Hedgehog (Hh) signalling has been implicated in the development of osteoblasts and osteoclasts whose balanced activities are critical for proper bone formation. As many mouse mutants in the Hh pathway are embryonic lethal, questions on the exact effects of Hh signalling on osteogenesis remain. Using zebrafish, we show that there are two populations of endochondral osteoblasts with differential sensitivity to Hh signalling. One, formed outside the cartilage structure, requires low levels of Hh signalling and fails to differentiate in Indian hedgehog mutants. The other derives from chondrocytes and requires higher levels of Hh signalling to form. This latter population develops significantly earlier in mutants with increased Hh signalling, leading to premature endochondral ossification, and also fails to differentiate in Indian hedgehog mutants, resulting in severely delayed endochondral ossification. Additionally, we demonstrate that the timing of first osteoclast activity positively correlates to Hh levels in both endochondral and dermal bone.

Bone formation in vertebrates is regulated by a balance between the activities of cells that secrete bone matrix, the osteoblasts, and those that reabsorb it, the osteoclasts. The tight regulation of the activities of these two cell types is of crucial importance.

Osteoclasts are of haematopoietic origin from the monocyte/macrophage lineage, and their differentiation is controlled by RANKL (Tnfsf11 - Mouse Genome Informatics) and Csf1 (Kong et al., 1999; Yasuda et al., 1998). Osteoblasts, by contrast, are derived from mesenchymal cells; Runx2 is the master regulator of bone and cartilage cell fate, whereas osterix (Sp7 - Mouse Genome Informatics) is the master regulator of osteoblastogenesis (Nakashima et al., 2002). Additionally there are two types of bone: endochondral bone, in which osteoblasts lead to mineralisation of an existing cartilage matrix; and dermal bone, in which bone is formed de novo.

Until recently the majority of the work on bone development has been undertaken either in the mouse, or using mammalian in vitro cell-culture systems; however, as in many other cases in developmental biology, the zebrafish is an excellent model in which to study bone formation, particularly as transgenic lines allow us to follow the behaviour of cells in vivo. Despite some differences, e.g. the appendicular skeleton, zebrafish and the related teleost medaka show remarkable similarities in bone formation to higher vertebrates, particularly in their craniofacial development (Renn and Winkler, 2009; Wagner et al., 2003; Yelick et al., 1996; Yelick and Schilling, 2002).

The Hedgehog (Hh) signalling pathway, principally responding to Indian hedgehog, has long been linked to endochondral bone formation. Ihh is expressed by chondrocytes; signals to both chondrocytes and the adjacent perichondral cells and exerts control over the timing of chondrocyte differentiation (St-Jacques et al., 1999; Vortkamp et al., 1996). Analysis of the knockout mouse showed that Indian hedgehog regulates chondrocyte proliferation and is required for osteoblast differentiation (St-Jacques et al., 1999). More recently, the Hedgehog receptor protein patched 1 (patched homolog 1, Ptch1 - Mouse Genome Informatics) has been implicated in the regulation of both osteoblast and osteoclast fate in the mouse (Mak et al., 2008a; Ohba et al., 2008). Two papers analysing different Ptch1 mouse mutants reached somewhat different conclusions from their analyses. In mouse, Ptch1 null animals die before bone formation; thus only heterozygous carriers of the Ptch1 gene or conditional knockouts are available for analysis. The Chung group, studying Ptch1 heterozygous animals saw increased bone mass in these animals, which they attributed to increased osteoblast differentiation, through a mechanism whereby reduction of the Gli3 repressor led to increased sensitivity to Runx2 expression (Ohba et al., 2008). By contrast, the Yang group, analysing a conditional knockout, in which Ptch1 was deleted in cells expressing osteocalcin (Bglap1 - Mouse Genome Informatics), i.e. mature osteoblasts, saw reduced bone mass in the adult animals, which they attributed to increased differentiation of osteoclasts under the control of RANKL (Mak et al., 2008a).

As many of the murine Hedgehog signalling pathway mutants are embryonic lethal, we chose to undertake studies in the zebrafish, which have a number of advantages that make them ideal for studies of this nature. Importantly, zebrafish carrying null mutations for a number of members of the Hedgehog signalling pathway survive long enough to undertake an analysis of early bone differentiation. These include three mutants thought to lead to activation of the pathway: patched 1, patched 2 and dre (suppressor of fused) (sufu - ZFIN) (Koudijs et al., 2008; Koudijs et al., 2005) and an ihha mutant carrying an early stop, which we believe to be a null mutant. As an independent approach we also made use of drugs that act directly on smoothened to titrate the effects on the pathway. Cyclopamine is a small molecule antagonist that directly binds to smoothened and inhibits its function (Chen et al., 2002). Purmorphamine, by contrast, is a smoothened agonist (Sinha and Chen, 2006), which has been previously observed to induce osteogenic differentiation in cell culture in some studies (Wu et al., 2004), while inhibiting osteoblast differentiation in others (Plaisant et al., 2009). Thus different culture systems can give conflicting results on the effect of activation of smoothened on osteoblast precursors. We therefore sought to investigate the effect of manipulating smoothened activity in vivo in the zebrafish.

Together the study of these mutants and titration of smoothened activity allows us to build up a comprehensive picture of the effects both of an increase or depletion of Hh signalling on early bone formation in vivo. We show that, at least in the early stages of development, alteration of Hh signalling has no effect on dermal bone formation. However, in chondral bone elements Hh signalling critically controls both the differentiation of osteoblasts and the onset of osteoclast activity. In addition we demonstrate that many cells that are located within the cartilage element retain a level of plasticity that allows them, on receipt of high levels of Hh signalling, to differentiate as osteoblasts.

In situ hybridisation

In situ labelling was performed as previously described (Schulte-Merker, 2002). The markers used were osterix, osteocalcin, RANKL and collagen1alpha2 (for primers, see Table S1 in the supplementary material). In all cases in situs were carried out with mutant and siblings, analysed blind and subsequently genotyped by DNA isolation and sequencing, using the primers shown in Table S1 in the supplementary material.

Mutant lines

ptc1, ptc2 and dre mutant lines are ptc1hu1602, lep(ptc2)tj222 and dretm146d, respectively (Koudijs et al., 2008; Koudijs et al., 2005). Ihhahu2131 stocks were obtained from the Sanger Centre (Cambridge, UK). Lines were crossed to the transgenic lines Tg(osterix:nuGFP) line (Spoorendonk et al., 2008) or Tg(kdr-l:gfp)s843, originally referred to as Tg(flk1:EGFP)s843 (Jin et al., 2005).

Drug treatment

Cyclopamine (Sigma-Aldrich) was used at a concentration of 75 μM, added directly to the E3 medium in which the larvae were grown. Purmorphamine (Calbiochem) was used at a concentration of 20 μM. For both drugs, treatment began at 2 days post-fertilisation (dpf) to prevent gross effects on patterning or heart formation, and solutions were replaced approximately every 12 hours. Controls were incubated in E3, to which appropriate amounts of ethanol (cyclopamine) or DMSO (purmorphamine) were added.

BrdU labelling

BrdU labelling was performed as previously described (Kimmel et al., 1998). In short, BrdU was diluted to a working concentration of 3 mM in E3 medium. Embryos were incubated in this solution overnight and fixed subsequently.

Immunohistochemistry

Embryos were briefly fixed in 4% PFA and stored in MeOH. Embryos were rehydrated, blocked in PBS with 5% lamb serum and incubated with 1/100 anti-BrdU primary antibody (DAKO), anti-GFP (1/500 Torrey Pines Biolabs) or anti-Collagen II (1/500 DSHB) overnight at 4°C. Embryos were washed extensively then incubated in Alexa-Fluor secondary antibodies (Molecular probes) diluted 1/500 in blocking solution for 3 hours at room temperature. Embryos were washed extensively in the dark, with DAPI (Sigma-Aldrich) added at 1/1000 to one wash, then mounted for analysis. For DAB staining the secondary antibody was anti-mouse IgG-biotin, followed by incubation in ABC reagent (DAKO) and development of the DAB stain.

