Search

The polarising region expresses the signalling molecule sonic hedgehog (Shh), and is an embryonic signalling centre essential for outgrowth and patterning of the vertebrate limb. Previous work has suggested that there is a buffering mechanism that regulates polarising activity. Little is known about how the number of Shh-expressing cells is controlled but, paradoxically, the polarising region appears to overlap with the posterior necrotic zone, a region of programmed cell death. We have investigated how Shh expression and cell death respond when levels of polarising activity are altered, and show an autoregulatory effect of Shh on Shh expression and that Shh affects cell death in the posterior necrotic zone. When we increased Shh signalling, by grafting polarising region cells or applying Shh protein beads, this led to a reduction in the endogenous Shh domain and an increase in posterior cell death. In contrast, cells in other necrotic regions of the limb bud, including the interdigital areas, were rescued from death by Shh protein. Application of Shh protein to late limb buds also caused alterations in digit morphogenesis. When we reduced the number of Shh-expressing cells in the polarising region by surgery or drug-induced killing, this led to an expansion of the Shh domain and a decrease in the number of dead cells. Furthermore, direct prevention of cell death using a retroviral vector expressing Bcl2 led to an increase in Shh expression. Finally, we provide evidence that the fate of some of the Shh-expressing cells in the polarising region is to undergo apoptosis and contribute to the posterior necrotic zone during normal limb development. Taken together, these results show that there is a buffering system that regulates the number of Shh-expressing cells and thus polarising activity during limb development. They also suggest that cell death induced by Shh could be the cellular mechanism involved. Such an autoregulatory process based on cell death could represent a general way for regulating patterning signals in embryos.

A central feature of the tetrapod body plan is that two pairs of limbs develop at specific positions along the head-to-tail axis. However, the potential to form limbs in chick embryos is more widespread. This could have implications for understanding the basis of limb abnormalities. Here we extend the analysis to mouse embryos and examine systematically the potential of tissues in different regions outside the limbs to contribute to limb structures. We show that the ability of ectoderm to form an apical ridge in response to FGF4 in both mouse and chick embryos exists throughout the flank as does ability of mesenchyme to provide a polarizing region signal. In addition, neck tissue has weak polarizing activity. We show, in chick embryos, that polarizing activity of tissues correlates with the ability either to express Shh or to induce Shh expression. We also show that cells from chick tail can give rise to limb structures. Taken together these observations suggest that naturally occurring polydactyly could involve recruitment of cells from regions adjacent to the limb buds. We show that cells from neck, flank and tail can migrate into limb buds in response to FGF4, which mimics extension of the apical ectodermal ridge. Furthermore, when we apply simultaneously a polarizing signal and a limb induction signal to early chick flank, this leads to limb duplications.

It has been proposed that digit identity in chick limb bud is specified in a dose-dependent fashion by a long-range morphogen, produced by the polarising region. One candidate is Sonic hedgehog (Shh) protein, but it is not clear whether Shh acts long or short range or via Bmps. Here we dissect the relationship between Shh and Bmp signalling. We show that Shh is necessary not only for initiating bmp2 expression but also for sustaining its expression during the period when additional digits are being specified. We also show that we can reproduce much of the effect of Shh during this period by applying only Bmp2. We further demonstrate that it is Bmps that are responsible for digit specification by transiently adding Noggin or Bmp antibodies to limbs treated with Shh. In such limbs, multiple additional digits still form but they all have the same identity. We also explored time dependency and range of Shh signalling by examining ptc expression. We show that high-level ptc expression is induced rapidly when either Shh beads or polarising regions are grafted to a host limb. Furthermore, we find that high-level ptc expression is first widespread but later more restricted. All these data lead us to propose a new model for digit patterning. We suggest that Shh initially acts long range to prime the region of the limb competent to form digits and thus control digit number. Then later, Shh acts short range to induce expression of Bmps, whose morphogenetic action specifies digit identity.

talpid3 is an embryonic-lethal chicken mutation in a molecularly un-characterised autosomal gene. The recessive, pleiotropic phenotype includes polydactylous limbs with morphologically similar digits. Previous analysis established that hox-D and bmp genes, that are normally expressed posteriorly in the limb bud in response to a localised, posterior source of Sonic Hedgehog (Shh) are expressed symmetrically across the entire anteroposterior axis in talpid3 limb buds. In contrast, Shh expression itself is unaffected. Here we examine expression of patched (ptc), which encodes a component of the Shh receptor, and is probably itself a direct target of Shh signalling, to establish whether talpid3 acts in the Shh pathway. We find that ptc expression is significantly reduced in talpid3 embryos. We also demonstrate that talpid3 function is not required for Shh signal production but is required for normal response to Shh signals, implicating talpid3 in transduction of Shh signals in responding cells. Our analysis of expression of putative components of the Shh pathway, gli1, gli3 and coupTFII shows that genes regulated by Shh are either ectopically expressed or no longer responsive to Shh signals in talpid3 limbs, suggesting possible bifurcation in the Shh pathway. We also describe genetic mapping of gli1, ptc, shh and smoothened in chickens and confirm by co-segregation analysis that none of these genes correspond to talpid3.

Homeobox gene XlHbox 1 is expressed in a mesodermal gradient in vertebrate forelimbs with maximal expression anteriorly and proximally and may encode positional values. In chick wing buds, anterior cells can be reprogrammed to form posterior structures by grafts of polarizing region tissue and by beads soaked in retinoic acid (RA), which is a good candidate for an endogenous morphogen. Applications of RA anteriorly or at the bud apex, treatments which produce duplicated digits or truncations respectively, substantially increase the extent of mesodermal XlHbox 1 expression. Polarizing region grafts that also produce additional digits lead to a moderate increase. The effects of RA application and the behaviour of transplanted tissue show that only anterior cells are competent to express XlHbox 1 and that expression is cell autonomous. Ectodermal expression in wing buds is enhanced by RA but not by polarizing region grafts and ectoderm/mesoderm recombinations show that the mesoderm is irreversibly affected. The changes in mesodermal expression do not fit the predictions of the simple model that XlHbox 1 encodes anterior positional values but are correlated with a series of novel malformations of the shoulder girdle which, in normal wing buds, is derived from cells expressing XlHbox 1.

