|Home | About | Journals | Submit | Contact Us | Français|
In the frog embryo, a sub-population of trunk neural crest (NC) cells undergoes a dorsal route of migration to contribute to the mesenchyme in the core of the dorsal fin. Here we show that a second population of cells, originally located in the dorsomedial region of the somite, also contributes to the fin mesenchyme. We find that the frog orthologue of Wnt11 (Wnt11-R) is expressed in both the NC and somite cell populations that migrate into the fin matrix. Wnt11-R is expressed prior to migration and persists in the mesenchymal cells after they have distributed throughout the fin. Loss of function studies demonstrate that Wnt11-R activity is required for an epithelial to mesenchymal transformation (EMT) event that precedes migration of cells into the fin matrix. In Wnt11-R depleted embryos, the absence of fin core cells leads to defective dorsal fin development and to collapse of the fin structure. Experiments using small molecule inhibitors indicate that dorsal migration of fin core cells depends on calcium signaling through calcium/calmodulin-dependent kinase II (CaMKII). In Wnt11-R depleted embryos, normal migration of NC cells and dorsal somite cells into the fin and normal fin development can be rescued by stimulation of calcium release. These studies are consistent with a model in which Wnt11-R signaling, via a downstream calcium pathway, regulates fin cell migration and, more generally, indicates a role for non-canonical Wnt signaling in regulation of EMT.
The dorsal fin is a prominent feature of the Amphibian embryo that extends along the dorsal surface of the animal, from behind the head to the tip of the tail. The mature dorsal fin is a simple keel-shaped out-pocketing of the dorsal epidermis that is supported by mesenchymal cells and extracellular matrix (Tucker, 1986). Development of the fin begins during the early tail bud stage and the structure remains present until metamorphosis when both the fin and tail regress. The dorsal fin develops as a result of inductive signals from the neural crest (NC) (DuShane, 1935; Bodenstein, 1952; Tucker and Slack, 2004). At first, epidermis of the nascent fin undergoes cell proliferation and extends dorsally, accompanied by an accumulation of extracellular matrix, but the interior of the fin is largely acellular. During tailbud stages, non-pigmented NC cells migrate into the fin matrix (Twitty and Bodenstein, 1941; Tucker and Erickson, 1986; Collazo et al., 1993), followed by a later migration of pigmented melanocytes occurring during the early tadpole stages. In contrast, the ventral fin of Xenopus is induced by mesoderm, not neural crest, and is populated primarily by mesoderm-derived cells (Tucker and Slack, 2004), although at least some neural crest cells are also present (Collazo et al., 1993). Recent studies of dorsal fin development in the axolotl indicate that, in addition to the well-characterized NC population, cells originally located in the somite also contribute to the fin mesenchyme (Sobkow et al., 2006).
Expression studies in Xenopus have shown that a Wnt11 related sequence, Wnt11-R, is expressed in the heart, the neural tube, the dorsal somite and mesenchymal cells within the dorsal fin (Garriock et al., 2005). Xenopus contains two Wnt11 genes and Wnt11-R is the orthologue of avian and mammalian Wnt11 (Garriock et al., 2005 and data not shown). Wnt11 proteins are members of the non-canonical family of Wnt-ligands which bind to cysteine-rich frizzled and ROR receptors and LRP5/6 coreceptors (Bhanot et al., 1996; Wang et al., 1996; Wehrli et al., 2000; Hikasa et al., 2002). After ligand binding, Wnt11 and related non-canonical Wnt-proteins signal through the Wnt/calcium pathway and the planar cell polarity (PCP) pathway, both of which involve the intracellular protein Dishevelled (Dsh) (Sokol, 1996; Sheldahl et al., 2003; reviewed in Fanto and McNeill, 2004; Kohn and Moon, 2005). Through the Wnt/calcium pathway, Wnt11 can elicit intracellular calcium fluctuations resulting in the activation of PKC, calmodulin and CaMKII (Sheldahl et al., 1999; Kuhl et al., 2000). The PCP pathway signals through RhoA and can result in the activation of JNK (Li et al., 1999; Yamanaka et al., 2002;Kim and Han, 2005). Non-canonical Wnt-signaling through the PCP and Wnt/calcium pathways modulates a variety of cell behaviors including convergent extension movements during gastrulation (Heisenberg et al., 2000; Tada and Smith, 2000; Choi and Han, 2002; Yamanaka et al., 2002, Wallingford et al., 2000; Wallingford and Harland, 2001), neural tube closure (Wallingford and Harland, 2002), dendritic outgrowth (Rosso et al., 2005), heart tube morphogenesis (Garriock et al., 2005) and cranial neural crest migration (De Calisto et al., 2005).
