SEARCH

SEARCH BY CITATION

Keywords:

  • chondrogenesis;
  • limb development;
  • pattern formation;
  • Wnt-5a

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Members of the Wnt family are known to play diverse roles in the organogenesis of vertebrates. The full-coding sequences of chicken Wnt-5a were identified and the role it plays in limb development was examined by comparing its expression pattern with that of two other Wnt members, Wnt-4 and Wnt-11, and by misexpressing it with a retrovirus vector in the limb bud. Wnt-5a expression is detected in the limb-forming region at stage 14, and in the apical ectodermal ridge and distal mesenchyme of the limb bud. The signal was graded along the proximal–distal axis at stages 20–28 and also along the anterior–posterior axis during early stages. It disappeared in the cartilage-forming region after stage 26, and was restricted to the region surrounding the phalanges at stage 34. Wnt-4 and Wnt-11, other members of the Wnt-5a-subclass, were expressed with a distinct spatiotemporal pattern during the later phase. Wnt-4 was expressed in the articular structure and Wnt-11 was expressed in the dorsal and ventral mesenchyme adjacent to the ectoderm. Wnt-5a expression was partially reduced after apical ectodermal ridge removal, whereas Wnt-11 expression was down-regulated by dorsal ectoderm removal. Therefore, expression of these Wnt was differentially regulated by the ectodermal signal. Misexpression of Wnt-5a in the limb bud with the retrovirus resulted in truncation of long bones predominantly in the zeugopod because of retarded chondrogenic differentiation. Distal elements, such as the phalanges and metacarpals, were not significantly reduced in size. These results suggest that Wnt-5a is involved in pattern formation along the proximal–distal axis by regulation of chondrogenic differentiation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Secretory proteins encoded by the Drosophila segment polarity genes play important roles in segmental patterning and polarity determination in Drosophila (Nüsslein-Volhard & Wieschaus 1980), and their vertebrate homologs are also known to play profound roles in embryonic dorsal–ventral patterning, neural tube patterning and limb formation. Although one of the segment polarity genes, wingless, is involved in pattern formation in the Drosophila appendage (Zecca et al. 1996), the role of the Wnt family in vertebrate limb development is less evident. Several Wnt family members are expressed in the developing limb with specific spatial and temporal patterns (Gavin et al. 1990; Dealy et al. 1993; Parr et al. 1993; Christiansen et al. 1995; Tanda et al. 1995b; Kengaku et al. 1998). In the mouse, Wnt-3, Wnt-4, Wnt-6 and Wnt-7b are expressed in the entire ectoderm, and Wnt-7a is expressed in the dorsal ectoderm. Wnt-5a is expressed in the ventral ectoderm and distal limb bud including the apical ectodermal ridge (AER), and Wnt-11 and Wnt-12 are also expressed in the limb mesenchyme. In the chicken, Wnt-3a, Wnt-5a, Wnt-7a and Wnt-11 are known to be expressed in the limb bud with specific spatiotemporal patterns. However, the expression patterns are not completely identical to those in the mouse. The role of Wnt-7a in dorsal–ventral patterning and the role of Wnt-3a in AER formation have been most extensively studied. Wnt-7a is known to be a signal from the dorsal ectoderm that dominantly specifies muscle and tendon patterns in the limb (Parr & McMahon 1995; Riddle et al. 1995), and Wnt-3a is a signal that regulates AER formation in the limb bud (Kengaku et al. 1998). Wnt-7a expressed in the dorsal ectoderm has a positive influence on Sonic hedgehog expression in the posterior limb mesenchyme and Fgf-4 expression in the posterior AER (Parr & McMahon 1995; Yang & Niswander 1995). However, the roles of other Wnt genes expressed in the developing limb remain elusive.

The Wnt family has been implicated in embryogenesis and organogenesis and play essential roles in animal development (Cadigan & Nusse 1997). Although Frzb encoding an endogenous secretory Wnt antagonist is expressed in the developing cartilage (Hoang et al. 1996), no Wnt member is known to be involved in cartilage differentiation. Ectopic expression of Wnt-1 in the limb bud of the chick and the mouse resulted in truncation of skeletal elements (Zakany & Duboule 1993; Rudnicki & Brown 1997), although Wnt-1 on its own is poorly expressed in the developing limb. Wnt-7a is expressed in the developing limb, and ectopic expression of Wnt-7a inhibits chondrogenesis in vitro (Rudnicki & Brown 1997). However, Wnt-7a transcript is restricted to the dorsal ectoderm (Dealy et al. 1993), and Wnt protein is easily trapped by extracellular matrix (Bradley & Brown 1990), Wnt-7a is unlikely to be an endogenous regulator of chondrogenesis. Therefore, other Wnt members expressed in the limb bud are thought to be involved in cartilage differentiation during limb development.

The Wnt family is currently divided into three major subgroups based on transforming activity in mouse mammary epithelial cells (Wong et al. 1994). Wnt-1, Wnt-3 and Wnt-3a have higher transformation activity; Wnt-2, Wnt-5b and Wnt-7b have lower activity; and Wnt-4, Wnt-5a, and Wnt-6 have no transforming activity. Differential activities are also recognized during early embryogenesis. Wnt-1 and Wnt-8 can induce a complete secondary axis in the Xenopus embryo by micro-injecting mRNA into the ventral half, whereas Wnt-5a, Wnt-4 and Wnt-11 are unable to induce a secondary structure, resulting in shortening of the body axis (Moon et al. 1993; Ungar et al. 1995). Therefore, Wnt-4, Wnt-5a and Wnt-11 constitute a subgroup, called the Wnt-5a-subclass (Du et al. 1995).

In the present study, we investigated the expression patterns of three Wnt-5a-subclass members, Wnt-4, Wnt-5a and Wnt-11, in the chick limb bud. We examined the role of Wnt-5a in cartilage differentiation by expressing it ectopically in vivo. Wnt-5a has the ability to reduce the size of the zeugopodal long bone, suggesting that it is involved in proximal–distal bone patterning through regulation of chondrogenic differentiation.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Chicken Wnt-5a

Total RNA was isolated from four-day chick embryos. Reverse transcription of the total RNA and polymerase chain reaction amplification of the cDNA with Wnt specific primers were carried out as described previously (Tanda et al. 1995a). The amplified cDNA was cloned into pBluescript SK(+) and identified by nucleotide sequencing. One of the cDNA clones encoding Wnt-5a was used for hybridization screening of the chick embryo cDNA library constructed in lambda gt10 to obtain full-length cDNA. Probe preparation and filter hybridization were carried out as described previously (Nohno et al. 1992). The cDNA fragments from the recombinant phages were subcloned into pGEM7Zf(+), and sequenced with an ABI DNA Sequencer (Perkin-Elmer, Pomona, CA, USA).

