Mechanisms of Development 121 (2004) 10191029 www.elsevier.

com/locate/modo

Review

The contribution of chicken embryology to the understanding of vertebrate limb development

C. Tickle*
Division of Cell and Developmental Biology, University of Dundee, Dow Street, Dundee DD1 5EH, UK Received 6 April 2004; received in revised form 25 May 2004; accepted 25 May 2004

Abstract The chicken is an excellent model organism for studying vertebrate limb development, mainly because of the ease of manipulating the developing limb in vivo. Classical chicken embryology has provided fate maps and elucidated the cell cell interactions that specify limb pattern. The rst dened chemical that can mimic one of these interactions was discovered by experiments on developing chick limbs and, over the last 15 years or so, the role of an increasing number of developmentally important genes has been uncovered. The principles that underlie limb development in chickens are applicable to other vertebrates and there are growing links with clinical genetics. The sequence of the chicken genome, together with other recently assembled chicken genomic resources, will present new opportunities for exploiting the ease of manipulating the limb. q 2004 Elsevier Ireland Ltd. All rights reserved.
Keywords: Chicken embryology; Vertebrates; Limb development

1. Outline of chick limb development The rst signs of chick limb development are swellings in the body wall at appropriate positions along the main head to tail (antero-posterior) axis of the embryo (Fig. 1). These swellings soon grow into well-dened buds, consisting of an apparently homogenous population of mesenchyme cells encased in ectoderm. The bud is vascularised from an early stage, but there is an avascular rim underneath the ectoderm, which is not seen in mouse limb buds. The ectoderm rimming the tip of the limb bud is thickened and known as the apical ectodermal ridge. Over a period of about 7 days, the bud elongates and undergoes detailed changes in shape to form a miniature limb-like structure. At the same time, mesenchyme cells within the bud differentiate into the various tissues of the limb-cartilage, muscle, etc.and this occurs progressively, laying down structures at the base of the limb, rst, and ending with those at the tip. The skeleton is laid down as cartilage, later replaced by bone, and the cartilaginous skeleton can be visualised in whole-mounts using Alcian Green staining. Fig. 2 illustrates
* Tel.: 44-1382-345817-5828; fax: 44-1382-345386. E-mail address: c.a.tickle@dundee.ac.uk 0925-4773/$ - see front matter q 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2004.05.015

the skeletons of a wing and a leg from a 9-day-old chick embryo. The plan of skeletal elements in wing and leg generally conforms to the basic vertebrate pattern, with a sequence of structures, from proximal to distal, consisting of shoulder/pelvic girdle elements, humerus/femur, radius and ulna/tibia and bula, and nally digits. There are only three digits in the chick wing, all with very distinct morphologies and numbered, from anterior to posterior, 2, 3 and 4 (this numbering system has been the subject of much debate), while the four digits in the leg, I, II, III, and IV have the more typical tetrapod morphology and each toe has a different number of phalanges. Limb development is initiated when the vertebrate body plan is being mapped out and is nished when precise arrays of organised cells and tissues have been generated. Both limbs and parts of limbssuch as digitsare repeated structures and the major questions about repeated structures concern number, position and identity. What determines the number of digits? What are the mechanisms that distinguish arm from leg? What ensures that a thumb arises at one edge of the hand, a little nger at the other? These are questions to which research on chicken limb development has helped to provide some answers.

1020

C. Tickle / Mechanisms of Development 121 (2004) 10191029

Fig. 1. Chick embryo with well-formed limb buds. Embryo viewed from right side.

2. Fate mapping and classical embryology Before one can begin to investigate the mechanisms involved in limb development, it is useful to know which cells in the early embryo give rise to the limb and what structures originate from particular parts of the bud. Initial fate maps were made by placing carbon marks on to the embryo and showed, for example, that a small region of

cells on either side of the embryo adjacent to the node contained the precursors of wing, interlimb region and leg; Saunders used the same technique to produce fate maps of early limb buds. These studies have been supplemented over the years with fate maps derived from chick/quail chimeras, a technique pioneered by Le Douarin or by the use of lipophilic dyes, such as DiI, to label small populations of cells and then tracking them and their progeny (Fig. 3). Fate maps of chick limb buds have revealed some unexpected features including the fact that all the digits of the limb come from the posterior half of the early limb bud (Bowen et al., 1989; Vargesson et al., 1997), that the ectoderm moves relative to the mesenchyme (Vargesson et al., 1997) and that the apical ectodermal ridge is a selfperpetuating structure (Saunders et al., 1976). One of the most extraordinary ndings to emerge from fate maps of the early embryo made, prior to limb development, by labelling small groups of cells with lipophilic dyes, is that the ectoderm covering the body in the trunk region consists of two lineage restricted compartments, one dorsal (back) and one ventral (front) (Altabef et al., 1997). The signicance of this compartmentalisation is that, in the limb-forming regions, the apical ectodermal ridge that will tip the bud forms at the compartment boundary. The boundary also exists in the interlimb region but is normally cryptic. When the interlimb is presented with a limb-inducing signal, an apical ectodermal ridge forms at this boundary and the ectopic limb bud that develops is thus in line with the normal limbs (Cohn et al., 1995). Fate maps of the trunkproduced

Fig. 2. Whole-mounts of wing and leg from 9-day-old chick embryos stained with Alcian Green to show the skeletal pattern together with line diagrams labelled to show the various skeletal elements. The limb axes are also illustrated.

Fig. 3. Example of a DiI labelling experiment in an early chick embryo. (A) Initial spot of DiI placed on a small region of cells in a 9-stage embryo. (B) Location of the labelled cells in the wing bud at stage 20.

