Development 101, 501-515 (1987) 501 Printed in Great Britain © The Company of Biologists Limited 1987
Posterior apical ectodermal ridge removal in the chick wing bud triggers a series of events resulting in defective anterior pattern formation
WILLIAM L. TODT* and JOHN F. FALLONt
Department of Anatomy, The University of Wisconsin, 1300 University Avenue, Madison, Wisconsin 53706, USA * Present address: Department of Physiology and Biophysics, University of California, Irvine, CA 92717. USA t Author for reprints
Summary
The ability of the anterior apical ectodermal ridge to by necrosis in the distal, anterior mesoderm at 48 h promote outgrowth in the chick wing bud when postoperatively and subsequent anterior truncation. disconnected from posterior apical ridge was exam- Clearly, healthy posterior limb bud mesoderm is ined by rotating the posterior portion of the stage- needed for anterior limb bud survival and develop- 19/20 to stage-21 wing bud around its anteroposterior ment. We propose that anterior truncation is the axis. This permitted contact between the anterior and direct result of anterior mesodermal cell death and posterior mesoderm, without removing wing bud that this may not be related to positional specification tissue. In a small but significant number of cases of anterior cells. In our view, cell death of anterior (10/54), anterior structures (digit 2) formed spatially mesoderm, after posterior mesoderm removal, should isolated from posterior structures (digits 3 and 4). not be used as evidence for a role in position specifi- Thus, continuity with posterior ridge is not a pre- cation by the polarizing zone during the limb bud requisite for anterior-ridge function in the wing bud. stages of development. We suggest that the posterior Nevertheless, posterior-ridge removal does result in mesoderm that maintains the anterior mesoderm need anterior limb truncation. To investigate events leading not be restricted to the mapped polarizing zone, but is to anterior truncation, we examined cell death pat- more extensively distributed in the limb bud. terns in the wing bud following posterior-ridge re- moval. We observed an abnormal area of necrosis Key words: limb development, apical ectodermal ridge, along the posterior border of the wing bud at 6-12 h wing bud, cell death, polarizing zone, chick, pattern following posterior-ridge removal. This was followed formation.
Introduction & Fallon, 1982) and the specification of distal limb elements ceases, resulting in a stage-dependent trunc- The avian limb bud consists of a mesodermal core ation of the formed limb (Saunders, 1948; Summer- covered with simple cuboidal epithelium on the bell, 1974; Rowe & Fallon, 1982). It is clear that dorsal and ventral surfaces and capped apically by a normal development of the limb depends on com- pseudostratified columnar ectodermal ridge. Inter- munication between the mesoderm and ectoderm. actions necessary for normal limb formation occur Communication is also required along the antero- between the mesoderm and apical epithelium. In- posterior axis of the limb. Posterior tissue develops itially, the limb mesoderm induces formation of the normally if the anterior tissue is removed. However, ridge (Kieny, 1960, 1968; Saunders & Reuss, 1974) anterior development is truncated if the posterior and subsequently maintains it as well (Zwilling & tissue is removed (Warren, 1934; Hinchliffe & Gum- Hansborough, 1956; Fallon, Boutin & Carrington, pel-Pinot, 1981). Insertion of a barrier between the 1986). Removal of the apical-ectodermal ridge results anterior and posterior tissue is also followed by in necrosis in the subjacent mesoderm (Rowe, Cairns anterior truncation (Warren, 1934; Summerbell, 502 W. L. Todt and J. F. Fallon
1979; Kaprio, 1981). In addition to anterior trunc- experimentally isolated from posterior ridge provided ation, cell death in the anterior mesoderm is observed that limb tissue is not removed and contact between following these operations (see Hinchliffe & Grif- the anterior and posterior mesoderm is reestablished. fiths, 1984). These data demonstrate that normal We also report specific cell death patterns in the anterior development is dependent on posterior tis- mesoderm following posterior-ridge removal. These sue. results permit insights into the ectodermal-mesoder- A series of experiments beginning with Saunders & mal and mesodermal-mesodermal interactions oc- Gasseling (1968) demonstrates that posterior border curring during limb development. mesoderm can bring about polarized duplications when a small piece is grafted into an anterior limb bud Materials and methods site. This has been interpreted to mean that posterior border mesoderm establishes a gradient of a morpho- Fertile White Leghorn eggs were incubated at 37-38°C for gen that prescribes the anteroposterior axis during 3 days, candled, windowed, sealed with Scotch magic tape normal limb development (see Tickle, Summerbell & and reincubated until they reached the desired stage Wolpert, 1975). Thus, in this view, normal limb (Hamburger & Hamilton, 1951). To operate on the em- development is dependent on active production of the bryo, a small hole was made in the membranes overlying morphogen and continuous maintenance of the gradi- the wing bud and two drops of sterile Tyrode's solution ent until all limb structures are specified. The pro- were added. The wing bud apex was stained lightly with posed morphogen has not been identified. However, Nile blue A and a camera-lucida drawing of the wing bud and adjacent somites was made (final magnification, X290). provocative results have been achieved with retinoic Following each operation described below, two drops acid (Tickle, Alberts, Wolpert & Lee, 1982; Summer- of penicillin-streptomycin (500 units ml"1; 500mcgml~'; bell, 1983; Eichele, Tickle & Alberts, 1985; Thaller & GIBCO) were added and eggs were resealed. We observed Eichele, 1987) in that it can mimic the action of better development of the anterior tissue in rotation exper- grafted posterior border mesoderm and cause polar- iments if the embryos were held at room temperature for a ized duplications. short period. Therefore, in all types of experiments, some Summerbell (1979) and Hinchliffe and co-workers embryos were left at room