In situ hybridization

Chick embryos were staged according to Hamburger and Hamilton (1951). Whole-mount in situ hybridization was carried out as described previously (Wilkinson 1992) using the following chicken cDNA. An antisense cRNA probe for Msx1 was synthesized as reported previously (Nohno et al. 1992). The 1.8-kb cDNA encoding full-length chicken Wnt-5a was used as a template. Chicken Wnt-4 and Wnt-11 cDNA was obtained as described previously (Yoshioka et al. 1994; Tanda et al. 1995a,b). Corresponding sense probes were used to estimate signal specificity and showed no significant hybridization signal under the same conditions as the antisense probes. Hybridized embryos were sectioned and examined for the presence of expression signals as described (Tanda et al. 1995b).

Removal of the apical ectodermal ridge and the dorsal ectoderm

To determine whether expression of Wnt-5a and Wnt-11 was regulated by ectodermal signals, the AER was removed from the right wing bud of stage 20 embryos with a sharpened tungsten needle. The embryos were incubated for 6 or 12 h after surgical manipulation and processed for in situ hybridization. The dorsal ectoderm was removed from a stage 20–21 wing bud without disturbing the underlying mesenchyme as described previously (Yang & Niswander 1995). The embryos were incubated for 36 h after the operation and processed for in situ hybridization.

Ectopic expression of Wnt-5a in the wing bud

Virus-free chicken embryonic fibroblasts (CEF) were transfected with Wnt-5a-expressing RCAS (Kawakami et al. 1996). The cells were grown for three days in KAv-1 medium (Kuwana et al. 1996) or in α-minimum essential medium (MEM) supplemented with 5% chicken serum and 2% fetal bovine serum. Then cells were scraped and grafted to the right wing bud of a virus-free embryo at stages 18–21 after cutting the cell pellet into small pieces. The grafted site was in the mid-distal to distal region along the proximal–distal axis, and in the central region along the anterior–posterior axis, where the cell pellet occupied about one-third to one-half of the length along the anterior–posterior axis. To observe phenotypic changes after misexpressing the Wnt-5a gene in the limb bud, the embryos were incubated at 38°C for 4–9 days, fixed in 10% (v/v) formalin, and stained with 0.1% Alcian green to visualize the bone pattern. Thin sections of unstained wing were prepared after paraffin embedding, and stained with hematoxylin-eosin to observe changes in the differentiation of the limb skeleton and joint tissues after Wnt-5a misexpression.

To determine the efficiency of transgene expression, RCAS-BPAP was used (Fekete & Cepko 1993). The embryos were fixed 7 days after implanting the CEF transfected with RCAS-BPAP and stained for alkaline phosphatase activity. As a negative control experiment, we grafted CEF without transfection to the right wing bud of a wild-type embryo to evaluate whether surgical manipulation on the wing bud affects limb morphology. After implantation, the embryos were fixed and stained with Alcian green.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Expression pattern of the Wnt family in the limb bud

We obtained a 1.8-kb cDNA from the chick embryo cDNA library by hybridization screening with a partial Wnt-5a fragment. As the entire chicken Wnt-5a sequence has not been reported, we first determined the entire nucleotide sequence. The deduced amino acid sequence indicates a total of 385 amino acids with 24 conserved cysteine residues observed in the Wnt family (Fig. 1). The chicken sequence is extensively similar to the mouse (87%), human (87%) and Xenopus (83%) Wnt-5a sequences, indicating that the cDNA encodes chicken Wnt-5a.

image

Figure 1. . Comparison of the deduced amino acid sequence of Wnt-5a in chicken (Gg), mouse (Mm), human (Hs) and Xenopus (Xl). Identical residues are indicated by dots, and gaps, indicated by dashes, are introduced to improve sequence alignment. Conserved cysteine residues are asterisked. The nucleotide sequence of chicken Wnt-5a has been deposited in the DDBJ/EMBL/GenBank databases with accession number AB006014.

Download figure to PowerPoint

Using the cDNA as a hybridization probe, we observed the expression pattern during chick embryogenesis throughout stages 13–34. Although the expression pattern of chick Wnt-5a at stages 19–32 has already been reported, the onset of the expression is not known. By examining the expression pattern of Wnt-5a in detail at stages 13–18, we detected a Wnt-5a signal in the prospective wing-forming region at stage 14 (Fig. 2A). Although the limb bud was not yet formed at this stage, Wnt-5a was confined to the limb-forming region, and the signal gradually intensified in the wing-forming region (Fig. 2B). When the limb bud was formed, Wnt-5a expression in mesenchyme was moderate in the distal posterior region, and weak in the proximal anterior region at stage 19 (Dealy et al. 1993). At stage 21 Wnt-5a was intensely expressed in the distal mesenchyme with posteriorly intensified signals as well as in the entire AER (Fig. 2C). These intense signals were also detected in the visceral arches, somitic mesoderm, front nasal process and telencephalon in a stage 20 embryo, as reported previously in mouse and chick embryos (Gavin et al. 1990; Dealy et al. 1993; Parr et al. 1993). The chick Wnt-5a signal faded out in the cartilage-forming region (Fig. 2D; Dealy et al. 1993), and was restricted to the region surrounding the phalanges at stage 34 (Fig. 2E–G). These expression patterns at later stages of limb development suggest involvement of Wnt-5a in cartilage formation and differentiation in the autopod region.