C. Tickle / Mechanisms of Development 121 (2004) 10191029

1021

by DiI labelling and using chick/quail chimeraeshowed, unexpectedly, that precursors of the apical ridge are initially spread over a large area of ectoderm and then later become concentrated at the dorso-ventral boundary of the limb bud (Altabef et al., 1997; Michaud et al., 1997). Fate maps can only reveal what cells will give rise to when left in place, they do not provide information about commitment. This has to be monitored by tissue transplantation and/or isolation. Thus, for example, transplantation of regions of the early chick body wall showed that cells are set aside to form limbs long before any outward signs of limb development are detectable (Hamburger, 1938). Furthermore, microsurgical operations including tissue ablation have revealed that cell cell interactions are necessary for limb bud development and identied signalling regions in

the bud. Ablation of the apical ectodermal ridge was found to lead to inhibition of bud outgrowth and result in limb truncations (Saunders, 1948; Summerbell, 1974; Fig. 4A); the mirror-image digit duplications resulting from transplanting cells from the posterior margin of the wing bud to the anterior margin of a second wing bud led to identication of the zone of polarizing activity (Saunders and Gasseling, 1968; Fig. 4B); rotation of limb bud tips revealed an activity emanating from the posterior part of the limb bud that maintains the apical ectodermal ridge (Zwilling and Hansborough, 1956) and recombining a left wing mesodermal core with a right wing ectodermal jacket from early limb buds resulted in reversed dorso-ventral polarity at the tip (MacCabe et al., 1974). Each of these main cell cell interactions in the developing chick wing are

Fig. 4. Two of the classical experiments carried out on the wing buds of chick embryos(A) experiment that revealed the properties of the apical ectodermal ridge and (B) the polarizing region; (C) summarises the main sets of cell cell interactions in the wing bud.

1022

C. Tickle / Mechanisms of Development 121 (2004) 10191029

thus related to the formation of the pattern of structures along one of the three axesthe reciprocal interaction between the apical ectodermal ridge and underlying mesenchyme with the proximo-distal axis, the interaction between the polarizing region and mesenchyme at the bud tip with the antero-posterior axis and the interaction between the ectoderm covering the sides of the limb bud and mesenchyme at the tip, with the dorso-ventral axis (Fig. 4C). Removal of the apical ridge halts limb bud outgrowth and the laying down of structures along the proximo-distal axis. Subsequent experiments showed that signalling by the apical ridge is permissive rather than instructive; when the apical ridge from an old chick wing bud is grafted in place of the apical ridge of a young wing, the wing develops normally (Rubin and Saunders, 1972), suggesting that pattern specication occurs in the mesenchyme. One of the issues still being debated is how outgrowth is linked to proximo-distal pattern. A long standing model, based on the results of experiments in chick embryos, is that the apical ridge maintains a progress zone at the tip of the limb bud and that the length of time mesenchyme cells spend in the progress zone determines whether they form proximal or distal structures (Summerbell et al., 1973). An alternative model has been put forward recently that suggests that all the structures along the proximo-distal axis are already specied in the early limb bud and that the role of the apical ridge is to promote expansion and differentiation of the different limb elements in a proximo-distal sequence (Dudley et al., 2002). With respect to antero-posterior patterning, similarly, it was experiments on chick limb development which provided evidence for a classical morphogen gradient model for conveying positional information across this axis (Tickle et al., 1975; Wolpert, 1969). Extensive series of experiments showed that polarizing region signalling is dose-dependent and long range, as predicted by the model. Grafting experiments also showed that (1) the polarizing region must be in contact to the apical ridge to produce additional digits, fore-shadowing the later discovery of a positive feedback loop in which apical ridge signalling maintains polarizing signalling (see later) and (2) that transplanting the polarizing region to the anterior margin of a host wing bud leads to an increase in bud width. This latter observation suggested that signalling by the polarizing region must be linked to production of the apical ridge maintenance factor thus ensuring that patterningspecication of digit identityis linked with morphogenesis and digit number. Even before any of the molecules involved in these signalling interactions had been identied, it was already clear from embryological experiments that the basis of these cell cell interactions was conserved in developing limbs of different vertebrates. Thus, for example, it was found that the posterior margin of a mouse limb bud could induce additional digits when implanted at the anterior margin of a chick wing bud (Tickle et al., 1976). In this experiment,

the additional digits produced are chick digits emphasising the fact that although the signal is the same, the interpretation depends on the character of the responding cells.

3. Molecular basis of cell cell interactions in chick limb development The rst dened chemical, that can mimic signalling by the polarizing region, was discovered in chick embryos. The formation of additional digits when a signalling source is placed anteriorly in the chick wing bud provides an excellent assay for polarizing region activity and this led to the discovery that retinoic acid, a vitamin A derivative, has polarizing activity. Initially, small pieces of paper soaked in retinoic acid were used; later, positively charged beads were loaded with the retinoid. About 10 years later, the pattern of expression of Sonic hedgehog, one of the vertebrate relatives of the Drosophila hedgehog gene, which encodes a cell cell signalling molecule was shown to match maps of polarizing region activity in chick limb buds (Riddle et al., 1993; Fig. 5A). Furthermore grafts of cells expressing Shh (Riddle et al., 1993) or beads soaked in Sonic Hedgehog protein placed at the anterior margin of chick limb buds (Yang et al., 1997; Fig. 5B) were shown to induce the formation of additional digits. It was also shown that retinoic acid acts by inducing ectopic expression of Shh (Riddle et al., 1993). There is more recent evidence that retinoic acid signalling is required for Shh expression in normal chick limb development (Stratford et al., 1996) and that it is involved in proximo-distal patterning (see later). Shh is also expressed at the posterior margin of mouse limb buds and this explains why posterior mouse limb tissue produces digit duplications in chick wing buds. This conservation of molecular signals has facilitated the powerful integration of results from various vertebrate modelschick and mouse embryos, and even zebrash. A good example of how such studies on different vertebrate embryos can re-inforce conclusions about the role of a particular gene in limb development comes from recent work in chick, mouse and zebrash on DHand, a gene that encodes a transcription factor which is required for localisation of Shh expression in early n/limb buds (Charite et al., 2000; Fernandez-Teran et al., 2000; Yelon et al., 2000). There is now substantial evidence that Shh signalling is pivotal in controlling both digit number and identity. Furthermore, Shh has the two main attributes of the postulated limb morphogen in that it acts in a dosedependent fashion and that it is diffusible. Nevertheless, this does not exclude the possibility that downstream consequences of Shh signalling might be mediated by second signals and several lines of evidence, from work on chick embryos, suggest that Bone morphogenetic proteins (Bmps), in particular Bmp2, acts downstream of Shh to

