The autopod: Its formation during limb ... - Wiley Online Library

PDFlib PLOP: PDF Linearization, Optimization, Protection Page inserted by evaluation version www.pdflib.com – [email protected]

Develop. Growth Differ. (2008) 50, S177–S187

doi: 10.1111/j.1440-169X.2008.01020.x

The autopod: Its formation during limb development Koji Tamura,* Sayuri Yonei-Tamura, Tohru Yano, Hitoshi Yokoyama and Hiroyuki Ide Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Aobayama, Aoba-ku, Sendai 980-8578, Japan

The autopod, including the mesopodium and the acropodium, is the most distal part of the tetrapod limb, and developmental mechanisms of autopod formation serve as a model system of pattern formation during development. Cartilage rudiments of the autopod develop after proximal elements have differentiated. The autopod region is marked by a change in the expression of two homeobox genes: future autopod cells are first Hoxa11/Hoxa13-double-positive and then Hoxa13-single-positive. The change in expression of these Hox genes is controlled by upstream mechanisms, including the retinoic acid pathway, and the expression of Hoxa13 is connected to downstream mechanisms, including the autopod-specific cell surface property mediated by molecules, including cadherins and ephrins/Ephs, for cell-to-cell communication and recognition. Comparative analyses of the expression of Hox genes in fish fins and tetrapod limb buds support the notion on the origin of the autopod in vertebrates. This review will focus on the cellular and molecular regulation of the formation of the autopod during development and evolutionary developmental aspects of the origin of the autopod. Key words: autopod, limb, pattern formation.

Introduction The hand and foot are distal components of a limb (forelimb/hindlimb) in tetrapods. A limb can be anatomically divided into three parts: the stylopod (upper arm/thigh), the zeugopod (forearm/shin), and the autopod (hand/foot, including palm/sole, thumb, and fingers/toes) (Fig. 1A). In the limb, the stylopod commonly consists of a long/thick bone named the humerus/femur, and there are two elements of bones in the zeugopod, the radius/tibia and ulna/ fibula. The autopod contains many small and thin elements of bones, including carpals/tarsals (mesopodium), metacarpals/metatarsals, and phalanges (acropodium = digit, each digit consisting of its metacarpal/metatarsal and a chain of phalanges) (Fig. 1B; Wagner & Chiu 2001; Johanson et al. 2007). The acropodium consists of long/thin bones, and the mesopodium is usually a complex of small nodular elements with the exception of those in some anurans, crocodiles, and primates (in which those elements have been transformed into two long bones) (Wagner *Author to whom all correspondence should be addressed. Email: [email protected] Received 25 January 2008; accepted 29 January 2008. © 2008 The Authors Journal compilation © 2008 Japanese Society of Developmental Biologists

& Chiu 2001). Morphology of the autopod is diverged, and every species in tetrapods has its own morphology of the autopod, with different numbers of digits and different types of skin derivatives such as ungues (claw, nail, hoof), pads, setae and webs. In spite of the diversification of its final morphology, the developmental process of autopod formation shares common mechanisms among tetrapod species, and autopod formation serves as an interesting model of pattern formation in vertebrate development.

1. Autopod formation along the proximo-distal axis A limb bud, the developmental primordium of a tetrapod limb, is composed of lateral plate-derived mesenchyme and ectoderm-derived epidermis (a structure called the somatopleure). The somatopleure protrudes from the lateral body, grows into a scooplike cylinder structure, and forms the limb skeleton inside and skin outside. Most of the other components of the limb are derived from the trunk body (muscle and vessels arising from somites, and axons of peripheral nerves invading the limb from the spinal cord and dorsal root ganglia). During limb bud development, the limb skeletal pattern is primarily established as a model (scaffold) of cartilage elements (Figs 2,3B) and the cartilage is then ossified (endochondral bone

S178

K. Tamura et al.

Fig. 1. Skeletal pattern of the autopod (terminology). Schematic representations of the entire limb (A) and the autopod (B) skeleton.

formation). Thus, the basic skeletal morphology in the autopod is determined in the early process of cartilage pattern formation. Cartilage elements in the autopod emerge subsequent to the formation of those in the stylopod and zeugopod (Fig. 3), and this sequence is well conserved among tetrapods. In the developing limb bud, the apical ectodermal ridge (AER) is a critical ectodermal thickness at the apex of the bud, and when the AER structure is removed or disrupted during the process of distal growth of the limb bud, a distal part of the cartilage elements is deleted while a proximal part develops. The earlier the stage of AER removal is, the more proximal are the elements that are defected, suggesting that specification of the pattern occurs in succession in the proximal-to-distal (stylopod-toautopod) direction and that the AER is essential for the distal progression of successive pattern formation (Saunders 1948; Summerbell 1974; see reviews by Martin 1998; Capdevila & Izpisua Belmonte 2001 and references therein). A series of AER-removal experiments has suggested that the developmental stage when the autopod starts to be determined is stage

21–22 (Hamburger & Hamilton 1951) in the chick limb bud (Summerbell & Lewis 1975). Interestingly, according to these experiments, it takes about 12 h for development of the mesopodium region (from stage 21 to stage 24), similar to the time required for zeugopod development (stage 18/19 to 21), and the time required to form each cartilage element therefore does not depend on how large the final bone is. The results of AER-removal experiments suggest the existence of a hypothetical zone in which cells distally change a property that determines which cartilaginous elements the cells will form with respect to the proximo-distal (PD) axis under the influence of the AER (Wolpert 1969; Summerbell & Lewis 1975; Wolpert et al. 1975). This zone, that ranges over a certain distal area, is called “the progress zone”, and the progress zone model hypothesizes that mesenchymal cells in the progress zone are unspecified in terms of the PD axis. As the limb bud grows, the unspecified area within a certain width beneath the AER shifts distally, and proximal cells left behind from the zone are specified and differentiate into first stylopod elements and then zeugopod elements. In this