The relationship between the expression of Hox‐4 genes in the mesenchyme and the apical ectodermal ridge was investigated in both normal chick wing buds and wing buds treated with retinoic acid. Two conclusions emerge. One is that the activation of Hox‐4 domains and the elaboration of Hox‐4 gene expression patterns involve cooperation with a signal from the apical ridge. The second is that the domains of expression of 5′‐located members of the complex correlate with the maintenance of the thickened ridge which is required for subsequent bud outgrowth.

A group of mesenchyme cells at the posterior margin of the developing wing—the polarizing region—can have a dramatic effect on the pattern of structures which develop across the antero-posterior axis of the limb. If one grafts a polarizing region to the anterior margin of a wing bud so that this bud now has one polarizing region at the anterior margin and one at the posterior margin, the wing that develops has duplicated structures across the antero-posterior axis in mirror-image symmetry1–3. That is, the normal pattern of digits 2 3 4 (reading from anterior to posterior) becomes 4 3 2 2 3 4. The ability of the polarizing region to specify additional digits from adjacent tissue can be progressively attenuated by γ-ray radiation4. If this attenuation is caused by progressively fewer cells remaining viable to signal, there should be a quantitative relationship between the number of polarizing cells used and the digit specified next to the graft. Here, two tests are reported which confirm this idea. The results suggest that the apical ridge cooperates with a small number of polarizing region cells in a monolayer to specify structures across the antero-posterior axis of the wing.

Considerable recent advances have been made in understanding the mechanisms of vertebrate limb development. New information about molecules governing cell interactions in embryonic limbs begins to bridge the gap between the experimental analysis and genetics of congenital limb defects. There are four main stages in vertebrate limb development: initiation, specification of limb pattern, tissue formation accompanied by limb morphogenesis, and growth. Although classical embryology focused on chick embryos and recent molecular analysis centres on limbs of both chickens and mice, most of the fundamental mechanisms that have been uncovered appear to be conserved between vertebrates and are likely to be directly applicable to human limb development.

Abstract:  Spatially-restricted expression domains of Msx 1 and Msx 2 in the developing chick face suggest that they may play a role in epithelial-mesenchymal interactions governing outgrowth of facial primordia. Retinoid application to developing chick faces reproducibly inhibits upper beak outgrowth but the lower beak is unaffected. In the normal face, high levels of Msx gene transcripts in upper and lower beak primordia correlate with regions of outgrowth. Following retinoid treatment, Msx 1 and Msx 2 transcripts are rapidly down-regulated in upper beak primordia where outgrowth is inhibited, but remain largely unchanged in lower beak primordia, where outgrowth is unaffected. Decreases in gene expression precede retinoid-induced morphological changes in the upper beak, suggesting that Msx gene products are involved in mediating the effect of retinoids on facial development.

The products of Hox-4 genes appear to encode position in developing vertebrate limbs. In chick embryos, a number of different signalling regions when grafted to wing buds lead to duplicated digit patterns. We grafted tissue from the equivalent regions in mouse embryos to chick wing buds and assayed expression of Hox-4 genes in both the mouse cells in the grafts and in the chick cells in the responding limb bud using species specific probes. Tissue from the mouse limb polarizing region and anterior primitive streak respecify anterior chick limb bud cells to give posterior structures and lead to activation of all the genes in the complex. Mouse neural tube and genital tubercle grafts, which give much less extensive changes in pattern, do not activate 5'-located Hox-4 genes. Analysis of expression of Hox-4 genes in mouse cells in the grafted signalling regions reveals no relationship between expression of these genes and strength of their signalling activity. Endogenous signals in the chick limb bud activate Hox-4 genes in grafts of mouse anterior limb cells when placed posteriorly and in grafts of mouse anterior primitive streak tissue. The activation of the same gene network by different signalling regions points to a similarity in patterning mechanisms along the axes of the vertebrate body.

Developing chicken wings homozygous for the talpid (ta3/ta3) mutation are polydactylous and have defects in the establishment of their anteroposterior polarity. We analysed the expression domains of the posteriorly restricted homeobox Hox-4 genes in such mutant wings. The Hox-4 genes are now expressed right across the anteroposterior axis instead of being expressed just posteriorly. This correlates well with the absence of clear morphological differences between the talpid3 digits and reinforces the idea that vertebrate Hox-4 genes are involved in setting up the limb anteroposterior asymmetry.

The Eph-related receptor tyrosine kinase gene, Cek-8, is expressed in mesenchyme at the tip of chick limb buds, with high levels of transcripts posteriorly and apically but fading out anteriorly. Expression of Cek-8 in distal mesenchyme is regulated by apical ridge- and FGF-polarising signals and retinoic acid, and is uniform across the anteroposterior axis in talpid3 mutants. These data indicate that Cek-8 expression responds to regulatory signals during limb patterning and suggest that this receptor tyrosine kinase may have a role in coordinating responses to signals in the progress zone of early buds. Later on in limb development, Cek-8 expression is associated with cell condensations that form tendons and their attachments to cartilage rudiments and then in developing feather buds.

We wish to understand how limbs are positioned with respect to the dorso-ventral axis of the body in vertebrate embryos, and how different regions of limb bud ectoderm, i.e. dorsal ectoderm, apical ridge and ventral ectoderm, originate. Signals from dorsal and ventral ectoderm control dorso-ventral patterning while the apical ectodermal ridge (AER) controls bud outgrowth and patterning along the proximo-distal axis. We show, using cell-fate tracers, the existence of two distinct ectodermal compartments, dorsal versus ventral, in both presumptive limb and flank of early chick embryos. This organisation of limb ectoderm is the first direct evidence, in vertebrates, of compartments in non-neural ectoderm. Since the apical ridge appears to be confined to this compartment boundary, this positions the limb. The mesoderm, unlike the ectoderm, does not contain two separate dorsal and ventral cell lineages, suggesting that dorsal and ventral ectoderm compartments may be important to ensure appropriate control of mesodermal cell fate. Surprisingly, we also show that cells which form the apical ridge are initially scattered in a wide region of early ectoderm and that both dorsal and ventral ectoderm cells contribute to the apical ridge, intermingling to some extent within it.

