The proportion and distribution of polarizing zone cells causing morphogenetic inhibition when coaggregated with anterior half wing mesoderm in recombinant limbs

The mesoderm along the posterior margin of the stage-17 to -28 (Hamburger & Hamilton, 1951) chick wing bud has been demonstrated under experimental conditions to influence polarity along the anteroposterior axis (Saunders & Gasseling, 1968; A. B. MacCabe, Gasseling & Saunders, 1973; Tickle, Summerbell & Wolpert, 1975; Summerbell & Tickle, 1977; Fallon & Crosby, 1975 a). Tickle, et al. (1975) grafted polarizing mesoderm to sequential positions along the distal rims of stage-19 to -21 wing buds, from anterior to posterior borders, and found that the digital pattern of duplicated structures induced was influenced by graft position along the anteroposterior axis. The polarity of the duplications was determined by distance of the graft from the mapped region (A. B. McCabe, et al. 1973) of polarizing activity. This suggested that cellular cues related to anteroposterior polarity emanate from the polarizing region via concentration of a diffusable substance, possibly a morphogen.

Asymmetries with respect to the anteroposterior axis disappear when dissociated-reaggregated whole wing mesoderm is placed in limb ectodermal jackets and allowed to grow (Zwilling, 1964). However, asymmetry can be restored by including a small piece of polarizing mesoderm along a border of such recombinants (J. A. MacCabe, Saunders & Pickett, 1973; Frederick, unpublished). Thus, addition of a small, intact block of polarizing mesoderm produces improved morphogenesis by conferring anteroposterior polarity to otherwise randomized limb mesoderm.

Anterior half wing mesoderm cells dissociated and assembled with ectodermal jackets to form recombinant limbs yield good limb-like outgrowths. However, the posterior half wing mesoderm cells in similar recombinants gives very poor development, or none at all (J. A. MacCabe et al. 1973; Crosby & Fallon, 1975). If polarizing mesoderm was removed from posterior half wing buds and dissociated cell recombinants made from the remaining posterior mesoderm, good limb-like development was achieved. Further, if dissociated polarizing mesoderm is added to dissociated anterior mesoderm and recombinant limbs made, these gave poor (i.e. lacking digit-like elements) or no development (Crosby & Fallon, 1975). It was concluded that the dissociated and dispersed polarizing mesoderm had an inhibitory effect on recombinant limb morphogenesis. This is in direct contrast with the inductive and polarizing capabilities of the polarizing mesoderm in the other experimental conditions described above. The aim of the present investigation was to determine the proportion and distribution of polarizing zone cells that inhibit recombinant limb morphogenesis.

Fertile White Leghorn chicken eggs and Japanese quail eggs were incubated at 38 °C for days. Chick eggs were windowed according to the technique of Zwilling (1959) and embryos of stages 21–22 used as donors. To rule out mesodermal contamination of the ectodermal jacket, recombinant limb buds in this study were assemblies of leg bud ectoderm with wing bud mesoderm.

Dissected anterior half wing mesoderm and polarizing zone pieces (Fig. 1) were placed in 2 % 1:300 trypsin and 1 % pancreatin in Ca²⁺- and Mg²⁺-free Hanks’ balanced salt solution (CMF Hanks’) for 18 min at 38 °C. These were transferred to Hanks’ balanced salt solution (Hanks’ BSS) and the ectoderm removed. The mesodermal pieces were rinsed for 5 min in CMF Hanks’ at room temperature, transferred to fresh CMF Hanks’, and incubated for 15 min at 38 °C. They were then placed in 2 % 1:300 trypsin and 1 % pancreatin for 20 min at 4 °C, followed by 16 min in the same solution at 38 °C. This was withdrawn and the fragments rinsed three or more times with a 1:1 mixture of foetal calf serum and Hanks’ BSS. A 1:2 mixture of foetal calf serum and Hanks’ BSS was added after the final rinse and the mesoderm dissociated by triturating through a fine-bore pipette. Dissociation was completed by vortexing at low speed for 5–10 seconds. The resulting mesodermal suspensions were composed of >95% single cells; the suspensions, of nonpolarizing (anterior half) and polarizing cell populations were volumetrically combined after cell counts for each were determined using a hemocytometer. Dead cells routinely accounted for < 2 % of either cell population as judged by trypan blue exclusion. The combined mesodermal suspension, vortexed to ensure adequate mixing of cells, was incubated for 30 min at 38 °C. Cells were pelleted by mild centrifugation for 6 min followed by incubation for about 1 h at 38 °C. The pellet was placed in 1:2 foetal calf serum and Hanks’ for 15 min at room temperature in preparation for recombinant assembly.