C. Tickle / Mechanisms of Development 121 (2004) 10191029

1023

Fig. 5. (A) Double in situ hybridisation of a chick embryo showing transcripts of Shh and Fgf8 in the early limb buds. (B) Example of a wing skeleton following implantation of a bead soaked in Sonic hedgehog protein to the anterior margin of a stage 20 chick wing bud. Digit pattern 432234. (C) Wing skeleton of a talpid 3 mutant. Wing bud was grafted onto a normal embryo to continue development. There are many digits but these all look the same and are fused together.

determine digit identity (Drossopoulou et al., 2000). Whether Bmp2 could be the polarizing region morphogen or not is very controversial and remains to be resolved. Another possibility, suggested from experiments in mice (Lewis et al., 2001), is that some digits require Bmp

signalling, others both Bmps and Shh and others just Shh. The other main downstream consequence of Shh signalling, control of digit number, is clearly mediated by a second separate signal. Work on mouse embryos has shown that the postulated apical ridge maintenance factor is the extracellular Bmp antagonist, gremlin, and expression of gremlin is induced in response to Shh signalling (Zuniga et al., 1999). In turn, signalling by the posterior apical ridge in chick embryos has been shown to maintain Shh expression in the polarizing region and this positive feedback loop between polarizing region and apical ridge ensures continued limb bud outgrowth (Laufer et al., 1994; Niswander et al., 1994). The importance of this feedback loop in mouse limb buds accounts for the limb phenotype in Shh 2 /2 null embryos; distal limb development is very severely impaired and, at best, a very rudimentary anterior digit, tipped by a nail, forms (Kraus et al., 2001). There are several chicken mutants in which Shh signalling has been shown to be abnormal. Two chicken mutants, talpid 2and talpid 3, have an increased number of digits (polydactylous) but all the digits look alike (Fig. 5C). In both cases, it has been shown that although Shh is expressed normally at the posterior margin of the wing bud, there are defects in Shh signalling (Caruccio et al., 1999; Francis-West et al., 1995). This contrasts with mouse polydactlyous mutants in which there is ectopic Shh (or Indian hedgehog) expression. In chicken talpid3 limb buds, some targets of Shh signalling are not expressed while others are ectopically expressed (Lewis et al., 1999). In another chicken mutant, oligozeugodactyly, ozd, which is unlike talpid mutants in that only the limbs are abnormal, posterior structures are missing. The phenotype is reminiscent of the limb phenotype of Shh 2 /2 mouse embryos and, in ozd chicken mutants, Shh expression cannot be detected in the limb buds (Ros et al., 2003). Signalling molecules produced by the apical ridge and dorsal ectoderm have also been identied. Genes encoding Fibroblast Growth Factors (Fgfs) were rst shown to be expressed in the apical ectodermal ridge of mouse limb buds and then experiments on chick embryos showed that a bead soaked in FGF could substitute for the apical ridge and sustain limb bud outgrowth (Fallon et al., 1994; Niswander et al., 1993). One of the most spectacular effects of FGF beads is the triggering of formation of an ectopic limb bud from the interlimb region of chick embryos (Cohn et al., 1995). It has since emerged that a cascade of FGF (and Wnt) signalling is involved in initiation of development of normal limbs (Kawakami et al., 2001). These initial ndings about the potent effects of FGFs in initiating and promoting limb bud outgrowth have been followed up by detailed analyses in mouse embryos. Four different Fgf genes are expressed specically in the apical ectodermal ridge of vertebrate limb buds and considerable effort has been devoted to dissecting out their individual and combined roles in conditional mouse knockouts (Mariani and Martin, 2003; Martin, 1998). When Fgf function is

1024

C. Tickle / Mechanisms of Development 121 (2004) 10191029

deleted in the apical ectodermal ridge from the very earliest stages of mouse limb bud development, no limbs form. Furthermore, functional inactivation of an Fgf gene which is expressed in the mesenchyme also leads to loss of limbs. Recent work has highlighted the importance of Fgf signalling from the tip of the limb bud and opposing retinoic acid signalling from the base in proximo-distal patterning (Mercader et al., 2000). Counteraction between these two signals has also been stressed recently in laying down the main body axis and the spinal cord (Diez del Corral et al., 2003; Moreno and Kintner, 2004; Sockanathan et al., 2003).