image

Figure 2. . Expression patterns of the Wnt family during limb development. In panels A,C,D,E,H,I,M, the anterior is to the top and the distal is right. Panels F,N,O show paraffin sections with 8 μm thickness, and panels G,J,K are vibratome sections with 50 μm thickness. (A) Dorsal view of a stage 14 embryo hybridized with a Wnt-5a probe. Wnt-5a is expressed in the wing-forming region (arrow). (B) Lateral view of a Wnt-5a-hybridized stage 16 embryo shows an intensified signal in the wing-forming region. (C) Dorsal view of a stage 21 wing bud shows Wnt-5a expression in the distal mesenchyme with a gradient along the anterior–posterior axis and in the apical ectodermal ridge (AER). (D) Dorsal view of a Wnt-5a-hybridized stage 26 wing bud shows that Wnt-5a expression has faded out in the cartilage-forming region (arrowheads), although it is expressed in the distal mesenchyme. (E) Dorsal view of a stage 34 wing shows Wnt-5a expression surrounding the phalanges. (F) Sagittal section through the autopod region of Wnt-5a-hybridized stage 34 wing shows the signal in the mesenchyme (arrowheads). (G) Sagittal section of Wnt-5a-hybridized stage 34 autopod shows the presence of the Wnt-5a signal surrounding the phalangeal bone. (H) Dorsal view of Wnt-4- hybridized stage 27 wing. Wnt-4 transcripts are detected in the elbow region at stage 27. (I) Dorsal view of Wnt-4-hybridized stage 30 wing. The Wnt-4 signals are detectable in the autopod region and presumptive joint region of the wrist in addition to the elbow region. (J,K) Sagittal sections of Wnt-4-hybridized stage 30 wing show Wnt-4 expression in the distal end of the humerus (J) and proximal end of the radius (K). Arrowheads indicate Wnt-4 signals, whereas ectodermal staining is an artifact from entrapped dye during fixation. (L) Anterior view of a stage 27 wing hybridized with a Wnt-11 probe. Wnt-11 is expressed in both the dorsal mesenchyme and the ventral mesenchyme. (M) Dorsal view of Wnt-11 expression in a stage 30 wing. Wnt-11 is weakly expressed in the zeugopod region and autopod region. (N) Longitudinal section of a stage 30 wing shows localization of the Wnt-11 signals both in the dorsal and ventral mesenchyme adjacent to the epidermis of the autopod region. (O) Cross- section of a Wnt-11-hybridized stage 30 leg autopod shows the signals in the mesenchyme anterior to digit II and the mesenchyme posterior to digit IV, as well as the dorsal mesenchyme. d, dorsal; v, ventral; pr, proximal; di, distal; h, humerus; r, radius; u, ulna; 2, digit II; 3, digit III, 4, digit IV.

Download figure to PowerPoint

We also observed expression patterns of the other members of the Wnt-5a-subclass, Wnt-4 and Wnt-11, in the limb bud. Previously, we reported Wnt-4 expression in the central nervous system but not in the limb bud, because Wnt-4 is not expressed in the limb bud until stage 24 (Yoshioka et al. 1994). In the present study, we examined the expression pattern during limb development at stage 24 and later to determine whether Wnt-4 is involved in limb cartilage formation. Wnt-4 was expressed in the central elbow region at stage 27 and afterwards (Fig. 2H). At stage 30, Wnt-4 expression was detectable in the autopod region as small spot-like signals and in the elbow and wrist- forming regions (Fig. 2I). Sagittal sections revealed that signals in the elbow region were associated with the margin of the humerus and radius, indicating localization of the Wnt-4 signals in the joint-forming region (Fig. 2J,K). Wnt-11 has already been proven to be expressed in the dorsal mesenchyme of the limb bud at stages 24–28 (Tanda et al. 1995b). From a detailed examination of Wnt-11 expression, we found that the signal was also detectable in the ventral mesenchyme at stage 27 (Fig. 2L). At stage 30, Wnt-11 expression became weak but still detectable in the autopod region surrounding the digits (Fig. 2M) as observed in the mouse (Christiansen et al. 1995). The Wnt-11 signals in the autopod region are localized in the dorsal and ventral mesenchyme (Fig. 2N). Wnt-11 signals in the autopod were also detected in the region adjacent to the digits (Fig. 2O). The different expression patterns of the Wnt family observed here suggest that the Wnt family plays distinct roles at late stages of limb development.

Wnt expression after ectoderm removal

The expression of Wnt-5a in the distal mesenchyme at stages 18–26 suggests an influence for AER on Wnt-5a expression. We examined this possibility by removing the AER from wing buds. As a result, expression of Wnt-5a in the distal mesenchyme was significantly reduced at 12 h after ridge removal, whereas expression in the posterior mesenchyme still remained (Fig. 3A). On the other hand, Msx1 expression, which is known to depend on AER, was completely reduced within 6 h after AER removal (Fig. 3B).

image

Figure 3. . Expression of Wnt-5a and Wnt-11 after removal of ectoderm. (A,B) Dorsal views show the effect of AER removal on Wnt-5a and Msx1 expression. The apical ectoderm between the arrowheads was removed from the right wing bud, and Wnt-5a expression 12 h later (A) and Msx1 expression 6 h later (B) were monitored. Wnt-5a expression in the distal mesenchyme was partially, but not completely, reduced compared with the non-treated left wing bud serving as a control (A). Expression of Msx1 was rapidly and completely decreased to the basal level within 6 h after AER removal (B). (C,D) Dorsal views show the effect of dorsal ectoderm removal on Wnt-5a (C) and Wnt-11 (D) expression monitored 36 h later. Wnt-5a expression was not affected by dorsal ectoderm removal (C). The Wnt-11 signals decreased to the basal level in the treated wing bud, while Wnt-11 expression remained in the contralateral left wing bud (D). The treated right wing bud is shown by an arrow, and the contralateral wing bud is placed in the reversed anterior–posterior orientation.

Download figure to PowerPoint

Removal of the dorsal ectoderm, which eventually eliminates the Wnt-7a signal, had no obvious effect on the expression of Wnt-5a (Fig. 3C). However, Wnt-11 expression in the dorsal mesenchyme was almost completely reduced to the basal level (Fig. 3D). Wnt-11 expression in the ventral mesenchyme was also reduced (data not shown).