To prepare the ectodermal jackets, whole leg buds were put in CMF Hanks’ for 10 min, then transferred to 2 % trypsin and 1 % pancreatin in CMF Hanks’ for 3–4 h at 4 °C. After washing, the leg buds were added to the dish containing the reaggregated mesodermal pellet. The ectoderm of each leg bud was removed and saved. Recombinants were made and grafted to hosts as described in Crosby & Fallon (1975). After 8 days, host embryos bearing the grafts were fixed in 10 % formalin and stained with Victoria blue for cartilage.

For histological examination, recombinants were harvested at 12 h intervals until 60 h following grafting. These recombinants were fixed in 0·02 % trinitrophenol, 2·0 % formaldehyde and 2·5 % gluteraldehyde in 0·075 M phosphate buffer, postfixed in 1·0 % osmium tetroxide in 0·1 M s-collidine, dehydrated and embedded in Epon 812; sections 2 μm thick were stained with methylene blue and azure II (Fallon & Kelley, 1977).

Chick/quail xenoplastic recombinants and autoradiography of homoplastic recombinants were used to assess the sorting pattern and the participation of polarizing mesoderm in recombinant limb development. Quail polarizing cells were distinguished in chick recombinants by the presence of one to three strongly Feulgen-positive heterochromatin clumps within their nuclei (LeDouarin & Barq, 1969). These xenoplastic recombinants were allowed to grow from 1–3 days or up to 8 days, fixed, embedded in paraffin, sectioned, and stained with the Feulgen reagent and fast green. Other xenoplastic recombinants allowed to grow for 8 days were stained with Victoria blue for cartilage.

In the homoplastic experiments, cells of polarizing mesoderm were 100 % labelled with tritiated thymidine. These were included in recombinants containing unlabelled anterior half mesoderm. Recombinant limbs containing labelled polarizing mesoderm were harvested and fixed at 16, 24, 48, and 72 h after grafting and processed for autoradiography by standard methods (cf. Pollak & Fallon, 1976).

It was first necessary to establish a reference of morphogenetic performance to which recombinants containing polarizing mesoderm were compared. When recombinants of anterior half mesoderm were made, distally complete limb-like outgrowths were obtained in 44 of 73 cases (60 %) (Fig. 2 a); this 60 % figure represented the baseline of performance, i.e. results obtained under optimal conditions when no polarizing mesoderm was present in recombinant limbs (Table 1). The incidence of digits reported here may reflect recent modifications of the protocol (cf. Crosby & Fallon, 1975; and Materials and Methods) required to produce complete mesodermal dissociation of stage-21 to -22 wings.

The occurrence of distally complete limb-like outgrowths decreased as the amount of dissociated polarizing mesoderm included with anterior half mesoderm increased (Table 1). When polarizing mesoderm comprised 16–18 % of the mesodermal component of a recombinant bud, most of the resulting outgrowths were limb-like in appearance but were missing jointed phalangeal-like structures in 28 of 43 (65 %) cases. Whereas dissociated anterior half alone produced digit-like structures in 60 % of the cases, inclusion of 16–18 % polarizing mesoderm resulted in a reduction of digit-like outgrowths to 10 of 43 cases (23%). Fewer limb-like outgrowths were able to be classified as ‘distally complete’ due to deletion or fusion of distal phalangeal-like elements.