In mouse, chicken, frog and zebrafish embryos, Wnt11/Wnt11-R expression marks populations of cells that undergo morphogenetic movements and cell shape change (Ku and Melton, 1993; Kispert et al., 1996; Heisenberg et al., 2000; Olivera-Martinez et al., 2002; De Calisto et al., 2005; Garriock et al., 2005). In this report we show that Wnt11-R expression marks a subset of somite cells and of trunk NC cells that will migrate dorsally to occupy the core of the Xenopus dorsal fin. The function of Wnt11-R is initially required for an EMT event that precedes migration of cells into the fin and ultimately for maintenance of fin structure. These Wnt11-R activities are mediated through a calcium sensitive pathway involving CamKII.
Xenopus laevis embryos were staged according to Nieuwkoop and Faber (1994) and cultured in 0.2×MMR. Microinjections occurred in 4% Ficoll in 0.4xMMR and embryos were maintained in this medium for the first 12 hours. Embryos were then cultured in 0.2×MMR until harvested. A morpholino oligo (MO) complementary to sequences in the 5′ UTR and shared by both pseudo-tetraploid copies of the Wnt11-R transcript has previously been shown to inhibit translation in vivo (Garriock et al., 2005). The sequence of the Wnt11-R MO1 is (5′-AATCATCTTCAAACCCAATAACAA-3′) and control mismatched MO is (5′-CTTGTACTTCTATAGCCTATAAGAA-3′). MOs were stored in 50 mM HEPES pH 8.0, diluted in water and heated to 65 °C for 10 minutes prior to injection. For neural crest targeting and somite targeting, 15-30 ng of MO was injected at the 8 and 16 cell stage targeted to the D1.2 and V1.2 blastomeres (Dale and Slack, 1987; Moody, 1987). KN-93, CaMKII inhibitor, was prepared as 10 mM stocks in DMSO and used immediately at 10 μM concentrations diluted in 0.2×MMR media (KN-93 Cat# S-2022, A.G. Scientific Inc). The media containing inhibitor was exchanged every three hours. Calcium signaling was activated with 100 nM thapsigargin (Sigma) or 10 nM A23187 ionophore (Sigma) for 1 hour.
Whole-mount in situ hybridization was carried out using a modification of the protocol by Harland, (1991), with antisense digoxigenin-labeled probes for Wnt11-R (Garriock et al., 2005). Plasmids were linearized with Not1 and transcribed with T7 RNA polymerase using the MEGAscript kit (Ambion). For serial sections, embryos were post-fixed in 4% paraformaldehyde, embedded in Paraplast and 10 μm transverse sections were prepared. DAPI was used to stain nuclei for cell counts.
Cell lineage studies utilized 2.5 mg/ml DiI (1,1′-dioctadecyl - 3,3,3′,3′-tetramethylindocarbocyanine perchlorate, Molecular probes, Cat# D282). Lineage tracer prepared and stored using standard methods (Collazo et al., 1993) were microinjected in a volume of 2.6 pl into the somite or the neural tube at St 19 and location of labeled cells was then followed directly by fluorescence. Neural tubes from St 19–20 embryos (i.e. prior to NC migration) were excised from donor embryos injected with 750 pg of synthetic GFP mRNA alone or in combination with 15 ng of Wnt11-R morpholino or 15 ng control morpholino. Tissue was implanted into the region of the dorsal somite, beneath the epidermis, of a similarly staged uninjected host embryo in 0.2×MMR. Embryos were cultured until St 37 and the dorsal fins of those embryos were examined for the presence of GFP-expressing cells within the fin. For neural tube ablations, a segment of the neural tube was carefully excised from St 19 embryos, which were then allowed to develop until St 27 when they were assayed.