temperature (25°C) for 1 to 1-5 h (Hinchliffe & Gumpel-Pinot, 1981; Hinchliffe & Grif- after the operation while others were reincubated immedi- fiths, 1984; Wilson & Hinchliffe, 1987) argue that ately. Only in the case of 180° posterior-bud rotation did anterior defects seen after posterior-bud removal or room temperature affect the results (see below). Embryos were checked several times in the 2 days following the barrier insertion are due to isolation of the anterior operation and prepared for cell death studies or allowed to mesoderm from the polarizing influence of the pos- develop until the skeletal patterns could be analysed. terior mesoderm. However, their work does not rule out the possibility that separation of the anterior Experimental manipulations (Fig. 1) ridge from the posterior ridge is responsible for Partial bud removal anterior truncation. If the anterior ridge became To remove part of the wing bud, a cut was made through nonfunctional when separated from the posterior the wing bud, parallel to the proximodistal axis, with a pair ridge, the results obtained would mimic the surgical of sharpened forceps. The location of the cut was marked removal of the anterior ridge, i.e. cell death would on the camera-lucida drawing. A cut was then made along occur in the distal, anterior mesoderm and truncation the base of either the anterior (Fig. 1A) or posterior of digital skeletal structures would result. As men- (Fig. IB) portion of the bud. The excised tissue was tioned above, cell death and truncation are seen removed from the egg. following barrier insertion and posterior-bud removal Partial ridge removal (Hinchliffe & Gumpel-Pinot, 1981; Hinchliffe & Grif- In order to remove a portion of the apical ectodermal ridge, fiths, 1984). Rowe & Fallon (1981) demonstrated that the region of ectoderm to be removed was stained lightly posterior-ridge removal, leaving the anterior apical with Nile blue A. Two cuts were made through the ridge and the posterior mesoderm in situ, results in ectodermal ridge using a sharpened tungsten needle. One truncated anterior development. To understand why cut was near the boundary between digits 2 and 3 (see such different manipulations (posterior-bud removal, below and Fig. 2). The second cut was made within 150/im barrier insertion and posterior-ridge removal) all of the anterior (Fig. 1C) or posterior (Fig. ID) junction of result in anterior truncation, two questions must be the limb bud and body wall. The locations of the cuts were marked on the camera-lucida drawing. Then the dorsal answered. (1) Is anterior apical ridge inactivated ectoderm proximal to the apical ridge was cut along the when separated from posterior ridge? (2) What cell extent of the ridge to be removed. The apical ridge was death patterns accompany posterior-ridge removal? teased off the apex of the bud. In the last step for ridge We examined the ability of anterior ridge to sup- removal, the ventral ectoderm proximal to the ridge was cut port normal anterior limb development and found and the apical ectoderm, including the apical ridge, was that anterior ridge can support development when removed from the egg. Posterior ridge and anterior wing bud patterning 503
Posterior-bud rotation and the posterior part of the bud was rotated 180° around To rotate the posterior portion of the wing bud, a cut was the pin. A second pin was inserted through the rotated made parallel to the proximodistal axis and the location was tissue and into the body wall, fixing the tissue in the rotated marked on the camera-lucida drawing. A platinum pin position (Fig. IF). (0025mm diameter, Goodfellow Metals, Cambridge) was inserted parallel to the anteroposterior axis (Fig. IE). A cut Cell death studies was made along the base of the posterior portion of the bud Regions of cell death were examined in ovo for some embryos at each stage (19/20 through 21) and for each type of operation. In addition, aseriesof wholemount wing buds was prepared to compare cell death patterns following posterior-ridge removal and posterior-bud removal at stage 20/21. Cell death was visualized using the vital dye neutral red (see Fallon & Saunders, 1968; Hinchliffe & Ede, 1973). To Partial-bud observe cell death in situ, a drop of neutral red solution removal (1/1000 in saline; 137mM-NaCl, 2-8mM-KCl) was dropped in the egg and followed by a 20min incubation at 37 °C. In ovo staining did not alter the percentage of embryos surviving until day 10 (approx. 85 %). Another way to observe cell death was to prepare wholemounts. Embryos were placed in a dilute neutral red solution (1/10000 in 0-075 M-phosphate buffer, pH7-4) and incubated for 20-30 min at 37°C. Following several rinses in phosphate buffer, embryos were fixed in aqueous formal- dehyde (4 %) at 4°C, overnight. Tissue was dehydrated in ethanol and the limbs were embedded in EMbed 812 (Electron Microscopic Sciences). Embedded tissue was viewed en bloc, with stereopair micrographs taken for permanent record; the blocks were then sectioned at 5 fim, stained with azure II-methylene blue and examined histo- logically (see Todt & Fallon, 1986). For comparing pos- terior-bud removal with posterior-ridge removal, between
Fig. 1. Semidiagrammatic representations of stage-20/21 Partial-ride right wing buds illustrating each type of operation. removal Anterior is up, posterior is down and distal is to the right. (A) Anterior-bud removal. Anterior tissue was removed at or slightly anterior to the boundary line between presumptive digits 2 and 3 (see Fig. 2). (B) Posterior-bud removal. Posterior tissue was removed at or slightly posterior to the boundary line in Fig. 2. (C) Anterior apical ectodermal ridge removal. The distal ectoderm, including the apical ridge, was removed from near the anterior junction of the wing bud and body wall to the level of the boundary line in Fig. 2 or slightly anterior to it. (D) Posterior apical ectodermal ridge removal. The distal ectoderm, including the apical ridge, was removed from near the posterior junction of the wing bud and body wall to the level of the boundary line in Fig. 2 or slightly posterior to it. Posterior-bud (E) First steps in posterior-bud rotation. A horizontal rotation cut was made as in (B) at or slightly posterior to the boundary line in Fig. 2 