Anteroposterior polarity in the vertebrate limb is thought to be regulated in response to signals derived from a specialized region of distal posterior mesenchyme, the zone of polarizing activity. Sonic Hedgehog (Shh) is expressed in the zone of polarizing activity and appears to mediate the action of the zone of polarizing activity. Here we have manipulated Shh signal in the limb to assess whether it acts as a long-range signal to directly pattern all the digits. Firstly, we demonstrate that alterations in digit development are dependent upon the dose of Shh applied. DiI-labeling experiments indicate that cells giving rise to the extra digits lie within a 300 microm radius of a Shh bead and that the most posterior digits come from cells that lie very close to the bead. A response to Shh involves a 12-16 hour period in which no irreversible changes in digit pattern occur. Increasing the time of exposure to Shh leads to specification of additional digits, firstly digit 2, then 3, then 4. Cell marking experiments demonstrate that cells giving rise to posterior digits are first specified as anterior digits and later adopt a more posterior character. To monitor the direct range of Shh signalling, we developed sensitive assays for localizing Shh by attaching alkaline phosphatase to Shh and introducing cells expressing these forms into the limb bud. These experiments demonstrate that long-range diffusion across the anteroposterior axis of the limb is possible. However, despite a dramatic difference in their diffusibility in the limb mesenchyme, the two forms of alkaline phosphatase-tagged Shh proteins share similar polarizing activity. Moreover, Shh-N (aminoterminal peptide of Shh)-coated beads and Shh-expressing cells also exhibit similar patterning activity despite a significant difference in the diffusibility of Shh from these two sources. Finally, we demonstrate that when Shh-N is attached to an integral membrane protein, cells transfected with this anchored signal also induce mirror-image pattern duplications in a dose-dependent fashion similar to the zone of polarizing activity itself. These data suggest that it is unlikely that Shh itself signals digit formation at a distance. Beads soaked in Shh-N do not induce Shh in anterior limb mesenchyme ruling out direct propagation of a Shh signal. However, Shh induces dose-dependent expression of Bmp genes in anterior mesenchyme at the start of the promotion phase. Taken together, these results argue that the dose-dependent effects of Shh in the regulation of anteroposterior pattern in the limb may be mediated by some other signal(s). BMPs are plausible candidates.

Tbx-2, Tbx-3, Tbx-4 and Tbx-5 chick genes have been isolated and, like the mouse homologues, are expressed in the limb regions. Tbx-2 and Tbx-3 are expressed in anterior and posterior domains in wings and legs, as well as throughout the flank. Of particular interest, however, are Tbx-5, which is expressed in wing and flank but not leg, and Tbx-4, which is expressed very strongly in leg but not wing. Grafts of leg tissue to wing and wing tissue to leg give rise to toe-like or wing-like digits in wing and leg respectively. Expression of Tbx-4 is stable when leg tissue is grafted to wing, and Tbx-5 expression is stable when wing tissue is grafted to leg. Induction of either extra wings or legs from the flank by applying FGF-2 in different positions alters the expression of Tbx-4 and Tbx-5 in such a way that suggests that the amount of Tbx-4 that is expressed in the limb determines the type that will form. The ectopic limb always displays a limb-like Tbx-3 expression. Thus Tbx-4 and Tbx-5 are strong candidates for encoding 'wingness' and 'legness'.

The developing limb is a good model for exploring mechanisms that establish cell and tissue plans during vertebrate development. A set of inductive cell interactions along each of the three limb axes specify the pattern of structures that involve the production of a signal by one group of cells and the response to this signal by a second group of cells. Signalling molecules have now been identified for each set of interactions and networks of such signals are beginning to be dissected. These not only co-ordinate cell responses that lead to patterning and morphogenesis, but also mutually maintain each other. Initiation of a limb signalling network can be induced by application of a single growth factor.

Bone morphogenetic proteins (BMPs) induce ectopic cartilage and bone when implanted intramuscularly in adult rats. Expression data suggest that BMPs signal skeletal development in embryos. An important question is which cells are targets of BMP signaling in adult and embryonic tissues. Here, we examined the effect of BMP-2 on micromass cultures of chick limb bud mesenchyme. We report that BMP-2 promotes formation of cartilage and simultaneously inhibits development of muscle cells. To follow the fate of presumptive muscle cells, we replaced chick somites with quail somites at the wing level and made micromass cultures from the chimeric wings. In untreated cultures, quail cells formed muscle but not cartilage. In BMP-2-treated cultures, quail cells disappeared altogether. This suggests that BMP-2 may simultaneously promote cartilage differentiation and reduce the presumptive myogenic cell populations in regions of skeletal development.

The invasiveness of mouse lymphocytes and thymocytes, rabbit peritoneal neutrophil granulocytes (PMNs), mouse peritoneal macrophages (both activated and non-activated) and pig endothelial cells was assayed by implanting these cells to the chick wing bud. Cells of each type moved into the wing mesenchyme, although activated macrophages invaded poorly. PMNs were the most invasive cells and had moved well into the limb after only a few hours. PMNs, lymphocytes and thymocytes were ingested by wing mesenchyme cells. Endothelial cells, however, ingested chick blood cells. The implanted cells showed differences in ability to survive in the limb: PMNs disappeared rapidly, lymphocytes and thymocytes sometimes persisted for 24 h, while grafts of macrophages and endothelial cells were present at 24 h. Mechanisms which might be involved in the invasiveness of these cells, and also in their different abilities to survive in the chick wing, are discussed with particular reference to the production of plasminogen activator.

The interpretation of positional information can provide the basis for pattern formation in limb morphogenesis. The gradient in positional information along the antero–posterior axis, which is specified with respect to a localised boundary region, can be modified by grafting this region to successive positions along the axis. The pattern of digits obtained is consistent with a model based on diffusion of a labile morphogen and is thus similar to models proposed for the development of pattern in invertebrates.

