Developmental Biology Section, Cartilage Biology and Orthopaedics Branch, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Maryland
Cartilage Biology and Orthopaedics Branch, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, 6 Center Drive, Bldg. 6, Room 324 MSC2745, Bethesda, MD 20892-2745
This article is a US Government work and, as such, is in the public domain in the United States of America.
Dach1 is a mouse homologue of the Drosophila dachshund gene, which is a key regulator of cell fate determination during eye, leg, and brain development in the fly. We have investigated the expression and growth factor regulation of Dach1 during pre- and postnatal skeletal development in the mouse limb to understand better the function of Dach1. Dach1 was expressed in the distal mesenchyme of the early embryonic mouse limb bud and subsequently became restricted to the tips of digital cartilages. Dach1 protein was localized to postmitotic, prehypertrophic, and early hypertrophic chondrocytes during the initiation of ossification centers, but Dach1 was not expressed in growth plates that exhibited extensive ossification. Dach1 colocalized with Runx2/Cbfa1 in chondrocytes but not in the forming bone collar or primary spongiosa. Dach1 also colocalized with cyclin-dependent kinase inhibitors p27 (Kip1) and p57 (Kip2) in chondrocytes of the growth plate and in the epiphysis before the formation of the secondary ossification center. Because fibroblast growth factors (FGF), bone morphogenetic proteins (BMP), and hedgehog molecules (Hh) regulate skeletal patterning of the limb bud and chondrocyte maturation in developing endochondral bones, we investigated the regulation of Dach1 by these growth and differentiation factors. Expression of Dach1 in 11 days postcoitus mouse limb buds in organ culture was up-regulated by implanting beads soaked in FGF1, 2, 8, or 9 but not FGF10. BMP4-soaked beads down-regulated Dach1 expression, whereas Shh and bovine serum albumin had no effect. Furthermore, FGF4 or 8 could substitute for the apical ectodermal ridge in maintaining Dach1 expression in the limb buds. Immunolocalization of FGFR2 and FGFR3 revealed overlap with Dach1 expression during skeletal patterning and chondrocyte maturation. We conclude that Dach1 is a target gene of FGF signaling during limb skeletal development, and Dach1 may function as an intermediary in the FGF signaling pathway regulating cell proliferation or differentiation. Published 2002 Wiley-Liss, Inc.
Dach1 is a murine member of the dachshund family of nuclear proteins that, thus far, includes one member in Drosophila (dac), one in medaka, three in zebrafish, and two each in chicken, mouse, and human (Mardon et al., 1994; Hammond et al., 1998, 2002; Heanue et al., 1999, 2002; Ayres et al., 2001; Davis et al., 2001b; Loosli et al., 2002). Although a DNA binding domain has not been identified in any of the dachshund proteins, data suggest that they participate in transcriptional regulatory complexes with members of the Pax and Eya families (Chen et al., 1997; Heanue et al., 1999, 2002; Ozaki et al., 2002). Studies in Drosophila and chicken indicate that dac/Dach family members regulate embryonic cell fate determination, and ectopic expression of dac in the developing fly causes ectopic retinal development (Mardon et al., 1994; Lecuit and Cohen, 1997; Shen and Mardon, 1997; Heanue et al., 1999). Previous studies have demonstrated the expression of Dach1 in early limb bud or fin bud development in several vertebrate species (mouse, chicken, zebrafish, and medaka) (Hammond et al., 1998, 2002; Caubit et al., 1999; Davis et al., 1999; Kozmik et al., 1999; Machon et al., 2000; Ayres et al., 2001; Heanue et al., 2002; Loosli et al., 2002). In addition to its expression in limb buds, Dach1 is expressed in multiple organs of the developing vertebrate embryo, including the eyes, ears, kidneys, and central nervous system (Hammond et al., 1998; Caubit et al., 1999; Davis et al., 1999, 2001a; Kozmik et al., 1999; Machon et al., 2000; Ayres et al., 2001; Heanue et al., 2002). Dach1 null mutant mice exhibit no abnormalities at the gross or histologic level; however, they develop respiratory distress and die within 32 hr of birth (Davis et al., 2001a). This finding suggests that Dach1 is not necessary for embryonic morphogenesis, but is necessary for some physiological function(s) required for postnatal life. Dach2, which overlaps in expression pattern with Dach1, may functionally compensate for the loss of Dach1 during embryonic morphogenesis (Heanue et al., 1999; Davis et al., 2001a,b). By studying the regulation of Dach1's cellular expression pattern in the developing limb skeleton, we aimed to understand better the function of Dach1 and the mechanisms of skeletal patterning and chondrocyte cell fate determination.