and a pin (dashed line) was inserted. (F) Final steps in posterior-bud rotation. Posterior tissue was cut free and rotated 180° around the first pin. Tissue was secured in the rotated orientation with a second pin (dashed line). Scale bar (F) represents 250jum. 504 W. L. Todt and J. F. Fallon three and six embryos were examined at each time dis- Controls cussed in the results. Abnormal areas of cell death were As a control for nonspecific effects of posterior-ridge defined as those present only in the operated wing bud removal, the posterior, dorsal ectoderm was removed. In when compared with the normally occurring zones of this way, posterior mesoderm was exposed to trauma and necrosis in the contralateral limb bud (see Saunders, the amniotic fluid as in posterior-ridge removal. Control Gasseling & Saunders, 1962). rotation experiments were done by rotating the posterior tissue 360°. The tissue was pinned in its original orientation. Finally, in the cell death studies, the unoperated left wing Analysis of skeletal patterns bud served as a control for the appearance of abnormal cell death in the operated right wing bud. When embryos reached 10-12 days of incubation they were fixed in aqueous formaldehyde (4 %), stained with Victoria blue B, dehydrated and cleared in methyl salicylate. The Results size and configuration of the skeletal structures in the wings were compared with the untreated controls (left wing). We Limb bud variability have designated structures in the operated limbs that were As one purpose of these studies was to examine the 95 % of the length of the corresponding structure in the left development of anterior tissue, specifically digit 2, wing to be normal in size (see Summerbell & Wolpert, following operations involving posterior tissue, it was 1973). Structures that were between 85% and 95% of necessary to determine the location of the boundary normal size were labelled as slightly reduced. Reduced between presumptive digits 2 and 3. After comparing structures were less than 85 % of the length of the corre- a large number of normal stage-19 to -22 embryos, we sponding structure in the left wing. Truncated skeletal noted variations in the absolute size and contour of structures had fewer elements than normal (e.g. a digit 3 the wing bud. The profiles matching the largest with two cartilaginous elements instead of the normal proportion of wing buds at each stage is illustrated in three). Finally, in extreme cases it was difficult to measure Fig. 2. Only operations on stage-19/20 to -21 wing the humerus, so the data for the humerus are qualitative buds matching these standard profiles are reported instead of quantitative. As there was variation in the here. In this way, direct comparison of the antero- absolute size of the stained and cleared wings, we have posterior location of different operations was possible included the left wings in Figs 3, 6. This allows for by superimposing the camera-lucida drawings and the quantitative comparison to the unoperated wing. appropriate standard profile.
19-20 20 20-21 21
16
20
B D
Fig. 2. Location of the boundary between presumptive digits 2 and 3. Standard profiles of right wing buds are shown for each stage used. Anterior is up, posterior is down and distal is to the right. Stage numbers (Hamburger & Hamilton, 1951) are (A) 19/20; (B) 20; (C) 20/21; (D) 21. On the left of each figure, the approximate somite levels are indicated (somites are labelled in A). The horizontal line in each wing bud represents the boundary between presumptive digits 2 and 3, determined from anterior-bud removal experiments. Scale bar (D, upper right) represents 250 jim. Bars associated with the boundary lines show the limits of the regions in which operations were done, determined by aligning the camera-lucida drawing with the standard profile (see text). For example, at stage 19/20, anterior-bud and anterior- ridge removals were done at a level between the boundary line and a parallel line 25 ^m anterior to the boundary line. Posterior-bud and posterior-ridge removal and rotation experiments were done between the boundary line and a parallel line 50/mi posterior to it. Posterior ridge and anterior wing bud patterning 505
Effects of anterior tissue removal By 36—48h following posterior-bud removal, abnor- Anterior-bud removal mal cell death was seen in the distal, anterior meso- A series of anterior-bud removals was done to deter- derm (stippled areas in Fig. 4C,D). In addition, it was mine the location of the boundary line between obvious that the proximodistal dimension of the wing presumptive digits 2 and 3. The anteroposterior level bud was shorter than the control, left wing bud. that resulted in consistent loss of digit 2, while not These observations are in agreement with Hinchliffe & Gumpel-Pinot (1981). adversely affecting digit 3, is illustrated in Fig. 2 (about the level of the junction of somites 18 and 19). When the wing bud tissue posterior to the bound- ary line between presumptive digits 2 and 3 was Abnormal cell death was not observed in the removed, the presumptive digit 2 tissue was left in posterior mesoderm following anterior-bud removal. place, yet digit 2 did not develop in 29/31 cases. Digit Anterior-bud removals at the levels indicated in 2 developed only when digit 3 was also present Fig. 2 resulted in deletion of the radius (23/23 cases) (2/31). The humerus was always reduced in size and and also digit 2 (22/23 cases; Fig. 3A-C). The humer- the radius was usually present (29/31 cases; see us was reduced in size and the glenohumeral joint was Fig. 3H). The ulna developed in one specimen with defective (see Fig. 3B). The ulna and digits 3 and 4 digits 2 and 3, and was reduced in another specimen were normal to reduced in size. In one case, a nodule with a cartilage nodule where digit 2 should form. of cartilage was attached by soft tissue to the anterior Finally, in one case, there were two unidentifiable surface of the middle third of the first phalangeal cartilaginous elements posterior to the radius. element of digit 3. Finally, one case had a reduced digit 2. Posterior-ridge removal: cell death patterns Removal of the apical ridge posterior to the boundary Anterior-ridge removal line between presumptive digits 2 and 3 resulted in an abnormal