A method for analysing the localization of the Hox-4 gene complex by three-dimensional reconstruction of serial cross-sections is presented. It allows the simultaneous visualization of three Hox expression domains on the same reconstructed embryo and the study of the effect of a graft of chicken polarizing region on the de novoexpression of the genes of the HOX-4 complex. The method consists of registering the sections, segmenting the biological patterns and finally displaying the reconstructed embryo. The precise relationships between the transcript domains of the three genes can thus be assessed in either the normal wing bud or upon pattern modification.

Development of the musculature involves generation of a precise number of individual muscles arranged in appropriate locations, each with the correct cellular patterning. To find out the rules that govern muscle number and arrangement, the forearm musculature of chick wing buds was analysed following grafts of the polarizing region or application of retinoic acid. Muscle patterns appear symmetrical with 'posterior' muscles now forming in the anterior part of the wing. When the number of muscles that develop is reduced, pattern symmetry is maintained, with loss of anterior muscles in the mid-line, especially dorsally. Strict anteroposterior ordering of muscles in duplicated patterns does not always occur. The number of muscles that develops bears some relationship to the number of forearm elements. Each muscle has a characteristic pattern of fast and slow fibres. In duplicated wings, each pair of symmetrically arranged muscles has the same fibre type pattern. Not only are proportions of fast and slow fibres similar, but local variations in fibre type arrangement within the muscle are also reproduced. This suggests that the cellular pattern within the new 'posterior' muscles at the anterior of the limb has been re-specified. In manipulated limb buds, which will develop a duplicated muscle pattern, there are no detectable changes in distribution and number of potentially myogenic cells, and fibre type patterning within early muscle masses also appears normal. In contrast, the splitting process that divides up muscle masses is altered. The appropriate fibre type arrangement only emerges after splitting is complete. This suggests that tissue patterning and cellular patterning occur at different times during muscle development.

We show here that bone morphogenetic protein 2 (BMP-2) is involved in patterning the developing chick limb. During early stages of limb development, mesenchymal expression of the Bmp-2 gene is restricted to the posterior part of the bud, in a domain that colocalizes with the polarizing region. The polarizing region is a group of cells at the posterior margin of the limb bud that can respecify the anteroposterior axis of the limb when grafted anteriorly and can activate expression of genes of the HoxD complex. We dissect possible roles of BMP-2 in the polarizing region signalling pathway by manipulating the developing wing bud. Retinoic acid application, which mimics the effects of polarizing region grafts, activates Bmp-2 gene expression in anterior cells. This shows that changes in anteroposterior pattern are correlated with changes in Bmp-2 expression. When polarizing region grafts are placed at the anterior margin of the wing bud, the grafts continue to express the Bmp-2 gene and also activate Bmp-2 expression in the adjacent anterior host mesenchyme. These data suggest that BMP-2 is part of the response pathway to the polarizing signal, rather than being the signal itself. In support of this, BMP-2 protein does not appear to have any detectable polarizing activity when applied to the wing bud. The pattern of Bmp-4 gene expression in the developing wing bud raises the possibility that BMP-2 and BMP-4 could act in concert. There is a close relationship, both temporal and spatial, between the activation of the Bmp-2 and Hoxd-13 genes in response to retinoic acid and polarizing region grafts, suggesting that expression of the two genes might be linked.

The polarizing region is a major signalling tissue involved in patterning the tissues of the vertebrate limb. The polarizing region is located at the posterior margin of the limb bud and can be recognized by its ability to induce additional digits when grafted to the anterior margin of a chick limb bud. The signal from the polarizing region operates at the tip of the bud in the progress zone, a zone of undifferentiated mesenchymal cells, maintained by interactions with the apical ectodermal ridge. A number of observations have pointed to a link between the apical ectodermal ridge and signalling by the polarizing region. To test this possibility, we removed the posterior apical ectodermal ridge of chick wing buds and assayed posterior mesenchyme for polarizing activity. When the apical ectodermal ridge is removed, there is a marked decrease in polarizing activity of posterior cells. The posterior apical ectodermal ridge is known to express FGF-4 and we show that the decrease in polarizing activity of posterior cells of wing buds that normally follows ridge removal can be prevented by implanting a FGF-4-soaked bead. Furthermore, we show that both ectoderm and FGF-4 maintain polarizing activity of limb bud cells in culture.

Mouse mesenchyme was grafted into chick embryos to investigate the control of mesenchymal expression of Msx-1 in the developing limb and face. In situ hybridization, using species-specific probes, allows a comparison between Msx-1 expression in the graft and the host tissue. The results show that Msx-1 expression in both limb-to-limb and face-to-face grafts corresponds closely with the level of Msx-1 expression in the surrounding chick mesenchyme. Cells in grafts that end up within the host domain of Msx-1 express the gene irrespective of whether they were from normally expressing, or non-expressing, regions. Therefore Msx-1 expression in both the developing limb and the developing face appears to be position-dependent. Mesenchyme from each of the three major facial primordia behaved in the same way when grafted to the chick maxillary primordium. Reciprocal grafts between face and limb gave a different result: Msx-1 expression was activated when facial mesenchyme was grafted to the limb but not when limb mesenchyme was grafted to the face. This suggests either that there are quantitative or qualitative differences in two local signalling systems or that additional factors determine the responsiveness of the mesenchyme cells.

Bone morphogenetic protein-2 (BMP-2) has been implicated in the polarizing region signalling pathway, which specifies pattern across the antero-posterior of the developing vertebrate limb. Retinoic acid and Sonic Hedgehog (SHH) can act as polarizing signals; when applied anteriorly in the limb bud, they induce mirror-image digit duplications and ectopic Bmp-2 expression in anterior mesenchyme. In addition, the two signals can activate Fgf-4 expression in anterior ridge and HoxD expression in anterior mesenchyme. We tested the role of BMP-2 in this signalling cascade by ectopically expressing human BMP-2 (hBMP-2) at the anterior margin of the early wing bud using a replication defective retroviral vector, and found that ectopic expression of Fgf-4 was induced in the anterior part of the apical ectodermal ridge, followed later by ectopic expression of Hoxd-11 and Hoxd-13 in anterior mesenchyme. This suggests that BMP-2 is involved in regulating Fgf-4 and HoxD gene expression in the normal limb bud. Ectopically expressed hBMP-2 also induced duplication of digit 2 and bifurcation of digit 3, but could not produce the mirror-image digit duplications obtained with SHH-expressing cells. These results suggest that BMP-2 may be involved primarily in maintenance of the ridge, and in the link between patterning and outgrowth of the limb bud.