We have carried out in situ hybridization analysis of Wnt11-R expression in the trunk region of the embryo, encompassing the time that mesenchymal cells migrate and populate the dorsal fin (Fig. 1). At late neurula (St 22) Wnt11-R-expressing cells were observed along the length of the neural tube (Fig. 1A). In transverse sections (Fig. 1B), Wnt11-R expression was located at the dorsal region of the neural tube corresponding to the location of NC progenitor cells (Davidson and Keller, 1999; Linker et al., 2000). At St 24, Wnt11-R expression was observed in the neural tube and at the dorsal margin of the somite (Fig. 1C) and this latter domain of expression became more prominent by St 26 (Fig. 1D, G), by which time it extends along almost the entire length of the embryo (Fig. 1E,F). At St 26 the presumptive dorsal fin is only a small raised keel of epidermis and does not contain any mesenchymal cells. Transverse sections showed that Wnt11-R was expressed in a group of cells located at the dorsal and lateral margin of the somite bundle and, at lower levels, in the dorsal neural tube (Fig. 1G). Approximately 2 hours later (St 27), the fin has become larger as the epidermis protrudes dorsally creating an ECM-filled space above the neural tube (Tucker, 1986) and, for the first time, a few detached Wnt11-R-expressing cells became visible within the fin (Fig. 1H,I). Transverse sections (Fig. 1J) show Wnt11-R expression in dorsal regions of the somite and neural tube and at higher levels in a subset of cells that have separated from surrounding tissues. At these early stages we sometimes observe cells within the fin that are not expressing Wnt11-R, but these are not detected at later stages. We have not pursued the identity of these cells, though it is possible that they represent precursors of melanocytes. At St 34 (Fig. 1K-L), Wnt11-R expression extends significantly to more lateral regions of the somite and numerous isolated cells occupy positions within the fin (Fig. 1M).
In the chick and mouse embryo, Wnt11 is expressed in the dorsomedial region of the somite, but not in neural crest cells (Kispert et al., 1996; Olivera-Martinez et al., 2002). In the Xenopus embryo, lineage-tracing studies have demonstrated that a minor population of trunk NC cells migrates from the neural tube and laterally across the dorsal surface of the somites (Collazo et al., 1993). This migration route raises the possibility that the domain of Wnt11-R expression in the Xenopus somite (Fig. 1D, G) represents NC cells in the process of lateral migration. To specifically address this question we have carried out ablation experiments to excise premigratory trunk NC. At St 19, well before any migration of trunk NC, the dorsal region of the neural tube was dissected from both sides of the embryo. The manipulated embryos were then allowed to develop until St 25/26, when they were assayed for Wnt11-R expression. As shown in Fig. 1O, even though these embryos completely lacked detectable neural tube in the region of ablation, the domain of Wnt11-R expression was still present in the dorsal somite at levels comparable to those observed in control regions of the embryo (Fig. 1N). We conclude that the Wnt11-R expressing cells in the dorsal somite are not of NC origin and that this domain of expression probably corresponds to the equivalent somitic region in the chick embryo (Olivera-Martinez et al., 2002).
We also detected Wnt11-R expressing cells at anterior locations of the embryo where no dorsal fin is present (Fig. 1P) and once again, the Wnt11-R expressing cells occupied a position dorsal to the neural tube. Furthermore, Wnt11-R marked fin core cells within the dorsal fin but not the ventral fin (Fig. 1Q), which contains cells primarily of mesodermal origin (Tucker and Slack, 2004). Finally, Wnt11-R expression only marked the mesenchymal population of fin cells and not the much larger pigmented melanocyte NC cells (Fig. 1R,S).
Numerous embryological studies have demonstrated that NC cells can migrate from the dorsal neural tube, into the matrix of the fin (Raven, 1931; DuShane, 1935; Krotoski et al., 1988; Collazo et al., 1993; Tucker and Slack, 2004). We examined serial sections to determine the location of Wnt11-R expressing cells when dorsal migration of fin mesenchyme first occurs (St 27) and observed cells detaching from the somite rather than the neural tube (Fig. 2A–C). This observation suggested that at least some fin mesenchymal cells were migrating from the dorsal somite rather than the neural tube. This would be consistent with observations of fin development in the axolotl embryo, where a significant proportion of fin mesenchyme cells originate from the somites (Sobkow et al., 2006). To specifically address this issue for the Xenopus embryo, we carried out lineage tracing experiments. At St 19, prior to any migration of trunk NC cells (Collazo et al., 1993), single somite bundles were labeled by injection of DiI into the dorsal surface (Fig. 2D). Embryos were allowed to develop for about 36 hours (until St 38) when the location of labeled cells was recorded. In 86% of embryos examined (n=22), DiI labeled cells were observed to migrate into the fin (Fig. 2E,F). We could demonstrate that the DiI injections were accurate because sectioning showed no DiI labeling within the neural tube (Fig. 3G,H). Based on these experiments, we conclude that cells of the dorsal somite contribute to the fin core cell population. When the neural tube was labeled at St 19 (Fig. 2I), subsequent tracing of fluorescent cells also showed contribution to the fin mesenchyme (Fig. 2J,K – 63% of embryos examined, n=11) and sectioning confirmed that the labeling was accurately directed to the neural tube (Fig. 2L,M). This experiment confirms previous results that dorsal migrating neural crest cells contribute to the fin mesenchyme (Raven, 1931; DuShane, 1935; Krotoski et al., 1988; Collazo et al., 1993; Tucker and Slack, 2004). Due to the limitations of lineage tracing methods however, these studies do not permit an estimate of the relative contributions of the somite and the neural crest to the final population fin core cells.