The regulation of dac expression in Drosophila is complex but relatively well characterized. dac is required for retinal cell fate determination in the posterior margin of the eye disc, for the formation of several leg segments, and for the formation of mushroom bodies in the fly brain (Mardon et al., 1994; Martini et al., 2000). Its expression is regulated in a concentration and sex-specific manner by the tissue patterning factors dpp and wg (Lecuit and Cohen, 1997; Chen et al., 1999; Curtiss and Mlodzik, 2000; Keisman and Baker, 2001). It participates in a network of genes, including ey, eya, and tsh, that regulate retinal determination (Chen et al., 1997, 1999; Pan and Rubin, 1998).
The expression of vertebrate homologues of genes in the Drosophila retinal determination network suggests that the interaction of these genes is functionally conserved. In mouse and chicken embryos, Dach proteins are coexpressed with Eya and Pax proteins in the developing eyes, ears, and somites (Hammond et al., 1998; Davis et al., 1999; Heanue et al., 2002). Pax6 is the vertebrate homologue of ey, which, along with eya, regulates dac expression in the fly eye. Dach1 expression in the vertebrate ear and eye, however, is not regulated by Eya or Pax genes and the expression of Dach1 in the developing mouse limb bud has minimal overlap with the expression of Pax genes (Hammond et al., 1998; LeClair et al., 1999; Heanue et al., 2002). Therefore, it is likely that the molecular mechanisms regulating Dach1 expression in mammals are significantly different from those regulating dac expression in the fly (Heanue et al., 2002).
Therefore, we have investigated the regulation of Dach1 expression in the developing mouse limbs to focus on opportunities to identify vertebrate-specific mechanisms for its regulation. Mesenchymal cell proliferation, skeletogenic cell fate determination, and patterning of the developing skeleton in the early limb bud (approximately 9–13 days postcoitus [dpc]) is regulated by the integration of signaling from several pathways, including fibroblast growth factors (FGF), bone morphogenetic proteins (BMP), and sonic hedgehog (Shh) (for reviews see Martin, 1998; Xu et al., 1999; Dahn and Fallon 1999; Capdevila and Belmonte, 2001; Tickle and Munsterberg, 2001). By 14 dpc, a cartilaginous template of the limb skeleton has formed and cartilage is replaced by bone through the process of endochondral ossification. Signaling by FGF, BMP, and a Shh-related molecule, Indian hedgehog (Ihh), also regulates this process of chondrocyte maturation during endochondral ossification (for reviews, see Olsen et al., 2000; Vortkamp, 2001; Shum and Nuckolls, 2002). This reiteration of signaling pathways during skeletal development provides opportunities to distinguish between gene expression associated with a developmental process and that associated with specific signaling pathways.
Therefore, we have examined the cellular and subcellular localization of Dach1 at stages of embryonic, fetal, and postnatal development that include early limb bud patterning, chondrogenesis, and chondrocyte maturation. Our data provide a better understanding of Dach1 function and the regulation of its expression by growth factor signaling. We have found that it was expressed in mesenchymal cells before chondrocyte differentiation but was excluded from proliferating chondrocytes. It was re-expressed in postmitotic, prehypertrophic, and early hypertrophic chondrocytes, but not in the terminal stages of hypertrophy. By implanting growth factor–soaked beads into early limb buds in organ culture, we found that Dach1 expression was regulated by FGF signaling, consistent with its coexpression with FGF receptors throughout limb development.
Previous studies have described the pattern of expression of Dach1 in mouse embryonic limb buds (Hammond et al., 1998; Caubit et al., 1999; Davis et al., 1999; Kozmik et al., 1999; Machon et al., 2000; and Ayres et al., 2001). However, the expression of Dach1 in fetal and postnatal skeletal development has not yet been described. Although the regulation of dac expression in Drosophila is well characterized, factors regulating the expression of Dach1 at any stage of vertebrate development have not been identified hitherto our studies.