BAC transgenesis

mCherry was recombined directly after the ATG site of the Collagen 2a1 on a bacterial artificial chromosome (BAC) clone, using similar principles to those previously (Kimura et al., 2006). The BAC was CH73-184B14, containing around 39.5 kb upstream of Col2 and around 11 kb downstream. Primer sequences for cloning available on request.

Alcian Blue and Alizarin Red staining

Bone and cartilage labelling was performed as described previously (Spoorendonk et al., 2008; Walker and Kimmel, 2007).

TRAP staining

Tartrate-resistant acidic phosphatase (TRAP) staining was performed as described (Albertson and Yelick, 2005), with the following modifications: zebrafish were fixed in cold methanol (−20°C) overnight, rehydrated with PBS, and incubated in freshly made TRAP medium, for 1.5 hours at 37°C. Fish were subsequently bleached in 10% H2O2 for 4 hours and post-fixed in 4% PFA.

Microscopy

In situ hybridisations were analysed and photographed with a Leica 480C camera on a Zeiss Axioplan microscope. For cell number analyses in transgenic lines, images were captured on a Leica TCS-SPE confocal system, and the stacks were analysed and cells counted using Velocity software. In the cell-counting experiments, a minimum of four different individuals for each genotype were counted. Results are presented as mean±1 standard deviation (s.d.), significance was ascertained by performing two-tailed paired Student's t-tests of each data set to the wild-type situation.

Increased Hh signalling leads to premature chondral mineralisation

Recently two papers have implicated the Hedgehog membrane receptor patched 1 (Ptch1) in mouse bone development, but came to somewhat contradictory conclusions. The first, studying heterozygous deficiency for Ptch1, concluded that decreased Ptch1 leads to increased bone deposition (Ohba et al., 2008), whereas the other found that conditional knockout of Ptch1 in osteocalcin-expressing cells leads to decreased bone density and increased osteoclast activity (Mak et al., 2008a). Owing to embryonic lethality before the onset of bone mineralisation, it is impossible to resolve these issues in homozygous Ptch1 mutant mice. In zebrafish, however, ptc1 mutants can typically survive to around 12 dpf, and ptc2 mutants are even sub-viable, with small numbers reaching 3 months of age; as ptc1, ptc2, dre and ihha are expressed in positions where cartilage later forms (see Fig. S1A in the supplementary material) (Avaron et al., 2006; Thisse and Thisse, 2005), and as bone development in zebrafish is first apparent from 3 dpf, this gives us a window of opportunity in which to study bone development in a variety of Hh mutants.

In wild-type zebrafish, the first evidence of mineralisation detected by Alizarin Red staining is at 3 dpf in the cleithrum (data not shown). By 4 dpf the dermal bones, cleithrum, operculum and notochord tip are mineralised, and at this stage ptc1, ptc2 and dre mutants were indistinguishable from their wild-type siblings (Fig. 1A). However, by 7 dpf all three mutants could be distinguished by premature mineralisation of the ceratohyal and, in the case of ptc2 and dre, the hyosimplectic - both bones of chondral origin (Fig. 1A, black arrowheads).

Osterix expression is increased in endochondral bones

In order to better understand why an increase in Hh signalling leads to increased bone deposition in chondral bones, we performed in situ hybridisations for osterix. Osterix is a marker of early osteoblast development and is thought to be the master regulator of osteoblast differentiation (Nakashima et al., 2002). In ptc1, ptc2 and dre mutants at 3.5 dpf, expression of osterix mRNA expression was upregulated in the endochondral bones (Fig. 1B, black arrowheads) while remaining at wild-type levels in dermal bone elements, such as the cleithrum (Fig. 1B, red arrowheads).

Fig. 1.

Levels of Hh signalling are critical for endochondral ossification and proliferation of chondrocytes. (A) Alizarin Red staining of representative Hh mutants reveals no differences at 4 dpf. At 7 dpf, however, premature mineralisation of endochondral bone elements can be seen (black arrowheads mark the ceratohyal, purple arrowheads mark the hyosymplectic), while the extent of ossification in the cleithrum is unchanged (red arrowheads). (B) In situ analysis of mutants at 4 dpf reveals increased expression of osterix in the endochondral bone, where premature mineralisation is later seen (black arrowheads point to ceratohyal, purple to hyosymplectic), while osterix expression is unchanged or slightly reduced in dermal bone elements such as the cleithrum (red arrowheads). Insets show ventral views of osterix expression in the ceratohyal. (C) Alcian Blue/Alazarin Red double staining of 17 dpf ihha and wild-type sibling fish. The second panel is an enlargement of the boxed area for each genotype. (D) Whole-mount antibody staining for BrdU incorporation between 4.5 and 5 dpf in wild type and ptc2 mutants. Excessive proliferation can be seen in the ptc2 mutant in the tectum (black arrowhead) and jaw cartilages (blue arrowhead). (E) Representative single confocal images of the Meckel's cartilage in 5 dpf zebrafish. Proliferating chondrocytes (blue arrowheads) are labelled with anti-BrdU (green), the dashed red line shows the shape of a single chondrocyte. (F) Quantitation of BrdU-positive cells in the Meckel's cartilage at 5 dpf. Data are mean±s.d. taken of at least five fish per genotype. *P<0.01 versus wild type. wt, wild type.

Fig. 1.

Levels of Hh signalling are critical for endochondral ossification and proliferation of chondrocytes. (A) Alizarin Red staining of representative Hh mutants reveals no differences at 4 dpf. At 7 dpf, however, premature mineralisation of endochondral bone elements can be seen (black arrowheads mark the ceratohyal, purple arrowheads mark the hyosymplectic), while the extent of ossification in the cleithrum is unchanged (red arrowheads). (B) In situ analysis of mutants at 4 dpf reveals increased expression of osterix in the endochondral bone, where premature mineralisation is later seen (black arrowheads point to ceratohyal, purple to hyosymplectic), while osterix expression is unchanged or slightly reduced in dermal bone elements such as the cleithrum (red arrowheads). Insets show ventral views of osterix expression in the ceratohyal. (C) Alcian Blue/Alazarin Red double staining of 17 dpf ihha and wild-type sibling fish. The second panel is an enlargement of the boxed area for each genotype. (D) Whole-mount antibody staining for BrdU incorporation between 4.5 and 5 dpf in wild type and ptc2 mutants. Excessive proliferation can be seen in the ptc2 mutant in the tectum (black arrowhead) and jaw cartilages (blue arrowhead). (E) Representative single confocal images of the Meckel's cartilage in 5 dpf zebrafish. Proliferating chondrocytes (blue arrowheads) are labelled with anti-BrdU (green), the dashed red line shows the shape of a single chondrocyte. (F) Quantitation of BrdU-positive cells in the Meckel's cartilage at 5 dpf. Data are mean±s.d. taken of at least five fish per genotype. *P<0.01 versus wild type. wt, wild type.