© 2008 The Authors Journal compilation © 2008 Japanese Society of Developmental Biologists

Autopod formation during development

Fig. 2. Schematic representations of cartilage development in the chick limb bud.

scenario, the autopod is specified at the final stage after the two proximal regions have been specified, in agreement with the proximal-to-distal direction of limb cartilage formation. Based on the progress zone model, it is assumed that an internal clock-like mechanism (measuring time that mesenchymal cells spend in the progress zone) controls which position along the PD axis a cell contributes to (Summerbell et al. 1973; Tickle & Wolpert 2002); however, the molecular nature of this model remains unsolved. Hairy2, a component of the molecular clock underlying the rhythm of somitogenesis, is a candidate of the time-counting molecule in the autopod. Hairy2 expression in the autopod region cycles with a 6 h periodicity at chick

S179

Fig. 3. Spatio-temporal change in Hoxa11/Hoxa13 expression along the proximo-distal axis. (A, C) Sketches showing the expression of Hoxa11/Hoxa13 in the chick limb bud (A) and the zebrafish fin bud (C). (B, D) Limb cartilage pattern of 10 days chick embryo (B) and fin cartilage pattern (radials) of 2 months zebrafish (D), visualized by Alcian blue staining. A, autopod; co, coracoid; fr, fin ray; S, stylopod; Z, zeugopod.

stage 24–26, at which time the acropodium is being specified, and the second phalange takes two periodical periods of hairy2, 12 h, to be formed (Pascoal et al. 2007). These results demonstrate a good correlation between cyclic expression of hairy2 and the chondrogenic differentiation program. Although the significance of the oscillated expression remains unclear, it is possible that cells in the progress zone record time as the number of oscillation of hairy2 expression, providing a temporal value along the PD axis (Pascoal et al. 2007). In addition, it is suggested that there is a regional difference in the molecular program within the autopod. Specification of the position-related value, by which the cells recognize the PD position to contribute to,


S180

K. Tamura et al.

can also be seen in the aspect of cell fate. Results of recent fate mapping studies (Arques et al. 2007; Pearse et al. 2007; Sato et al. 2007) have revealed that at the early stage of limb development there is a large overlap between the prospective autopod and zeugopod in the distal limb bud of mouse and chick embryos. These results suggest that the value for the autopod has not been specified as the cell fate at that time point. A clearer regionality of cell fate specifying the prospective autopod is observed at later stages (later than chick stage 23). Our fate map, however, showed that regional heterogeneity along the PD axis exists even in the distal region, indicating that there are cells in the autopod region that have different values for the PD axis (Sato et al. 2007), as suggested by the oscillated hairy2 expression in the autopod region. The fact that an unspecified and ubiquitous area, which is postulated by the progress zone model, cannot be observed in the distal limb bud suggests that cells are specified and differentiated from proximal to distal sequentially and gradually. The above interpretation for the specification of the autopod region, in which the autopod is specified following the specification of proximal elements, can explain results of classical embryological experiments and agrees with the notion of sequential progression of cartilage formation along the PD axis, but it is not conclusive and is still controversial (see a recent review by Tabin & Wolpert 2007). The reason for this issue remaining unsolved is that the specification progresses gradually, and, therefore, quantitative changes in molecular mechanisms such as gene and protein expression should be further investigated.

2. Molecular mechanisms for specification of the autopod 2-1. Retinoic acid and its molecular cascade How the autopod region is specified in the process of pattern formation along the PD axis is still vague, as described above, but there are some molecular clues for this issue. Retinoic acid is known to show various functions in many developmental systems, regardless of whether the function is endogenous or mimic, and limb development is one of the targets for its function. Artificial deficiency of retinoic acid affects limb growth and patterning in all three limb axes (Stratford et al. 1999); thus, this molecule seems to be essential for normal limb development. Excess administration of retinoic acid, on the other hand, gives rise to an ectopic limb or formation of digits in various species. Retinoic acid was reported to be a morphogen (a substance that functions in pattern formation in a

concentration-dependent manner) for digit identity (Tickle et al. 1982, 1985; Thaller & Eichele 1987), but the idea of direct function of its concentration in the formation of antero-posterior heterogeneity inside the autopod has been challenged (Tamura et al. 1990; Noji et al. 1991; Wanek et al. 1991; Tamura et al. 1993); a different series of studies has suggested its function in autopod formation itself. The function of retinoic acid in autopod formation was originally shown in limb regeneration of amphibians (Niazi & Saxena 1978; Maden 1982; Thoms & Stocum 1984). When a limb in a urodele amphibian is amputated, the animal can regenerate a complete replica of the lost structure. When the autopod is removed at the wrist/ankle level, the autopod is reconstructed, whereas a complete set of PD components, the stylopod, zeugopod and autopod, is regenerated if the limb is amputated at the proximal end of the stylopod. Interestingly, application of a high concentration of retinoic acid concomitantly with the wrist/ankle amputation results in a duplicated regenerate along the PD axis; all three elements are regenerated at the original wrist/ankle site. This odd phenomenon is called “proximalization”. Evaluation of both concentration (Scadding & Maden 1994) and activity (Brockes 1992) of retinoic acid in the normally regenerating limb blastema indicated that distal and proximal blastemas contain low and high levels of activity of retinoic acid, respectively, supporting the idea that retinoic acid endogenously regulates the limb pattern along the PD axis during limb regeneration. According to this idea, the autopod is determined by the lowest concentration of retinoic acid, and administration of a high dose of retinoic acid converted the autopod blastema into the stylopod one, resulting in duplicated stylopod formation at the wrist/ankle level. Similar PD duplication by retinoic acid cannot be seen in the developing limb bud (Tickle & Crawley 1988). However, the combination of local application of retinoic acid and tissue transplantation in the chick limb bud revealed that retinoic acidtreated distal cells acquire the ability to make more proximal structures, including the stylopod, suggesting that retinoic acid proximalizes the distal property of cells into a proximal one (Tamura et al. 1997). Subsequent studies further demonstrated the endogenous function of retinoic acid in formation of the autopod in the developing limb bud. Retinoic acid is an active derivative of retinoids, and it is metabolized from retinol through retinal. The second step of retinoic acid synthesis (oxidation of retinal to retinoic acid) is mediated by cytosolic retinaldehyde dehydrogenases (RALDH). The raldh2 gene, one of the enzyme family members, is expressed in the lateral plate mesoderm at the base of the limb bud field