Dynamic Expression of Dach1 in Early Limb Bud
The expression pattern of Dach1 message in the mouse forelimb bud changed with the stage of differentiation of the mesenchymal cells. Consistent with the pattern described by Caubit et al. (1999), we observed at least three patches of Dach1 message accumulation in the mouse limb bud at 10 dpc (Fig. 1A): (1) extending from the anterior margin to the midline; (2) overlapping the zone of polarizing activity (ZPA) on the posterior margin; and (3) on the posterior margin, proximal to area number 2. In addition, we observed an area with increased staining on the anterior margin but proximal to area number 1 (Fig. 1A). Aside from the region that overlaps the ZPA, signaling centers that correspond to these sites of Dach1 expression have not been characterized. However, the paired anterior and posterior proximal Dach1 expression areas may correspond to the anterior and posterior necrotic zones that develop by 11 dpc. There was a conspicuous absence of Dach1 expression at the distal tip of the limb bud at 10 dpc. By 11 dpc, Dach1 expression occupied a broad band of tissue along the subridge mesenchyme of the limb bud and was excluded from the central region of prechondrogenic cell condensation (Fig. 1B). Dach1 expression became restricted to the tip of each developing digit at 12 dpc with weaker staining surround the distal phalangeal cartilages and in the interdigital web space (Fig. 1C). The limb buds were the most prominent sites of accumulation of Dach1 message in the whole-mount–stained 12 dpc embryo (Fig. 1D).
The distribution of Dach1 protein was similar to that of its message in early limb bud development. Immunohistochemistry with an affinity purified anti-Dach1 antibody revealed staining in 10 dpc forelimb buds concentrated in mesenchymal cells along the anterior margin of the bud and more widely distributed in proximal, posterior cells (Fig. 2A), corresponding to the areas of highest expression of Dach1 message (Fig. 1A). At 12 dpc, Dach1 protein, like the message, was localized at the tips of developing digits (arrow in Fig. 2B), and surrounded the distal ends of the digital cartilages (open arrowhead in Fig. 2B). Relatively intense staining was also observed in the subectodermal mesenchyme associated with the first and last digits (filled arrowheads in Fig. 2B).
To determine whether Dach1 expression was associated with proliferating or cell cycle arrested cells, adjacent sections of the 12 dpc limb bud were stained with anti-proliferating cell nuclear antigen (PCNA), a marker of cells in S-phase. Like Dach1, PCNA was localized to the tips of the digits and in the subectodermal mesenchyme of the first and last digits (Fig. 2C). PCNA was also expressed in the mesenchyme surrounding the digital cartilages but was less tightly associated with the centers of chondrogenic condensation than was Dach1. Immunofluorescent staining of Dach1 and PCNA in the subectodermal mesenchyme of 12 dpc limb buds revealed numerous cells expressing both antigens (Fig. 2D).
Dach1 Expression in Developing Chondrocytes
The 17 dpc fetal mouse hind limb exhibited immunohistochemical staining for Dach1 in chondrocytes and cells of the dermis. Chondrocytes of the proximal, middle, and distal phalanges were immunopositive for Dach1 (arrows in Fig. 3A,B), although staining in other developing bones of the hind limb was not evident at this stage. Dermal folds of the plantar foot pads and digit pads also exhibited prominent staining (arrowheads in Fig. 3A,B). In the proximal phalanges, chondrocytes at two stages of development expressed Dach1; those in the center of the anlagen, which are prehypertrophic or early hypertrophic cells (arrow in Fig. 3C), and chondrocytes adjacent to the developing joint space (arrowheads in Fig. 3C). The specificity of the anti-Dach1 antibody was confirmed by incubating the antibody with purified Dach1 antigen before staining, and no immunoreactivity was detected (Fig. 3D).