region of cell death in the mesoderm. At When only the anterior apical ridge was removed to 6-12 h following posterior-ridge removal, a large area the level of the boundary line between presumptive of abnormal cell death was seen in the posterior digits 2 and 3, abnormal cell death was seen in the mesoderm of all six wholemount specimens (stippled anterior mesoderm within 5h (see Rowe, Cairns & area in Fig. 4E; see also Fig. 5A). This cell death was Fallon, 1982). In the skeletal patterns examined after predominantly in the dorsal half of the wing bud. In anterior-ridge removal, digit 2 failed to develop in addition, in some wing buds there were macrophages 18/23 cases while digits 3 and 4 were not adversely scattered throughout the rest of the wing bud (com- affected (Fig. 3D-F). In the five cases where digit 2 pare Fig. 5A and 5B), although these were not seen was present, it was always slightly reduced to reduced consistently. The pattern of extensive posterior nec- in size. The radius was absent in 18/23 cases. Two of rosis 6-12 h after posterior-ridge removal was also the five cases with a radius also had a slightly reduced observed in five embryos stained in ovo. to reduced digit 2. Thus, removal of anterior ridge The posterior cell death pattern just described results in anterior truncation in most cases (see also suggests that cell death following posterior-ridge Rowe & Fallon, 1981; Iten, 1982). removal eliminates a significant portion of the pos- In summary, we have located the boundary be- terior wing bud mesoderm. This is supported by the tween presumptive digits 2 and 3. In addition, we obvious lack of development in the posterior half of confirmed that removal of the wing bud or apical the operated wing bud compared to the contralateral ridge anterior to the boundary line resulted in an- wing bud (see Fig. 4). Additional support comes from terior deletions, but did not affect the formation of the observation that at 24 h following posterior-ridge the remaining posterior skeletal structures (Hinch- removal (stage 25) the normal posterior necrotic zone liffe & Gumpel-Pinot, 1981; Rowe & Fallon, 1981). (Saunders et al. 1962) was not seen in three right wing buds but was evident in the unoperated left wing buds (Fig. 4F). Cell death at 24 h following posterior-ridge Effects of posterior tissue removal removal was restricted to the normal anterior necrotic Posterior-bud removal zone (ANZ) (Fig. 4F). Following posterior-bud removal to the level indi- At 36— 38 h following posterior-ridge removal a cated in Fig. 2, an area of abnormal cell death was small increase in the amount of cell death along the seen along the anterior border of the remaining wing anterior border was seen (n = 4; Fig. 4G). By 48 h, an bud, beginning at 18 to 24 h (stippled area in Fig. 4B). area of abnormal cell death was seen extending into 506 W. L. Todt and J. F. Fallon the distal mesoderm (n = 3; Fig. 4H) beneath the Posterior-ridge removal: skeletal patterns remaining apical ectoderm. By 48h following either Removal of the apical ridge posterior to the boundary posterior-ridge removal or posterior-bud removal, line between presumptive digits 2 and 3 resulted in the shape and cell death pattern of the resulting wing truncated anterior development (Fig. 3J-L). Even buds were very similar (compare Fig. 4D and 4H). though some of the metacarpal for digit 2 was often
3A B
H Posterior ridge and anterior wing bud patterning 507 present, digit 2 failed to form in all 29 cases. The Dorsal ectoderm removal radius was always present and development of the As a control for posterior-ridge removals, the pos- ulna was variable (Fig. 3K). We stress that the ridge terior dorsal ectoderm was removed, leaving the normally associated with digit 2 and all of the wing apical ridge intact. Massive necrosis was never ob- bud mesoderm, including polarizing zone, were left in served in the posterior mesoderm. Neutral red stain- situ in these experiments, yet digit 2 did not form. ing was seen in the ectoderm around the edge of the In another nineteen embryos, we observed prox- wound at 6—12h. In addition, occasional, superficial imodistal outgrowth along the posterior portion of macrophages were observed in the mesoderm. The the wing bud at 24-48 h following posterior-ridge wing buds maintained normal dimensions and devel- oped normal skeletal patterns in all ten cases (see also removal. We regard this outgrowth as evidence of Saunders, 1948; Rowe et al. 1982; Martin & Lewis, posterior-ridge activity and we infer that the posterior 1986). This control makes unlikely the possibility that ridge had not been completely removed. In fact, the amniotic fluid or surgical trauma caused the digits 3 and/or 4 were present, albeit reduced or mesodermal cell death and truncation observed fol- truncated, in 11/19 cases. These cases are not in- lowing posterior-ridge removal. cluded in our analysis except to note that, as in posterior-bud removals, digit 2 only formed in cases Rotation of the posterior wing bud tissue where posterior digits also developed. The rotation experiments described in the Materials Therefore, we have confirmed that removal of the and methods section were designed to investigate the posterior bud (Hinchliffe & Gumpel-Pinot, 1981) or ability of the anterior ridge to function when separ- only the posterior ridge in the wing bud (Rowe & ated from the posterior ridge. When the posterior Fallon, 1981) results in truncated anterior develop- wing bud tissue was rotated 180° around its antero- ment; digit 2 does not form. These results are in posterior axis (Fig. IF), continuity was maintained between the anterior and posterior mesoderm. In contrast to the development of digits 3 and 4 from the addition, no tissue was removed from the wing bud. posterior tissue in the absence of anterior tissue. However, the anterior apical ectodermal ridge was spatially separated from the posterior-apical ridge. In Fig. 3. Skeletal patterns resulting from preliminary this way, it is possible to test directly the ability of experiments at stage 20/21. The first column (A,D,G,J) anterior ridge to function when it is not in contact illustrates the type of operation. The second column with the posterior apical ridge while permitting con- (B,E,H,K) illustrates