S181

Fig. 4. Molecular and cellular mechanisms of the proximodistal (PD) axis formation during limb development and regeneration. (A) Expression pattern of molecules in the retinoic acid signaling pathway for the PD axis formation. In this model, proximal synthesis and distal degradation of retinoic acid (RA) produces a gradation of RA activity along the PD axis (left). The graded activity results in the distinct activation of transcription factor genes (right) along the PD axis. (B) Distal displacement of the distal blastema in the regenerating urodele limb. Blastemas derived from different levels of the PD axis always regenerate at the original position. (C) Sorting-out behavior of chick limb mesenchymal cells. Cells taken from distal limb bud at stage 24 were labeled with fluorescent dye and mixed with non-labeled stage 24 cells (left) or stage 22 cells (right). Hetero-mixture gives rise to sorting-out behavior, resulting in patchy cell clusters (right).

(Fig. 4A; Niederreither et al. 1997; Swindell et al. 1999). In the limb field, therefore, retinoic acid is synthesized at the base of the bud, and the proximal region of the limb bud is exposed to a high level of activity of retinoic

acid (Niederreither et al. 2002; Mic et al. 2004; Yashiro et al. 2004). Then retinoic acid is degraded by a P450 enzyme, termed CYP26, and the activity of retinoic acid is balanced by synthesis and degradation of the molecule (see a recent review by McCaffery & Simons 2007). The cyp26b1 gene, one of the CYP26 enzymes, is expressed in the distal limb bud, and the distal region has a low level of activity of retinoic acid (Fig. 4A; MacLean et al. 2001; Yashiro et al. 2004). Targeted disruption of cyp26b1 in the mouse gives rise to a defect of the distal region mainly in the autopod (Yashiro et al. 2004), suggesting that a lower level of activity of retinoic acid balanced by its degradation is critical for autopod formation. The idea that a low concentration of retinoic acid specifies the autopod corresponds to the proximalization phenomena caused by exogenously applied retinoic acid described above. Both raldh2 and cyp26b1 start to be expressed in the limb field at the early stage of limb development, and the gradient activity of retinoic acid along the PD axis is established in the early stage (Niederreither et al. 2002; Mic et al. 2004; Yashiro et al. 2004). In this sense, specification of the autopod is probably initiated a long time before the cell fate for the autopod region is determined and the cartilage elements are differentiated. The activity of retinoic acid is translated into region-specific gene expression along the PD axis. Available gene markers for the highest level of retinoic acid activity are Meis1 and Meis2, which encode homeodomain transcription factors and are expressed at the most proximal domain of the limb bud corresponding to the presumptive stylopod region (Fig. 4A; Capdevila et al. 1999; Mercader et al. 2000). Disruption of retinoic acid synthesis diminishes the expression of Meis genes (Niederreither et al. 2002), and disruption of retinoic acid degradation expands the expression domain of these genes distally (Yashiro et al. 2004). In this context, a gene(s) expression for the presumptive autopod region regulated by the lowest


S182

K. Tamura et al.

concentration of retinoic acid is postulated, and Hoxa13 is a reliable molecule specific for the autopod lineage.

2-2. Hoxa13 as a marker for the future autopod Hoxa13, one of the homeobox genes that are found in a special gene cluster, starts to be expressed in the limb bud at chick stage 22 (Yokouchi et al. 1991; Nelson et al. 1996; Vargesson et al. 2001), in agreement with the beginning of autopod formation (Fig. 2A). The domain of Hoxa13 expression is initially restricted to a very small distal-posterior region and expands proximally and anteriorly as limb outgrowth proceeds. At later stages (chick stage 25 and later), the Hoxa13 domain corresponds with the autopod region, which is visible as cartilage rudiments. The onset, expansion, and final location of Hoxa13 expression strongly suggest that the Hoxa13 domain at all stages marks the future autopod. Hoxd13 is also expressed in the autopod region, although the most anterior digit in the autopod does not express Hoxd13 (Nohno et al. 1991; Yokouchi et al. 1991; Nelson et al. 1996; Vargas & Fallon 2005). It should be noted that Hoxd13 is not expressed only in the autopod region but is also expressed at the posterior half of the more proximal region. Hoxa13 expression in the limb bud starts inside the domain of Hoxa11 expression, and for a while during the process of expansion of the Hoxa13 domain, cells in the distal limb bud are double-positive for Hoxa11/Hoxa13 (Yokouchi et al. 1991; Nelson et al. 1996; Sato et al. 2007). Hoxa11 expression disappears in the most distal region, the overlapping domain of Hoxa11/Hoxa13 becomes narrow, and the autopod cells finally become single-positive for Hoxa13. Double knocking-out Hoxa13/Hoxd13 gives rise to loss of autopod skeletons (Fromental-Ramain et al. 1996), suggesting that these genes are more than just markers for the autopod. Other genedisruption studies in mice have demonstrated that Hoxa11 is not essential for autopod formation (Davis et al. 1995; Davis & Capecchi 1996). However, it is notable that ectopic expression of Hoxa11 that has expanded into the autopod region (produced by overexpression of Meis1) disrupts the normal skeletal pattern in the autopod (Mercader et al. 1999) and that ectopic expression of Hoxa13 in the zeugopod gives rise to an abnormal skeletal pattern therein (Yokouchi et al. 1995). It is therefore thought that, with respect to the expression of Hoxa genes, disappearance of Hoxa11 expression and a Hoxa13-single-positive situation in the autopod region are crucial for normal skeletal formation in the autopod.