To characterize the stage of differentiation of the chondrocytes expressing Dach1, we analyzed the mitotic status of these cells, and compared the Dach1 staining with that of Runx2/Cbfa1, a marker for prehypertrophic and hypertrophic cells in the chondrocyte lineage (Kim et al., 1999). Mouse fetuses at 17.5 dpc were labeled in utero with bromodeoxyuridine (BrdU), and the staining pattern of labeled nuclei was compared with the Dach1 expression pattern by immunofluorescence (Fig. 4A). The majority of BrdU-labeled nuclei were among the columnar proliferating cells, and labeled nuclei were rarely detected in the more mature chondrocytes that expressed Dach1. There were very few nuclei that stained both for Dach1 and BrdU, which confirmed that Dach1 was expressed almost exclusively in postmitotic cells, which were at the prehypertrophic stage or later. Dach1 and Runx2/Cbfa1 both stained maturing chondrocytes in the developing proximal phalanges (Fig. 4B,C). In addition to the chondrocyte staining, Runx2/Cbfa1 was localized to osteoblasts in the developing bone collar, where Dach1 was not expressed. Endogenous alkaline phosphatase (AP) activity was detected in hypertrophic chondrocytes and in the developing bone collar, and Dach1 expression overlapped with AP in the early hypertrophic chondrocytes but not at other sites (data not shown). These data suggest that Dach1 was expressed in postmitotic, prehypertrophic, and early hypertrophic chondrocytes of the developing phalanges. However, Dach1 was not detected at this time point in chondrocytes in more proximal bones of the limb where ossification had begun, even though prehypertrophic and hypertrophic cells were present in these structures. Therefore, we concluded that Dach1 expression was associated with the initial onset of chondrocyte hypertrophy and was not maintained in established growth plates.
To examine further the association of Dach1 expression with the initiation of chondrocyte hypertrophy, we assayed for its expression over a time course of secondary ossification center formation in the distal femoral epiphysis. At 17.5 dpc, Dach1 expression in the chondrocytes of the distal femoral growth plate was barely detectable (Fig. 5A). However, at 4 days of postnatal development (P4), Dach1 expression was evident in epiphyseal chondrocytes at the site where the secondary ossification center was to form (Fig. 5B). By P9, the epiphyseal ossification center had formed, and the staining of Dach1 had returned to background levels (Fig. 5C). This finding suggests that Dach1 expression is associated with the morphogenetic event of initial chondrocyte hypertrophy in both primary and secondary ossification centers and is not associated with chondrocyte maturation after ossification has begun.
Dach1 Coexpression With p27(Kip1) and p57(Kip2)
Studies suggest that the initiation of chondrocyte hypertrophy is dependent on arrest of the cell cycle. To investigate the association of Dach1 expression in chondrocytes with changes in cell cycle regulation, we compared the pattern of Dach1 protein accumulation with that of two cyclin-dependent kinase inhibitors, p27(Kip1) and p57(Kip2). Both p27 and p57 are known to be expressed in chondrocytes, and disruption of the Kip2 gene interferes with hypertrophic chondrocyte differentiation (Zhang et al., 1997; Sunters et al., 1998; Nagahama et al., 2001). Dach1 staining of the proximal tibial growth plate of P3 mice revealed labeled chondrocytes in the prehypertrophic and early hypertrophic zones of the diaphyseal ossification center (flanked by arrows in Fig. 6A) as well as chondrocytes involved in the initiation of the epiphyseal ossification center (flanked by arrowheads in Fig. 6A). Staining of adjacent sections with antibodies to p27 (Fig. 6B) and p57 (Fig. 6C) revealed significant overlap in the accumulation of these cell cycle regulators with Dach1.
Growth Factor Regulation of Dach1 Expression
Several signaling pathways are known to regulate chondrocyte proliferation and maturation in the developing long bones, including FGF, BMP, and Hh proteins. To determine whether one of these pathways is rate limiting for Dach1 expression in the developing limb, we implanted growth factor-soaked beads into 11 dpc mouse forelimb buds and assayed for changes in Dach1 expression after 24 hr of organ culture. FGF1-, FGF2-, and FGF8-soaked beads implanted in the distal mesenchyme of the limb buds up-regulated the expression of Dach1 around the implanted bead, and the FGF9-soaked bead up-regulated Dach1 expression throughout the subridge mesenchyme of the limb bud (Fig. 7F–I). Placement of the bead in the center of the bud, where chondrogenic condensation first occurs resulted in little or no detectable up-regulation of Dach1 expression. FGFs 1, 2, 8, and 9 all bind to the IIIc splice variants of FGFR2 and FGFR3 that are known to be expressed in the mesenchyme and chondrocytes of developing limbs (Szebenyi et al., 1995; Ornitz et al., 1996; Xu et al., 1998). However, the specificity of these splice variants does not include FGF10, which preferentially associates with the IIIb splice variants expressed in the overlying limb epithelium and the apical ectodermal ridge (AER) (Xu et al., 1998). Consistent with these differences in ligand specificity and expression pattern of FGF receptors, beads soaked in FGF10 did not alter Dach1 expression when implanted into developing limb buds (Fig. 7J). Of the other signaling molecules known to regulate patterning in the limb bud and development of cartilage, we tested BMP4 and Shh. BMP4 caused a down-regulation of Dach1 expression in the region immediately surrounding the bead (Fig. 7D), whereas Shh had no noticeable effect on Dach1 expression (Fig. 7E).