Wnt11-R is expressed in both the neural crest and the dorsal somite cell populations that contribute to the fin mesenchyme. In the subsequent studies, we will treat the dorsal migrating Wnt11-R expressing cells as a single population, even though we acknowledge the distinct germ layer origins of the migrating cells. To determine whether Wnt11-R function is required for fin core cell migration and fin morphogenesis we used a morpholino oligomer (MO) that has previously been shown to inhibit translation of Wnt11-R (Garriock et al., 2005). Targeting of the MO to dorsal trunk tissues produced embryos that were phenotypically normal at early stages, but which showed defects in fin morphology that were first visible at about St 33 and became pronounced by St 35. When assayed at St 34/35 the most common defects were collapsed or deficient fin tissue, especially in more anterior regions of the embryo (Fig. 3C, D). Severe disruption of fin morphology was observed in 37% of Wnt11-R MO treated embryos but only rarely (4%) in control MO injected embryos (Fig. 3J). Co-injection with 100 pg of Wnt11-R mRNA that does not contain the MO target sequences results in partial rescue of fin morphological defects (defects reduced to 19%) (Fig. 3E, F, I, J). This reduction is statistically significant, p < 0.02. Examination of transverse sections through Wnt11-R MO-treated embryos showed a greatly reduced number of mesenchymal cells within the fin matrix, with the remaining cells located proximal to the neural tube and somites (Fig. 3H). This contrasts with control fins, and mRNA rescued fins, where mesenchymal cells were distributed throughout the extent of the fin (Fig. 3G, I). Quantitation showed that inhibition of Wnt11-R expression reduced the number of isolated mesenchyme cells in the fin matrix to 48% of control levels (Fig. 3K). The reduction in fin core cell number in MO-treated embryos was restored to near wild type numbers by the co-injection of Wnt11-R mRNA (Fig. 3K). As can be seen in Fig. 3H, the overall fin structure of Wnt11-R knockdown embryos also appeared shorter than controls (Fig. 3G). Counting of cells in the epidermal layer of the fin in knockdown embryos showed a significant reduction relative to controls (29.8 +/− 1.3 for MO-treated compared to 44.3 +/− 1.2 for controls). Since Wnt11-R is not expressed in fin epidermis, this observation suggests the possibility of interactions between fin core cells and the nearby epidermal tissue.
There are two likely explanations for the absence of fin mesenchymal cells following MO treatment. First, that cells failed to migrate into the fin or second, that cells migrated normally but failed to survive in the absence of Wnt11-R activity. To distinguish between these options, we examined MO-treated embryos at St 32, soon after initial migration of cells into the fin matrix. Examination of MO treated embryos (Fig. 4D–F), revealed a severe inhibition of separation of Wnt11-R expressing cells from the somite and from the dorsal neural tube, compared to control MO-treated embryos (Fig. 4A–C). Although not ideal, it was necessary to use Wnt11-R as an in situ marker for these experiments because we are not aware of any other sequence that marks the migratory population and that maintains expression for the duration of the delamination and migration process. As presented in graphical form in Fig. 4G, approximately 62% of MO-treated embryos showed reduced migration of cells into the fins. Since both the neural tube and the somites are epithelialized structures after early neurula stages (Keller, 2000), we propose that inhibition of Wnt11-R function is blocking the EMT event that allows separation of premigratory cells from the somite and dorsal neural tube.