Region-specific expression of Hoxa13 is a downstream target of the retinoic acid signaling pathway (Fig. 4A). Administration of excess retinoic acid results in reduction of Hoxa13 expression in the autopod region (Mercader et al. 2000), and also in cyp26b1knockout mouse embryos, in which the level of retinoic acid activity is high in the entire limb bud, the Hoxa13 domain in the distal limb bud is severely reduced (Yashiro et al. 2004). A low level of retinoic acid activity caused by disruption of its synthetic enzyme gene, raldh2, results in elimination of the entire limb skeleton including autopod elements (Niederreither et al. 2002; Mic et al. 2004), although it remains unclear whether the severe limb phenotype is due to defects in PD axis formation. 2-3. Cell surface property The autopod region specified by the region-specific Hoxa13 expression develops cartilage modules of the skeletal pattern, including the mesopodium and acropodium. A characteristic trait of autopod skeletal elements is their smaller and thinner morphologies than those in the zeugopod and stylopod. For phalange elements, periodical cycles of distal growth of cartilage condensation and segmentation result in the formation of a chain of small cartilage elements that are separated by joints (see a review by Casanova & Sanz-Ezquerro 2007). Prechondrogenic condensation of mesenchymal cells formed in the autopod is smaller than that in the proximal region, implying that mesenchymal cells in the future autopod possess a certain cell property in order to make skeletal morphology characteristic of the autopod. Indeed, evidence for the idea that cell surface property is one of the characteristics has emerged from cell/tissue aggregation and transplantation studies. In urodeles, as mentioned above, regenerating limb blastema formed at the proximal stylopod level sequentially regenerates the stylopod, zeugopod and autopod, while the blastema at the wrist level regenerates only the autopod structure. If the wrist blastema is implanted into the thigh blastema, the wrist blastema does not regenerate immediately but is displaced distally (Fig. 4B). When the host regeneration reaches ankle level, the grafted donor blastema begins regenerating the autopod structure (Crawford & Stocum 1988). These facts suggest that distal blastema cells for the autopod have their own cell surface property different from that of proximal ones and that the distal cells can start morphogenesis when the surface property matches the surrounding tissue. This is supported by results obtained by using a tissue culture system (Nardi & Stocum 1983). When two blastemas from the



same level of origin are jointed and positioned together, they fuse and evenly distribute. However, when the wrist blastema is jointed with the upper arm blastema, the wrist blastema is engulfed by the upper arm one, demonstrating that the wrist blastema to form the autopod has greater adhesive property than the proximal one. Different cell adhesive property is thought to cause such a tissue sorting behavior (Steinberg 1970). Developing limb mesenchymal cells show sorting behavior similar to that of regenerating limb blastema cells. In a recombinant limb bud that contains dissociated and reaggregated limb mesenchymal cells with an ectodermal jacket, future autopod cells are redistributed to the distal part of the resultant limb (Wada et al. 1993). In a monolayer culture system of chick limb mesenchymal cells, the future autopod cells are segregated from more proximal cells, and they are eventually sorted out from each other (Fig. 4C; Wada et al. 1993; Ide et al. 1994, 1998; Wada & Ide 1994). Similar results are obtained using Xenopus hindlimb bud cells (Koibuchi & Tochinai 1998). The position-dependent cell sorting behavior is completed within a short period (about 18 h) in culture, suggesting that the sorting is due to different of cell surface properties that the cells initially possess. Further studies have demonstrated that the positiondependent cell sorting is closely correlated to positional identity along the PD axis. Retinoic acidtreated autopod cells do not segregate from the proximal cells in culture, and these treated cells are converted to contribute to a more proximal structure (Tamura et al. 1997). Displacement of the autopod cells into the proximal structure can also be seen in Meis1-overexpressed autopod cells (Mercader et al. 2000, 2005) and in cyp26b1-knockout limb bud (Yashiro et al. 2004). Likewise, autopod cells of Hoxa13-knockout mouse embryos fail to segregate from the proximal cells (Stadler et al. 2001). Positionspecific cell affinity is thought to be a cell property for the formation of autopod-specific morphology. These studies have demonstrated that different limb regions use different amount/kinds of cell surface molecules such as cell adhesion molecules. Further investigations revealed several molecules, including cadherins and Ephs, as candidates responsible for the cell sorting. Cadherin-11 transcripts are expressed in the distal mesenchyme of limb buds (Kimura et al. 1995). In culture, cadherin-11-positive cells are sorted out from the negative cell population, suggesting that cadherin-11 is involved in selective association of mesenchymal cells. N-cadherin transcripts and proteins are distributed abundantly in the distal limb bud, and the amount of N-cadherin in the distal bud