Effect of misexpression of Wnt-5a

To clarify the role of the Wnt family in limb bone formation in vivo, we ectopically expressed Wnt-5a in the limb buds. Embryos grafted with RCAS-Wnt-5a virus-producing fibroblasts at stages 18–21 were allowed to develop until stages 32–36. The morphology of a treated right wing bud was compared with that of a non-treated left wing bud that served as a control. In the treated right wing, the length of the radius and ulna was profoundly reduced, whereas the length of the metacarpal was rarely affected (Fig. 4). The length of the humerus was slightly reduced. When zeugopods were extensively shortened to less than 50% of the control in the severely affected embryos, the hypertrophic zone characteristic of the long bone was absent in the diaphyseal region usually observed at stage 28 and later (Fig. 4B). The length of the radius and ulna was reduced to 50–79% of that of the control in mildly affected embryos (Fig. 4C). Formation of articular cartilage and joint tissue was also affected by Wnt-5a misexpression, and fusion between the humerus and radius or ulna was occasionally observed (Fig. 4D). As a control experiment, we grafted CEF producing the RCAS-BPAP virus that bears the human placental alkaline phosphatase gene as a marker to the right wing bud. The treated right wing could be stained by alkaline phosphatase activity, whereas the contralateral left wing could not be stained (Fig. 4E,F). Among the 10 treated wings, the stylopod region was stained in seven wings, the zeugopod region was stained in all wings, and the autopod region was stained in six wings. Therefore, the transgene was efficiently expressed in treated wings, although the expression was not even throughout the entire wing bud.

image

Figure 4. . Phenotypic changes in the wing bone after Wnt-5a misexpression. (A) The skeletal pattern of the non-treated left wing, which is reversed photographically for comparison, served as a control. (B) Severe phenotype at stage 36 after Wnt-5a misexpression at stage 18. The radius and ulna were shortened significantly (to about 30% of the control), whereas the humerus was shortened slightly (to about 90% of the control). The metacarpals and phalanges were not reduced in size. (C) Mild phenotype after Wnt-5a misexpression. The radius and ulna were shortened to about 70% and 60% of the control, respectively, but the humerus was shortened to only about 90% of the control. Central bending of the humerus was not observed in the treated wing. (D) The articular structure did not form properly, and the long bones were fused (arrow) in the Wnt-5a-misexpressed wing. In this sample, the radius was shortened to about 65% of the control and fused to the humerus. The ulna was also shortened slightly, to about 80% of the control. (E,F) As a control experiment, chicken embryonic fibroblasts (CEF) transfected with RCAS bearing the alkaline phosphatase gene were grafted. The whole right wing could be stained for alkaline phosphatase activity (F), whereas the contralateral left wing, which is reversed for comparison, could not be stained 7 days after implantation (E). h, humerus; r, radius; u, ulna; 2, digit II; 3, digit III; 4, digit IV.

Download figure to PowerPoint

Table 1 summarizes the results of misexpression of Wnt-5a in the wing bud from seven independent experiments. As a negative control, we grafted CEF without transfection to the wing bud of wild-type embryos. We can exclude the possibility that truncation of the radius and ulna is produced by a surgical artifact. As another control, we examined the phenotype of RCAS-BPAP-treated wings used for monitoring transgene expression, as described earlier. In the 10 treated wings, only one wing showed mild reduction (50–79% of the non-treated contralateral wing) in zeugopod and autopod length. Thus, the frequency was equivalent to that of non-transfected CEF grafting. Therefore, infection and spread of the RCAS virus on its own had a subtle effect on the wing length. However, size reduction of the humerus may result from manipulation, as the humerus was shortened in the control experiments with similar frequency. A few embryos exhibited humerus bifurcation both in the Wnt-5a misexpression and the control grafting experiment, suggesting that humerus bifurcation may be artifactual possibly through interference in chondrogenic condensation by grafted cells. Nevertheless, the results summarized in Table 1 suggest that Wnt-5a overexpression preferentially affects zeugopod length. Among a total of 41 Wnt-5a-misexpressed embryos, the radius and ulna became shortened to less than 80% in about one-third of the embryos, respectively, whereas the humerus and metacarpal of digit III became shortened with much lower frequencies. Furthermore, severe shortening of the long bone to less than 50% of the contralateral wing was observed in six and seven embryos for the radius and ulna, respectively, but not for the humerus and metacarpal.

Table 1.  . The number of embryos with long bones of indicated percentage after Wnt-5a-misexpression and control manipulation Thumbnail image of

Histological analysis

Longitudinal sections of the wings severely affected by Wnt-5a at stages 36, 38 and 39 revealed retardation of chondrogenic differentiation in the limb skeleton (Fig. 5). In the control ulna, there are three distinct zones along the proximal–distal axis characteristic of the long bone (Fig. 5A). Between the epiphyseal and diaphyseal regions, a quiescent zone with round-shaped cells (Fig. 5B), a proliferative zone composed of growing, flat-shaped chondrocytes (Fig. 5C), and a hypertrophic zone composed of hypertrophic chondrocytes are noted (Fig. 5D). The growing chondrocytes are arranged vertically in the same orientation. The long bone is surrounded by diaphyseal periosteum, and calcification starts at the middle of the ulna at this stage. Although the quiescent, proliferative and hypertrophic zones were present in the Wnt-5a-treated wing, these zones were rather equivocal compared with those in the control (Fig. 5E). In the Wnt-5a-misexpressed wing (Fig. 5F–H), chondrocytes were orbitally arranged surrounding hypertrophic chondrocytes. The periosteum layer was poorly organized with low cell density (Fig. 5H). The number of chondrogenic cells in the hypertrophic zone that had left the proliferative zone was significantly decreased as compared with the control limb. At stages 38 and 39, disintegration of hypertrophic cells proceeded from the diaphyseal to epiphyseal region in the ulna (Fig. 5I,K). By misexpressing Wnt-5a, the number of hypertrophic cells in the ulna at stage 38 was decreased to the number of hypertrophic cells equivalent to the normal ulna at stages 32 and 33 (Fig. 5J). Hypertrophic chondrocytes in the diaphyseal region were poorly disintegrated after Wnt-5a misexpression. The diaphyseal region of the Wnt-5a-misexpressed ulna at stage 39 was histologically identical to the normal ulna at stages 35–36 (Fig. 5D,L). These results suggest that chondrocyte proliferation and/or differentiation was significantly retarded by Wnt-5a treatment, although mesenchymal condensation and chondrocyte differentiation actually proceeded as observed in the normal limb bud at much earlier stages.

