implants of fragments A, B and C following regular and irregular healing are presented in Table 2. It is clear from these results that the mode of pattern regulation in these fragments is dependent upon the mode of wound healing. In these three fragments regular wound healing (Figs 4B and E) results in the regeneration of those structures removed by fragmentation. The only exception was one implant of fragment A which in addition to single copies of HP, PCo and M/DCo (Fig. 1), structures almost always deleted by fragmentation (Table 2), also differentiated duplicated copies of the AS1 and AS2 (Fig. 1). Irregular wound healing resulted in variable pattern regulation but the frequency of duplication was greater than that found following regular wound healing (Table 2). The frequency of duplication in fragment A was 42 % (8/19), only 16 % (3/19) regenerated and two of these also duplicated ventral hinge structures (PWP, PS, AP). Following irregular wound healing in fragment B equal frequencies of regeneration and duplication were observed, 24% (4/17), both often occurring in the same fragment, 35 % (6/17). Similarly in fragment C 20 % (4/20) of implants had regenerated either one or two dorsal hinge structures (UP, Scd), 30% (6/20) duplicated (AS3, AC, Teg, YC, PVR) and a further 30 % both duplicated (AS3, AC, AL) and regenerated (UP, Scd, DR, PWP, PS, AP). Pattern regulation in all these fragments was usually extensive, many structures far from the wound edge being regenerated or duplicated. Indistinguishable wound healing in these three fragments usually results in regeneration although duplication is also frequent (data not shown; Dale, 1983). These results suggest that variations in the mode of wound healing are responsible for the variable pattern regulation observed in fragments A, B and C.

The results of culturing the remaining 12 types of fragment are consistent with this conclusion. Table 1 shows that for six of these fragments (D, E, G, I, J, K; Fig. 2) regular healing (Fig. 5B) was the most common mode of wound healing. All six of these fragments regenerated extensively following culture, duplication being very rare (Tables 3 and 4). For one type of fragment (N) the mode of wound healing was identified as irregular in all cases and duplication was observed in all those implants that regulated (Table 5). Only two examples of regeneration were observed in this last fragment (M/DCo, TR) and both also duplicated extensively (Not, Scut, Teg, HP, UP, ASI, AS2, AS3, AL, Ac). The only other fragment in which wound healing could be clearly identified in a large number of cases was fragment 0. Regular and irregular healing were observed with approximately equal frequency in this fragment (Table 1) and following culture equal frequencies of both regeneration and duplication were observed (Table 5). A further three implants regenerated extensively and duplicated a small number of structures. For the remaining four types of fragment (F, H, L, M), the mode of wound healing could not be clearly distinguished in the majority of cases. All four classes of fragment are capable of both regeneration and duplication, three of these predominantly duplicate (L, H, F) while one predominantly regenerates (M; Tables 3–5).

The results presented in this paper demonstrate that in the imaginal wing disc of Drosophila different modes of wound healing result in the formation of different structures during pattern regulation. This is most clearly demonstrated by fragment A in which the two modes of wound healing described by Reinhardt et al. (1977) can easily be distinguished. Reinhardt et al. have suggested that these two modes of wound healing, in accordance with recent models of epimorphic pattern regulation (French et al. 1976; Kauffman & Ling, 1981; Lewis, 1981) should result in the formation of different structures during intercalation. An examination of Table 2 confirms that this is the case; regular healing results in regeneration of those structures removed by fragmentation, while irregular healing results in duplication of those structures present in the fate map of the fragment. It is clear that fragment A is capable of forming alternative structures during intercalation and that it is the mode of wound healing which determines which ones are formed.