S183

increases as limb development proceeds (Yajima et al. 1999, 2002). As in the case of cadherin-11, N-cadherin-positive cells are segregated from negative cells in culture. Both inhibition of N-cadherin function in the distal region and overexpression of N-cadherin in the proximal cells disturb the normal cell sorting behavior. Interestingly, N-cadherin protein shows a gradient distribution along the PD axis, and hence the gradient of N-cadherin amount might be involved in the identity of each element in the autopod. Ephs belong to a large family of receptor tyrosine kinases regulating cell shape, movements, and attachment in a cell–cell contact-dependent manner. Interaction with their ligands, ephrins, triggers a wide range of cellular responses, including cell recognition, adhesion, boundary formation, and repulsion. Ephrin ligands are categorized into two groups: GPIanchored type (ephrin-A) and membrane-bound type (ephrin-B). Sorting and engulfment are disturbed by PI-PLC, which removes GPI-anchored membranebound proteins (Wada et al. 1998; da Silva et al. 2002), and overexpressed ephrin-A2 indeed neutralizes the affinity of autopod cells (Wada et al. 2003). In normal limb development, ephrin-A2 expression is absent from the autopod region and some of the receptors, EphA4 and EphA7, are expressed in the autopod region, suggesting that these molecules play roles in the autopod-specific cell surface property and position-dependent cell sorting (Stadler et al. 2001; Wada et al. 2003). EphA7 expression is decreased in mesenchymal cells in Hoxa13 mutant homozygous mouse embryos (Stadler et al. 2001), and there is evidence that EphA7 is a direct downstream target of Hoxa13 (Salsi & Zappavigna 2006). Cadherins and Ephs/Ephrins, both families of which are important for cell surface property, have different functions in cell communications, cell–cell adhesion and recognition, and multiple mechanisms to determine autopod-specific cell properties may function as the final step for the autopod formation downstream of the retinoic acid pathway and Hoxa transcription cascade.

3. Origin of the autopod Limbs of all tetrapods, including both extant and extinct species, have autopods. Limbs and fish fins are believed to be homologous appendages, and the limb has evolved from the paired fins of lobe-finned fish (sarcopterygians). In the evolutional process of fin-tolimb transition, the autopod was assembled as a novelty, but this does not necessarily mean that the autopod has been newly added in the earliest tetrapod (extinct amphibians such as Acanthostega and


S184

K. Tamura et al.

Ichthyostega). Rather, some sarcopterygians such as Panderichthys and Tiktaalik were already equipped with autopod-like radial structures as the distal endoskeleton of their fins, although the structures appear incomplete (Shubin et al. 2006; Johanson et al. 2007). Developmental biological information on fin-to-limb transition is very limited because embryos of only a few species of sarcopterygians are available for laboratory use. Hoxd13 expression is one of the available clues. As mentioned, Hoxd13 is expressed in future autopod region. In the tetrapod limb bud, its expression domain in the early phase is located at the posterior half of the bud, corresponding to the posterior zeugopod. The expression expands anteriorly as the limb bud grows, and the late phase of the expression domain includes a large part of future autopod region (Nelson et al. 1996). Hoxd13 expression has been reported in the pectoral fin bud of Australian lungfish, Neoceratodus forsteri, a rare living sarcopterygian fish available as a laboratory animal (Johanson et al. 2007). According to that study, expression of Hoxd13 closely matches the late phase of the expression pattern (anterior-expanded domain) observed in the autopod of a tetrapod, suggesting that fin radials in Neoceratodus and tetrapod digits have common developmental mechanisms different from those for proximal fin/limb elements. A basal actinopterygian (ray-finned fish), the paddlefish (Polyodon spathula), also exhibits a late phase of Hoxd13 expression, a pattern of expression thought to be a developmental feature of the autopod (Davis et al. 2007). It appears that the late phase of Hoxd13 expression is not a tetrapod novelty correlated with the evolution of the complete autopod and that gene regulation of Hoxd13 expression for the autopod is ancient to tetrapods. In the zebrafish, a teleost, pectoral fin buds appear to lack this late phase of HoxD expression (Sordino et al. 1995), suggesting that teleosts have lost the gene regulation mechanism of ancestral HoxD expression that is retained in basal actinopterygians, sarcopterygians and tetrapods (osteichthyans). HoxD expression has been reported also in chondrichthyans (cartilaginous fish), a sister group of osteichthyans, and the results suggest that the mechanism for the late phase of HoxD expression is a primitive condition present the common ancestor of chondrichthyans and osteichthyans (Freitas et al. 2007). Data on Hoxa11/Hoxa13 expression in fin buds are available for some teleost species. In zebrafish, both Hoxa11/Hoxa13 are expressed in the developing pectoral fin bud, and the expression domains are largely overlapped as seen in the early stage of the tetrapod limb bud. The overlapped expression of these genes continues to later stages, and the domains are never separated (Fig. 2C; Sordino et al.

1995, 1996; Neumann et al. 1999; Grandel et al. 2000). Based on the failure of spatio-temporal change in Hoxa11/Hoxa13 expression along the PD axis, the resultant endoskeletal elements in the zebrafish paired fin cannot be classified into any part of the tetrapod limb (radials, Fig. 2D). In this sense, it could be said that the zebrafish pectoral fin does not have any elements corresponding to the autopod. The paddlefish fin also has overlapped expression of Hoxa11/Hoxa13 (Metscher et al. 2005; Davis et al. 2007), indicating that the fin does not establish a PD axis to distinguish the autopod and zeugopod, although Hoxd13 regulation for the autopod may be set up. It would be interesting to know whether the Hoxa11/Hoxa13 separation program was lost in actinopterygians or whether the program was newly evolved in sarcopterygians. Studies on Hoxa11/ Hoxa13 expression in chondrichthyans might provide clues for elucidating this issue.