the resulting skeletal pattern of the tact between anterior and posterior wing bud meso- operated right wing. The third column (C,F,I,L) illustrates the left wing from each embryo for comparison derm. purposes. The normal wing skeletal elements are labelled in (C) and include the humerus (h), radius (r), ulna (u), Skeletal patterns following posterior-bud rotation digit 2, digit 3 and digit 4. In addition, pectoral girdle Control 360° rotations elements are shown: coracoid (c) and scapula (s). When tissue posterior to the boundary line between (A,B) Anterior-bud removal. In the specimen presumptive digits 2 and 3 was rotated 360° and illustrated here, the ulna and digits 3 and 4 were slightly pinned in its original orientation, there was no loss of reduced. Digit 2 and the radius were missing. The skeletal structures. Five of seven wings had normal humerus was reduced and the glenohumeral joint defective. The scapula and coracoid failed to separate. skeletal patterns; digits 3 and 4 were slightly reduced (C) Left wing for B with skeletal elements labelled (see in size (Fig. 6A,B), while digit 2 was normal to above). reduced in size. Two of seven cases had a duplication (D,E) Anterior-ridge removal. Digit 2 was deleted in of the distal portion of digit 3. In addition, immediate 18/23 cases. The radius was reduced or absent. reincubation or holding at room temperature for (F) Left wing for E. Note that both left and right wing about lh did not alter the skeletal pattern, nor did buds were smaller than the others pictured. leaving the pins in or removing them after healing of (G,H) Posterior-bud removal. The humerus and radius the graft. Thus, the manipulations involved in ro- were reduced in size. The ulna and digits 3 and 4 were tation of the posterior tissue did not cause deletion of eliminated. In addition, digit 2 failed to develop even any skeletal structures. though presumptive digit 2 tissue was not removed. (I) Left wing for H. Posterior rotation and immediate reincubation (J,K) Posterior-ridge removal. The humerus and radius When the posterior tissue was rotated 180° (Fig. IF) developed normally. The radius was normal to slightly reduced in size. Digits 3 and 4 as well as digit 2 did not and the eggs were immediately reincubated at 37°C, develop. Development of the ulna was variable. the most common result (16/22 cases) was a wing with (L) Left wing for K. Scale bar represents 2 mm for a humerus and radius growing from the anterior illustration of skeletal patterns. tissue. The ulna and digits 3 and 4 grew posteriorly 508 W. L. Todt and J. F. Fallon Posterior-bud removal Posterior-ridge removal
10 h
24 h
36h
D
Fig. 4. Camera-lucida drawings of wing buds at the indicated times following posterior-bud removal (A-D) and posterior-ridge removal (E-H). The outer solid line of each figure is a reversed outline of the control left wing bud. The inner solid line is the profile of the operated right wing bud superimposed on the reversed left wing bud profile. The solid black areas are regions of cell death common to both wing buds (ANZ, anterior necrotic zone, Saunders, Gasseling & Saunders, 1962). The stippled areas are regions of abnormal, excessive cell death seen in the right wing bud resulting from the operation. (A-D) Posterior-bud removal. (A) At 10 h following posterior-bud removal, no cell death was observed in the right wing bud. (B) By 24 h, there was excessive cell death along the anterior border of the bud. In the left wing bud, the normal posterior border necrosis was seen (dashed line). (C,D) At 36 and 48h following posterior-bud removal, massive cell death remained visible in the distal mesoderm. Also note that proximodistal outgrowth was reduced. (E-H) Posterior-ridge removal. (E) At lOh following posterior-ridge removal, massive cell death was seen in posterior mesoderm (compare with Fig. 5A,B). (F) By 24 h, posterior cell death was no longer visible, even though normally occurring necrosis was seen along the posterior border of the left wing bud (dashed line). The ANZ was slightly enlarged. (G) At 36h after posterior-ridge removal, abnormal cell death was seen distal to the ANZ. (H) At 48 h after posterior-ridge removal the entire distal tip of the wing bud was necrotic (compare with Fig. 5C). Note the similarity in size of the bud and pattern of necrosis in C, D, and H. Scale bar (bottom) represents 1 mm. Posterior ridge and anterior wing bud patterning 509 from the region of the elbow (similar to Fig. 6C). The Table 1. Development* of presumptive digit 2 distal portion of the humerus could also be identified Stages 'proximal' to the ulna in some cases. The rotated Experiment (criteria) 19/20 20 20/21 21 (posterior) elements were often reduced in size. Three of 22 cases showed very reduced development Posterior tissue removalt of the posterior tissue, and one case showed develop- (digit 2; no digit 3) ment of digit 3 in both the anterior and posterior Posterior-bud 0%(6) 0% (6) 0% (13) 0% (6) removal tissue (similar to Fig. 6G). In two of 22 cases, a Posterior-ridge 0%(8) 0%(6) 0% (12) 0% (3) reduced digit 2 was seen at the distal tip of the radius removal (similar to Fig. 6E). This is in dramatic contrast to the complete failure of digit 2 to develop from the Posterior-bud rotation t corresponding tissue following posterior-bud or pos- (anterior digit 2; no anterior digit 3) terior-ridge removal. Immediate 0 % (5) 0 % (9) 33 % (6) 0 % (2) reincubation The healing time of these grafts, as evidenced by Brief 25°C 20% (5) 0% (6) 27% (11) 40%-(10) reestablishment of the venous circulation and drain- incubation ing of the pooled blood, was variable, being around * Data are expressed as percent (number of observations) of 24 to 30 h in cases where it was documented. In an wings with digit 2 developing isolated from other digits. attempt to facilitate healing, some embryos were left t Difference in development of isolated digit 2 following at room temperature (25°C) for about lh following posterior tissue removal (0/60 cases) and postenor-bud rotation the operation, prior to reincubation at 37°C. These (10/54 cases) is statistically significant (*2; 0-001