To determine whether cell death might contribute to the reduction in number of fin core cells, we counted the number of Wnt11-R expressing cells located within the fin or remaining adjacent to the somite (Fig. 4H). While MO treatment decreased the number of cells within the fin matrix, we observed a corresponding increase in the number of Wnt11-R expressing cells adjacent to, or in, the somite. The total number of Wnt11-R-expressing cells was unchanged in MO knockdown embryos relative to controls (Fig. 4H) and therefore, cell death was not a primary cause of the reduction in fin core cells. We conclude that inhibition of Wnt11-R activity did not lead to reduction in cell number, but instead inhibited cell separation and subsequent cell migration into the fin.
To directly test whether Wnt11-R function is required within the migrating cell population we carried out transplant experiments using tissue isolated from control and Wnt11-R MO-treated embryos. In these studies, neural tube tissue from donor embryos was implanted into the equivalent location of recipient wild type embryos. Implanted tissue was distinguished from host cells by the presence of GFP tracer. In control experiments, implants of un-manipulated tissue showed migration of labeled cells in 91% of embryos (n=11) (Fig. 4I, J). On the other hand, inhibition of Wnt11-R function in transplanted neural tubes reduced observed migration to 50% of embryos (n=10) (Fig. 4K). These results suggest that Wnt11-R activity is required within the migrating cells, rather than in adjacent tissues.
Previous studies have shown that Wnt11 is capable of signaling through the Wnt/calcium pathway, which results in the activation of CaMKII (Kuhl et al., 2000). We utilized a specific inhibitor for CaMKII (KN-93) to determine if calcium signaling via CaMKII was required for migration of cells into the fin. Previous studies have demonstrated that KN-93 effectively inhibits CaMKII activation in Xenopus embryos (Wu and Cline, 1998). Embryos were cultured with KN-93 from St 26 to St 34/35 during which time migration of cells into the fin is taking place (Fig. 1). As previously observed for Wnt11-R knockdown embryos (Fig. 3C, D, Fig. 5C), KN-93 treated embryos exhibited dorsal fin defects (62%, n=60) (Fig. 5D, E). Examination of transverse sections from KN-93 treated embryos showed a reduced number of mesenchymal cells within the fin matrix (Fig. 5H), again similar to the situation in Wnt11-R MO-treated embryos (Fig. 5G). Counts of cells in serial sections (Fig. 5I) showed that inhibition of CaMKII-signaling reduced the number of mesenchyme cells within the fin to 36% of controls, slightly more efficient than the reduction observed with Wnt11-R MO treatment (48%).
We used in situ hybridization to examine the distribution of presumptive fin core cells between the fin matrix and adjacent neural tube and somite tissues. As in previous experiments, Wnt11-R was used as a marker for fin mesenchyme and KN-93 treatments produced no detectable effect on expression levels of Wnt11-R transcript (compare Fig. 5J to Fig. 5L). Sectioning showed that inhibition of CaMKII prevented Wnt11-R expressing cells from undergoing EMT from the somite, resulting in an accumulation of cells in the somite and neural tube (compare Fig. 5K and M). This appears to be very similar to the effect observed in Wnt11-R MO treated embryos (Fig. 4F). To determine whether CaMKII activity is required for subsequent dorsal migration of cells into the fin matrix we treated embryos with CaMKII inhibitor commencing at St 28, after cells have delaminated from the dorsal somite (Fig. 1H–J). Late stage inhibition of CaMKII produced no detectable effect on migration and mesenchymal cells were distributed throughout the fin (Fig. 5P, Q), indistinguishable from carrier treated controls (Fig. 5N, O). This series of experiments implies that CaMKII activity is required for EMT of Wnt11-R cells from the somite, but not for subsequent migration within the fin matrix.