The regulation of Dach1 by FGF can be interpreted with respect to the anatomic structure of the developing limb buds. The AER is a prominent site of FGF2, 4, 8, and 9 production, and the underlying mesenchyme, which expresses Dach1 is the target tissue for FGF's effects on cell proliferation and patterning (for review, see Martin, 1998). To test whether endogenous FGFs from the AER are necessary for the regulation of Dach1 expression in the early limb bud, we removed the AER from 11.5 dpc mouse forelimbs, or replaced the AER with FGF-soaked beads, and assayed for Dach1 message accumulation. Within 24 hr after removing the AER, Dach1 expression was significantly down-regulated (Fig. 8B). However, when the AER was replaced with beads soaked in FGF4 or FGF8, which are known to functionally substitute for the AER in maintaining limb bud outgrowth and patterning (Niswander et al., 1993; Vogel and Tickle, 1993; Crossley et al., 1996; Vogel et al., 1996), Dach1 expression was maintained (Fig. 8C,D). Beads soaked in FGF10 did not maintain Dach1 expression (data not shown). These data suggest that specific FGF signaling is necessary for the expression of Dach1 in the early limb bud and the AER is the source of this FGF signal.
Dach1 Coexpression with FGF Receptors
To explore further the link between FGF signaling and Dach1 expression, we compared the pattern of Dach1 expression in the limb bud and chondrocytes with that of FGFR2 and FGFR3, receptors known to participate in limb development (for reviews, see Martin, 1998; Xu et al., 1999; Ornitz, 2001). In the 12 dpc limb bud, there was overlap in expression between Dach1 and both of these FGF receptors (Fig. 9). FGFR3 expression was more widely distributed than that of FGFR2, and features of the pattern of FGFR3 staining more resembled the Dach1 expression pattern. Like Dach1, FGFR3 accumulated at the tips of the digits (arrow in Fig. 9C), in the mesenchyme anterior to the first digit and posterior to the last digit (arrowhead in Fig. 9C) and surrounding but not including the cell condensations of the digits (open arrowhead in Fig. 9C). In the phalanges of 17.5 dpc hind paws, Dach1 expression in the chondrocytes overlapped with that of FGFR2 and FGFR3, although FGFR2 was expressed in chondrocytes of the resting and proliferating zones (Fig. 9D–F). In the skin of the footpad at this stage, Dach1 was localized in cells of the dermis (Fig. 9G), whereas the FGFR2 antibody labeled the basal cells of the epidermis most strongly (Fig. 9H). This epidermal staining of FGFR2, which did not correlate with Dach1 expression, most likely indicated the distribution of the keratinocyte growth factor receptor. This FGFR2 antibody cross-reacts with the keratinocyte growth factor receptor, which is a IIIb splice variant of FGFR2. FGFR3 was more widely distributed in the skin, with some expression in the dermal cells that express Dach1 (Fig. 9I). These data suggest that Dach1 is coexpressed with IIIc splice variants of FGF receptors in the developing limb, which further supports our model of FGF signaling regulating Dach1 expression.
We conclude from our data that FGF signaling regulates the expression of Dach1 in the developing limb bud and subsequent cartilage tissue. As shown by our bead implantation experiments, Dach1 expression in the limb bud was defined by the local concentration of FGF in undifferentiated mesenchymal tissue. The normal expression of Dach1 in the limb bud was dependent on an intact AER, which is a source of FGF2, 4, 8, and 9. Furthermore, FGFs can substitute for the AER in maintaining Dach1 expression. The colocalization of Dach1 and FGF receptors in the limb bud and during chondrocyte maturation suggests that the dependence of Dach1 expression on FGF signaling is maintained in cartilage tissue.