Loss of Indian hedgehog leads to a severe retardation of endochondral mineralisation

In mouse loss of Indian hedgehog (Ihh) leads to a reduction in chondrocyte proliferation and a failure of osteoblast differentiation in endochondral bones (St-Jacques et al., 1999), and further studies have shown Ihh to promote chondrocyte hypertrophy before bone formation (Mak et al., 2008b). The zebrafish ihhahu213allele contains a premature stop codon at amino acid 44, and is thus likely to be a functional null. We crossed ihhahu2131 carriers and examined bone development at a variety of stages (Fig. 1C and data not shown). At stages before the onset of endochondral mineralisation, no difference between siblings and ihhahu2131 mutants was seen. However, from 8 dpf the mutants were distinguished by a near-complete lack of endochondral ossification. This became more pronounced over time, until 17 dpf, by which point all branchial arches were mineralised in siblings, whereas in the mutant only very small areas of endochondral mineralisation could be detected (Fig. 1C). Surprisingly, the mutant fish could survive and from 21 dpf began to mineralise all bone elements, perhaps owing to partial redundancy between the ihha and ihhb genes, such that by 30 dpf the fish were almost indistinguishable in the extent of their mineralisation from siblings, although the fish were frequently much smaller (data not shown). These results show that, in the zebrafish ihha is crucial for timely onset of endochondral mineralisation, analogous to the situation in mammals.

Loss of ptc1 or ptc2 leads to increased, and loss of ihha leads to decreased, chondrocyte proliferation

Hh signalling has long been implicated in maintaining the balance between proliferation and differentiation of a number of cell types (Agathocleous et al., 2007). Indeed, the zebrafish ptc2 and dre mutants were identified in a screen for changes in proliferation (Koudijs et al., 2005). We therefore decided to study proliferation in the jaw cartilages in the various Hh mutants. In addition to the previously reported increased proliferation in the ciliary marginal zones and tecta of the ptc2 mutants (Fig. 1D, black arrowhead), we detected a significant increase in proliferation in the jaw cartilages at all stages studied (Fig. 1D-F and data not shown), whereas in ihha mutants we saw a reduction in the levels of proliferation in the cartilage (Fig. 1F and data not shown), mirroring the situation in the mouse (St-Jacques et al., 1999).

Fig. 2.

Mutants with altered Hh signalling show no differences in the number of osterix-positive cells in dermal bone elements. (A) Confocal stacks of the cleithra of osterix-nuclear GFP fish at 4 dpf. Inset image on left is a merged image of Tg(osx:nuGFP) (green), Alazarin Red and brightfield images. (B) Confocal stacks of the operculum of osterix-nuclear GFP fish at 4 dpf. Inset image on left is a merged image of Tg(osx:nuGFP) (green), Alazarin Red and brightfield images. (C) Quantification of the number of osterix-positive nuclei in the cleithrum (below) and operculum (above) at 3 and 4 dpf. Data is shown as mean±s.d. taken of at least four fish per genotype. *P<0.05 versus wild type. wt, wild type.

Fig. 2.

Mutants with altered Hh signalling show no differences in the number of osterix-positive cells in dermal bone elements. (A) Confocal stacks of the cleithra of osterix-nuclear GFP fish at 4 dpf. Inset image on left is a merged image of Tg(osx:nuGFP) (green), Alazarin Red and brightfield images. (B) Confocal stacks of the operculum of osterix-nuclear GFP fish at 4 dpf. Inset image on left is a merged image of Tg(osx:nuGFP) (green), Alazarin Red and brightfield images. (C) Quantification of the number of osterix-positive nuclei in the cleithrum (below) and operculum (above) at 3 and 4 dpf. Data is shown as mean±s.d. taken of at least four fish per genotype. *P<0.05 versus wild type. wt, wild type.

The number of osterix-expressing osteoblasts is independent of Hedgehog signalling levels in dermal bones

To investigate the effects of Hh signalling on osteoblast numbers, we crossed ptc1, ptc2 and ihha carriers to a transgenic line in which nuclear GFP is fused to the medaka osterix promoter (Renn and Winkler, 2009; Spoorendonk et al., 2008), allowing us to count the number of osterix-expressing cells in different bone elements.

Comparison of the number of osterix-expressing cells in dermal bone elements such as the cleithrum (Fig. 2A) and the operculum (Fig. 2B) between the different lines revealed no significant differences at the stages examined (Fig. 2C), although in ihha mutants at later stages the size of the operculum was reduced despite the number of cells remaining the same (see Fig. S2A in the supplementary material), perhaps suggesting that the osteoblasts are less active. However, in a number of the ptc1 mutants ectopic tissue between the eyes was marked by osterix (see Fig. S2B in the supplementary material) and subsequently mineralised (data not shown). We conclude that Hh signalling levels are not critical for numbers of early dermal bone osteoblasts.

Two distinct populations of osterix-expressing cells in endochondral bone with different sensitivity to Hedgehog signalling

These data suggested a differential effect of Hh signalling on dermal versus endochondral bone. In order to better understand the way in which altered Hh signalling was influencing endochondal osterix expression, we investigated the ceratohyal and Meckel's cartilage, which mineralise prematurely in mutants with increased Hh signalling. We observed that there are two different populations of osterix-expressing cells associated with these cartilage elements. The first appears to be the equivalent of the mammalian bone collar. We see this population as a host of cells strongly expressing osterix and surrounding the edge of the cartilage element (red arrowheads in Fig. 3A-D,F). These cells were first seen at 5 dpf in wild type, ptc1 and ptc2 mutants but not in ihha-/-, and increased in number over days 6-7. In the ptc1 and ptc2 mutants a slight increase in the number of these cells is seen at 5 dpf but by 6 dpf the difference in number was no longer significant (Fig. 3F). The second population of osterix-expressing cells retain chondrocyte morphology and are interspersed with other cells in the cartilage elements (blue arrows in Fig. 3A,C,D,E). This second population of cells was first seen in ptc mutants at 3 dpf (Fig. 3C). Significantly in siblings the first time the equivalent cells were observed was 6 dpf when around half of the siblings had one or two osterix-positive cells (Fig. 3A,E). At this stage these cells were present in significantly higher numbers in the ptc mutants (blue arrows in Fig. 3A,E). By contrast, the first stage in which multiple osterix-positive cells were reliably seen in siblings was 9 dpf (Fig. 3D). The presence of osteoblasts within the endochondral bone around 3 days earlier in mutants than siblings corresponds with the position and timing of endochondral ossification (detected by Alizarin Red) in the different fish genotypes (Fig. 3 and Fig. 1A).

The number of both populations of osterix-expressing osteoblasts is significantly decreased in ihha mutants in endochondral bones

We also counted the number of both populations of osterix-positive cells in ihha mutants. We found that both populations of cells were absent in all stages studied, to 9 dpf (Fig. 3E,F and data not shown); by which time osteoblasts were populating the maxilla, a dermal bone that lies closely behind the Meckel's cartilage (white arrows in Fig. 3E). Again, the lack of both types of osterix-expressing cells corresponds with the failure to make endochondral bone in these mutants. From this we conclude that differentiation of the bone collar cells also require Hh signalling, as although their numbers were unaffected in the mutants with increased Hh signalling they were absent in the ihha mutant. However, as the morphology of the bone collar cells is similar to other cells in the vicinity, we could not distinguish between loss of the cells and presence of the cells but loss of the markers.

The changes to the number of both populations of cells can be phenocopied by titrating smoothened activity by using small molecule inhibitors and activators

To demonstrate the dose dependence for Hedgehog signalling on the differentiation of the different populations of osteoblasts, we used drugs to titrate smoothened activity. Cyclopamine is a plant-derived alkyloid demonstrated to bind smoothened and block its activity (Chen et al., 2002), whereas purmorphamine also binds smoothened but leads to its activation (Sinha and Chen, 2006). Addition of cyclopamine (75 μm) to ptc1,2 double heterozygous incrosses led to rescue of the premature mineralisation (data not shown) and reduction of the number of chondrocytes expressing osterix to near wild-type levels (Fig. 3H) while not significantly altering the number of bone collar osteoblasts in mutants or siblings (Fig. 3H).

Fig. 3.