Increasing the concentration of polarizing zone cells to proportions between 20 and 36 % gave results that were similar, and for convenience, are discussed together. In this percentage range, morphogenetic inhibition became more pronounced and two effects were observed: (1) deletion of distal elements, and (2) reduction in size and discreteness of cartilagenous models that did form as defined by the uptake of stain. A distinctive morphology became progressively evident within this percentage range (Fig. 2b). Cartilaginous models appeared squat, compact, and were surrounded by more fleshy tissue than usual. This type of outgrowth never appeared when dissociated anterior border mesoderm was included with anterior half mesoderm in comparable proportions. For example, of 13 cases where anterior border (derived from quail) represented 15–42% of recombinant mesoderm, distally complete limb-like outgrowths were obtained in 10 cases (77 %), with the remaining 3 (23 %) demonstrating good, but imperfect skeletal patterns.

Where the proportion of dissociated polarizing mesoderm was above 36 %, the outgrowths that occurred were terminally deficient or fused (Table 1). Proportions above 60 % suppressed limb-like outgrowth almost completely (Fig. 2 c). Of 38 cases in which polarizing mesoderm was included with anterior half in excess of 36 %, no distally complete outgrowths resulted. Of 7 cases in which polarizing mesoderm comprised 60–100 % of the recombinant mesoderm, all were distally incomplete with 6 of the 7 yielding small cartilage nodules or tufts of soft tissue only.

The specimens discussed above represent all grafts that, at 24–36 h post-operatively, were vascularized and whose hosts survived the 8-day growth period. Although initially the graft may have appeared curled and white, by 12 h a typical graft became translucent and plump in appearance; proximal and peripheral areas appeared translucent first while the centre remained opaque. From combined histological and gross observation, it appeared that there was a necrotic core which was cleared by ingestion of debris by phagocytes. At the same time, a peripheral terminal blood vessel formed. A proximal arterial vessel penetrated the graft centrally, and soon became continuous with the already formed terminal vessel. There was a characteiistic avascular area beneath the apical ridge. Subsequently, there seemed to occur a rapid proliferation of mesodermal cells characterized in histological section at 48 h by numerous mitotic figures, and elongation primarily along the proximodistal axis. Grafts that contained even substantial proportions (e.g. 30 %) of polarizing mesoderm were grossly indistinguishable at 48 h from those that did not contain polarizing cells. However, between and 3 days, blood stasis, blood vessel enlargement, and breakdown of the vascularization began in recombinants containing higher proportions of polarizing mesoderm. By 3 days, very large sinus-like vessels made up a significant part of such recombinants. The mesoderm was loosely packed, and the apical ectodermal ridge had become simple cuboidal in morphology (Figs. 3 a, b). In contrast, in recombinants having no polarizing zone cells, blood vessels remained fine-calibered channels through which the blood cells flowed. The mesodermal cells were tightly packed with prominent nucleoli in their nuclei. The apical ectodermal ridge had the tall pseudostratified columnar morphology seen in normal limb buds (Figs. 3 c, d).

When polarizing mesoderm comprised 20–36 % of a recombinant, distal deletions, fusion, and a squat morphology were characteristic of resulting outgrowths. Therefore, 30 % was the proportion of tritiatedthymidine-labelled polarizing mesoderm chosen to observe the distribution of polarizing mesoderm in the recombinant. This percentage was beyond the amount required to produce minimal inhibitory effects, and contained enough cells to be easily visualized.

Of the recombinant grafts containing 30 % tritiated-thymidine-labelled polarizing mesoderm, only the 24 h specimens warrant reporting in detail because the label was diluted after this time. These cases (4) displayed a random distribution of radioactive-thymidine-labelled cells (polarizing mesoderm) with the original 30 % proportion being approximately maintained. Grafts harvested at 48 and 72 h showed a random distribution of very lightly labelled cells.

Results obtained using chick-quail xenoplastic recombinants corroborated those of autoradiography with the advantage that the quail Feulgen-positive marker does not suffer from successive dilution with cellular proliferation. Therefore, the location of the marked cells could readily be discerned until they reached an advanced stage of differentiation. Quail posterior border mesoderm possesses polarizing activity as assayed by grafting an intact piece to the anterior border of a host chick wing (Fallon & Thoms, 1979). Further, quail posterior border mesoderm is capable of inhibiting recombinant outgrowth and morphogenesis when dissociated and mixed with cells of chick anterior half mesoderm and follows the same pattern in Table 1. In a baseline series (8 cases), the distribution of quail mesoderm cells derived from anterior border was observed among cells of chick anterior half wing mesoderm in recombinant limbs. Pockets of quail cells were randomly distributed within such recombinants at all times sampled, using 15–17% and 42% quail anterior border mixed with anterior half wing.