image

Figure 5. . Histologic changes in Wnt-5a-affected limb bud. Limb buds with a severe phenotype at stages 36 (A–H), 38 (I,J) and 39 (K,L) were fixed in 10% formalin, embedded in paraffin, sectioned, and stained with hematoxylin-eosin to compare the histological differences. Histological views of the control (A) and Wnt-5a-treated ulna (E) at low magnification were compared. Panels B–D and F–H are high magnification views of the control and Wnt-5a-treated ulna, respectively. In the normal ulna at stage 36 (A), a quiescent zone (B), proliferative zone (C) and hypertrophic zone (D) characteristic of the long bone were observed. However, in the Wnt-5a-affected ulna (E), these three layers were formed (F–H) but not organized completely. The formation of periosteum was incomplete (H). In the normal ulna at stages 38 (I) and 39 (K), the vein invaded (arrowhead) and hypertrophic cells were gradually calcified in the diaphyseal zone. In the Wnt-5a-affected ulna, calcification of hypertrophic cells in the diaphyseal zone was not detected at stage 38 (J), although the quiescent zone, proliferative zone and hypertrophic zone were formed. In the Wnt-5a-affected ulna at stage 39 (L), calcification of hypertrophic cells and invasion of the vein had just started, which is identical to the normal view of the ulna at stage 36 (D). Bars, 100 μm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Wnt family exhibits distinct expression pattern in the limb bud

Among the Wnt family members, Wnt-3a, Wnt-5a, Wnt-7a, and Wnt-11 have distinct expression patterns in the developing chick limb. To clarify the roles of Wnt members that belong to the same subgroup, we first compared the expression patterns of Wnt-5a-class members, Wnt-4, Wnt-5a and Wnt-11.

Previous work on the expression of chick Wnt-5a at stages 19–32 suggests that Wnt-5a may be involved in proximal–distal patterning of the limb bud (Dealy et al. 1993). However, the onset of expression was unknown, and expression at stage 32 and later had not been previously examined. Our detailed expression analysis at the early stages revealed Wnt-5a expression in the prospective wing-forming region as early as stage 14. At this stage, the wing-field is already established (Altabel et al. 1997; Michaud et al. 1997), and wing bud outgrowth begins at stage 16. The Wnt-5a signal gradually intensifies in the limb-forming region (Fig. 2B). After the limb bud is formed, Wnt-5a is expressed in the AER and distal mesenchyme with a gradient along both the proximal–distal axis and the anterior–posterior axis (Fig. 2C and Dealy et al. 1993). This expression pattern indicates that Wnt-5a is involved in the initiation and early patterning of the limb bud. At later stages, absence of the Wnt-5a signal in the proximal cartilage-forming region (Fig. 2D and Dealy et al. 1993) and intense expression in the distal mesenchymal region surrounding the cartilage (Fig. 2E,F) suggest possible involvement of Wnt-5a in proliferation and differentiation of chondrocytes in the limb bud.

To address the functional role of Wnt-5a, we compared the expression pattern of Wnt-5a with the other Wnt family members of this subclass, Wnt-4 and Wnt-11. Previously Wnt-4 expression has not been observed in the chick limb bud until stage 24 (Yoshioka et al. 1994). In the present study, we found Wnt-4 expression in the developing limb at stage 27 and later. This expression pattern suggests involvement of Wnt-4 in joint and articular structure formation. We also found Wnt-11 expression in the ventral mesenchyme, in addition to the dorsal mesenchyme in the developing limb. The distinct expression patterns of these Wnt members of the same subclass suggest that these Wnt family members play distinct roles during limb development.

Wnt family members are also expressed in the mouse limb bud. Wnt-3, Wnt-4, Wnt-6, Wnt-5a, Wnt-7a, Wnt-7b, Wnt-11 and Wnt-12 are expressed in the limb bud of 9.5–11.5 days post-coitum (dpc) mice. This is equivalent to the stage 19–21 chicken limb (Parr et al. 1993; Christiansen et al. 1995). Wnt-10a and Wnt-10b are also expressed at later stages in the mouse limb (Wang & Shacklefold 1996). The expression pattern of Wnt-5a in the mouse limb is similar but not identical to that of the chick Wnt-5a presented here. Mouse Wnt-5a is initially expressed in the ventral ectoderm and distal mesenchyme (Gavin et al. 1990), whereas chicken Wnt-5a showed graded expression along the anterior– posterior axis at early stages. Wnt-4 expression is only detectable at stage 27 and later in the chick limb bud, whereas uniform expression of Wnt-4 in the entire limb ectoderm has been observed in the mouse embryo (Parr et al. 1993). The Wnt-11 expression pattern also differs between the chicken and the mouse. Wnt-11 is expressed in the mouse limb bud from the onset of limb formation (Christiansen et al. 1995; Kispert et al. 1996), whereas chicken Wnt-11 signals are only detectable at stage 24 and later. Expression of Wnt-11 in the limb bud at later stages is equally observed in the chicken and the mouse. Differential expression patterns of several Wnt genes in mouse and chicken embryos are also observed in other organs. Wnt-4, for example, is expressed in a different region of the diencephalon in the chicken and the mouse (Yoshioka et al. 1994). Interspecies differences in the expression pattern between cognate Wnt genes suggest the presence of additional Wnt members that have identical functions. Several Wnt members seem to play different roles in organogenesis depending on the species, and therefore a cognate Wnt shows a different expression pattern.