There are two populations of endochondral osteoblasts with differential sensitivity to Hh signalling. (A) Confocal stacks from wild-type, ptc1-/- and ptc2-/- 6 dpf larvae in an osterix-nuclear GFP background, imaged for GFP expression. The far right image is a single confocal plane to allow distinstintion of the cell morphology. At this stage in both ptc mutants and wild-type larvae, similar numbers of bone collar osteoblasts are present (red arrowheads). However, in ptc1-/- and ptc2-/- larvae, there is a significant increase in the number of chondrocytes that express osterix (blue arrows). A single chondrocyte is outlined in black in the image of the ptc1 mutant. (B) A ventral view of one side of the Meckel's cartilage and ceratohyal of a 6 dpf wild-type larva stained with Alcian Blue and Alizarin Red; a single chondrocyte is outlined in black to demonstrate the morphology of the chondrocyte cells. (C) Confocal stacks of the ceratohyal from ptc1-/- and wild-type embryos at 3 dpf. In ptc1-/- embryos (but not in wild-type siblings) cells can be found that retain the shape of chondrocytes but that express osterix (blue arrow). (D) Confocal stack of the Meckel's cartilage of a 9 dpf wild type. Expression of osterix is visible in cells within the cartilage element (blue arrows) and outside (red arrowhead). (E) Confocal stacks of the Meckel's cartilage of wild-type and ihha-/- 9 dpf larvae. By this stage the wild-type embryo has osterix- expressing bone collar cells (red arrowhead) and transdifferentiating chondrocytes (blue arrows), while the ihha-/- larvae has neither. White arrowheads mark the maxilla, a dermal bone, which forms adjacent to the Meckel's cartilage. (F) Quantification of number of osteoblasts at 6 dpf showing number of osteoblasts in the bone collar of the Meckle's cartilage and the number of chondrocytes differentiating as osteoblasts in the Meckel's cartilage. Data are presented as means±s.d. taken from at least six fish per genotype; asterisk represents P<0.05 in paired Student's t-tests of mutant versus wild type. (G) Quantification of osteoblast number at 6 dpf in the Meckel's cartilage of embryos treated with 20 μm purmorphamine or controls. Data are presented as means±s.d. taken from at least four fish per genotype; * represents P<0.05 in paired Student's t-tests of treated versus untreated. (H) Quantification of osteoblast number at 6 dpf in the Meckel's cartilage of embryos treated with 75 μM cyclopamine or controls. Data are presented as means±s.d. taken from at least four fish per genotype; * represents P<0.05 in paired Student's t-tests of treated versus untreated. wt, wild type.

Fig. 3.

There are two populations of endochondral osteoblasts with differential sensitivity to Hh signalling. (A) Confocal stacks from wild-type, ptc1-/- and ptc2-/- 6 dpf larvae in an osterix-nuclear GFP background, imaged for GFP expression. The far right image is a single confocal plane to allow distinstintion of the cell morphology. At this stage in both ptc mutants and wild-type larvae, similar numbers of bone collar osteoblasts are present (red arrowheads). However, in ptc1-/- and ptc2-/- larvae, there is a significant increase in the number of chondrocytes that express osterix (blue arrows). A single chondrocyte is outlined in black in the image of the ptc1 mutant. (B) A ventral view of one side of the Meckel's cartilage and ceratohyal of a 6 dpf wild-type larva stained with Alcian Blue and Alizarin Red; a single chondrocyte is outlined in black to demonstrate the morphology of the chondrocyte cells. (C) Confocal stacks of the ceratohyal from ptc1-/- and wild-type embryos at 3 dpf. In ptc1-/- embryos (but not in wild-type siblings) cells can be found that retain the shape of chondrocytes but that express osterix (blue arrow). (D) Confocal stack of the Meckel's cartilage of a 9 dpf wild type. Expression of osterix is visible in cells within the cartilage element (blue arrows) and outside (red arrowhead). (E) Confocal stacks of the Meckel's cartilage of wild-type and ihha-/- 9 dpf larvae. By this stage the wild-type embryo has osterix- expressing bone collar cells (red arrowhead) and transdifferentiating chondrocytes (blue arrows), while the ihha-/- larvae has neither. White arrowheads mark the maxilla, a dermal bone, which forms adjacent to the Meckel's cartilage. (F) Quantification of number of osteoblasts at 6 dpf showing number of osteoblasts in the bone collar of the Meckle's cartilage and the number of chondrocytes differentiating as osteoblasts in the Meckel's cartilage. Data are presented as means±s.d. taken from at least six fish per genotype; asterisk represents P<0.05 in paired Student's t-tests of mutant versus wild type. (G) Quantification of osteoblast number at 6 dpf in the Meckel's cartilage of embryos treated with 20 μm purmorphamine or controls. Data are presented as means±s.d. taken from at least four fish per genotype; * represents P<0.05 in paired Student's t-tests of treated versus untreated. (H) Quantification of osteoblast number at 6 dpf in the Meckel's cartilage of embryos treated with 75 μM cyclopamine or controls. Data are presented as means±s.d. taken from at least four fish per genotype; * represents P<0.05 in paired Student's t-tests of treated versus untreated. wt, wild type.

By contrast, treatment of wild-type embryos with 20 μM purmorphamine from 2 dpf to 6 dpf led to a significant increase in the number of osterix-positive cells located within the collagen matrix (Fig. 3G and Fig. 4C), without a significant concomitant increase in the number of bone collar osteoblasts, suggesting that there may be a maximum number of these cells, which is already reached in the mutants. Treatment of ihha mutants from 2 to 6 dpf with 20 μm purmorphamine led to a near-complete rescue of bone collar osteoblast number to wild-type levels (Fig. 3G and Fig. 4C); however, the effect of this dose of purmorphamine in the mutants led to only slight increases in the number of osteoblasts within the cartilage matrix (Fig. 3G and Fig. 4C).

Fig. 4.

The osterix-expressing cells with chondrocyte morphology express markers of chondrocyte and osteoblast differentiation. (A) Confocal stacks taken of the ceratohyal of 5 dpf embryos stained for collagen II in red, GFP (osterix) in green and DAPI in blue. In wild-type embryos osteoblasts can be seen lying on the surface of the collagen II-positive matrix (white arrows). In ptc1 mutant embryos, similar cells can be seen (white arrows), and also osterix-positive cells, which are completely surrounded by collagen II-positive matrix can be seen (blue arrows). In ptc1/2 double mutants the cartilage structure is disorganised, but all cells surrounded by collagen II-positive matrix are also positive for osterix (blue arrows). In the ihha-/- mutants, neither population of osterix-positive cells are seen, despite the normal appearance of the collagen II-positive cartilage matrix. (B) Confocal stack from a 4 dpf ptc2 mutant osterix-nuGFP larva, injected with a BAC containing mCherry inserted after the ATG of collagen2a1, leading to mosaic expression through the cartilage. White arrows mark cells expressing osterix, which do not contain col2 mCherry. Yellow arrow marks a cell expressing col2 which is negative for osterix, whereas blue arrows mark cells expressing both mCherry and GFP. (C) Confocal stacks from the ceratohyal of collagen 2 (red) and anti-GFP osterix (green) antibody stained 6 dpf larvae treated from 2-6 dpf with purmorphamine. White arrows indicate bone collar osteoblasts overlying the cartilage matrix. Blue arrows mark cells expressing osterix, which are surrounded by cartilage matrix. (D) Col1a2 in situ hybridisation in 9 dpf wild-type embryo, showing the palatoquadrate. Two populations of col1a2 expressing-osteoblasts can be seen. One population is located outside but adjacent to the cartilage matrix (black arrows), whereas the other comprises cells located within the cartilage matrix (red arrows). (E) Ventral views of 5 dpf embryos in situ hybridised for osteocalcin, anterior to right. In all genotypes osteocalcin staining can be seen in the teeth (black arrows). Particularly strong staining in the osteoblasts of the bone collar (red arrowheads) is present in wild-type and ptc2-/-, embryos but is absent for ihha-/-. In ptc2 mutant embryos weaker osteocalcin expression can be seen in cells within the cartilage (blue arrows), which are not seen in wild-type embryos at this stage. (F) High magnification image of Meckel's cartilage from a ptc2-/- 5 dpf embryo showing osteocalcin expression in cells within the cartilage element (blue arrows). wt, wild type.