A total of 17 recombinant outgrowths containing up to 40 % quail polarizing mesoderm was sectioned and examined. Of 5 recombinants containing 8 % quail polarizing mesoderm, all demonstrated a random and sparse distribution of the quail cells. At early times, i.e. up until 24 h, the quail cells were detected only with careful examination since they most often appeared as isolated, individual, intensely staining nuclei. At later times, however (e.g. 48–72 h and 8 days), detection of quail cells was facilitated by the fact that they were present in clusters, even though the clusters were few in number.

Recombinant limbs containing quail polarizing mesoderm in the 17– 19 % range (7 cases) were sectioned and examined at intervals until 8 postoperative days. As depicted in Fig. 4, quail cells were evident in small, randomly scattered pockets both at 62 h and 8 days. Quail cell clusters were never consistently associated with any particular zone of the developing recombinant; rather, their distribution was random. Quail cells, obtained from wing posterior border, dissociated, and mixed with chick anterior half wing mesoderm, were found in all mesodermal tissue types - dermis, hypodermis, muscle and its precursors, and cartilage.

With inclusion of greater proportions of mesoderm derived from quail posterior border, i.e. in the 34–40 % range (5 cases), at 58 and 82 h, the random disposition of quail cell pockets appeared to be maintained. However, at these high percentages, there were enough quail cells present that boundaries between the pockets became obscured.

When dissociated flank mesoderm was tested in the same manner as cells of the polarizing zone, comparable results were not obtained. Using 16 % dissociated flank cells, distally complete limb-like outgrowths occurred in 5 of 6 cases. Anterior half wing mesoderm with 26–30 % dissociated flank mesodermal cells yielded distally complete outgrowths in 8 of 19 (42 %) cases, distally incomplete limb-like structures in another 8 cases, and small mounds or less in the remaining 3. Histological examination of grafts containing labelled (quail or radioactive chick) flank cells revealed that at 24 h the flank cells were randomly distributed. However, it appeared that the original percentage was not maintained. Specifically, when 32 % tritiated-thymidine-labelled chick flank cells were initially added to the mesodermal aggregate, only a small fraction of this total amount could be visualized in recombinants harvested after 24 h of growth. These few cells were invariably intensely labelled and located in the middle or distal portion of the recombinant; this suggests that flank cells failed to proliferate in the recombinant limb and possibly were eliminated. Similarly, after 8 days of recombinant growth, quail flank cells were detected in proximal dermal papillae and hypodermis only. Differences in the results using labelled flank cells, which distinguish the persistent and random sorting behaviour of polarizing region cells under similar conditions, are: (1) the absence of labelled cells in 48 and 72 h grafts initially containing 32 % labelled flank cells, save for a rare and intensely labelled cell; and (2) the fine calibre of distal vasculature of 72 h recombinant outgrowths.

The data reported in this paper demonstrate that polarizing zone cells will cause detectable changes in the morphogenetic performance of recombinant limbs at relatively low percentages. The deleterious effects increase as the percentage of polarizing zone cells increase, leading ultimately to the complete failure of recombinant outgrowth and morphogenesis. The use of cell markers has demonstrated that the polarizing zone cells remain randomly distributed throughout the 8-day development of the recombinant for all percentages tested.

The role of polarizing zone during normal limb development has not yet been determined. Some investigators question the existence of such a zone or minimize its importance in any part of limb development (Saunders, 1977; Iten & Javois, 1981). Others propose polarizing zone as the source of a diffusible morphogen required at the time of the establishment of anteroposterior polarity during the limb-bud stages of development (e.g. Summerbell, 1979; Tickle, 1980). Several investigators have advanced the hypothesis that polarizing zone may have a role in the initial induction of limb outgrowth and establishment of its anteroposterior polarity (Fallon & Crosby, 1975 b, 1977; Smith, 1979; Slack, 1979). This last possibility suggests the stabilization of mitotic rate in the limb field (cf. Searls & Janners, 1971) which results in its outgrowth.