Regulation of Wnt-5a expression by the ectodermal signal

Expression of Wnt-5a in the posterior and distal mesenchyme seems to be differentially regulated, because Wnt-5a expression in the posterior mesenchyme does not depend on the AER, and expression in the distal mesenchyme partially depends on it. The AER- independent expression in the posterior mesenchyme is consistent with the expression pattern of Wnt-5a in limbless mutant limb buds (Ros et al. 1996). Limb buds arise normally and are morphologically normal until stage 19 in the limbless mutant, although the apical ridge is not formed and the entire limb mesoderm undergoes cell death (Fallon et al. 1983). Graded expression in the posterior mesenchyme in normal embryos was observed at early stages, suggesting that Wnt-5a is involved in anterior–posterior patterning in the early limb bud. However, Wnt-5a seems to be unnecessary for anterior–posterior patterning at later stages, as the expression of Wnt-5a in the posterior mesenchyme is maintained in the limbless mutant limb bud, where Sonic hedgehog in the posterior mesenchyme is not detected (Ros et al. 1996). In addition, misexpression of Wnt-5a did not change the skeletal pattern along the anterior–posterior axis (Fig. 4). The expression of Wnt-5a in the distal mesenchyme is also differentially regulated by signals from the AER. However, another ectodermal signaling center, the dorsal ectoderm, which expresses Wnt-7a, had no effect on Wnt-5a expression (Fig. 3C). The signals that determine AER- independent expression of Wnt-5a are unclear at present, although many signaling molecules are produced in the limb bud (Tickle & Eichele 1994). Wnt-11 expression, on the contrary, requires signals from the dorsal ectoderm, presumably Wnt-7a. Thus, expression of Wnt-5a and Wnt-11 is differentially regulated by the ectodermal signals.

Limb bone patterning through the Wnt family

Although, based on its expression pattern, Wnt-5a is involved in cartilage development, its physiological role is less evident during limb development. In the present study, we examined the role of Wnt-5a by misexpressing it in vivo. We ectopically expressed Wnt-5a by implanting CEF expressing RCAS-Wnt-5a. As the Wnt-5a gene is normally expressed in the distal mesenchyme of the limb bud, implantation of CEF producing RCAS-Wnt-5a to the distal limb bud had a subtle effect on the bone pattern and size of the metacarpals and phalanges. Severe size reduction was consistently observed in the radius and ulna, and less frequently in the humerus, accompanying loss of the diaphyseal hypertrophic zone. Overexpression of Wnt-5a in the distal mesenchyme seems to have different effects on cartilage differentiation depending on the site, because the distal autopod elements were rarely affected by Wnt-5a misexpression. Otherwise, the effect in vivo may vary according to the duration of ectopic expression. In the autopod region, endogenous Wnt-5a expression continues until stage 30, while in the zeugopod region, its expression continues until stage 24. Another possibility is that the zeugopod may be most sensitive to the grafting experiments, as the zeugopod region was more efficiently infected than the stylopod and autopod region after RCAS-BPAP infection. Surgical damage in the progress zone is known to cause truncation of the stylopod and zeugopod (Summerbell 1977), raising the possibility that implantation of CEF alone in the limb bud might cause truncation of the zeugopod. However, control grafting experiments excluded this possibility (Table 1). Furthermore, wing length was not changed significantly in the RCAS-BPAP-treated wing, and therefore the reduction of the cartilage length is unlikely to have resulted from virus infection itself. There are a few examples of joint fusion in Wnt-5a-misexpressed limbs, probably resulting from Wnt-4-like activity, which will be discussed later. Histologic views of the ulna at stages 36, 38 and 39 suggest inhibition of chondrocyte proliferation and/or retardation of chondrocyte differentiation after Wnt-5a misexpression. The overall process from mesenchymal condensation to chondrocyte differentiation eventually leading to endochondral calcification is therefore delayed but actually processed by Wnt-5a-misexpression even though chondrogenesis is inhibited. Long bone shortening might occur through improper proliferation and differentiation of chondrocytes. Therefore, Wnt-5a may be involved in proximal– distal bone patterning by regulating the temporal process of chondrogenic differentiation.

Misexpression of Wnt-1 causes truncation of limb cartilage in the mouse and chicken (Zakany & Duboule 1993; Rudnicki & Brown 1997), although Wnt-1 is poorly expressed in the limb bud. There are similarities and differences between the present results using Wnt-5a and previous experiments using Wnt-1 (Zakany & Duboule 1993; Rudnicki & Brown 1997). The present results show that misexpression of Wnt-5a results in preferential shortening of the zeugopodal long bone. On the other hand, Rudnicki and Brown (1997) injected RCAS-Wnt-1 at stage 17 and observed truncation of long bone especially in the zeugopod and autopod. Zakany and Duboule (1993) misexpressed Wnt-1 using Hoxa-5 promoter in the mouse limb. The transgene is expressed only in the distal peripheral region and preferentially affected the autopod. The zeugopod and autopod are affected by ectopically expressed Wnt-1, where the duration of the misexpression was sufficient to result in shortening of the long bones. As Wnt-5a is expressed in the autopod region until stage 30, there was not enough time for misexpression in the autopod. Misexpressed Wnt-5a would affect the zeugo-pod by acting for a longer period than in the autopod. From this point of view, misexpression of Wnt-1 and Wnt-5a might have a different effect but exhibited a similar phenotype on the long bone. However, lower frequency on the stylopod phenotype by Wnt-5a-misexpression and Wnt-1-misexpression cannot exclude the possibility that the short period of misexpression on undifferentiated mesenchyme of the stylopod region resulted in the insufficient effect and frequency. When Wnt-7a and Wnt-1 are expressed in micromass culture, they can inhibit chondrogenesis at early chondroblast maturation stage (Rudnicki & Brown 1997). Addition of limb ectoderm on micromass cultures inhibits chondrogenesis in the underlying mesenchymal cells (Solursh et al. 1981). These reports suggested that Wnt-7a, expressed in the dorsal ectoderm, may be an endogenous cartilage-inhibitory factor. Lmx-1 expression in the limb mesenchyme is induced by Wnt-7a, and Lmx-1 expression can be a marker for cells receiving Wnt-7a signaling (Riddle et al. 1995). Lmx-1-expressing cells are located in a region adjacent to the dorsal ectoderm. In addition, hyperchondroplasia was not observed in Wnt-7a knockout mice (Parr & McMahon 1995). Therefore Wnt-7a is unlikely to be an endogenous factor that directly regulates chondrogenesis. Wnt-1 and Wnt-7a may mimic an unknown Wnt member and result in shortening of long bones.