4. Molecular responses to patterning signals Among the rst developmentally important genes to be discovered in vertebrates were relatives of the homeotic genes of Drosophila and some of theseXlhbox1 (now known as Hoxc6), Hox4 (Hoxd9-13) and Hox7, Hox8 (Msx1, Msx2)were found to be expressed in developing limbs. Application of retinoic acid to chick wing buds increased the extent of Xlhbox1 expression (Oliver et al., 1990) and retinoic acid, in combination with signalling by the apical ridge, mediated induction of sequential expression of Hoxd genes, thus establishing a second axis (Izpisua-Belmonte et al., 1992); grafts of mouse tissue into chick limb buds showed that expression of Msx1 and Msx2 was position dependent (Davidson et al., 1991). The nested patterns of expression of Hoxd genes, in which transcripts of the more 50 genes are seen more posteriorly and distally, suggested that these could encode position in the developing limb. This hypothesis was tested using, for the rst time, replication competent retroviral vectors to manipulate gene expression in chick embryos (Morgan et al., 1992). When Hoxd11 (then known as Hox4.6) was expressed throughout the limb bud, by infection with an RCAS retrovirus containing an insert coding for Hoxd11, extra digits or other pattern changes were produced (Morgan et al., 1992; Fig. 6). At the time, the results were interpreted as being consistent with the idea that 50 Hox genes encode position across the anteroposterior axis of the limb thus controlling digit identity. Recent data suggest that other interpretations are likely and that the basis for the observed changes in digit pattern is more complicated. Nevertheless this was a landmark in that it established a new method for testing the function of genes, including genes that encode transcription factors, in chicken embryos and this transgenic technology has been widely used. Subsequent painstaking analysis of the roles of Hoxd and Hoxa genes in developing mouse limbs, again involving both single and multiple knockouts, has led to the conclusion that different combinations of Hoxa and Hoxd genes are responsible for patterning the different limb segmentsHoxd13 and a13 the digits, Hoxd11 and a11 the lower arm/shank and Hoxd9 and a9 the upper arm

Fig. 6. Diagrams showing misexpression of a Hox gene in chick embryos and affects on limb pattern. Drawing on the left shows injection of virus in the early embryo and, on the right, drawings illustrate some of the limb phenotypes obtained. After Morgan et al., 1992.

and Hox10 paralogs the upper leg (Wellik and Capecchi, 2003). More recently, members of the Tbx family of genes encoding transcription factors have been demonstrated by three groups (Logan and Tabin, 1999; Rodriguez-Esteban et al., 1999; Takeuchi et al., 1999) to play a role in encoding digit or limb identity by means of this retroviral misexpression in chick embryos, with one group using electroporation to introduce the retroviral vectors into the presumptive limb-forming regions (Takeuchi et al., 1999). Two Tbx genes, Tbx2 and Tbx3, are expressed in stripes in chick limb buds, initially along the posterior margin, and then both anteriorly and posteriorly (Fig. 7C). When these genes are overexpressed in chick leg buds, this leads to a posteriorisation of toe morphology, supporting the idea that posterior expression of Tbx2 and Tbx3 regulates posterior digit identity (Suzuki et al., 2004). Two other Tbx genes, Tbx5 and Tbx4, are expressed in regions of early chick embryos that will give rise to wings and legs, respectively, and this expression is maintained in the limb buds that subsequently develop (Fig. 7A,B). Their identication was exciting because transplantation experiments had previously suggested that cells possessed wingness and legness qualities based on the nding that grafts of cells from the leg placed in the wing, still develop into leg rather than wing structures and vice versa (Saunders and Gasseling, 1959). Retroviral misexpression has provided evidence that Tbx4 and Tbx5 do indeed contribute to limb-type specic development. Thus for example when Tbx5, the gene normally expressed in the wing, is expressed in the leg, the leg that develops is very feathery and there can be skeletal changes too.

C. Tickle / Mechanisms of Development 121 (2004) 10191029

1025

Fig. 7. Expression of Tbx genes in chicken embryos. (A) Tbx5 expressed in wing bud but not leg bud. (B) Tbx4 expressed in leg but not wing. (C) Tbx3 expressed in stripes in both wing buds and leg buds. (D) Sketch of an X-ray of the hands of a human patient with ulnar-mammary syndrome (after Bamshad et al., 1997).

5. Making digits As described above, the results of innumerable experimental manipulations on early chick limb buds show that cell cell interactions in the developing limb bud determine the number of digit primordia and provide them with identity. Nevertheless, it has become clear that the development of the primordia to give the detailed

anatomy of each digit is surprisingly plastic. Operations carried out when the digital primordia have been laid downbut not undergone segmentation to form phalangesand are still joined by intervening soft tissue webs can lead to formation of elongated digits with an additional phalange or shorter digits with a phalanx missing. These operations include grafting interdigital mesenchyme from between two particular toes to another

1026

C. Tickle / Mechanisms of Development 121 (2004) 10191029

interdigital space or implanting beads soaked in Sonic Hedgehog or various inhibitors (Dahn and Fallon, 2000; Sanz-Ezquerro and Tickle, 2003). Both Bmp and Fgf signalling in mesenchyme and apical ectodermal ridge, respectively, have been implicated in these changes in digit anatomy. It has been suggested that a signalling feedback loop, rather like the feedback loop that operates earlier in the limb bud, involving Hedgehogs, Bmps and Bmp antagonists might operate at the tip of each digit (Sanz-Ezquerro and Tickle, 2003) and regulate Fgf signalling. When beads soaked in Shh are added, Fgf signalling is prolonged and an additional phalanx is produceda more posterior toe phenotypewhile when beads soaked in Fgf inhibitors are added, Fgf signalling is halted and the toe lacks a phalanxa more anterior phenotype. In both cases, the tip of the toe forms normally suggesting that there is a special mechanism for nishing off the digit and making the tip and that this comes into play when Fgf signalling ceases (Sanz-Ezquerro and Tickle, 2003). The separation of individual digits is mediated by cell death of the soft tissue webs between the digits (Hurle and Ganan, 1986; Saunders and Fallon, 1966). Although this interdigital mesenchyme normally disappears, it has been shown that it still has the potential to form digits. Thus when the ectoderm in the interdigital region in chick legs is damaged, this can lead to formation of a rudimentary digit with 1 2 phalanges and a tip (Hurle and Ganan, 1986). One class of molecules that has been shown to be involved in regulation of this cell death comprises Bmps (Zou and Niswander, 1996). Interdigital cell death is a particularly well-known example of the programmed cell death that occurs during normal embryonic development; other regions of cell death occur in earlier limb bud stages at both anterior and posterior margins of the limb buds (anterior and posterior necrotic zones) and more centrally (the opaque patch). The posterior necrotic zone overlaps with the polarizing region and experimental manipulations show that cell death acts as a buffering mechanism to regulate the number of cells that express Shh in the limb bud (Sanz-Ezquerro and Tickle, 2000).