Fig. 4.

The osterix-expressing cells with chondrocyte morphology express markers of chondrocyte and osteoblast differentiation. (A) Confocal stacks taken of the ceratohyal of 5 dpf embryos stained for collagen II in red, GFP (osterix) in green and DAPI in blue. In wild-type embryos osteoblasts can be seen lying on the surface of the collagen II-positive matrix (white arrows). In ptc1 mutant embryos, similar cells can be seen (white arrows), and also osterix-positive cells, which are completely surrounded by collagen II-positive matrix can be seen (blue arrows). In ptc1/2 double mutants the cartilage structure is disorganised, but all cells surrounded by collagen II-positive matrix are also positive for osterix (blue arrows). In the ihha-/- mutants, neither population of osterix-positive cells are seen, despite the normal appearance of the collagen II-positive cartilage matrix. (B) Confocal stack from a 4 dpf ptc2 mutant osterix-nuGFP larva, injected with a BAC containing mCherry inserted after the ATG of collagen2a1, leading to mosaic expression through the cartilage. White arrows mark cells expressing osterix, which do not contain col2 mCherry. Yellow arrow marks a cell expressing col2 which is negative for osterix, whereas blue arrows mark cells expressing both mCherry and GFP. (C) Confocal stacks from the ceratohyal of collagen 2 (red) and anti-GFP osterix (green) antibody stained 6 dpf larvae treated from 2-6 dpf with purmorphamine. White arrows indicate bone collar osteoblasts overlying the cartilage matrix. Blue arrows mark cells expressing osterix, which are surrounded by cartilage matrix. (D) Col1a2 in situ hybridisation in 9 dpf wild-type embryo, showing the palatoquadrate. Two populations of col1a2 expressing-osteoblasts can be seen. One population is located outside but adjacent to the cartilage matrix (black arrows), whereas the other comprises cells located within the cartilage matrix (red arrows). (E) Ventral views of 5 dpf embryos in situ hybridised for osteocalcin, anterior to right. In all genotypes osteocalcin staining can be seen in the teeth (black arrows). Particularly strong staining in the osteoblasts of the bone collar (red arrowheads) is present in wild-type and ptc2-/-, embryos but is absent for ihha-/-. In ptc2 mutant embryos weaker osteocalcin expression can be seen in cells within the cartilage (blue arrows), which are not seen in wild-type embryos at this stage. (F) High magnification image of Meckel's cartilage from a ptc2-/- 5 dpf embryo showing osteocalcin expression in cells within the cartilage element (blue arrows). wt, wild type.

The osterix-positive cells located in the cartilage are surrounded by collagen II-positive matrix, and go on to express bone markers

To investigate whether this second population of endochondral osteoblasts displayed molecular differences to the bone collar osteoblasts in addition to their morphological differences, we studied a number of markers of cartilage and bone. Type II collagen is the main constituent of cartilage matrix (Mackie et al., 2008), we therefore stained the various mutants for collagen type II and GFP to show the osterix-positive cells. Again, it was clear that there are two populations of cells: one entirely enclosed by the cartilage matrix, and the other on the surface (Fig. 4A). Maximal activation of the Hedgehog signalling pathway is seen in embryos that are mutant for both ptc1 and ptc2 (Koudijs, 2008). Ptc1/2 double mutants survived only to 4 dpf and had gross malformation of the head, such that individual bone elements could not be distinguished; however, by 4 dpf there were disorganised collagen II-positive condensations (Fig. 4A). Staining for osterix-positive cells revealed that every cell in the collagen II-positive condensation also expressed osterix (Fig. 4A).

To ascertain whether the cells encircled by cartilage matrix were themselves secreting it, we made BAC constructs in which the mCherry fluorescent protein is driven by the promoter of collagen2a1. Injection of this construct led to mosaic expression of mCherry in the collagen2-expressing chondrocytes. When injected into a ptc2 heterozygous incross, cells could be seen in cartilage elements that contained both GFP and cherry in mutant embryos (Fig. 4B). This shows that the osterix-positive cells located within the cartilage matrix themselves express collagen2, making it unlikely that the matrix surrounding these cells is derived solely from neighbouring cells.

Because in rare cases hypertrophic chondrocytes can express Osterix (Shibata et al., 2006; Shibata and Yokohama-Tamaki, 2008), we also studied later markers of osteoblast maturation that do not overlap with hypertrophic chondrocytes. One such marker is osteocalcin, which is exclusively expressed in osteoblasts (Hoffmann et al., 1996). We assayed osteocalcin expression in mutants and saw that in ptc1 and ptc2 mutants in addition to strong osteocalcin expression in the teeth (Fig. 4E, black arrows) and bone collar (Fig. 4E, red arrowheads), there was expression in cells in the cartilage elements (Fig. 4E,F, blue arrows). Further supporting this concept that these chondrocytes are differentiating as osteoblasts we also saw expression of type 1 collagen, another marker specific to osteoblasts, both in cells of the bone collar and in this population of cells in the cartilage (Fig. 4D and data not shown), demonstrating that these cells are indeed mature osteoblasts. These results taken together demonstrate that cells that are differentiating as chondrocytes retain a degree of plasticity that allows them, on receipt of high levels of Hh, to differentiate as osteoblasts.

Fig. 5.

Osteoclasts are active earlier in ptc mutants and later in ihha mutants than in the wild-type situation. (A) Whole mount stained for tartrate resistant acid phosphatases (TRAP). Activity is first present in the fifth branchial arch and in teeth (black arrows). Additionally, by 14 dpf activity can be seen in the Meckel's cartilage (blue arrow in inset, region enlarged is marked by a dashed box) and operculum (red arrow) in the ptc2-/- mutant. Dashed box in the top right image denotes the region shown for 10 and 12 dpf. (B-D) Close-up views of the regions of TRAP activity: (B) dissected teeth and fifth branchial arch, (C) neural arch, (D) tail fin. wt, wild type.

Fig. 5.

Osteoclasts are active earlier in ptc mutants and later in ihha mutants than in the wild-type situation. (A) Whole mount stained for tartrate resistant acid phosphatases (TRAP). Activity is first present in the fifth branchial arch and in teeth (black arrows). Additionally, by 14 dpf activity can be seen in the Meckel's cartilage (blue arrow in inset, region enlarged is marked by a dashed box) and operculum (red arrow) in the ptc2-/- mutant. Dashed box in the top right image denotes the region shown for 10 and 12 dpf. (B-D) Close-up views of the regions of TRAP activity: (B) dissected teeth and fifth branchial arch, (C) neural arch, (D) tail fin. wt, wild type.

Interestingly, in situ hybridisations for gli transcription factors showed increased expression of both gli1 and gli3 in chondro-progenitors of mutants with increased Hh signalling at 3 dpf, and a slight reduction in expression of the repressive gli2 (see Fig. S1B in the supplementary material). This shift in favour of activator forms such as gli1 over repressive factors such as gli2 bears similarities to mouse, in which altered processing of Gli3 to its activator form was seen (Ohba et al., 2008). Taken together, these data suggest that the two populations of osterix-expressing cells require different levels of Hh signalling for their differentiation: bone collar cells require low levels of ihh, whereas the chondro-osteoblasts require higher levels of the Hh signal.