The role of the polarizing zone has been proposed giving the polarizing zone a Polarizing zone mesoderm is defined by its ability to central role in establishment of the anteroposterior induce the formation of polarized supernumerary pattern of the limb (see Introduction). In addition to structures from anterior wing bud tissue (see Mac- the patterning function, a second function has been Cabe, Gasseling & Saunders, 1973; Honig & Sum- attributed to the polarizing region, namely, that of merbell, 1985; Hinchliffe & Sansom, 1985). A model maintaining anterior mesodermal viability (Hinchliffe
B Posterior ridge and anterior wing bud patterning 511
& Gumpel-Pinot, 1981). However, it is not clear cases digit 2 or digits 2 and 3 developed anterior to the if cell death in the distal, anterior mesoderm is barrier (see table 2, Summerbell, 1979). Thus, separ- caused by removal of the putative morphogen. Stated ation of the anterior tissue from the polarizing zone another way: is anterior cell survival dependent on does not always cause anterior truncation. positional specification? The difficulty in answering The possibility has been raised that residual the question is that all assays for positional specifi- polarizing zone mesoderm is responsible for the cation require cell survival. However, the fact that development of anterior structures in the exper- cells die does not necessarily indicate they have lost, iments mentioned above (Hinchliffe & Gumpel- or do not have, positional specification. Pinot, 1981). However, Fallon & Crosby (1975) Experiments carried out specifically to remove only assayed limbs for polarizing activity after polarizing polarizing zone mesoderm can result in anterior zone removal and found none. These experiments development (MacCabe etal. 1973; Fallon & Crosby, should be extended to the other operations described 1975). Similarly, Iten & Murphy (1980) showed that above. In addition, posterior skeletal structures are removal of a large rectangular piece of tissue from often missing following removal of posterior tissue along the posterior border always resulted in develop- (MacCabe et al. 1973; Fallon & Crosby, 1975; Iten & ment of digit 2 (15/15 cases) and usually digit 3 (13/15 Murphy, 1980; Hinchliffe & Gumpel-Pinot, 1981). cases; see fig. 1 of Iten & Murphy, 1980). Digit 4 was We infer that presumptive tissue for the deleted eliminated in 9/15 cases and the ulna was absent posterior structures was removed. Therefore, to in 5/15 cases. Hinchliffe & Gumpel-Pinot (1981) maintain that some polarizing mesoderm was left in removed the posterior one third of the wing bud situ means that in those cases with posterior deletions (experiment P2). In 13/35 cases, one or two digits and anterior development, polarizing zone must ex- formed from the anterior two thirds of the wing bud. tend anteriorly into the presumptive skeletal tissue. Finally, when a barrier is inserted at the level of the Current maps of the wing bud do not indicate such an junction of somites 18 and 19 at stages 20-22, in 5/6 extensive distribution of polarizing tissue (see fig. 3, Hinchliffe & Gumpel-Pinot, 1984). Fig. 5. (A-C) Steropair micrographs of wholemount There is another possible explanation for anterior wing buds following posterior-ridge removal. Embryos development following polarizing zone removal. If were stained with neutral red to visualize cell death the putative morphogen were stable, i.e. had a long patterns. Viewing through a stereoscope permits three- half-life, residual morphogen in the anterior tissue dimensional visualization of the phagocytosed dead cells. could continue to signal after removal of the source. Anterior is up, posterior is down. (A) Stereopair However, we believe this is an unlikely explanation. micrographs of a right wing bud 10 h following posterior ridge removal. A large number of macrophages are seen First, implicit in the polarizing zone gradient model is in the posterior mesoderm (compare with Fig. 4E). In a signalling molecule with a short half-life. The addition, macrophages are dispersed throughout the concentration of a long-lived molecule would in- entire wing bud. The arrow shows the level to which the crease with constant production and diminish the ridge was removed. The dashed line indicates the plane gradient. Second, retinoic acid, a molecule that can of section for D. (B) Stereopair micrographs of the mimic the action of polarizing mesoderm, has a short contralateral left wing bud for A. The anterior necrotic half-life in the wing bud (approx. 20min, Eichele et zone is visible. There is very little cell death in the al. 1985). Finally, if the signalling molecule were posterior mesoderm. The dashed line indicates the plane stable, anterior truncation should not occur, yet it of section for E. (C) Stereopair micrographs of a wing often does. bud at 48 h after posterior-ridge removal. There is cell death in the distal anterior mesoderm. In addition, proximodistal outgrowth has been reduced (compare with A proposal that anterior maintenance (survival) and Fig. 4H). polarizing activity are separable (D,E) Histological sections from the wing buds Anterior mesodermal cell death is associated with pictured in A and B, respectively, at the level indicated anterior truncation in all experiments separating an- by the dashed lines (A,B). Micrographs were taken of the terior from posterior tissue. Therefore, the anterior dorsal, posterior margin of the wing bud. Mesoderm is to mesoderm is dependent on the posterior tissue for the right, ectoderm is on the left. Fixation procedures for survival. However, it is also argued above that wholemount preparations (e.g. A-C) were not optimal anterior development can proceed in the absence of for histology. However, cell death is visible. (D) polarizing activity. To reconcile these observations, Numerous macrophages are evident in the posterior mesoderm (open arrows). (E) Area corresponding to D we propose that the posterior tissue function respon- in the unoperated contralateral limb bud has only a small sible for anterior survival is distinct from, and has a amount of necrosis of the normal posterior necrotic zone. larger spatial distribution than, the polarizing zone. Scale bar in C represents 200\im for A-C. Scale bar in Thus, polarizing zone removal could leave posterior D represents 20^m for D,E. maintenance mesoderm in situ and permit anterior 512 W. L. Todt and J. F. Fallon development in the absence of polarizing activity. Wilson & Hinchliffe (1987) demonstrated that Implicit in this proposal is that mesoderm with the grafting polarizing mesoderm to the distal tip of an maintenance function, unlike the polarizing zone, anterior-half wing bud prevents cell death. However, extends anteriorly at least through presumptive digit we believe these experiments do not demonstrate 4 tissue. conclusively that the Dolarizine zone mornhnaen is