Final words Hoxa13 expression is a good marker for the autopod, but its expression does not directly mean the autopod. What is important for autopod formation is the sequential change in the Hoxa13-expressing domain during limb development, that is, change from Hoxa11/ Hoxa13-double positive to Hoxa13-single positive, as shown in Figure 2A. Even in the Hoxa13-expressing domain, there is evidence of clear regionality of cell fate and affinity, and the regionality exists gradually along the PD axis. Qualitative difference in key molecules is probably responsible for the gradual regionality, and accumulation of data from fine-tuned studies with newly developed methodology is opening the next window to investigate this complex process. Studies on autopod formation, which focus on traditional developmental biological problems, the most advanced molecular issues and evolutional topics, will continue to provide general ideas on pattern formation during development.

Acknowledgment Our research results referred to in this paper, and all of the authors, are supported by research grants from the Ministry of Education, Science, Sports and Culture of Japan.

Conflict of Interest No conflict of interest has been declared by K. Tamura, S. Yonei-Tamura, T. Yano, H. Yokoyama or H. Ide.



References Arques, C. G., Doohan, R., Sharpe, J. & Torres, M. 2007. Cell tracing reveals a dorsoventral lineage restriction plane in the mouse limb bud mesenchyme. Development 134, 3713– 3722. Brockes, J. P. 1992. Introduction of a retinoid reporter gene into the urodele limb blastema. Proc. Natl Acad. Sci. USA 89, 11 386–11 390. Capdevila, J. & Izpisua Belmonte, J. C. 2001. Patterning mechanisms controlling vertebrate limb development. Ann. Rev. Cell Dev. Biol. 17, 87–132. Capdevila, J., Tsukui, T., Rodriquez Esteban, C., Zappavigna, V. & Izpisua Belmonte, J. C. 1999. Control of vertebrate limb outgrowth by the proximal factor Meis2 and distal antagonism of BMPs by Gremlin. Mol. Cell 4, 839–849. Casanova, J. C. & Sanz-Ezquerro, J. J. 2007. Digit morphogenesis: is the tip different? Dev. Growth Differ. 49, 479– 491. Crawford, K. & Stocum, D. L. 1988. Retinoic acid coordinately proximalizes regenerate pattern and blastema differential affinity in axolotl limbs. Development 102, 687–698. Davis, A. P. & Capecchi, M. R. 1996. A mutational analysis of the 5’ HoxD genes: dissection of genetic interactions during limb development in the mouse. Development 122, 1175– 1185. Davis, A. P., Witte, D. P., Hsieh-Li, H. M., Potter, S. S. & Capecchi, M. R. 1995. Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11. Nature 375, 791–795. Davis, M. C., Dahn, R. D. & Shubin, N. H. 2007. An autopodiallike pattern of Hox expression in the fins of a basal actinopterygian fish. Nature 447, 473–476. Freitas, R., Zhang, G. J. & Cohn, M. J. 2007. Biphasic Hoxd gene expression in shark paired fins reveals an ancient origin of the distal limb domain. PLOS One 2 (8), e754. Fromental-Ramain, C., Warot, X., Messadecq, N., LeMeur, M., Dolle, P. & Chambon, P. 1996. Hoxa-13 and Hoxd-13 play a crucial role in the patterning of the limb autopod. Development 122, 2997–3011. Grandel, H., Draper, B. W. & Schulte-Merker, S. 2000. Dackel acts in the ectoderm of the zebrafish pectoral fin bud to maintain AER signaling. Development 127, 4169–4178. Hamburger, V. & Hamilton, H. L. 1951. A series of normal stage in the development of the chick embryo. J. Morphol. 88, 49–92. Ide, H., Wada, N. & Uchiyama, K. 1994. Sorting out of cells from different parts and stages of the chick limb bud. Dev. Biol. 162, 71–76. Ide, H., Yokoyama, H., Endo, T., Omi, M., Tamura, K. & Wada, N. 1998. Pattern formation in dissociated limb bud mesenchyme in vitro and in vivo. Wound Repair Regen. 6, 398– 402. Johanson, Z., Joss, J., Boisvert, C. A., Ericsson, R., Sutija, M. & Ahlberg, P. E. 2007. Fish fingers: digit homologues in sarcopterygian fish fins. J. Exp. Zoolog. B Mol. Dev. Evol. 308, 757–768. Kimura, Y., Matsunami, H.Inoue, T. et al. 1995. Cadherin-11 expressed in association with mesenchymal morphogenesis in the head, somite, and limb bud of early mouse embryos. Dev. Biol. 169, 347–358. Koibuchi, N. & Tochinai, S. 1998. Existence of gradient in cell adhesiveness along the developing Xenopus hind limb bud, shown by a cellular sorting-out experiment in vitro. Dev. Growth Differ. 40, 355–362.