If activation of CaMKII is one of the primary steps in the pathway downstream of Wnt11-R during regulation of EMT and cell migration, then increasing calcium signaling may be able to rescue the fin defects observed in Wnt11-R knockdown embryos. To directly test this possibility, we cultured Wnt11-R MO treated embryos in thapsigargin or A23187 to increase intracellular calcium signaling. Thapsigargin and A23187 are well characterized calcium activators, previously shown to be effective in Xenopus oocytes and embryos (Osborn et al., 1997; Lupu-Meiri et al., 1993; Matifat et al., 1997). Brief treatment with either thapsigargin or A23187 at St 26–27 decreased the incidence of dorsal fin defects in Wnt11-R MO treated embryos (Fig. 6A–D). Thapsigargin rescued fin defects by 57% while A23187 rescued fin defects by 19% (Fig. 6E). The modest rescue by A23187 was probably due to toxicity because A23187 alone caused fin malformations in 25% of treated embryos (n=24 – data not shown). Examination of transverse sections through Wnt11-R MO knockdown embryos treated with thapsigargin or A23187 MO showed a rescue in the number of mesenchymal cells within the fin when compared to Wnt11-R MO alone (Fig. 6F–I). Quantitation of cells in serial sections showed that the reduction in fin core cell number in Wnt11-R knockdown embryos was completely rescued by thapsigargin or A23187 (Fig. 6J). Although treatment with thapsigargin alone resulted in a 20% increase in fin core cells, this number was not statistically different from controls (p = 0.08).
Wnt11-R is a marker for a unique population of cells that populate the dorsal fin matrix. The Wnt11-R expressing cells originate in two distinct locations, the dorsal region of the neural tube that will contribute trunk neural crest and the dorsomedial region of the somites. The Wnt11-R expressing cells populating the fin matrix are quite distinct from, and much more numerous than, the better characterized melanocyte NC population (Fig. 1R, S). Loss of function studies show that Wnt11-R activity is required for normal formation and maintenance of fin structure in the embryo. MO inhibition of Wnt11-R expression results in a severe block of dorsal migration of presumptive fin core cells into the fin and, presumably as a result of this failure, the fin fails to undergo correct morphogenesis. First, the fin lacks rigidity and is often seen to be collapsed to one side of the embryonic midline (Fig. 3C–D). Second, the epidermal layer of the fin in Wnt11-R deficient embryos contains less total cells than control embryos (Fig. 3G,H). It is possible that both morphological defects are due to lack of adequate matrix deposition within the fin. If the Wnt11-R expressing population is a source of matrix components, then failure to distribute throughout the fin, may lead to a lack of space filling structures, especially in more distal regions.
Non-canonical Wnt-signaling is mediated through the PCP, Wnt/calcium and Wnt/JNK pathways (reviewed in Pandur et al, 2002; Kohn and Moon, 2005). PCP signaling involves several proteins, including Flamingo, Strabismus and Prickle that are required for convergent extension movements during gastrulation (reviewed in Ueno and Greene, 2003; Fanto and McNeill, 2004; Takeuchi et al., 2003). However, none of these factors appear to be expressed in the NC cells or dorsal somite cells that express Wnt11-R (Darken et al., 2002;Wallingford et al., 2002; Morgan et al., 2003). This strongly suggests that PCP signaling, as it occurs during gastrulation, is not required for dorsal migration of fin core cells. Using small molecule inhibitors, we found that blocking calcium signaling through CaMKII prevented EMT of cells from the dorsal somite (Fig. 5H). The effects of CaMKII inhibition were apparently identical to those obtained by blocking Wnt11-R signaling by MO knockdown. Considering that CaMKII is a target of Wnt11-signaling (Kuhl et al, 2000), we conclude that CaMKII activity is required downstream of Wnt11-R and is essential for the delamination event when cells leave the neural tube and dorsal somite. In contrast, CaMKII activity is not required for subsequent migration of isolated mesenchymal cells into the fin (Fig. 5P, Q). In support of a calcium dependent pathway, stimulation of calcium signaling using thapsigargin or A23187 releases cells from the somite when Wnt11-R function is blocked (Fig. 6). It is important to note that, although these studies are consistent with a mechanism in which EMT of Wnt11-R expressing cells is regulated by the Wnt/calcium pathway, our experiments cannot distinguish whether the calcium signaling event is a direct or an indirect response to Wnt11-signaling.
Taken together, these studies suggest that Wnt11-R regulates delamination of cells from the neural tube and the dorsal somite through a calcium dependent pathway. This would represent a novel role for non-canonical Wnt signaling, involving CaMKII, in the separation of cells during EMT. Similar Wnt-mediated, calcium dependent EMT events might be involved in regulation of other developmental processes. For example, cells of the endothelial layer of the atrio-ventricular (AV) canal undergo a similar calcium-sensitive EMT to populate the matrix layer of the cardiac cushion (Runyan et al., 1990). While there is currently no evidence that non-canonical Wnt activity is required for EMT of the AV canal, several Wnt ligands capable of activating non-canonical Wnt-signaling are expressed in the relevant heart tissues at the appropriate time (Zakin et al., 1998; Yamaguchi et al., 1999; Le Floch et al., 2005).