Rhombomeres are segmental units of the hindbrain that are separated from each other by a specialised zone of boundary cells. Retinoic acid application to a recently segmented hindbrain leads to disappearance of posterior rhombomere boundaries. Boundary loss is preceded by changes in segmental expression of Krox-20 and Cek-8 and followed by alterations in Hox gene expression. The characteristic morphology of boundary cells, their expression of follistatin and the periodic accumulation of axons normally associated with boundaries are all lost. In the absence of boundaries, we detect no change in anteroposterior dispersal of precursor cells and, in most cases, no substantial cell mixing between former rhombomeric units. This is consistent with the idea that lineage restriction can be maintained by processes other than a mechanical barrier composed of boundary cells. Much of the early organisation of the motor nuclei appears normal despite the loss of boundaries and altered Hox expression.

We have produced detailed fate maps for mesenchyme and apical ridge of a stage 20 chick wing bud. The fate maps of the mesenchyme show that most of the wing arises from the posterior half of the bud. Subapical mesenchyme gives rise to digits. Cell populations beneath the ridge in the mid apical region fan out into the anterior tip of the handplate, while posterior cell populations extend right along the posterior margin. Subapical mesenchyme of the leg bud behaves similarly. The absence of anterior bending of posterior cell populations has implications when considering models of vertebrate limb evolution. The fatemaps of the apical ridge show that there is also a marked anterior expansion and cells that were in anterior apical ridge later become incorporated into non-ridge ectoderm along the margin of the bud. Mesenchyme and apical ridge do not expand in concert--the apical ridge extends more anteriorly. We used the fatemaps to investigate the relationship between cell lineage and elaboration of Hoxd-13 and Fgf-4 domains. Hoxd-13 and Fgf-4 are initially expressed posteriorly until about the mid-point of the early wing bud in mesenchyme and apical ridge respectively. Later in development, the genes come to be expressed throughout most of the handplate and apical ridge respectively. We found that at the proximal edge of the Hoxd-13 domain, cell populations stopped expressing the gene as development proceeded and found no evidence that the changes in extent of the domains were due to initiation of gene expression in anterior cells. Instead the changes in extent of expression fit with the fate maps and can be attributed to expansion and fanning out of cell populations initially expressing the genes.

The pattern of connexin43 expression in developing chick limb buds was examined using a site-specific polyclonal antibody and confocal microscopy. Connexin43 is expressed at stages of limb development when epithelial-mesenchymal interactions are occurring that mediate morphogenesis. Extensive labeling was observed in the apical ectodermal ridge and labeling was also found in underlying mesenchyme cells at the tip of the bud. In mouse limb buds, the same gap junction protein is expressed only in the apical ridge. Manipulations of developing chick wing buds show that mesenchymal expression of connexin43 appears to be controlled by the apical ectodermal ridge. When the apical ridge is surgically removed and limb truncations result, mesenchymal labeling is markedly reduced and conversely the grafting of an additional ridge induces connexin43 expression between underlying mesenchymal cells which do not normally show expression at this stage of development. In addition, a treatment with retinoic acid that flattens the apical ridge and inhibits bud outgrowth reduces expression in both mesenchymal and epithelial tissues. The abolition of connexin43 expression in mesenchymal and epithelial domains when bud outgrowth is halted suggests that synthesis of this gap junction protein is related to the epithelial-mesenchymal interactions that mediate morphogenesis of the bud. Copyright 1994, 1999 Academic Press

When retinoic acid is locally applied to the anterior margin of developing chick wing buds on ion-exchange beads, dose-dependent changes in the skeletal pattern result. At low doses, additional digits develop. At high doses, there is thinning of the symmetrical wing. Local application of retinoic acid to the apex of the bud also leads to pattern changes, but in contrast normal wing patterns are almost always obtained following application posteriorly. These effects are manifest at 6–7 days after the operation although only a brief exposure (14–20 h) to retinoic acid is required. Therefore the morphology of wing buds was studied at shorter times after the start of treatment. The local application of retinoic acid to the wing bud margin leads to changes in extent of the apical ridge that can be detected at 24 h after application. The behaviour of the apical ridge with varying doses and positions of retinoic acid application has been analysed quantitatively and dose response curves obtained. At low doses of retinoic acid, the length of the apical ridge increases or remains constant, but then progressively decreases with higher doses. The progressive obliteration of the ridge starts first near the bead and then involves more distant parts of the bud. Thus the region of the ridge affected depends on the position at which the retinoic acid is applied. We propose that these effects on the apical ridge reflect dose-dependent responses to the local concentration of retinoic acid that varies with distance from the source. At high doses, the apical ridge disappears but at low doses it is maintained. Since grafts of polarizing region tissue also have a graded effect on ridge morphology, a possible interpretation of the retinoic acid effects is that tissue adjacent to the source is converted into polarizing region tissue. Alternatively, retinoic acid may act directly on the ridge cells. The changes in the extent of the apical ridge produced by retinoic acid lead to different forms of bud outgrowth. The form of the outgrowth depends on the dose of retinoic acid, the position of application and the interaction between the effects of the local source of retinoic acid and those of the polarizing region of the host bud. These considerations give some insights into why anterior application of retinoic acid leads to the development of additional digits whereas posterior application generally gives normal wings.