S185

MacLean, G., Abu-Abed, S., Dolle, P., Tahayato, A., Chambon, P. & Petkovich, M. 2001. Cloning of a novel retinoic-acid metabolizing cytochrome P450, Cyp26B1, and comparative expression analysis with Cyp26A1 during early murine development. Mech. Dev. 107, 195–201. Maden, M. 1982. Vitamin A and pattern formation in the regenerating limb. Nature 295, 672–675. Martin, G. R. 1998. The roles of FGFs in the early development of vertebrate limbs. Genes Dev. 12, 1571–1586. McCaffery, P. & Simons, C. 2007. Prospective teratology of retinoic acid metabolic blocking agents (RAMBAs) and loss of CYP26 activity. Curr. Pharm. Design. 13, 3020– 3037. Mercader, N., Leonardo, E., Piedra, M. E. et al. 2000. Opposing RA and FGF signals control proximodistal vertebrate limb development through regulation of Meis genes. Development 127, 3961–3970. Mercader, N., Leonardo, E., Azpiazu, N. et al. 1999. Conserved regulation of proximodistal limb axis development by Meis1/ Hth. Nature 402, 425–429. Mercader, N., Tanaka, E. M. & Torres, M. 2005. Proximodistal identity during vertebrate limb regeneration is regulated by Meis homeodomain proteins. Development 132, 4131– 4142. Metscher, B. D., Takahashi, K., Crow, K., Amemiya, C., Nonaka, D. F. & Wagner, G. P. 2005. Expression of Hoxa-11 and Hoxa-13 in the pectoral fin of a basal ray-finned fish, Polyodon spathula: implications for the origin of tetrapod limbs. Evol. Dev. 7, 186–195. Mic, F. A., Sirbu, I. O. & Duester, G. 2004. Retinoic acid synthesis controlled by Raldh2 is required early for limb bud initiation and then later as a proximodistal signal during apical ectodermal ridge formation. J. Biol. Chem. 279, 26 698– 26 706. Nardi, J. B. & Stocum, D. L. 1983. Surface properties of regenerating limb cells: evidence for a gradient along the proximodistal axis. Differentiation 27, 27–31. Nelson, C. E., Morgan, B. A., Burke, A. C. et al. 1996. Analysis of Hox gene expression in the chick limb bud. Development 122, 1449–1466. Neumann, C. J., Grandel, H., Gaffield, W., Schulte-Merker, S. & Nüsslein-Volhard, C. 1999. Transient establishment of anteroposterior polarity in the zebrafish pectoral fin bud in the absence of sonic hedgehog activity. Development 126, 4817–4826. Niazi, I. A. & Saxena, S. 1978. Abnormal hind limb regeneration in tadpoles of the toad, Bufo andersoni, exposed to excess vitamin A. Folia Biol. (Krakow) 26, 3–8. Niederreither, K., McCaffery, P., Dräger, U. C., Chambon, P. & Dolle, P. 1997. Restricted expression and retinoic acid-induced downregulation of the retinaldehyde dehydrogenase type 2 (RALDH-2) gene during mouse development. Mech. Dev. 62, 67–78. Niederreither, K., Vermot, J., Schuhbaur, B., Chambon, P. & Dolle, P. 2002. Embryonic retinoic acid synthesis is required for forelimb growth and anteroposterior patterning in the mouse. Development 129, 3563–3574. Nohno, T., Noji, S., Koyama, E. et al. 1991. Involvement of the Chox-4 chicken homeobox genes in determination of anteroposterior axial polarity during limb development. Cell 64, 1197–1205. Noji, S., Nohno, T., Koyama, E. et al. 1991. Retinoic acid induces polarizing activity but is unlikely to be a morphogen in the chick limb bud. Nature 350, 83–86.


S186

K. Tamura et al.

Pascoal, S., Carvalho, C. R., Rodriguez-Leon, J. et al. 2007. A molecular clock operates during chick autopod proximaldistal outgrowth. J. Mol. Biol. 368, 303–309. Pearse, R. V., II, Scherz, P. J., Campbell, J. K. & Tabin, C. J. 2007. A cellular lineage analysis of the chick limb bud. Dev. Biol. 310, 388–400. Salsi, V. & Zappavigna, V. 2006. Hoxd13 and Hoxa13 directly control the expression of the EphA7 Ephrin tyrosine kinase receptor in developing limbs. J. Biol. Chem. 281, 1992–1999. Sato, K., Koizumi, Y., Takahashi, M., Kuroiwa, A. & Tamura, K. 2007. Specification of cell fate along the proximal-distal axis in the developing chick limb bud. Development 134, 1397– 1406. Saunders, J. W. 1948. The proximo–distal sequence of origin of the parts of the chick wing and the role of the ectoderm. J. Exp. Zool. 108, 363–403. Scadding, S. R. & Maden, M. 1994. Retinoic acid gradients during limb regeneration. Dev. Biol. 162, 608–617. Shubin, N. H., Daeschler, E. B. & Jenkins, F. A., Jr 2006. The pectoral fin of Tiktaalik roseae and the origin of the tetrapod limb. Nature 440, 764–771. da Silva, S. M., Gates, P. B. & Brockes, J. P. 2002. The newt ortholog of CD59 is implicated in proximodistal identity during amphibian limb regeneration. Dev. Cell 3, 547– 555. Sordino, P., Duboule, D. & Kondo, T. 1996. Zebrafish Hoxa and Evx-2 genes: cloning, developmental expression and implications for the functional evolution of posterior Hox genes. Mech. Dev. 59, 165–175. Sordino, P., van der Hoeven, F. & Duboule, D. 1995. Hox gene expression in teleost fins and the origin of vertebrate digits. Nature 375, 678–681. Stadler, H. S., Higgins, K. M. & Capecchi, M. R. 2001. Loss of Eph-receptor expression correlates with loss of cell adhesion and chondrogenic capacity in Hoxa13 mutant limbs. Development 128, 4177–4188. Steinberg, M. S. 1970. Does differential adhesion govern self-assembly processes in histogenesis? Equilibrium configurations and the emergence of a hierarchy among populations of embryonic cells. J. Exp. Zool. 173, 395–433. Stratford, T., Logan, C., Zile, M. & Maden, M. 1999. Abnormal anteroposterior and dorsoventral patterning of the limb bud in the absence of retinoids. Mech. Dev. 81, 115–125. 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. 1975. Time, place and positional value in the chick limb-bud. J. Embryol. Exp. Morphol. 33, 621–643. Summerbell, D., Lewis, J. H. & Wolpert, L. 1973. Positional information in chick limb morphogenesis. Nature 244, 492– 496. Swindell, E. C., Thaller, C., Sockanathan, S., Petkovich, M., Jessell, T. M. & Eichele, G. 1999. Complementary domains of retinoic acid production and degradation in the early chick embryo. Dev. Biol. 216, 282–296. Tabin, C. & Wolpert, L. 2007. Rethinking the proximodistal axis of the vertebrate limb in the molecular era. Genes Dev. 21, 1433–1442. Tamura, K., Aoki, Y. & Ide, H. 1993. Induction of polarizing activity by retinoic acid occurs independently of duplicate formation in developing chick limb buds. Dev. Biol. 158, 341–349.