While there is no persuasive evidence for any of the various positions just alluded to, it seems clear that the mesoderm of the posterior limb bud border is a high point of morphogenetic activity. Under experimental conditions, this zone has either stimulatory or inhibitory properties not found in any other part of the limb bud. Qualitatively similar (stimulatory) activity has been found elsewhere in the embryo (Saunders, 1977) viz. somite and flank. However, it is always of reduced activity requiring a more sensitive assay than for the polarizing zone itself (Hornbruch, personal communication). We suggest it is more than likely that the mechanism(s) involved with the action of the polarizing zone are not unique to the developing limb. Rather, the posterior border may be the first recognized polarizer in the embryo, and it is likely there are other zones of such morphogenetic activity elsewhere. It is reasonable to assume that the morphogenetic activity may diminish gradually from such high points. In the case of the polarizing zone, the gradual decline would be what can be measured in flank and somites by the more sensitive assay. It is worth stressing that our study shows that flank cells do not have the inhibitory properties that polarizing zone cells display.

There is another report of inhibition using dissociated-reaggregated mesodermal cells and recombinant limbs. Singer (1972) constructed recombinant limbs composed of stage-30 leg chondrocytes combined with the stage-19 dissociated leg mesoblast. Graft size and perfection of distal structures improved proportionally to the amount of stage-19 mesoblast included in the recombinant. In a second series of experiments (Singer, 1972), stage-19 mesoblast cells were combined with proximal limb-bud cells of stage 23, 24, or 25. Compared with the stage-30 chondrocyte recombinants, these recombinants required more stage-19 mesoderm to reach an equivalent stage of development. We point out the fact that polarizing zone cells were included in the proximal limb mesoderm used. Consequently, the inhibitory contribution of dissociated and dispersed polarizing zone cells among nonpolarizing proximal limb and stage-19 mesoderm must be recognized in the analysis of the data and could account for the increased amount of stage-19 mesoderm required for good development.

In a recombinant, mesodermal cells are randomized and simultaneously polarity along the anteroposterior axis is abolished (J. A. MacCabe et al. 1973; Zwilling, 1964). Polarity, or asymmetry can be restored by inclusion of an intact piece of polarizing zone in the recombinant at the proximal and posterior margin of the ectodermal jacket. Thus polarity can be reinstated in the potentially symmetric recombinant limb by polarizing zone. Recognizing that polarizing zone is located at one edge of the 3- to 6-day embryonic chick limb, it seems reasonable that at some time during normal development this zone may specify ‘posterior’ to those prospective limb cells capable of responding to its instruction. This instruction, or message, defines the posterior border, and is essentially unidirectional. As already noted, the unidirectional signal also may be responsible for the initial outgrowth of the limb. We assume that dissociated, single polarizing zone cells produce the same signal as polarizing zone collectively in a piece of posterior border mesoderm. An interpretation of the results of these studies is that the persistence of random distribution of polarizing cells within a recombinant produces a multidirectionality of message; cells capable of responding in a prescribed manner to a given unidirectional signal fail to respond properly when the message comes from many directions and there is no clear concentration gradient along a single vector. This implies that ‘polarization’ is a vector quantity having properties of direction and magnitude. Indeed, there is evidence this is the case (Tickle, 1981). Alternatively, it is possible that polarizing zone cells in a dissociated and dispersed condition may produce an altered message. Although these are reasonable interpretations of the data described in this report, further work must be done to determine how polarizing zone cells can inhibit growth and differentiation of other limb cells when dispersed in a recombinant, and stimulate growth and differentiation of the same cells when grafted to the anterior border.

This investigation was supported by NSF Grant no. PCM7903980 and NIH Training Grant no. T32HD7118. We thank Drs Allen W. Clark, David B. Slautterback, Ms Jeanie Boutin, Ms Donene A. Rowe, and Ms B. Kay Simandl for their constructive criticism of the manuscript. Special thanks are due Dr Gayle M. Crosby for her advice and assistance in the early phases of this project. We thank Ms Lucy Taylor for the drawing and Ms Sue Leonard and Mrs Julie Meixelsperger for typing the manuscript.