An endogenous secretory Wnt antagonist, Frzb, which can bind to Wnt proteins and eventually inhibit Wnt activity, is expressed in chondroblasts in the appendicular cartilage (Hoang et al. 1996; Moon et al. 1997). Cartilage extracts containing Frzb have in vivo chondrogenic activity (Hoang et al. 1996). The possibility is therefore suggested that an unknown Wnt member exists that could be expressed in chondrocytes and inhibit chondrogenesis. Frzb is known to bind the Wnt-1-subclass members, Wnt-1 and Xwnt-8 (Leyns et al. 1997; Wang et al. 1997). From these reports, we think misexpressed Wnt-1-subclass members, Wnt-1 and Wnt-7a (Rudnicki & Brown 1997), might bind Frzb and inhibit the Frzb activity that would repress the unknown Wnt. As it is not known whether Frzb can bind Wnt-5a, we cannot judge whether Frzb is involved in the Wnt-5a-phenotype.

Based upon the expression pattern and distinct signaling pathway (Slusarski et al. 1997), Wnt-5a is a candidate among known Wnt members as a regulator of cartilage proliferation and differentiation, although it is not expressed in developing chondrocytes.

Roles of Wnt-4 and Wnt-11 during limb development

The expression pattern of Wnt-4 suggests that it may be involved in joint cartilage formation in the limb bud. In the mouse, GDF-5, a member of the bone morphogenetic protein family, is expressed in the joint-forming region (Storm & Kingsley 1996). Brachypodism mutant mice having a mutation in the GDF-5 gene have joint defects as revealed by fusion of joint elements in the limb bud. The joint-fusion occasionally found in Wnt-5a-misexpressed embryos may be the result of Wnt-4-like activity.

The met gene encoding the hepatocyte growth factor receptor is expressed in the dermatomyotome of the somite and in the limb bud of the mouse (Yang et al. 1996) with a pattern similar to that of chick Wnt-11. The met expression in the 10.5 dpc mouse limb is restricted to the dorsal and ventral sides and later to the digit-forming area, and it has been proposed that the met gene is required for myoblast migration in the limb bud. Pax-3 activates met expression and is required for muscle patterning in the limb bud. Pax-3 is expressed in the dorsal and ventral mesenchyme adjacent to the ectoderm (Yang et al. 1996; Amthor et al. 1998) as well as Wnt-11. Patterning of the musculature in the chick wing (Robson et al. 1994) in association with the expression pattern of Wnt-11 has raised the possibility that Wnt-11 participates in myoblast migration and patterning in the limb bud.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank C. Komaguchi and R. Nakao for their technical assistance. This work was supported in part by Grants-in-Aid from the Japanese Ministry of Education, Science and Culture, by Research Project Grants from Kawasaki Medical School, and by a grant from the Ryobi-Teien Memorial Foundation.