Tamura, K., Ohsugi, K. & Ide, H. 1990. Distribution of retinoids in the chick limb bud: analysis with monoclonal antibody. Dev. Biol. 140, 20–26. Tamura, K., Yokouchi, Y., Kuroiwa, A. & Ide, H. 1997. Retinoic acid changes the proximodistal developmental competence and affinity of distal cells in the developing chick limb bud. Dev. Biol. 188, 224–234. Thaller, C. & Eichele, G. 1987. Identification and spatial distribution of retinoids in the developing chick limb bud. Nature 327, 625–628. Thoms, S. D. & Stocum, D. L. 1984. Retinoic acid-induced pattern duplication in regenerating urodele limbs. Dev. Biol. 103, 319–328. Tickle, C. & Crawley, A. 1988. The effects of local application of retinoids to different positions along the proximo-distal axis of embryonic chick wings. Roux’s Arch. Dev. Biol. 197, 27–36. Tickle, C. & Wolpert, L. 2002. The progress zone – alive or dead? Nat. Cell Biol. 4, E216–E217. Tickle, C., Alberts, B., Wolpert, L. & Lee, J. 1982. Local application of retinoic acid to the limb bond mimics the action of the polarizing region. Nature 296, 564–566. Tickle, C., Lee, J. & Eichele, G. 1985. A quantitative analysis of the effect of all-trans-retinoic acid on the pattern of chick wing development. Dev. Biol. 109, 82–95. Vargas, A. O. & Fallon, J. F. 2005. Birds have dinosaur wings: the molecular evidence. J. Exp. Zoolog. B Mol. Dev. Evol. 304, 86–90. Vargesson, N., Kostakopoulou, K., Drossopoulou, G., Papageorgiou, S. & Tickle, C. 2001. Characterisation of hoxa gene expression in the chick limb bud in response to FGF. Dev. Dyn. 220, 87–90. Wada, N. & Ide, H. 1994. Sorting out of limb bud cells in monolayer culture. Int. J. Dev. Biol. 38, 351–356. Wada, N., Kimura, I., Tanaka, H., Ide, H. & Nohno, T. 1998. Glycosylphosphatidylinositol-anchored cell surface proteins regulate position-specific cell affinity in the limb bud. Dev. Biol. 202, 244–252. Wada, N., Tanaka, H., Ide, H. & Nohno, T. 2003. Ephrin-A2 regulates position-specific cell affinity and is involved in cartilage morphogenesis in the chick limb bud. Dev. Biol. 264, 550–563. Wada, N., Uchiyama, K. & Ide, H. 1993. Cell sorting and chondrogenic aggregate formation in limb bud recombinants and in culture. Dev. Growth Differ. 35, 421– 430. Wagner, G. P. & Chiu, C. H. 2001. The tetrapod limb: a hypothesis on its origin. J. Exp. Zool. 291, 226–240. Wanek, N., Gardiner, D. M., Muneoka, K. & Bryant, S. V. 1991. Conversion by retinoic acid of anterior cells into ZPA cells in the chick wing bud. Nature 350, 81–83. Wolpert, L. 1969. Positional information and the spatial pattern of cellular differentiation. J. Theor. Biol. 25, 1– 47. Wolpert, L., Lewis, J. & Summerbell, D. 1975. Morphogenesis of the vertebrate limb. Ciba Found Symp. 0, 95–130. Yajima, H., Hara, K., Ide, H. & Tamura, K. 2002. Cell adhesiveness and affinity for limb pattern formation. Int. J. Dev. Biol. 46, 897–904. Yajima, H., Yoneitamura, S., Watanabe, N., Tamura, K. & Ide, H. 1999. Role of N-cadherin in the sorting-out of mesenchymal cells and in the positional identity along the proximodistal axis of the chick limb bud. Dev. Dyn. 216, 274–284.



Yashiro, K., Zhao, X., Uehara, M. et al. 2004. Regulation of retinoic acid distribution is required for proximodistal patterning and outgrowth of the developing mouse limb. Dev. Cell 6, 411–422. Yokouchi, Y., Sasaki, H. & Kuroiwa, A. 1991. Homeobox gene expression correlated with the bifurcation process of limb cartilage development. Nature 353, 443–445.

S187

Yokouchi, Y., Nakazato, S., Yamamoto, M., Goto, Y., Kameda, T., Iba, H., & Kuroiwa, A. 1995. Misexpression of Hoxa-13 induces cartilage homeotic transformation and changes cell adhesiveness in chick limb buds. Genes Dev. 9, 2509– 2522.