Implanting inert carriers soaked in retinoic acid into the anterior margin of the developing limb of chicken embryos leads to orofacial malformations as well as affecting pattern formation in the limb. Using anion-exchange beads as carriers, and soaking solutions of 1–10 mg/ml retinoic acid, almost 100% of the embryos have malformations of the face. The effects on the treated limbs range from symmetrical patterns of duplicated digits (maximum number of digits being four) to truncations in which no digits were formed at all. Typically, in the malformed faces the upper beak is completely absent, no nostrils are present and the front of the face forms a scalloped rim of tissue above the mouth. By reference to normal beak development, the seven bulges of tissue that make up the rim can be identified as derivatives of the masses of tissue that normally would fuse to form the upper beak. The roof of the mouth consists of three bulges of tissue flanked by widely separated palatal shelves. The defect can thus be classified as severe bilateral clefting of the primary palate. By examining the morphology of the faces of treated embryos, the origin of the defect can be traced to failure of the frontonasal mass to enlarge. Thus, the oronasal fissures are very wide and fusion across them to form the primary palate cannot occur. The way in which retinoic acid brings about the defect is discussed in relation to possible mechanisms involved in the production of cleft palate. The parallel is noted between the associated effects of retinoic acid on beak and limb morphogenesis and the chick mutation cpp, that also affects both face and limbs.

The chick limb bud is a powerful experimental system in which to study pattern formation in vertebrate embryos. Exogenously applied retinoic acid, a vitamin A derivative, can bring about changes in pattern and, on several grounds, is a good candidate for an endogenous morphogen. As such, the local concentration of retinoic acid might provide cells with information about their position in relation to one axis of the limb. Alternatively, retinoic acid may be part of a more complex signalling system. Homeobox genes are possible target genes for regulation by retinoic acid in the limb. In particular, one homeobox gene, XlHbox 1 is expressed locally in the mesenchyme of vertebrate forelimbs and might code for an anterior position. When the pattern of the chick wing is changed by retinoic acid or by grafts of signalling tissue such that anterior cells now form posterior structures, the domain of XlHbox 1 expression expands rather than contracts. The expansion of XlHbox 1 expression correlates with shoulder girdle abnormalities. Retinoic acid application leads to visible changes in bud shape and this allows dissection of the way in which patterning is co-ordinated with morphogenesis. Results of recombination experiments and studies of changes in the apical ridge and proliferation in the mesenchyme suggest the following scheme: retinoic acid is involved in specification of position of mesenchyme cells; this specification determines their local interaction with the ridge that controls ridge morphology; the thickened apical ridge permits local proliferation in the underlying mesenchyme. The recent advances in molecular biology that permit analysis of the expression of various interesting genes in developing limbs hold out the promise that further investigation may soon allow a complete account of the patterning process in one part of the vertebrate embryo.

Pattern formation is the developmental process that leads to the spatial ordering of cell differentiation. We have explored the problem of pattern formation in the development of the face of chick embryos. At early stages, the developing face consists of a series of small buds of tissue, the facial primordia that encircle the primitive mouth. The concepts of positional information provide a framework for considering how the patterns of differentiated cells are generated in the face. We suggest that the cranial neural crest cells must first be informed to which facial primordium they belong and then of their position within that primordium. The cells of the early primordia appear indistinguishable. However, when the mesenchyme cells are placed in high-density culture, cartilage differentiates. The extent and pattern of cartilage differentiation is characteristic for the cell population of each facial primordium. Myogenic cells also differentiate in the cultures, but the proportion of myogenic cells is independent of the extent of chondrogenesis. Within the facial primordia, a set of epithelial–mesenchymal interactions appears to be required for outgrowth and pattern formation along the proximodistal axis of the chick beaks. In culture, face epithelium locally inhibits cartilage differentiation and suggests that another set of epithelial–mesenchymal interactions may be involved in cell patterning. The mechanisms involved in specifying the mediolateral axis of the face, for example, the midpoint of the upper beak, are not known. Vitamin A derivatives, collectively known as retinoids, affect the development of the face of chick embryos and lead to a specific facial defect. Upper beak development is inhibited but the lower beak develops normally. The response to retinoids could be related to the specification of cells to belong to the facial primordium that will form the upper beak. Alternatively, retinoids may interfere with positional cues that operate to inform cells of their position within that primordium.

All-trans-retinoic acid causes striking digit pattern changes when it is continuously released from a bead implanted in the anterior margin of an early chick wing bud. In addition to the normal set of digits (234), extra digits form in a mirror-symmetrical arrangement, creating digit patterns such as a 432234. These retinoic acid-induced pattern duplications closely mimic those found after grafts of polarizing region cells to the same positions with regard to dose-response, timing, and positional effects. To elucidate the mechanism by which retinoic acid induces these pattern duplications, we have studied the temporal and spatial distribution of all-trans-retinoic acid and its potent analogue TTNPB in these limb buds. We find that the induction process is biphasic: there is an 8-h lag phase followed by a 6-h duplication phase, during which additional digits are irreversibly specified in the sequence digit 2, digit 3, digit 4. On average, formation of each digit seems to require between 1 and 2 h. The tissue concentrations, metabolic pattern, and spatial distribution of all-trans-retinoic acid and TTNPB in the limb rapidly reach a steady state, in which the continuous release of the retinoid is balanced by loss from metabolism and blood circulation. Pulse-chase experiments reveal that the half-time of clearance from the bud is 20 min for all-trans-retinoic acid and 80 min for TTNPB. Manipulations that change the experimentally induced steep concentration gradient of TTNPB suggest that a graded distribution of retinoid concentrations across the limb is required during the duplication phase to induce changes in the digit pattern. The extensive similarities between results obtained with retinoids and with polarizing region grafts raise the possibility that retinoic acid serves as a natural "morphogen" in the limb.

One of the ways in which Fundulus deep cells move in vivo is by putting out long, fingerlike protrusions. This involves a change in the shape of the cell as a whole, with cytoplasmic flow, and is not just a local phenomenon. Moreover, particles on the cell surface move toward a protrusion as it is forming, suggesting surface flow. The role of surface flow is discussed both on a gross level and in respect to molecular fluidity. Long, stable protrusions can be pulled from cells by the application of negative pressure at a constant rate and these behave in a similar way to those formed during cell locomotion. Such long protrusions must be structured. The importance of contractile properties of the cytoplasm in the formation of protrusions was studied by treating cells with media that modify cellular contractility.