acid (Robson et al., 1994) or Sonic Hedgehog expressing cells (Duprez et al., 1999) lead to precise duplications of the muscle pattern in chick wings, with individual wing muscles being recognised not only by shape but also by the patterns of fast and slow bres. The cells that mediate muscle patterning in the limb in response to developmental cues appear to be the muscle connective cells. Chick/quail chimeras have demonstrated that the myogenic cells of the muscles have a separate origin in the somites and migrate into the developing limb bud region (Christ et al., 1977). These myogenic cells do not appear to have intrinsic positional information with respect to the limb muscles. When somites that would normally supply the myogenic cells for the muscles of the leg are grafted opposite the forming wing region, this still results in development of a normal wing musculature. Recent experiments using replication decient retroviral vectors in order to carry out cell lineage analysis also show that myogenic cells in the somites and proximal limb are not committed to form particular muscles (Kardon et al., 2002).

7. Human connections Over the last ve years or so, there has been increasing convergence between developmental biology and clinical medicine. Among the rst connections, with respect to genes involved in limb development, was the discovery that mutations in broblast growth factor receptors are responsible for achondroplasia and several other conditions with digit anomalies such as Apert syndrome (reviewed Wilkie, 2003). Soon after, mutations in Hoxd13 were found to be associated with human synpolydactyly (Muragaki et al., 1996) and mutations in a Gli gene with Grieg cephalopolysyndactyly (Hui and Joyner, 1993). It is now known that Gli genes encode transcription factors that are the effectors of Shh signalling. Tbx genes have also been implicated in Holt-Oram syndrome (Tbx5) and ulnar-mammary syndrome (Tbx3). In ulnar-mammary syndrome, which is associated with haploinsufciency of Tbx3, posterior structures, including digits, are missing (Fig. 7D). Thus the conclusion from the chick overexpression experiments (Suzuki et al., 2004) that Tbx3 plays an important role in development of posterior digits seems consistent with the human phenotype. The list of genes that are known to be responsible for limb defects in human patients is growing (see for example Cohn and Bright, 1999; Wilkie, 2003). In addition to completing this list, a major challenge is to understand how specic mutations in a gene lead to a precise phenotype and chicken embryos are being used for this purpose (Caronia et al., 2003). Experiments on chick embryos have also given insights into limb teratogenesis. The best known example of a chemical teratogen that affects limb development in human embryos is thalidomide, which leads to loss of proximal structures while distal structures are relatively unaffected.

6. Muscles, tendons and other tissues Much of the analysis of patterning mechanisms in chick limb development has concentrated on the skeleton which is readily visualised in whole-mounts, but the limb has a complex pattern of muscles, tendons, ligaments, etc. In addition, the limbs must become supplied with blood vessels and nerves. The development of all these other tissues must be co-ordinated with skeletal development. For muscles, this appears to be achieved using the same positional cues as those which pattern the skeletal elements. Thus, for example, grafts of the polarizing region (Shellswell and Wolpert, 1977) or beads soaked in retinoic

C. Tickle / Mechanisms of Development 121 (2004) 10191029

1027

Fig. 8. Two OPT images of a chick embryo following in situ hybridisation for Fgf8.

Similar phenotypes can be produced in chick embryos by X-irradiating early limb buds and killing mesenchyme cells (Wolpert et al., 1979). The fact that surviving cells give rise to distal structures rather than proximal structures ts well with the progress zone model. This outcome would be predicted because cells will spend longer at the tip of the limb bud in order to replenish the zone. Thalidomide has more recently been shown to affect angiogenesis (DAmato et al., 1994) and defective vascularisation of the limb could lead to mesenchymal cell death.

8. Conclusions and future prospects This aim of this brief review was to give an overall picture of how work on chicken embryos has contributed to understanding of vertebrate limb development rather than a catalogue of embryological manipulations and genes that have been studied. Despite the rapid and accelerating advances, the knowledge of limb development is still fragmentary. Even some of the underlying biological principles are still being disputed and there are yawning gaps in knowledge of the molecular basis of patterning. The elucidation of the chicken genome and the increasing availability of resources such as ESTs (Boardman et al., 2002) should allow new approaches. It should now be possible, for example, to identify all the genes required to make a limb and several sophisticated tools, such as Optical projection tomography (Sharpe et al., 2002); Fig. 8) and MRI (Chudek et al., work in progress), are becoming available to map gene expression in three-dimensions and follow limb development in living embryos, respectively. Gene identication is just the start and unravelling function is much more challenging. Another issue will be dealing with large amounts of data and it seems certain that the current view of limb development which is based on

knowledge of a relatively few genes will require modication. The chicken genome sequence also presents tremendous opportunities for identifying conserved gene regulatory sequences and it may be possible to adapt the methods for enhancer bashing developed for chick neural tube (Timmer et al., 2001) for the limb. Future prospects also include overcoming one of the most serious drawbacks of working in chickens, the difculty of functionally inactivating genes. This has been addressed in the past in the developing limb by the use of the engrailed repressor sequence and/or dominant negative constructs. This seems likely to change with the advent of new RNAi strategies and this, together with the ability to electroporate expression constructs into limb buds, may open up possibilities for high throughput screens of gene function (Brown et al., 2003). Finally, a personal view point; although this review extols the chicken as a model system, one of the strengths of working on limb development in chickens has been relating this to limb development in other vertebrates.