Similar results were obtained with fragments B and C. Although the frequency of regeneration following irregular wound healing in these fragments is high, this is not entirely unexpected. Recent models of epimorphic pattern regulation suggest that following irregular wound healing the two wound edges should regulate independently. The mode of pattern regulation, i.e. regeneration or duplication, would be dependent upon those positional values which are apposed by wound healing. If, as suggested by Karlsson (1981), positional values are unevenly distributed in the wing disc then it is conceivable that irregular wound healing could create positional discrepancies along one of the wound edges sufficient for regeneration. Indeed fragments with wound edges similar to either the anterior or posterior wound edge in fragments B and C, have been found to both regenerate and duplicate (Bryant, 1975a; Karlsson, 1981; Karlsson & Smith, 1981; Kauffman & Ling, 1981; Dale, 1983). On the other hand it is inconceivable that wing disc fragments as large as B and C would duplicate following regular wound healing. Given this the critical observation in these fragments is that regular wound healing results in regeneration alone, while irregular wound healing results in duplication as well as regeneration (Table 2). Following irregular wound healing in these two fragments 60 % of implants duplicated, most of these extensively including many structures far from the wound edge. Along with the permanency of irregular wound healing, the frequency and extent of duplication in these fragments precludes an explanation based on small-scale regulative events involving transient wound healing (Reinhardt et al. 1977).

The results obtained after culturing the remaining 12 types of fragment (Tables 3–5) are consistent with those obtained for fragments A, B and C. Frequent and extensive duplication is only observed in those fragments in which wound healing is either clearly irregular or indistinguishable. Those fragments in which wound healing is clearly regular predominantly regenerate, duplication being rare and restricted to a small number of structures.

Variations in the mode of wound healing cannot explain all the variable pattern regulation observed in imaginal disc fragments. Recently it has been shown that many wing disc fragments which initially duplicate are also capable of regenerating if the culture period is extended (Duranceau et al. 1980; Kauffman & Ling, 1981; Kirby et al. 1982). Similar results have been obtained in the genital disc (Lüónd, 1961). It is unlikely that the mode of wound healing could explain this phenomenon, additional explanations are required.

During wound healing in imaginal discs heterotypic contacts are established between the wound edges of the epithelium proper and those of the peripodial membrane lying directly beneath them (Reinhardt et al. 1977; Reinhardt & Bryant, 1981), it is unlikely that this mode of wound healing would be variable. Heterotypic wound healing is transient, contact between the two cell types are broken once stable homotypic wound healing has occurred. This latter type of wound healing is clearly variable, resulting in alternative modes of pattern regulation, suggesting that it is stable contact between cells of the epithelium proper that determines which structures are formed during pattern regulation.

The results presented above are consistent with those obtained from grafting experiments in the imaginal discs of Drosophila and the larval limbs of cockroaches. Imaginal disc fragments which when cultured alone duplicate, can be induced to regenerate by mixing them with fragments from different positions of either the same or different discs (Haynie & Bryant, 1976; Adler & Bryant, 1977; Wilcox & Smith, 1977; Bryant et al. 1978; Adler, 1979; Karlsson, 1979; Littlefield & Bryant, 1979b; Haynie & Schubiger, 1979; Haynie, 1982). Following amputation of an insect limb, both proximal and distal stumps regenerate a new set of distal limb structures (Bohn, 1965; Bullière, 1970; Shaw & Bryant, 1975). If epidermal cells from different proximal-distal levels are confronted, wound healing is followed by growth and intercalary regeneration of structures normally intermediate between the host graft junction (Bohn, 1970; Bullière, 1970; French, 1976). Since the intercalary regenerate is derived from both host and graft tissue (Bullière, 1971; Bohn, 1971; French, 1976), cells can form either more proximal or distal structures during intercalation. Growth and intercalary regeneration also occur when strips of cockroach leg epidermis are grafted into different circumferential positions (French, 1978, 1980). Cells confronted at the host-graft junction are able to regenerate in either direction around the circumference (French, 1980); the direction taken is governed by their normal position in the limb, such that intercalary regeneration always occurs via the shortest section of the circumference normally separating the confronted cells. The developmental fate of cells can clearly be modified by cell-cell interactions.

In conclusion the results presented in this paper suggest that pattern regulation in the imaginal wing disc of Drosophila melanogaster, is the result of local cell-cell interactions between normally non-adjacent cells apposed by wound healing. These interactions result in localized cell proliferation and the formation of intermediary structures by intercalation.

We thank Graham Brown for photographic assistance and the S.E.R.C. for financial support.