The formation of the nasal passages involves complex morphogenesis and their lining develops a spatially ordered pattern of differentiation, with distinct domains of olfactory and respiratory epithelium. Using antibodies to the neural cell adhesion molecule (N-CAM), keratan sulphate and heparan sulphate proteoglycan (HSPG) and a panel of lectins (agglutinins of Canavalia ensiformis (ConA), Dolichos biflorus (DBA), peanut (PNA), Ricinis communis (RCA1), soybean (SBA), Ulex europaeus (UEA1), and wheatgerm (WGA)), we have documented cell surface characteristics of each epithelial domain. Binding of antibodies to N-CAM and to keratan sulphate, and the lectins ConA, PNA, RCA1, SBA and WGA marks the olfactory epithelial domain only. The restriction of N-CAM to the sensory region of the epithelium has also been reported in the developing ear. This striking similarity is consistent with the idea that N-CAM may be involved in the division of functionally and histologically distinct cell groups within an epithelium. We traced the olfactory-specific cell markers during development to gain insights into the origin of the epithelial lining of the nasal passages. AU reagents bind at early stages to the thickened nasal placode and surrounding head ectoderm and then become progressively restricted to the olfactory domain. The expression of these characteristics appears to be modulated during development rather than being cell autonomous. The distribution of keratan sulphate was compared with collagen type II in relation to the specification of the chondrocranium. Keratan sulphate and collagen type II are only colocalized at the epithelial-mesenchymal interface during early nasal development. At later stages, only collagen type II is expressed at the interface throughout the nasal passages, whereas keratan sulphate is absent beneath the respiratory epithelium.

Local application of retinoic acid to wing buds of chick embryos leads to dose- and position-dependent changes in the pattern of cellular differentiation. Early effects of retinoid treatment on the apical ectodermal ridge coor- dinate pattern changes and morphogenesis. The length of the apical ridge increases when additional digits will form but decreases when digits are lost. These changes in length can be understood in terms of a threshold response to the local retinoid concentration that results in either disappearance or maintenance of the ridge (Lee & Tickle, J. Embryol. exp. Morph. 90, 139-169 (1985)). Here, we have analysed the mechanisms involved in ridge disappearance by locally applying retinoic acid to the apex of stage 20 chick wing buds. With this treat- ment regime, low doses give duplicated digit patterns and higher doses truncations. The height of the apical ridge is progressively reduced with increasing doses of retinoid and the time course of ridge flattening indicates that the height of the ridge is correlated with bud outgrowth. With high doses of retinoic acid, the typical ridge, a pseudostratified epithelium in which the colum- nar cells are tightly packed, disappears and the epi- thelium at the tip of the bud consists of loosely packed cuboidal cells. Shortly after treatment, there is a de- crease in the number of gap junctions between ridge cells. This early change in cell contacts suggests that gap junctions may be involved in maintaining epithelial morphology. When treated epithelium is recombined with untreated mesenchyme, an apical ridge is re- established and distal structures can be generated. In contrast, when treated mesenchyme is recombined with the epithelium from normal buds, only proximal struc- tures are formed. Therefore, retinoids can lead to a reorganization of the apical ectodermal ridge which is mediated and maintained by the mesenchyme.

Retinoic acid was locally applied to presumptive limb regions of chick embryos to find out the earliest time at which the limb pattern can be reprogrammed. When beads soaked in retinoic acid were placed in the appropriate positions in embryos at stage 10 or older, duplicated or reduced leg patterns resulted. To pin point the time at which the cells in the limb rudiment respond to the retinoid, beads were removed at various times and the lengths of exposure required to reprogramme limb development found. The early limb rudiments require longer exposures to give duplications than late rudiments. The effective treatment periods last at least until stage 17 when the limb bud and apical ectodermal ridge develop. In contrast, the length of exposure to reduce the limb is constant at early stages. Retinoids first start acting to produce duplicated structures between stages 10 and 13. Therefore, retinoids appear to begin to reprogramme the cells as soon as they are determined to give rise to a limb.

The mammary gland is a system of hollow interconnecting tubes which develops from an invasive branching cord of epithelial cells. This ultrastructural study of the developing mammary gland focuses on how the lumen forms and establishes the polarized epithelial lining of the gland. The earliest signs of lumen formation are many small cavities and crevices lined with microvilli which appear at scattered sites throughout the branching cords and neck of the gland. It is suggested that these initial small lumina form quite simply by separation of cells whose opposing faces are non-adhesive. The continuous central lumen of the gland develops by fusion and enlargement of the many small lumina. The cells adjacent to the developing lumen will form the polarized epithelial lining of the gland. Excess, more basal, epithelial cells degenerate. The lumen begins to appear when the branching pattern is almost complete. Thus, during morphogenesis, invasion by the mammary gland epithelium involves penetration of the mesenchyme by a solid cord of cells. We suggest that this cellular organization may be a fundamental characteristic of invasive epithelia and that a crucial step in the development of malignant epithelial tumours is a change in cell organization from a polarizedcell sheet to a solid cord of cells which can invade.

It has been suggested that positional information along the proximo-distal axis of the limb-bud is specified by time spent in the progress zone. Mesenchyme cells have been killed by X-irradiation, reducing the rate cells leave the zone. The time spent there by some cells is thus increased. When limbs, stage 18/19, stage 21, or tips of stage 24, are treated with increasing doses of X-irradiation, from 1000 rads to 2500 rads proximal structures are progressively lost, whereas distal ones— the digits—are relatively unaffected. There was no evidence for intercalation of missing parts. These effects are due to killing or damage of mesenchyme cells: the ectoderm is not affected at these doses. The results are consistent with a quantitative analysis based on the progress zone model, in which viable cells repopulate the progress zone and gradually restore it to normal as non-dividing cells are diluted out. It is suggested that any treatment causing damage to the mesenchyme at early stages will give similar results. The mesenchyme cells appear to be surprisingly resistant to radiation damage. The form of the limb-bud is not altered by damaging the mesenchyme. Differences in the development of structures at similar proximo-distal levels, following irradiation, is considered in terms of the requirement of a threshold number of cells.