Our data further suggests that BMP4 and other factors are also important in regulating the expression of Dach1. There are several sites in the developing limb bud and limb cartilage where Dach1 was not expressed, even though FGF signaling was active. For examples, FGF8 is expressed in the AER at 10 dpc, but at this stage Dach1 expression was not associated with the mesenchyme adjacent to this structure. Furthermore, FGF-soaked beads implanted into the center of the limb bud at 11 dpc, where chondrogenesis was under way, did not induce Dach1 expression. Other factors may actively suppress Dach1 expression or may be required to activate the responsiveness of the Dach1 gene to FGF signals. BMP4 inhibited endogenous Dach1 expression in our bead implantation experiments, which suggests that BMP signaling may inhibit Dach1 expression at sites in the limb bud and developing cartilage. BMP signaling and other factors related to the stage of differentiation of mesenchymal cells and chondrocytes may account for the differences between the patterns of FGF signaling and Dach1 expression.
The expression pattern of Dach1 in the limb bud has similarities to that of Mtsh1, another mouse homologue of a Drosophila retinal determination gene. Like Drosophila dac, the ectopic expression of the zinc finger protein tsh can induce ectopic retinal development (Pan and Rubin, 1998). tsh is higher in the genetic hierarchy than dac, because tsh can induce ey and visa versa. Mtsh1 is one of two mouse homologues of tsh, and like Dach1, it is expressed in the distal mesenchyme of the limb buds, first and second branchial arches, and neural tube at 11 dpc, the whisker follicles at 12.5 dpc and at the tips of the digits at 13 dpc (Hammond et al., 1998; Davis et al., 1999; Caubit et al., 2000; Long et al., 2001). Furthermore, the expression of Mtsh1 in the limb bud is up-regulated by FGF8 and down-regulated BMP4-soaked beads, although only in the proximal portion of the bud (Long et al., 2001). The similarities of expression and regulation of these two genes and the genetic interactions of their Drosophila counterparts suggest that Dach1 and Mtsh1 may participate in the same developmental or cellular pathways during morphogenesis of the limb and other structures.
Although some features of the genetic interactions of dac/Dach1 have diverged between Drosophila and mice, other features are conserved. As discussed previously, dac expression is regulated by ey and eya, whereas Dach1 is not dependent on Pax or Eya expression (Chen et al., 1997; Shen and Mardon, 1997; Hammond et al., 1998; Chen et al., 1999; Heanue et al., 2002). Interestingly, the regulation of Dach1 expression by BMP4, which is a vertebrate homologue of Drosophila dpp, does appear to be evolutionarily conserved. BMP4-soaked beads suppressed Dach1 expression in the limb buds, and dpp signaling suppresses dac expression in the Drosophila legs and female genital discs (Lecuit and Cohen, 1997; Keisman and Baker, 2001). The coexpression of Dach family members with Eya, Pax, and Mtsh genes suggests that they interact in a regulatory network. Even though some of the interactions within this network may be different between Drosophila, mouse, and other organisms, the network may regulate similar processes in each organism and in multiple structures of each embryo.
The regulation of Dach1 expression by FGF signaling suggests that Dach1 functions as an intermediary in the FGF signaling pathways regulating limb development. In the limb bud and skeletal tissue, FGF signaling contributes to the regulation of cell proliferation and also to cell cycle arrest, depending on the cell type or stage of differentiation. Treatment with exogenous FGF promotes proliferation of limb bud mesenchymal cells in vivo and in tissue culture (Kaplowitz et al., 1982; Niswander et al., 1993). Up-regulation of FGF signaling in chondrocytes, either by treatment with growth factor or by activating mutations in FGFR3 causes cell cycle arrest (Su et al., 1997; Sahni et al., 1999; Aikawa et al., 2001). Consistent with its association with FGF signaling and cell cycle regulation, Dach1 was expressed in proliferating mesenchyme in the limb buds and Dach1 was expressed in cell cycle arrested chondrocytes of the growth plates. The association of Dach1 with cell cycle control is further supported by its colocalization with the cyclin-dependent kinase inhibitors p27(Kip1) and p57(Kip2) in postmitotic chondrocytes. Dach1 is coexpressed with p27 and p57 in podocytes of the developing kidney, and with p27 but not p57 in the neuroepithelium of the organ of Corti, retinoblasts, and retinal pigment epithelium cells (Ayres et al., 2001; Davis et al., 2001a; Nagahama et al., 2001). Although we did not compare the pattern of expression of Dach1 with that of p21(CIP1/WAF1) or other cell cycle regulators in our experiments, p21 is expressed in maturing chondrocytes and likely overlaps in its expression with Dach1 (Stewart et al., 1997). Furthermore, Dach1 associates with the ubiquitin-conjugating enzyme Ubc9, which is involved in the degradation of nuclear proteins that regulate cell cycle progression (Seufert et al., 1995; Loveys et al., 1997; Machon et al., 2000). Although we have not observed changes in proliferation as a result of transient or stable overexpression of Dach1 in cells in culture, results from localization studies are consistent with the hypothesis that Dach1 functions in the FGF pathway and contributes to the regulation of proliferation and/or cell cycle arrest, perhaps in conjunction with other genes of the retinal determination network.