3-«* Posterior ridge and anterior wing bud patterning 513 required for normal anterior survival and develop- insertion, posterior-bud removal and posterior-ridge ment. First, if polarizing mesoderm has both main- removal. In these experiments, and in contrast to tenance and polarizing abilities, these experiments posterior-bud rotation, the anterior ridge is separated fail to separate them. Second, using quail grafts, from healthy posterior mesoderm. If separation from Wilson & Hinchliffe (1987) demonstrated that the healthy posterior mesoderm inactivated the anterior anterior tissue is respecified and often forms a com- ridge, it could lead to anterior cell death and trunc- plete set of digits from tissue that normally forms, at ation. most, digit 2. The distinction between normal devel- MacCabe and co-workers demonstrated that an- opment and development with a normal outcome terior wing tip mesoderm cultured with its over- needs to be emphasized (see also page 11, Slack, lying ectoderm becomes necrotic in vitro unless co- 1983). These experiments do not reveal the normal cultured with posterior border mesoderm (MacCabe development of anterior tissue, but are an assay for & Parker, 1975) or extracts from posterior limb tissue polarizing activity (see also Saunders & Gasseling, (Calandra & MacCabe, 1978; MacCabe & Richard- 1968) as they show that anterior mesoderm that son, 1982). These procedures may be helpful in would normally form digit 2 can be respecified to unravelling the relationship between anterior survival form a full set of polarized digits. and posterior mesoderm. However, as in the exper- iments mentioned above, they do not differentiate How does posterior mesoderm keep anterior between direct mesoderm-mesoderm signalling and mesoderm alive? indirect signalling via the apical ridge (see Hinchliffe At present, it is not possible to distinguish between a & Griffiths, 1984). direct or indirect effect of posterior mesoderm on In sum, it has been established that (1) posterior anterior mesoderm survival. Anterior ridge can func- tissue is necessary to keep anterior wing bud tissue tion to promote outgrowth of digit 2 following pos- alive prior to stage 22; and (2) posterior, polarizing terior-bud rotation. We attribute the low proportion mesoderm can respecify anterior wing bud tissue of digit 2 formation (10/54) to the length of time of when grafted near the apex. Results from several disruption of communication between posterior and laboratories lead us to propose that these two func- anterior tissue, because, as the time of graft reattach- tions may be separable and that the ability to keep ment and healing is decreased, digit 2 formation is anterior tissue alive is not confined to the polarizing facilitated. However, even though anterior ridge can zone. While our proposal suggests the existence of function isolated from the posterior ridge, it is another factor(s) for anterior mesoderm mainten- possible that anterior ridge fails following barrier ance, we believe our viewpoint is consistent with the Fig. 6. Skeletal patterns resulting from posterior bud experimental results, some of which conflict with a rotation. All operations were done on stage-20/21 strict interpretation of the polarizing zone model. embryos with brief room temperature incubation (see We hope that our proposal will stimulate further text). A, C, E and G illustrate the right wing skeletal experiments to investigate the proposed difference patterns. B, D, F and H show the control left wing between cell survival and anteroposterior axis specifi- skeletal patterns. The normal wing skeletal elements are cation in the limb. labelled as in Fig. 3: humerus (//), radius (r), ulna (u), digit 2, digit 3 and digit 4. In addition, pectoral girdle elements are shown: coracoid (c) and scapula (s). This investigation was supported by NIH Grant # (A) Right wing following 360° rotation of the posterior T32HD07118 and NSF Grant # PCM84O6338. We are tissue. All skeletal elements are present, although the grateful to Robert Auerbach, Allen Clark, Jim Cook, Leah digits are slightly reduced in size. (B) Left wing for A. Dvorak, Rebecca Fuldner, Karen Krabbenhoft, Jim Petter- (C) Right wing following 180° rotation. Digit 2 was not sen, Mary Savage, Kay Simandl, David Slautterback and present. The radius formed from the anterior tissue. A Tony Stretton for their constructive criticism of this work. cartilage nodule (n) probably represents the rotated distal Special thanks are due to Kay Simandl for expert technical portion of the humerus. The ulna and digits 3 and 4 assistance in all aspects of this work. developed in their rotated orientation and face the 'wrong' direction. In addition, they were slightly reduced to reduced in size. (D) Left wing for C. (E) Right wing References bud with a reduced digit 2 and the distal end of the radius. This degree of development was never observed CALANDRA, A. J. & MACCABE, J. A. (1978). The in vitro from the anterior tissue following posterior bud removal. maintenance of the limb-bud apical ridge by cell-free The ulna and digits 3 and 4 are normal to reduced in size. preparations. Devi Biol. 62, 258-269. (F) Left wing for E. (G) Right wing bud with digit 3 in EICHELE, G., TICKLE, C. & ALBERTS, B. M. (1985). the anterior and posterior portions of the wing. The Studies on the mechanism of retinoid-induced pattern identity of the posterior forewing element was not certain duplications in the early chick limb bud: Temporal and (?). (H) Left wing for G. Scale bar represents 2 mm. spatial aspects./ Cell Biol. 101, 1913-1920. 514 W. L. TodtandJ. F. Fallon