Acknowledgements I thank members of the lab for their help and in particular Andy Bain for Figs. 1,2,7, Eva Tiecke for Figs. 2,3,5,6, Elizabeth Farrell for Fig. 4, and Allyson Clelland for Figs. 5B,8. OPT on the specimen in Fig. 8 was carried out by Richard Buckland and Duncan Davidson. I thank The Royal Society and also the MRC and BBSRC for supporting my work on limb development and chickens.

References
Altabef, M., Clarke, J.D.W., Tickle, C., 1997. Dorso-ventral compartments and origin of apical ectodermal ridge in developing chick limb. Development 124, 45474556.

1028

C. Tickle / Mechanisms of Development 121 (2004) 10191029 bone morphogenetic proteins and sonic hedgehog in talpid (ta3) limb buds: their relationships in the signalling cascade involved in limb patterning. Dev. Dyn. 203, 187197. Hamburger, V., 1938. Morphogenetic and axial self-differentiation of transplanted limb primordia of 2-day chick embryo. J. Exp. Zool. 77, 379 400. Hui, C.C., Joyner, A.L., 1993. A mouse model of greig cephalopolysyndactyly syndrome: the extra-toesJ mutation contains an intragenic deletion of the Gli3 gene. Nat. Genet. 3, 241246. Hurle, J.M., Ganan, Y., 1986. Interdigital tissue chondrogenesis induced by surgical removal of the ectoderm in the embryonic chick leg bud. J. Embryol. Exp. Morphol. 94, 231244. Izpisua-Belmonte, J.C., Brown, J.M., Duboule, D., Tickle, C., 1992. Expression of Hox-4 genes in the chick wing links pattern formation to the epithelialmesenchymal interactions that mediate growth. Eur. Mol. Biol. Org. J. 11, 14511457. Kardon, G., Campbell, J.K., Tabin, C.J., 2002. Local extrinsic signals determine muscle and endothelial cell fate and patterning in the vertebrate limb. Dev. Cell 3, 533 545. Kawakami, Y., Capdevila, J., Buscher, D., Itoh, T., Rodriguez Esteban, C., Izpisua Belmonte, J.C., 2001. WNT signals control FGF-dependent limb initiation and AER induction in the chick embryo. Cell 104, 891 900. Kraus, P., Fraidenraich, D., Loomis, C.A., 2001. Some distal limb structures develop in mice lacking Sonic hedgehog signaling. Mech. Dev. 100, 4558. Laufer, E., Nelson, C.E., Johnson, R.L., Morgan, B.A., Tabin, C., 1994. Sonic hedgehog and Fgf-4 act through a signaling cascade and feedback loop to integrate growth and patterning of the developing limb bud. Cell 79, 9931003. Lewis, K.E., Drossopoulou, G., Paton, I.R., Morrice, D.R., Robertson, K.E., Burt, D.W., et al., 1999. Expression of ptc and gli genes in talpid3 suggests bifurcation in Shh pathway. Development 126, 23972407. Lewis, P.M., Dunn, M.P., McMahon, J.A., Logan, M., Martin, J.F., StJacques, B., McMahon, A.P., 2001. Cholesterol modication of sonic hedgehog is required for long-range signaling activity and effective modulation of signaling by Ptc1. Cell 105, 599 612. Logan, M., Tabin, C.J., 1999. Role of Pitx1 upstream of Tbx4 in specication of hindlimb identity. Science 283, 17361739. MacCabe, J.A., Errick, J., Saunders, J.W. Jr., 1974. Ectodermal control of the dorsoventral axis in the leg bud of the chick embryo. Dev. Biol. 39, 69 82. Mariani, F.V., Martin, G.R., 2003. Deciphering skeletal patterning: clues from the limb. Nature 423, 319325. Martin, G.R., 1998. The roles of FGFs in early development of vertebrate limbs. Genes Dev. 12, 1571 1586. Mercader, N., Leonardo, E., Piedra, M.E., Martinez, A.C., Ros, M.A., Torres, M., 2000. Opposing RA and FGF signals control proximodistal vertebrate limb development through regulation of Meis genes. Development 127, 39613970. Michaud, J.L., Lapointe, F., Le Douarin, N.M., 1997. The dorsoventral polarity of the presumptive limb is determined by signals produced by the somites and by the lateral somatopleure. Development 124, 14531463. Moreno, T.A., Kintner, C., 2004. Regulation of segmental patterning by retinoic acid signaling during Xenopus somitogenesis. Dev. Cell 6, 205 218. Morgan, B.A., Izpisua-Belmonte, J.C., Duboule, D., Tabin, C.J., 1992. Targeted misexpression of Hox-4.6 in the avian limb bud causes apparent homeotic transformations. Nature 358, 236 239. Muragaki, Y., Mundlos, S., Upton, J., Olsen, B.R., 1996. Altered growth and branching patterns in synpolydactyly caused by mutations in HOXD13. Science 272, 548551. Niswander, L., Tickle, C., Vogel, A., Booth, I., Martin, G.R., 1993. FGF-4 replaces the apical ectodermal ridge and directs outgrowth and patterning of the limb. Cell 75, 579 587.