FGF signaling also regulates chondrocyte maturation in the growth plates, and it is currently unclear whether this regulation is an indirect effect of cell cycle control or a separate, parallel function for FGF signaling in chondrocytes. The expression of Dach1 in chondrocytes during the initiation of hypertrophy and its coexpression with FGF receptors suggests that Dach1 may be involved in the mechanisms that regulate chondrocyte maturation.
Runx2/Cbfa1 is a runt domain transcription factor that is a key regulator of chondrocyte hypertrophy. Loss of function mutations in the Runx2/Cbfa1 gene inhibit or delay chondrocyte hypertrophy, whereas overexpression of Runx2/Cbfa1 promotes hypertrophy (Komori et al., 1997; Mundlos et al., 1997; Otto et al., 1997; Kim et al., 1999; Takeda et al., 2001; Ueta et al. 2001). Runx2/Cbfa1 expression is up-regulated by activated FGFR1 in the calvaria and by FGF2 or FGF8 treatment of skeletogenic tissue culture cells (Zhou et al., 2000). Because Dach1 was coexpressed with Runx2/Cbfa1 in prehypertrophic and early hypertrophic chondrocytes, we hypothesize that Dach1 may be involved in the mechanisms by which FGF signaling regulates Runx2/Cbfa1 expression and chondrocyte hypertrophy.
Although the developmental, cellular, and molecular functions of Dach1 are still unclear, our data and studies by other groups suggest that Dach1 contributes to the fundamental mechanisms of morphogenesis that are conserved among diverse embryonic structures in divergent organisms. We have demonstrated that Dach1 expression is associated with FGF signaling at multiple stages of mesenchymal cell and chondrocyte differentiation during morphogenesis of the limb bud and subsequent limb skeleton. Further study of the function of Dach1 and its role in FGF signaling will contribute to an understanding of the mechanisms of skeletal morphogenesis and may clarify the connections between cell cycle control and cell differentiation.
Cell Culture and Purification of Murine Recombinant Dach1
A recombinant adenovirus for the expression of Dach1 was constructed (Qbiogene, Carlsbad, CA) by using the full-length, FLAG-tagged cDNA from pDacCF (Ayres et al., 2001). C3H10T1/2 cells cultured in BME media supplemented with 10% fetal bovine serum and glutamine, and antibiotics were infected with the DacCF adenovirus at a multiplicity of infection of 100. After 2 days of culture, cells were lysed and the recombinant Dach1 purified on and anti–FLAG-M2-Sepharose column (Sigma, St. Louis, MO). Dach1 was eluted from the column with Gentle Elution Buffer (Pierce, Rockford, IL), dialyzed against PBS and freeze dried.
Affinity purified rabbit anti-Dach1 antisera SV-3 raised against a human Dach1 fusion protein and affinity purified with mouse Dach1 has been described previously (Ayers et al., 2001). A chicken anti-Dach1 immunoglobulin (Ig) Y, HEN-AH-1, was raised against recombinant murine Dach1. Before use, the chicken IgY preparation was diluted 1 in 10 and absorbed against undifferentiated mesenchymal cells. For immunolocalization, the absorbed antibody was further diluted 1:100. The affinity purified rabbit anti-Dach1 and the chicken anti-Dach1 gave identical staining patterns on mouse tissues. The specificity of both the rabbit and chicken anti-Dach1 antisera was confirmed by preadsorption of the diluted antibody with 10 μg/ml purified, recombinant murine Dach1. Nonimmune rabbit IgG and chicken IgY were also used as controls in immunohistochemical staining.