The increased mineralisation in ptc1 and ptc2 mutants cannot be accounted for by a deficit in the differentiation of osteoclasts

As bone deposition and modelling is a balance in the activities of both osteoblasts and osteoclasts, there is an alternative hypothetical explanation for the increase in mineralisation: namely a decrease or a delay in osteoclast differentiation. In the mouse conditional Ptch1 knockout in osteocalcin-expressing cells, decreased bone density was observed and attributed to increased osteoclastogenesis (Mak et al., 2008a). We sought to test whether this was the case in fish. In zebrafish two different populations of osteoclasts are believed to exist: a mononuclear population forms earlier than the multinucleated osteoclasts; both populations have been described as forming relatively late in development between 12 and 20 dpf (Witten et al., 2001). To understand the effect Hh signalling has on osteoclast activity, we performed TRAP staining on mutants at a variety of ages, from 8 to 14 dpf (Fig. 5 and data not shown). In accordance with published literature, in wild-type larvae we first observed osteoclast activity at 12 dpf in the fifth branchial arch, at the position where the teeth join the bone (Fig. 5A,B). In the ptc2 mutant we could consistently observe TRAP activity from 10 dpf, indicating that by this stage osteoclasts are already differentiated and active in the mutant (Fig. 5A). Over time both the levels of activity and the number of bone elements displaying osteoclast activity increased, such that at 14 dpf in the ptc2 mutants activity was observed in the Meckel's cartilage (blue arrow in Fig. 5A), the neural arches (Fig. 5C) and the tail fin (Fig. 5D). By contrast, in ihha mutants TRAP activity was not seen until 14 dpf (Fig. 5A). Thus increased systemic Hedgehog activity, in addition to promoting osteoblastogenesis, also promotes increased osteoclast activity.

Fig. 6.

Model illustrating the different populations of osterix-expressing cells during cartilage ossification in situations of normal, high or low levels of hedgehog signalling during zebrafish development.

Fig. 6.

Model illustrating the different populations of osterix-expressing cells during cartilage ossification in situations of normal, high or low levels of hedgehog signalling during zebrafish development.

Interestingly, TRAP activity was also seen in the operculum, a dermal bone, earlier in the ptc2 mutant than in siblings or wild-type fish (Fig. 5A, red arrow), despite there being no difference in the number of dermal bone osteoblasts at early stages between the genotypes (Fig. 2). This suggests that the effect on Hh signalling on osteoclastogenesis is direct, and is independent of osteoblast number. Additionally, we studied adult fish heterozygous for ptc1, ptc2 or both. In contrast to mouse (Ohba et al., 2008), there were no obvious morphological or skeletal changes in the heterozygous fish compared to siblings (data not shown); this could be because fish are less dose-sensitive to loss of ptc1 and 2, or because zebrafish do not have trabecular bone, in which the greatest differences were observed in mice (Ohba et al., 2008). We also generated an in situ probe to RANKL (TNFSF11), and assayed expression levels in the various mutants; however, we could not detect expression in or around bone elements before 8 dpf (data not shown) consistent with our TRAP staining, which suggests that osteoclast formation begins around 10-12 dpf in the zebrafish. We could, however, detect expression in the brain from 6 dpf (data not shown). We therefore conclude that differences in RANKL expression are not responsible for the early changes to bone deposition seen in the hedgehog mutants.

The difference in the number of osteoblasts in endochondral bone elements cannot be explained by increased vascularisation of the cartilage elements

One hypothesis for how osteoblasts come to be in endochondral bone is that they are brought in by blood vessels that invade the cartilage. In some cases anti-angiogenic agents can block ectopic ossification (Mori et al., 1998). To test whether the increased mineralisation and increased osteoblast number in the endochondral bones in ptc mutants is preceded by increased or premature vascularisation of the cartilage, we crossed the ptc mutants into the kdr-l:eGFP transgenic line that marks blood vessels (Jin et al., 2005). In neither ptc1 nor ptc2 mutants, ectopic or premature vascular sprouting was seen when compared to wild-type embryos (see Fig. S3A in the supplementary material). Moreover, in the ptc1 mutants, vascularisation close to the endochondral bone mineralisation was actually somewhat reduced, with the gill vessels failing to loop around the branchial arches (see Fig. S3A in the supplementary material). Indeed, the first cartilage elements to mineralise were not closely associated with any blood vessels at onset of osteoblast appearance (see Fig. S3B in the supplementary material). This result, taken together with the previous data, strongly suggests that the osteoblasts that contribute to premature endochondral ossification in ptc mutants derive from chondrocytes that then differentiate as osteoblasts in situ; we have summarised the data in a model (Fig. 6).

Our data demonstrate that there are two populations of osteoblasts in zebrafish that make endochondral bone: those lying outside the cartilage matrix, which we refer to as the bone collar, and those that lie within the cartilage, have chondrocyte morphology and on receipt of Hh signalling express markers of osteoblast differentiation. A key finding of our study is that the two populations of osteoblasts have differential sensitivity to Hh signalling. Osteoblasts of the bone collar require Hh signalling, as they do not form in the ihha mutant and their numbers are reduced following cyclopamine treatment. However, they are less sensitive to increased Hh signalling than the latter population. We presume that for the osteo-chondrocyte precursors a certain threshold of Hh signalling needs to be reached before they are able to differentiate as osteoblasts.

As chondrocytes mature, they cease to proliferate and differentiate into prehypertrophic chondrocytes expressing Ihh (Kobayashi et al., 2005; Maeda et al., 2007). In zebrafish, ihh expression is readily detectable at the relevant stages in chondrocytes (Avaron et al., 2006). We suggest that during wild-type chondrocyte maturation these cells begin to secrete Ihh, which, upon reaching a certain threshold, acts both in an autocrine and paracrine manner to allow a number of these cells and those surrounding them to differentiate as osteoblasts. Consistent with this idea, in ptc mutants, which are sensitised to Hh signalling, these cells express osteoblast markers earlier and in larger numbers than in siblings. An alternative model postulates hypertrophic chondrocytes to secrete a signal that induces osteoblast differentiation and bone formation close to the position of the hypertrophic chondrocytes (Chung, 2004). We would argue that, a number of the chondrocytes themselves begin to express markers of osteoblast differentiation such as osterix and osteocalcin. This, indeed, appears also to be true in other teleosts (Gavaia et al., 2006), and the amphibian Xenopus tropicalis, in which weak expression of osterix is seen in the lacuna of chondrocytes at the same time as the surrounding bone collar forms (Miura et al., 2008). In amniotes a number of chondrocytes display properties typically seen in osteoblasts (Galotto et al., 1994). Furthermore, there is also evidence that the hypertrophic chondrocytes located adjacent to the perichondrum differentiate as osteoblasts and contribute to the earliest endochondral bone matrix (Bianco et al., 1998).

Lending further credence to the idea that some osteoblasts develop from the cartilage itself are a number of results, including the fact that mice null for Runx2 fail to make bone (Kim et al., 1999; Otto et al., 1997). In situations in which Runx2 is knocked out in cells of the chondrocyte lineage endochondral bone formation is severely delayed (Ueta et al., 2001), whereas overexpression of Runx2 in cells of the same lineage leads both to rescue of the Runx2 null phenotype and to premature ossification of cartilage elements; showing that expression of Runx2 in cells of the chondrocyte lineage alone can re-establish endochondral bone formation in a cell-autonomous fashion (Takeda et al., 2001). This argues against a mechanism by which all osteoblasts are brought in by the blood vessels, and is suggestive of a mechanism by which some of the osteoblasts differentiate in situ. We do not debate the close physiological connection between cartilage and blood vessels. Vasculogenesis and endochondral bone formation are closely linked, with hypertrophic chondrocytes able to stimulate vasculogenesis via Vegf, and Vegf itself also able to stimulate proliferation of osteoblasts and osteogenesis (Takeda et al., 2001; Wang et al., 2007). Moreover, Runx2 mutant mice do not express Vegf in the cartilage, a feature that is fully revertable upon introduction of Runx2, demonstrating that Runx2 regulates cartilage expression of Vegf (Zelzer et al., 2004; Zelzer and Olsen, 2005). However, from our data it is clear that the presence of blood vessels within the cartilage itself is not a prerequisite for all endochondral osteogenesis, although proximity to vasculature to maintain normoxic conditions is likely to be a requirement for normal bone formation.