Bamshad, M., Lin, R.C., Law, D.J., Watkins, W.S., Krakowiak, P.A., Moore, M.E., et al., 1997. Mutations in human TBX3 alter limb, apocrine, and genital development in ulnar-mammary syndrome. Nat. Genet. 16, 311316. Boardman, P.E., Sanz-Ezquerro, J., Overton, I.M., Burt, D.W., Bosch, E., Fong, W.T., et al., 2002. A comprehensive collection of chicken cDNAs. Curr. Biol. 12, 19651969. Bowen, J., Hinchliffe, J.R., Horder, T.J., Reeve, A.M., 1989. The fate map of the chick forelimb-bud and its bearing on hypothesized developmental control mechanisms. Anat. Embryol. 179, 269 283. Brown, W.R., Hubbard, S.J., Tickle, C., Wilson, S.A., 2003. The chicken as a model for large-scale analysis of vertebrate gene function. Nat. Rev. Genet. 4, 87 98. Caronia, G., Goodman, F.R., McKeown, C.M., Scambler, P.J., Zappavigna, V., 2003. An I47L substitution in the HOXD13 homeodomain causes a novel human limb malformation by producing a selective loss of function. Development 130, 17011712. Caruccio, N.C., Martinez-Lopez, A., Harris, M., Dvorak, L., Bitgood, J., Simandl, B.K., Fallon, J.F., 1999. Constitutive activation of sonic hedgehog signaling in the chicken mutant talpid(2): Shh-independent outgrowth and polarizing activity. Dev. Biol. 212, 137149. Charite, J., McFadden, D.G., Olson, E.N., 2000. The bHLH transcription factor dHAND controls Sonic hedgehog expression and establishment of the zone of polarizing activity during limb development. Development 127, 24612470. Christ, B., Jacob, H.J., Jacob, M., 1977. Experimental analysis of the origin of the wing musculature in avian embryos. Anat. Embryol. (Berl) 150, 171 186. Cohn, M.J., Bright, P.E., 1999. Molecular control of vertebrate limb development, evolution and congenital malformations. Cell Tissue Res. 296, 317. Cohn, M.J., Izpisua-Belmonte, J.C., Abud, H., Heath, J.K., Tickle, C., 1995. Fibroblast growth factors induce additional limb development from the ank of chick embryos. Cell 80, 739 746. Dahn, R.D., Fallon, J.F., 2000. Interdigital regulation of digit identity and homeotic transformation by modulated BMP signaling. Science 289, 438 441. DAmato, R.J., Loughnan, M.S., Flynn, E., Folkman, J., 1994. Thalidomide is an inhibitor of angiogenesis. Proc. Natl Acad. Sci. USA 91, 40824085. Davidson, D.R., Crawley, A., Hill, R.E., Tickle, C., 1991. Positiondependent expression of two related homeobox genes in developing vertebrate limbs. Nature 352, 429431. Diez del Corral, R., Olivera-Martinez, I., Goriely, A., Gale, E., Maden, M., Storey, K., 2003. Opposing FGF and retinoid pathways control ventral neural pattern, neuronal differentiation, and segmentation during body axis extension. Neuron 40, 6579. Drossopoulou, G., Lewis, K.E., Sanz-Ezquerro, J.J., Nikbakht, N., McMahon, A.P., Hofmann, C., Tickle, C., 2000. A model for anteroposterior patterning of the vertebrate limb based on sequential long- and short-range Shh signalling and Bmp signalling. Development 127, 13371348. Dudley, A.T., Ros, M.A., Tabin, C.J., 2002. A re-examination of proximodistal patterning during vertebrate limb development. Nature 418, 539544. Duprez, D., Lapointe, F., Edom-Vovard, F., Kostakopoulou, K., Robson, L., 1999. Sonic hedgehog (SHH) species muscle pattern at tissue and cellular chick level in the chick limb bud. Mech. Dev. 82, 151163. Fallon, J.F., Lopez, A., Ros, M.A., Savage, M.P., Olwin, B.B., Simandl, B.K., 1994. FGF-2: apical ectodermal ridge growth signal for chick limb development. Science 264, 104107. Fernandez-Teran, M., Piedra, M.E., Kathiriya, I.S., Srivastava, D., Rodriguez-Rey, J.C., Ros, M.A., 2000. Role of dHAND in the anteriorposterior polarization of the limb bud: implications for the Sonic hedgehog pathway. Development 127, 21332142. Francis-West, P.H., Robertson, K.E., Ede, D.A., Rodriguez, C., IzpisuaBelmonte, J.C., Houston, B., et al., 1995. Expression of genes encoding