Studies exploring the role of Wnt11 and related genes during gastrulation and heart morphogenesis have suggested that the primary effects of Wnt signaling may be cell- autonomous (Heisenberg et al., 2000; Tada and Smith, 2000; Garriock et al., 2005). In these examples, the cells expressing Wnt11 are also the cells that exhibit altered behavior when Wnt function is reduced. Similarly, our transplant studies showed that neural crest cells in which Wnt11-R function had been inhibited, failed to undergo migration into the fin (Fig. 4K) even though the transplanted tissue was immediately adjacent to normal somite tissue expressing Wnt11-R. On the other hand, studies of migrating cranial neural crest cells implied that signaling by Wnt11 is required cell non-autonomously (De Calisto et al., 2005). This is based on the observation that Wnt11 is initially expressed only at the leading edge of the migrating NC cells with the trailing cells expressing the Wnt receptor, Frizzled-7. The suggestion is that cranial NC cells may be migrating along a chemotactic gradient towards the source of Wnt11 (De Calisto et al., 2005). Our results appear to indicate that Wnt11-R function is required cell-autonomously during dorsal migration into the fin because the cells actively undergoing migration are the only cells in the vicinity that are expressing Wnt11-R (Fig. 1 and and4).4). These somewhat differing viewpoints can be reconciled of if we propose that Wnt11/Wnt11-R signaling is required to alter the properties of the local environment around the cell. This is consistent with the established role of non-canonical Wnt signaling to cell-autonomously facilitate cell movements through regulation of polarity and adhesion (Tada and Smith, 2000; Wallingford et al., 2000). It is also consistent with models in which non-canonical Wnt signaling regulates the polarity of fibronectin fibril deposition (Goto et al., 2005). Overall, failure to modulate the environment surrounding the cells, including the structure of the fibronectin network, may explain the inhibition of EMT and migration following inhibition of Wnt11-R expression. It seems likely that additional, currently uncharacterized factors, direct migration of fin core cells towards distal regions of the fin, although it is possible that the distribution of cells within the matrix is stochastic.
Recent studies of the axolotl have shown that fin core cells are derived from both neural crest and from the somites (Sobkow et al., 2006). Our studies indicate that the same cell populations contribute to the mesenchyme of the dorsal fin in Xenopus and that these cells are marked by expression of Wnt11-R. Although birds do not possess a structure equivalent to a dorsal fin it is intriguing that a dorsal migrating cell population, originating in the dorsomedial somites (but with no neural crest component) also expresses Wnt11, the chick orthologue of frog Wnt11-R (Olivera-Martinez et al., 2002). During avian embryogenesis, cells from the medial dermatome undergo an EMT and then migrate dorsally from the somites to populate the dorsal feather field and it has been proposed that this process might be regulated by Wnt11 (Olivera-Martinez et al., 2004). Our studies provide strong experimental evidence that Wnt11 signaling is indeed mechanistically involved in delamination of somite cells prior to migration. The mouse embryo also expresses Wnt11 in the dorsomedial somite (Kispert et al., 1996) but it has not been determined whether this expression serves a developmental function. The fact that neural crest cells contribute to the mesenchyme of the Amphibian dorsal fin has been known for some time (Twitty and Bodenstein, 1941; Tucker and Erickson, 1986; Krotoski et al., 1986; Collazo et al., 1993; Tucker and Slack, 2004). Furthermore, it has been observed that neural crest cells of the lamprey and neural crest-like cells of the ascidian, a urochordate, also migrate dorsally into the fin structures of the embryo (Newth, 1956; McCauley and Bronner-Fraser, 2003; Jeffery et al, 2004). It seems likely therefore, that the NC component of the fin core cells in Xenopus has preserved an extremely ancient function of the NC lineage, and it will be interesting to determine whether dorsal migration of lamprey and ascidian NC cells is also regulated by Wnt11.
Special thanks to Florence Broders and Roberto Mayor for sharing results prior to publication. P.A.K. is the Allan C. Hudson and Helen Lovaas Endowed Professor of the Sarver Heart Center at the University of Arizona College of Medicine and is supported by the Sarver Heart Center and by the NHLBI of the NIH, grants #HL63926 and HL74184.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.