The interaction of cells from embryonic liver, neural retina and mesonephros with cells from limb-bud mesenchyme has been investigated in vivo by grafting these tissues into the developing chick wing-bud. The implanted cells were in all cases from quail tissue which can be recognized histologically. As embryonic liver and neural tube are tissues that sort externally to limb-bud mesenchyme in mixed aggregates, it would be expected, from a differential adhesiveness hypothesis, that heterotypic adhesions along the borders of graft and host would be favoured over cell-cell adhesions in the graft. No morphological signs of this were evident: rather the grafted cells maximized like-like contacts. The cells of the grafts, including those from control mesenchyme, did not invade into the wing. The results were the same irrespective of whether the graft was a fragment of tissue or a pellet of reaggregated cells. This supports the idea that cells within tissues are not actively moving around and also provides controls for assaying the invasiveness of other cell types, such as malignant cells into the wing.

Cells from normal rat brain tissue did not survive and few cells could be found 1 d after grafting. In contrast, cells from a glioma and a carcinogen-treated rat brain survived well and many mitoses were observed. These malignant cells also invaded the limb. The behaviour of normal and malignant cells was followed at shorter times after grafting and some invasion by the normal cells was detected. The first signs of degeneration of normal cells were apparent around 7 h after grafting, and after this the grafts progressively deteriorated. These results support the ideal that the ability of cells to survive and grow in embryonic tissues is a characteristic of malignant cells. The findings are discussed in relation to mechanisms of tumour formation.

A survey of the behaviour of a variety of normal and malignant tumours and cells has been carried out to gain insights into the mechanisms of tumour invasiveness. The tumours and cells were implanted into the developing chick wing bud, which is a loose mesenchyme bounded by ectoderm. The distribution of the grafted cells was examined histologically after one or two days. The special feature of this assay is that the behaviour of cells is tested in a 3-dimensional tissue. Cells from 3 different carcinomas, mouse lung tumour, rat bladder tumour and human breast tumour did not invade the mesenchyme, whereas trophoblast, sarcoma 180, cultured hamster fibroblasts (BHK, PyBHK, Nil 8, HSV Nil 8) and neuroblastoma cells did. Cells from embryonic pigmented retina and heart ventricle were non-invasive. These results suggest that cell movement may not be a common feature of all invasive tumours. The cells that did move into the mesenchyme appeared to do so by various mechanisms. Lack of contact inhibition of movement, although probably involved in the invasiveness of sarcoma 180 cells, does not appear to be necessary for invasion: cells that have been shown to exhibit contact inhibition of movement (BHK and PyBHK) also invade. Both normal and transformed cells (BHK and PyBHK; Nil 8 and HSV Nil 8) moved into the mesenchyme. Other invading cells, such as trophoblast, neuroblastoma and to a small extent, HSV Nil 8 cells, destroy the adjacent host tissue and this may be important in the invasiveness of these cells. The patterns of invasion and interactions with the host tissue were varied. Trophoblast and the fibroblasts were often elongated along the basement membrane at the ectoderm/mesenchyme border and also closely apposed to the endothelial linings of blood vessels. Sarcoma 180 and neuroblastoma cells clustered around nerves. The embryonic tissues and neuroblastoma cells were often associated with blood vessels. These results are discussed in relation to tumour invasion. A striking finding was that the carcinoma cells were frequently found positioned within the wing ectoderm on the basement membrane. This affinity of carcinoma cells for the epithelium rather than the mesenchyme leads to a reappraisal of the mechanisms involved in the invasiveness of carcinomas.

Three different kinds of carcinoma cell, human breast tumours, mouse lung tumour and rat bladder tumours have been implanted into the developing chick wing to assay their invasive behaviour. We found that the majority of the carcinoma cells did not invade the mesenchyme but were positioned in the ectoderm. We examined the ultrastructure of the interaction of the carcinoma cells with both the ectoderm and the mesenchyme to see if there are any differences. We found no specialized adhesive junctions between carcinoma cells and mesenchyme cells. In contrast, the carcinoma cells seem to be adhesive to the ectoderm cells; desmosomes between the two have been found and also frequent parallel alignment of membranes over long distances. This difference between the interaction of the carcinoma cells with the mesenchyme and the ectoderm may explain why the carcinoma cells can penetrate into the ectoderm but rarely do so into the mesenchyme. The carcinoma cells formed a smooth border with the mesenchyme even in the absence of a basal lamina. These results are discussed in relation to the mechanisms of invasion of epithelial tumours in vivo. In particular, the relative importance of the basal lamina, and cell adhesion within the tumour, in limiting invasiveness arc assessed.

This work presents a quantitative measure, α, of the degree of segregation of two cell types in sections of aggregates, and some results obtained with the measure relating to ‘sorting out’. The method is designed particularly for the case where labelling of one type of cell is incomplete, and the importance of this effect is assessed. Possible problems in formulating such a model are discussed. The measure α is compared with methods used in investigations of segregation in plant communities. Segregation of chick heart and limb-bud cells in mixed aggregates has been analysed using α. In control aggregates of mixtures of labelled and unlabelled cells of one type, α is near to its random value of 1, and we suggest that the departure from random can be adequately accounted for by cell division. In mixed aggregates, significant segregation is consistently found, even in aggregates formed after 2 and 4 h. Both disaggregation procedures (EDTA, trypsin or trypsin+ EDTA) and reaggregation methods (reciprocating or gyratory shaker) are found to have an effect on the degree of segregation. Possible reasons for these findings are discussed. Positioning of the cells relative to the outside of aggregates is also investigated for some of the aggregates.

The polarizing region, a small group of cells at the posterior margin of the limb bud, acts as a signalling region to specify the pattern of structures which develop across the anteroposterior axis of the limb1. When a polarizing region is grafted to the anterior margin of a second limb, a mirror-image symmetrical limb develops1,2. There is indirect evidence that the signal from the polarizing region is a diffusible morphogen3. In an attempt to identify the nature of the putative morphogen, we have developed a method whereby extracts of polarizing region cells and chemicals can be locally applied to the wing bud by being bound to implanted beads and other inert carriers. The chemicals tested included hyaluronidase, dibutyryl cyclic AMP and thalidomide. All experiments have been negative until, at the suggestion of Dr J. Pitts, we tested retinoic acid because of its effects on cell to cell communication and cell differentiation, and we have now for the first tune been able to mimic the action of the polarizing region with a defined chemical.