In the mouse, two recent studies, both interfering with Ptch1 levels and focusing on bone homeostasis, led to somewhat contradictory results. Ohba et al. (Ohba et al., 2008) reported increased bone deposition through increased osteogenesis upon lowering Ptch1 levels by 50%, whereas Mak et al. (Mak et al., 2008a) showed osteopenia and reduced bone mass through stimulation of osteoclastogenesis. Our studies in zebrafish help in clarifying the role of Ptch and Hh signalling, as we can study the effect of systemic loss of ptc1 and ptc2 and the titration of smoothened activity. We show that increased Hh signalling promotes the formation of both osteoblasts and osteoclasts, confirming the main results from each paper. Our data, though, more closely reflect the findings of Ohba et al. (Ohba et al., 2008), who showed an increase in bone mass upon increased Hedgehog signalling, despite the fact that we did not see any obvious defects in heterozygotes. We believe that by knocking out Ptch1 only in mature osteoblasts Mak et al. (Mak et al., 2008a) would not have observed the Hh-sensitive differentiation of a population of chondrocytes into osteoblasts. In order to see this it would require Ptch1 to be conditionally knocked out under the control of an earlier marker expressed in both osteoblasts and chondrocytes such as Runx2. Thus, we suspect that their model would underestimate the number (and activity) of mature osteoblasts that would be present in a mouse with systemic loss of Ptch1. We believe that our findings help to explain the discrepancies between the two papers.

It appears that the increased activity of osteoclasts upon increased Hh signalling is not limited to endochondral bone but also occurs in dermal bones such as the opercula, where we see no difference in osteoblast number. Thus, we conclude that misregulation of Hh leads to a situation where the number of osteoblasts and osteoclasts are no longer proportional, particularly in dermal bone, in which osteoblast activity is increased with no concomitant increase in osteoblast number. This is particularly interesting in the context of the recent finding that Ihh is expressed during dermal bone development in the mouse (Abzhanov et al., 2007). As a consequence, the effect on dermal bones is likely to be a net loss of bone mass over time, as seen with osteopenia observed in membranous bones as well as endochondral ones in the mouse (Mak et al., 2008a).

In summary we demonstrate in a number of different ways that systemically increased levels of Hh signalling lead both to increased differentiation of osteoblasts from cells of a chondrogenic lineage, and to increased activity from osteoclasts, confirming and extending the findings from mice (Mak et al., 2008a; Ohba et al., 2008). Importantly we show that not only does loss of ptc1 cause this phenotype but also a number of other manipulations that increase Hh signalling, demonstrating that the effects seen in mice are not due to intrinsic effects of Ptch1 but rather to a general increase in Hh signalling. We unite these results with experiments in ihha mutants and cyclopamine-treated larvae, which suggest that the effects of increased craniofacial endochondral osteoblast and osteoclastogenesis are likely to be mediated via increased Indian hedgehog signalling.

Many thanks to Josi Peterson-Maduro for her help and expertise in cloning the BAC transgenic, to members of the Schulte-Merker laboratory and C. Winkler for helpful discussions, and to the Sanger Institute for providing the Ihhahu2131 mutant fish. This work was supported by an EMBO long-term fellowship (C.L.H.). S.S.-M. gratefully recognizes the support of the Smart Mix Programme of the Netherlands Ministry of Economic Affairs and the Netherlands Ministry of Education, Culture and Science.