FALLON, J. F., BOUTIN, E. L. & CARRINGTON, J. L. MACCABE, J. A. & PARKER, B. W. (1975). The in vitro (1986). Apical ectodermal ridge maintenance in ovo maintenance of the apical ectodermal ridge of the chick and in vitro. In Cellular Endocrinology, From Embryos embryo wing bud: An assay for polarizing activity. Devi to Cells: Hormonal Control of Differentiation (ed. G. H. Biol. 45, 349-357. Sato et ai), pp. 103-113. New York: Alan R. Liss. MACCABE, J. A. & PARKER, B. W. (1979). The target FALLON, J. F. & CROSBY, G. M. (1975). Normal tissue of limb-bud polarizing activity in the induction of development in the chick wing following removal of the supernumerary structures. J. Embryol. exp. Morph. 53, polarizing zone. J. exp. Zool. 193, 449-455. 67-73. FALLON, J. F. & SAUNDERS, J. W., JR (1968). In vitro MACCABE, J. A. & RICHARDSON, K. E. Y. (1982). Partial analysis of the control of cell death in a zone of characterization of a morphogenetic factor in the prospective necrosis from the chick wing bud. Devi developing chick limb. J. Embryol. exp. Morph. 67, Biol. 18, 553-570. 1-12. HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the development of the chick embryo. MARTIN, P. & LEWIS, J. (1986). Normal development of J. Morph. 88, 49-92. the skeleton in chick limb buds devoid of dorsal HINCHLIFFE, J. R. & EDE, D. A. (1973). Cell death and ectoderm. Devi Biol. 118, 233-246. the development of limb form and skeletal pattern in ROWE, D. A., CAIRNS, J. M. & FALLON, J. F. (1982). normal and wingless (ws) chick embryos. J. Embryol. Spatial and temporal patterns of cell death in limb bud exp. Morph. 30, 753-772. mesoderm after apical ectodermal ridge removal. Devi HINCHLIFFE, J. R. & GRIFFITHS, P. J. (1984). In Cell Biol. 93, 83-91. Ageing and Cell Death (ed. I. Davies & D. C. Sigee), ROWE, D. A. & FALLON, J. F. (1981). The effect of pp. 223-242. Cambridge University Press, Cambridge. removing posterior apical ectodermal ridge of the chick HINCHLIFFE, J. R. & GUMPEL-PINOT, M. (1981). Control wing and leg bud on pattern formation. J. Embryol. of maintenance and anteroposterior skeletal exp. Morph. 65 Supplement, 309-325. differentiation of the anterior mesenchyme of the chick ROWE, D. A. & FALLON, J. F. (1982). The proximodistal wing bud by its posterior margin (the ZPA). J. determination of skeletal parts in the developing chick Embryol. exp. Morph. 62, 53-82. leg. J. Embryol. exp. Morph. 68, 1-7. HINCHLIFFE, J. R. & GUMPEL-PINOT, M. (1984). SAUNDERS, J. W., JR (1948). The proximo-distal sequence Experimental analysis of avian limb morphogenesis. In of origin of the parts of the chick wing and the role of Current Ornithology, vol. 1 (ed. R. F. Johnston), pp. the ectoderm. J. exp. Zool. 108, 363-403. 293-327. New York: Plenum Publishing Corporation. HINCHLIFFE, J. R. & SANSOM, A. (1985). The distribution SAUNDERS, J. W., JR & GASSELING, M. T. (1968). of the polarizing zone (ZPA) in the legbud of the chick Ectodermal-mesenchymal interactions in the origin of embryo. J. Embryol. exp. Morph. 86, 169-175. limb symmetry. In Epithelial-Mesenchymal Interactions HONIG, L. S. & SUMMERBELL, D. (1985). Maps of strength (ed. R. Fleischmajer & R. Billingham), pp. 78-97. of position signalling activity in the developing chick Baltimore: Williams and Wilkins. wing bud. J. Embryol. exp. Morph. 87, 163-174. SAUNDERS, J. W., JR, GASSELING, M. T. & SAUNDERS, L. ITEN, L. E. (1982). Pattern specification and pattern C. (1962). Cellular death in morphogenesis of the avian regulation in the embryonic chick limb bud. Amer. wing. Devi Biol. 5, 147-178. Zool. 22, 117-129. SAUNDERS, J. W., JR & REUSS, C. (1974). Inductive and ITEN, L. E. & MURPHY, D. J. (1980). Supernumerary limb axial properties of prospective wing-bud mesoderm in structures with regenerated posterior chick wing bud the chick embryo. Devi Biol. 38, 41-50. tissue. J. exp. Zool. 213, 327-335. SLACK, J. M. W. (1983). From Egg to Embryo. Cambridge KAPRIO, E. A. (1981). Transfilter evidence for a zone of University Press, Cambridge. polarizing activity participating in limb morphogenesis. SUMMERBELL, D. (1974). A quantitative analysis of the J. Embryol. exp. Morph. 65, 185-197. effect of excision of the AER from the chick limb-bud. KIENY, M. (1960). Role inducteur du m^soderme dans la J. Embryol. exp. Morph. 32, 651-660. differentiation prdcoce du bourgeon de membre chez SUMMERBELL, D. (1979). The zone of polarizing activity: I'embryon de poulet. J. Embryol. exp. Morph. 8, Evidence for a role in normal chick limb 457-467. morphogenesis. J. Embryol. exp. Morph. 50, 217-233. KIENY, M. (1968). Variation de la capacity inductrice du SUMMERBELL, D. (1983). The effect of local application of m6soderme et de la competence de l'ectoderme au cours de l'induction primaire du bourgeon de membre, retinoic acid to the anterior margin of the developing chez I'embryon de poulet. Archs Anat. microsc. Morp. chick limb. J. Embryol. exp. Morph. 78, 269-289. exp. 57, 401-418. SUMMERBELL, D. & WOLPERT, L. (1973). Precision of MACCABE, A. B., GASSELING, M. T. & SAUNDERS, J. W., development in chick limb morphogenesis. Nature, JR (1973). Spatiotemporal distribution of mechanisms Lond. 244, 228-230. that control outgrowth and anteroposterior polarization THALLER, C. & EICHELE, G. (1987). Identification and of the limb bud in the chick embryo. Mech. Ageing spatial distribution of retinoids in the developing chick Develop. 2, 1-12. limb. Nature, Lond. 327, 625-628. Posterior ridge and anterior wing bud patterning 515
TICKLE, C, SUMMERBELL, D. & WOLPERT, L. (1975). WARREN, A. E. (1934). Experimental studies on the Positional signalling and specification of digits in chick development of the wing in the embryo of Gallus limb morphogenesis. Nature, Lond. 254, 199-202. domesticus. Am. J. Anat. 54, 449-485. TICKLE, C, ALBERTS, B., WOLPERT, L. & LEE, J. (1982). WILSON, D. J. & HINCHLIFFE, J. R. (1987). The effect of Local application of retinoic acid to the limb bud the zone of polarizing activity (ZPA) on the anterior mimics the action of the polarizing region. Nature, half of the chick wing bud. Development 99, 99-108. ZWILLING, E. & HANSBOROUGH, L. A. (1956). Interaction Lond. 296, 564-566. between limb bud ectoderm and mesoderm in the chick TODT, W. L. & FALLOW, J. F. (1986). Development of the embryo. III. Experiments with polydactylous limbs. J. apical ectodermal ridge in the chick leg bud and a exp. Zool. 132, 219-239. comparison with the wing bud. Anat. Rec. 215, 288-304. (Accepted 7 July 1987)