C. Tickle / Mechanisms of Development 121 (2004) 10191029 Niswander, L., Jeffrey, S., Martin, G., Tickle, C., 1994. A positive feedback loop coordinates growth and patterning in the limb. Nature 371, 609612. Oliver, G., De Robertis, E.M., Wolpert, L., Tickle, C., 1990. Expression of a homeobox gene in the chick wing bud following application of retinoic acid and grafts of polarizing region tissue. Eur. Mol. Biol. Org. J. 9, 30933099. Riddle, R.D., Johnson, R.L., Laufer, E., Tabin, C., 1993. Sonic hedgehog mediates the polarizing activity of the ZPA. Cell 75, 14011416. Robson, L.G., Kara, T., Crawley, A., Tickle, C., 1994. Tissue and cellular patterning of the musculature in chick wings. Development 120, 12651276. Rodriguez-Esteban, C., Tsukui, T., Yonei, S., Magallon, J., Tamura, K., Izpisua Belmonte, J.C., 1999. The T-box genes Tbx4 and Tbx5 regulate limb outgrowth and identity. Nature 398, 814818. Ros, M.A., Dahn, R.D., Fernandez-Teran, M., Rashka, K., Caruccio, N.C., Hasso, S.M., et al., 2003. The chick oligozeugodactyly (ozd) mutant lacks sonic hedgehog function in the limb. Development 130, 527537. Rubin, L.R., Saunders, J.W., 1972. Ectodermal mesodermal interactions in the growth of limb buds in the chick embryo: constancy and temporal limits of the ectodermal induction. Dev. Biol. 28, 94112. Sanz-Ezquerro, J.J., Tickle, C., 2000. Autoregulation of Shh expression and Shh induction of cell death suggest a mechanism for modulating polarising activity during chick limb development. Development 127, 48114823. Sanz-Ezquerro, J.J., Tickle, C., 2003. Fgf signaling controls the number of phalanges and tip formation in developing digits. Curr. Biol. 13, 18301836. Saunders, J.W., 1948. The proximo-distal sequence of origin of parts of the chick wing and the role of the ectoderm. J. Exp. Zool. 108, 363404. Saunders, J., Fallon, J., 1966. Cell death in morphogenesis. In: Locke, M., (Ed.), Major Problems of Developmental Biology, Academic Press, New York, pp. 289314. Saunders, J.W., Gasseling, M.T., 1959. The differentiation of prospective thigh mesoderm grafted beneath the apical ectodermal ridge of the wing bud in the chick embryo. Dev. Biol. 1, 281 301. Saunders, J.W., Gasseling, M.T., 1968. Ectodermalmesenchymal interactions in the origin of limb symmetry. In: Fleischmeyer, R., Billingham, R.E. (Eds.), Epithelial Mesenchymal Interactions, Williams & Wilkins, Baltimore, pp. 7897. Saunders, J.W. Jr., Gasseling, M.T., Errick, J.E., 1976. Inductive activity and enduring cellular constitution of a supernumerary apical ectodermal ridge grafted to the limb bud of the chick embryo. Dev. Biol. 50, 1625. Sharpe, J., Ahlgren, U., Perry, P., Hill, B., Ross, A., Hecksher-Sorensen, J., et al., 2002. Optical projection tomography as a tool for 3D microscopy and gene expression studies. Science 296, 541 545. Shellswell, G., Wolpert, L., 1977. The pattern of muscle and tendon development in the chick wing. In: Ede, D.A., et al. (Eds.), Vertebrate Limb and Somite Morphogenesis, Cambridge University Press, Cambridge, pp. 7186.

1029

Sockanathan, S., Perlmann, T., Jessell, T.M., 2003. Retinoid receptor signaling in postmitotic motor neurons regulates rostrocaudal positional identity and axonal projection pattern. Neuron 40, 97 111. Stratford, T., Horton, C., Maden, M., 1996. Retinoic acid is required for the initiation of outgrowth in the chick limb bud. Curr. Biol. 6, 11241133. Summerbell, D., 1974. A quantitative analysis of the effect of excision of the AER from the chick limb-bud. J. Embryol. Exp. Morphol. 32, 651 660. Summerbell, D., Lewis, J.H., Wolpert, L., 1973. Positional information in chick limb morphogenesis. Nature 244, 492 496. Suzuki, T., Takeuchi, J., Koshiba-Takeuchi, K., Ogura, T., 2004. Tbx genes specify posterior digit identity through Shh and BMP signaling. Dev. Cell 6, 4353. Takeuchi, J.K., Koshiba-Takeuchi, K., Matsumoto, K., Vogel-Hopker, A., Naitoh-Matsuo, M., Ogura, K., et al., 1999. Tbx5 and Tbx4 genes determine the wing/leg identity of limb buds. Nature 398, 810814. Tickle, C., Summerbell, D., Wolpert, L., 1975. Positional signalling and specication of digits in chick limb morphogenesis. Nature 254, 199 202. Tickle, C., Shellswell, G., Crawley, A., Wolpert, L., 1976. Positional signalling by mouse limb polarising region in the chick wing bud. Nature 259, 396 397. Timmer, J., Johnson, J., Niswander, L., 2001. The use of in ovo electroporation for the rapid analysis of neural-specic murine enhancers. Genesis 29, 123 132. Vargesson, N., Clarke, J.D.W., Vincent, K., Coles, C., Wolpert, L., Tickle, C., 1997. Cell fate in the chick limb bud and relationship to gene expression. Development 124, 1909 1918. Wellik, D.M., Capecchi, M.R., 2003. Hox10 and Hox11 genes are required to globally pattern the mammalian skeleton. Science 301, 363 367. Wilkie, A.O., 2003. Why study human limb malformations? J. Anat. 202, 27 35. Wolpert, L., 1969. Positional information and the spatial pattern of cellular differentiation. J. Theor. Biol. 25, 1 47. Wolpert, L., Tickle, C., Sampford, M., 1979. The effect of cell killing by xirradiation on pattern formation in the chick limb. J. Embryol. Exp. Morphol. 50, 175 193. Yang, Y., Drossopoulou, G., Chuang, P.T., Duprez, D., Marti, E., Bumcrot, D., et al., 1997. Relationship between dose, distance and time in Sonic Hedgehog-mediated regulation of anteroposterior polarity in the chick limb. Development 124, 43934404. Yelon, D., Ticho, B., Halpern, M.E., Ruvinsky, I., Ho, R.K., Silver, L.M., Stainier, D.Y., 2000. The bHLH transcription factor hand2 plays parallel roles in zebrash heart and pectoral n development. Development 127, 25732582. Zou, H., Niswander, L., 1996. Requirement for BMP signaling in interdigital apoptosis and scale formation. Science 272, 738741. Zuniga, A., Haramis, A.P., McMahon, A.P., Zeller, R., 1999. Signal relay by BMP antagonism controls the SHH/FGF4 feedback loop in vertebrate limb buds. Nature 401, 598 602. Zwilling, E., Hansborough, L., 1956. Interaction between limb bud ectoderm and mesoderm in the chick embryo. III Experiments with polydactylous limbs. J. Exp. Zool. 132, 219239.