Timed pregnant Swiss Webster mice were purchased (Harland, Indianapolis, IN) and used according to Animal Study Protocols approved by the NIH. Tissue samples from embryonic, fetal, and postnatal stages were embedded in OCT compound (Electron Microscope Sciences, Warrington, PA), frozen, and stored at −80°C until sectioned. Sections, 10 μm thick, were air-dried for 10–15 min and fixed for 20 min in 4% (w/v in phosphate buffered saline) paraformaldehyde pH 7.4. Indirect enzyme linked and fluorescence immunolocalization were performed as previously described (Horner et al., 1998). Immunofluorescence stained sections were counterstained with Hoechst 33342 (Molecular Probes, Eugene, OR). Diaminobenzidine- and Vector VIP-stained sections were counterstained with Gill's hematoxylin (Vector Laboratories). All sections were mounted in aqueous mounting medium and stored in the dark.
Affinity purified rabbit polyclonal IgG preparations specific for the cyclin-dependent kinase inhibitors p27(Kip1) (C-19) and p57(Kip2) (H91) were purchased from Santa Cruz Biotech, Inc. (Santa Cruz, CA). Comparison of p27 and p57 expression with Dach1 expression was performed on adjacent sections by using the SV-3 and on the same section by using HEN-AH-1. Colocalization experiments using HEN-AH-1 used fluorescein isothiocyanate (FITC) -conjugated donkey anti-chicken and rhodamine-conjugated donkey anti-rabbit secondary antibodies. Sections were stained with rabbit anti-Cbfa-1/Runx2/AML-1/PEBP2αA (M70, Santa Cruz). Alkaline phosphatase activity was detected by using nitro blue tetrazolium/5-bromo-4-chloro-3-indoxyl phosphate (NBT/BCIP; Roche Molecular Biochemicals, Indianapolis, IN). Antibodies specific for the cytoplasmic domains of FGF receptors FGFR2 (C-17) and FGFR3 (C-15) were purchased from Santa Cruz Biotechnology.
Proliferation was investigated in day 10 dpc forelimb limb buds and day 17 dpc hind feet by bromodeoxyuridine (BrdU) incorporation and PCNA staining. For BrdU incorporation, mice were labeled in utero by intraperitoneal injection of the dam with BrdU in PBS (2 mg/ml) at a dose of 30 mg/kg, 4 hr before killing. BrdU incorporation was detected by using a mouse monoclonal anti-BrdU (Sigma) after microwave treatment of the sections in 0.1 M citrate buffer pH 6 for 10 min before nonspecific protein block (Dover and Patel, 1994). Colocalized experiments for BrdU incorporation and Dach-1 expression were performed by using SV-3. Dach1 expression was visualized by using FITC-conjugated donkey anti-rabbit antiserum and BrdU incorporation using rhodamine-conjugated sheep anti-mouse antiserum. PCNA was stained by using a goat polyclonal (C-20, Santa Cruz Biotechnology), following the directions of the supplier.
Bead Implantation and In Situ Hybridization
Heparin sulfate beads (Sigma) were soaked in 1 mg/ml FGF1, 2, 4, 8, 9, or 10, or Affigel Blue beads (Bio-Rad, Richmond, CA) were soaked in 0.5 mg/ml BMP4 or 500 nM Shh for 2 hr at room temperature (all growth factors from R&D Systems, Minneapolis, MN). Beads soaked in PBS or 1 mg/ml bovine serum albumin (BSA) were used as controls. Forelimb buds were dissected from 11 or 11.5 dpc mouse embryos. For some experiments the AER was removed from the limb buds by using a tungsten needle. Beads were implanted into the limb buds, and the limb buds were cultured at 37°C for 24 hr in a serum-free medium (BGJb, 2 mg/ml BSA, 100 g/ml ascorbate, antibiotics, and antimycotics) in a humid, 5% CO2 environment. After incubation, they were fixed and processed for the detection of Dach1 message by whole-mount in situ hybridization as described previously (Semba et al., 2000), by using a digoxigenin-labeled probe containing 1 kb of the coding sequence and 500 bp of the 3′ untranslated region of mouse Dach1.
We thank Harold Slavkin for his inspiration and guidance, and Peter Lipsky and Rocky Tuan for their encouragement and continued support of our research.