Footnotes
  1. Author to whom all correspondence should be addressed.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Altabel, M., Clarke, J. D. W., Tickle, C. 1997. Dorso-ventral ectodermal compartments and origin of apical ectodermal ridge in developing chick limb. Development 124, 45474556.
  • 2
    Amthor, H., Christ, B., Weil, M., Patel, K. 1998. The importance of timing differentiation during limb muscle development. Curr. Biol. 8, 642652.
  • 3
    Bradley, R. S. & Brown, A. M. C. 1990. The proto-oncogene int-1 encodes a secreted protein associated with the extracellular matrix. EMBO J. 9, 15651575.
  • 4
    Cadigan, K. M. & Nusse, R. 1997. Wnt signaling: A common theme in animal development. Genes Dev. 11, 32863305.
  • 5
    Christiansen, J. H., Dennis, C. L., Wicking, C. A., Monkley, S. J., Wilkinson, D. G., Wainwright, B. J. 1995. Murine Wnt-11 and Wnt-12 have temporally and spatially restricted expression patterns during embryonic development. Mech. Dev. 51, 341350.
  • 6
    Dealy, C. N., Roth, A., Ferrari, D., Brown, A. M., Kosher, R. A. 1993. Wnt-5a and Wnt-7a are expressed in the developing chick limb bud in a manner suggesting roles in pattern formation along the proximodistal and dorsoventral axes. Mech. Dev. 43, 175186.
  • 7
    Du, S. J., Purcell, S. M., Christian, J. L., McGrew, L. L., Moon, R. T. 1995. Identification of distinct classes and functional domains of Wnts through expression of wild-type and chimeric proteins in Xenopus embryos. Mol. Cell. Biol. 15, 26252634.
  • 8
    Fallon, L. K., Frederick, J. M., Carrington, J. L., Lanser, M. R., Simandl, B. K. 1983. Studies on a limbless mutant in the chick embryo. In Limb Development and Regeneration, Part A (Eds J. F. Fallon & A. I. Caplan), pp. 33–43. Alan R. Liss, New York.
  • 9
    Fekete, D. M. & Cepko, C. L. 1993. Retroviral infection coupled with tissue transplantation limits gene transfer in the chicken embryo. Proc. Natl Acad. Sci. USA 90, 23502354.
  • 10
    Gavin, B. J., McMahon, J. A., McMahon, A. P. 1990. Expression of multiple novel Wnt-1/int-1-related genes during fetal and adult mouse development. Genes Dev. 4, 23192332.
  • 11
    Hamburger, V. & Hamilton, H. L. 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 4992.
  • 12
    Hoang, B., Moos M.Jr, Vukicevic, S., Luyten, F. P. 1996. Primary structure and tissue distribution of FRZB, a novel protein related to Drosophila frizzled, suggest a role in skeletal morphogenesis. J. Biol. Chem. 271, 26 13126 137.
  • 13
    Kawakami, Y., Ishikawa, T., Shimabara, M. et al. 1996. BMP signaling during bone pattern determination in the developing limb. Development 122, 35573566.
  • 14
    Kengaku, M., Capdevila, J., Rodriguez-Esteban, C. et al. 1998. Distinct WNT pathways regulating AER formation and dorsoventral polarity in the chick limb bud. Science 280, 12741277.
  • 15
    Kispert, A., Vainio, S., Shen, L., Rowitch, D. H., McMahon, A. P. 1996. Proteoglycans are required for maintenance of Wnt-11 expression in the ureter tips. Development 122, 36273637.
  • 16
    Kuwana, T., Hashimoto, K., Nakanishi, A., Yasuda, Y., Tajima, A., Naito, M. 1996. Long-term culture of avian embryonic cells in vitro. Int. J. Dev. Biol. 40, 10611064.
  • 17
    Leyns, L., Bouwmeester, T., Kim, S. H., Piccolo, S., DeRobertis, E. M. 1997. Frzb-1 is a secreted antagonist of Wnt signaling expressed in the Spemann organizer. Cell 88, 747756.
  • 18
    Michaud, J. L., Lapointe, F., Le Douarin, N. M. 1997. The dorsoventral polarity of the presumptive limb is determined by the signals produced by the somite and by the lateral somatopleure. Development 124, 14531463.
  • 19
    Moon, R. T., Brown, J. D., Yang-Snyder, J. A., Miller. J. R. 1997. Structurally related receptors and antagonists compete for secreted Wnt ligands. Cell 88, 725728.
  • 20
    Moon, R. T., Campbell, R. M., Christian, J. L., McGrew, L. L., Shih, J., Fraser, S. 1993Xwnt-5A: A maternal Wnt that affects morphogenetic movements after overexpression in embryos of Xenopus laevis.Development 119, 97111.
  • 21
    Nohno, T., Noji, S., Koyama, E. et al. 1992. Differential expression of two msh-related homeobox genes Chox-7 and Chox-8 during chick limb development. Biochem. Biophys. Res. Commun. 182, 121128.
  • 22
    Nüsslein-Volhard, C. & Wieschaus, E. 1980Mutations affecting segment number and polarity in Drosophila. Nature 287, 795801.
  • 23
    Parr, B. A. & McMahon, A. P. 1995. Dorsalizing signal Wnt-7a required for normal polarity of D-V and A-P axes of mouse limb. Nature 374, 350353.
  • 24
    Parr, B. A., Shea, M. J., Vassileva, G., McMahon, A. P. 1993. Mouse Wnt genes exhibit discrete domains of expression in the early embryonic CNS and limb buds. Development 119, 247261.
  • 25
    Riddle, R. D., Ensini, M., Nelson, C., Tsuchida, T., Jessell, T. M., Tabin, C. 1995. Induction of the LIM homeobox gene Lmx1 by WNT7a establishes dorsoventral pattern in the vertebrate limb. Cell 83, 631640.
  • 26
    Robson, L. G., Kara, T., Crawley, A., Tickle, C. 1994. Tissue and cellular patterning of the musculature in chick wings. Development 120, 12651276.
  • 27
    Ros, M. A., Lopez-Martinez, A., Simandl, B. K. et al. 1996. The limb field mesoderm determines initial limb bud anteroposterior asymmetry and budding independent of sonic hedgehog or apical ectodermal gene expressions. Development 122, 23192330.
  • 28
    Rudnicki, J. A. & Brown, A. M. C. 1997. Inhibition of chondrogenesis by Wnt gene expression in vivo and in vitro. Dev. Biol. 185, 104118.
  • 29
    Slusarski, D. C., Yang-Snyder, J., Busa, W. B., Moon, R. T. 1997. Modulation of embryonic intracellular Ca2+ signaling by Wnt-5A. Dev. Biol. 182, 114120.
  • 30
    Solursh, M., Singley, C. T., Reiter, R. S. 1981. The influence of epithelia on cartilage and loose connective tissue formation by limb mesenchyme cultures. Dev. Biol. 86, 471482.
  • 31
    Storm, E. E. & Kingsley, D. M. 1996. Joint patterning defects caused by single and double mutations in members of the bone morphogenetic protein (BMP) family. Development 122, 39693979.
  • 32
    Summerbell, D. 1977. Regulation of deficiencies along the proximal–distal axis of the chick wing-bud: A quantitative analysis. J. Embryol. Exp. Morphol. 41, 137159.
  • 33
    Tanda, N., Kawakami, Y., Saito, T., Noji, S., Nohno, T. 1995a. Cloning and characterization of Wnt-4 and Wnt-11 cDNAs from chick embryo. DNA Seq. 5, 277281.
  • 34
    Tanda, N., Ohuchi, H., Yoshioka, H., Noji, S., Nohno, T. 1995b. A chicken Wnt gene, Wnt-11, is involved in dermal development. Biochem. Biophys. Res. Commun. 211, 123129.
  • 35
    Tickle, C. & Eichele, G. 1994. Vertebrate limb development. Annu. Rev. Cell Biol. 10, 121152.
  • 36
    Ungar, A. R., Kelly, G. M., Moon, R. T. 1995. Wnt4 affects morphogenesis when misexpressed in the zebrafish embryo. Mech. Dev. 52, 153164.
  • 37
    Wang, S., Krinks, M., Lin, K., Luyten, F. P., Moors, M. J. 1997. Frzb, a secreted protein expressed in the Spemann organizer, binds and inhibits Wnt-8. Cell 88, 757766.
  • 38
    Wang, J. & Shacklefold, G. M. 1996. Murine Wnt10a and Wnt10b: Cloning and expression in developing limbs, face and skin of embryos and in adults. Oncogene 13, 15371544.
  • 39
    Wilkinson, D. G. 1992. Whole-mount in situ hybridization of vertebrate embryos. In In Situ Hybridization (Ed. D. G. Wilkinson), pp. 75–83. IRL Press, Oxford.
  • 40
    Wong, G. T., Gavin, B. J., McMahon, A. P. 1994. Differential transformation of mammary epithelial cells by Wnt genes. Mol. Cell. Biol. 14, 62786286.
  • 41
    Yang, Y. & Niswander, L. 1995. Interaction between the signaling molecules WNT7a and SHH during vertebrate limb development: Dorsal signals regulate anteroposterior patterning. Cell 80, 939947.
  • 42
    Yang, X-M., Vogan, K., Gros, P., Park, M. 1996. Expression of the met receptor tyrosine kinase in muscle progenitor cells in somites and limbs is absent in Splotch mice. Development 122, 21632171.
  • 43
    Yoshioka, H., Ohuchi, H., Nohno, T. et al. 1994. Regional expression of the Cwnt-4 gene in developing chick central nervous system in relationship to the diencephalic neuromere D2 and a dorsal domain of the spinal cord. Biochem. Biophys. Res. Commun. 203, 15811588.
  • 44
    Zakany, J. & Duboule, D. 1993. Correlation of expression of Wnt-1 in developing limbs with abnormalities in growth and skeletal patterning. Nature 362, 546549.
  • 45
    Zecca, M., Basler, K., Struhl, G. 1996. Direct and long-range action of a wingless morphogen gradient. Cell 87, 833844.