Abzhanov
A.
,
Rodda
S. J.
,
McMahon
A. P.
,
Tabin
C. J.
(
2007
).
Regulation of skeletogenic differentiation in cranial dermal bone
.
Development
134
,
3133
-
3144
.
Agathocleous
M.
,
Locker
M.
,
Harris
W. A.
,
Perron
M.
(
2007
).
A general role of hedgehog in the regulation of proliferation
.
Cell Cycle
6
,
156
-
159
.
Albertson
R. C.
,
Yelick
P. C.
(
2005
).
Roles for fgf8 signaling in left-right patterning of the visceral organs and craniofacial skeleton
.
Dev. Biol.
283
,
310
-
321
.
Avaron
F.
,
Hoffman
L.
,
Guay
D.
,
Akimenko
M. A.
(
2006
).
Characterization of two new zebrafish members of the hedgehog family: atypical expression of a zebrafish indian hedgehog gene in skeletal elements of both endochondral and dermal origins
.
Dev. Dyn.
235
,
478
-
489
.
Bianco
P.
,
Cancedda
F. D.
,
Riminucci
M.
,
Cancedda
R.
(
1998
).
Bone formation via cartilage models: the “borderline” chondrocyte
.
Matrix Biol.
17
,
185
-
192
.
Chen
J. K.
,
Taipale
J.
,
Cooper
M. K.
,
Beachy
P. A.
(
2002
).
Inhibition of Hedgehog signaling by direct binding of cyclopamine to Smoothened
.
Genes Dev.
16
,
2743
-
2748
.
Chung
U. I.
(
2004
).
Essential role of hypertrophic chondrocytes in endochondral bone development
.
Endocr. J.
51
,
19
-
24
.
Galotto
M.
,
Campanile
G.
,
Robino
G.
,
Cancedda
F. D.
,
Bianco
P.
,
Cancedda
R.
(
1994
).
Hypertrophic chondrocytes undergo further differentiation to osteoblast-like cells and participate in the initial bone formation in developing chick embryo
.
J. Bone Miner. Res.
9
,
1239
-
1249
.
Gavaia
P. J.
,
Simes
D. C.
,
Ortiz-Delgado
J. B.
,
Viegas
C. S.
,
Pinto
J. P.
,
Kelsh
R. N.
,
Sarasquete
M. C.
,
Cancela
M. L.
(
2006
).
Osteocalcin and matrix Gla protein in zebrafish (Danio rerio) and Senegal sole (Solea senegalensis): comparative gene and protein expression during larval development through adulthood
.
Gene Expr. Patterns
6
,
637
-
652
.
Hoffmann
H. M.
,
Beumer
T. L.
,
Rahman
S.
,
McCabe
L. R.
,
Banerjee
C.
,
Aslam
F.
,
Tiro
J. A.
,
van Wijnen
A. J.
,
Stein
J. L.
,
Stein
G. S.
, et al. 
. (
1996
).
Bone tissue-specific transcription of the osteocalcin gene: role of an activator osteoblast-specific complex and suppressor hox proteins that bind the OC box
.
J. Cell Biochem.
61
,
310
-
324
.
Jin
S. W.
,
Beis
D.
,
Mitchell
T.
,
Chen
J. N.
,
Stainier
D. Y.
(
2005
).
Cellular and molecular analyses of vascular tube and lumen formation in zebrafish
.
Development
132
,
5199
-
5209
.
Kim
I. S.
,
Otto
F.
,
Zabel
B.
,
Mundlos
S.
(
1999
).
Regulation of chondrocyte differentiation by Cbfa1
.
Mech. Dev.
80
,
159
-
170
.
Kimmel
C. B.
,
Miller
C. T.
,
Kruze
G.
,
Ullmann
B.
,
BreMiller
R. A.
,
Larison
K. D.
,
Snyder
H. C.
(
1998
).
The shaping of pharyngeal cartilages during early development of the zebrafish
.
Dev. Biol.
203
,
245
-
263
.
Kimura
Y.
,
Okamura
Y.
,
Higashijima
S.
(
2006
).
alx, a zebrafish homolog of Chx10, marks ipsilateral descending excitatory interneurons that participate in the regulation of spinal locomotor circuits
.
J. Neurosci.
26
,
5684
-
5697
.
Kobayashi
T.
,
Soegiarto
D. W.
,
Yang
Y.
,
Lanske
B.
,
Schipani
E.
,
McMahon
A. P.
,
Kronenberg
H. M.
(
2005
).
Indian hedgehog stimulates periarticular chondrocyte differentiation to regulate growth plate length independently of PTHrP
.
J. Clin. Invest.
115
,
1734
-
1742
.
Kong
Y. Y.
,
Yoshida
H.
,
Sarosi
I.
,
Tan
H. L.
,
Timms
E.
,
Capparelli
C.
,
Morony
S.
,
Oliveira-dos-Santos
A. J.
,
Van
G.
,
Itie
A.
, et al. 
. (
1999
).
OPGL is a key regulator of osteoclastogenesis, lymphocyte development and lymph-node organogenesis
.
Nature
397
,
315
-
323
.
Koudijs
M. J.
,
den Broeder
M. J.
,
Keijser
A.
,
Wienholds
E.
,
Houwing
S.
,
van Rooijen
E. M.
,
Geisler
R.
,
van Eeden
F. J.
(
2005
).
The zebrafish mutants dre, uki, and lep encode negative regulators of the hedgehog signaling pathway
.
PLoS Genet.
1
,
e19
.
Koudijs
M. J.
,
den Broeder
M. J.
,
Groot
E.
,
van Eeden
F. J.
(
2008
).
Genetic analysis of the two zebrafish patched homologues identifies novel roles for the hedgehog signaling pathway
.
BMC Dev. Biol.
8
,
15
.
Mackie
E. J.
,
Ahmed
Y. A.
,
Tatarczuch
L.
,
Chen
K. S.
,
Mirams
M.
(
2008
).
Endochondral ossification: how cartilage is converted into bone in the developing skeleton
.
Int. J. Biochem. Cell Biol.
40
,
46
-
62
.
Maeda
Y.
,
Nakamura
E.
,
Nguyen
M. T.
,
Suva
L. J.
,
Swain
F. L.
,
Razzaque
M. S.
,
Mackem
S.
,
Lanske
B.
(
2007
).
Indian Hedgehog produced by postnatal chondrocytes is essential for maintaining a growth plate and trabecular bone
.
Proc. Natl. Acad. Sci. USA
104
,
6382
-
6387
.
Mak
K. K.
,
Bi
Y.
,
Wan
C.
,
Chuang
P. T.
,
Clemens
T.
,
Young
M.
,
Yang
Y.
(
2008a
).
Hedgehog signaling in mature osteoblasts regulates bone formation and resorption by controlling PTHrP and RANKL expression
.
Dev. Cell
14
,
674
-
688
.
Mak
K. K.
,
Kronenberg
H. M.
,
Chuang
P. T.
,
Mackem
S.
,
Yang
Y.
(
2008b
).
Indian hedgehog signals independently of PTHrP to promote chondrocyte hypertrophy
.
Development
135
,
1947
-
1956
.
Miura
S.
,
Hanaoka
K.
,
Togashi
S.
(
2008
).
Skeletogenesis in Xenopus tropicalis: characteristic bone development in an anuran amphibian
.
Bone
43
,
901
-
909
.
Mori
S.
,
Yoshikawa
H.
,
Hashimoto
J.
,
Ueda
T.
,
Funai
H.
,
Kato
M.
,
Takaoka
K.
(
1998
).
Antiangiogenic agent (TNP-470) inhibition of ectopic bone formation induced by bone morphogenetic protein-2
.
Bone
22
,
99
-
105
.
Nakashima
K.
,
Zhou
X.
,
Kunkel
G.
,
Zhang
Z.
,
Deng
J. M.
,
Behringer
R. R.
,
de Crombrugghe
B.
(
2002
).
The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation
.
Cell
108
,
17
-
29
.
Ohba
S.
,
Kawaguchi
H.
,
Kugimiya
F.
,
Ogasawara
T.
,
Kawamura
N.
,
Saito
T.
,
Ikeda
T.
,
Fujii
K.
,
Miyajima
T.
,
Kuramochi
A.
, et al. 
. (
2008
).
Patched1 haploinsufficiency increases adult bone mass and modulates Gli3 repressor activity
.
Dev. Cell
14
,
689
-
699
.
Otto
F.
,
Thornell
A. P.
,
Crompton
T.
,
Denzel
A.
,
Gilmour
K. C.
,
Rosewell
I. R.
,
Stamp
G. W.
,
Beddington
R. S.
,
Mundlos
S.
,
Olsen
B. R.
, et al. 
. (
1997
).
Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development
.
Cell
89
,
765
-
771
.
Plaisant
M.
,
Fontaine
C.
,
Cousin
W.
,
Rochet
N.
,
Dani
C.
,
Peraldi
P.
(
2009
).
Activation of hedgehog signaling inhibits osteoblast differentiation of human mesenchymal stem cells
.
Stem Cells
27
,
703
-
713
.
Renn
J.
,
Winkler
C.
(
2009
).
Osterix-mCherry transgenic medaka for in vivo imaging of bone formation
.
Dev. Dyn.
238
,
241
-
248
.
Schulte-Merker
S.
(
2002
).
Looking at embryos
. In
Zebrafish, A Practical Approach
(ed.
Nüsslein-Volhard
C.
,
Dahm
R.
), pp.
41
-
43
.
New York
:
Oxford University Press
.
Shibata
S.
,
Yokohama-Tamaki
T.
(
2008
).
An in situ hybridization study of Runx2, Osterix, and Sox9 in the anlagen of mouse mandibular condylar cartilage in the early stages of embryogenesis
.
J. Anat.
213
,
274
-
283
.
Shibata
S.
,
Suda
N.
,
Suzuki
S.
,
Fukuoka
H.
,
Yamashita
Y.
(
2006
).
An in situ hybridization study of Runx2, Osterix, and Sox9 at the onset of condylar cartilage formation in fetal mouse mandible
.
J. Anat.
208
,
169
-
177
.
Sinha
S.
,
Chen
J. K.
(
2006
).
Purmorphamine activates the Hedgehog pathway by targeting Smoothened
.
Nat. Chem. Biol.
2
,
29
-
30
.
Spoorendonk
K. M.
,
Peterson-Maduro
J.
,
Renn
J.
,
Trowe
T.
,
Kranenbarg
S.
,
Winkler
C.
,
Schulte-Merker
S.
(
2008
).
Retinoic acid and Cyp26b1 are critical regulators of osteogenesis in the axial skeleton
.
Development
135
,
3765
-
3774
.
St-Jacques
B.
,
Hammerschmidt
M.
,
McMahon
A. P.
(
1999
).
Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation
.
Genes Dev.
13
,
2072
-
2086
.
Takeda
S.
,
Bonnamy
J. P.
,
Owen
M. J.
,
Ducy
P.
,
Karsenty
G.
(
2001
).
Continuous expression of Cbfa1 in nonhypertrophic chondrocytes uncovers its ability to induce hypertrophic chondrocyte differentiation and partially rescues Cbfa1-deficient mice
.
Genes Dev.
15
,
467
-
481
.
Thisse
B.
,
Thisse
C.
(
2005
).
Fast Release Clones: A High Throughput Expression Analysis
.
ZFIN Direct Data Submission
.
Ueta
C.
,
Iwamoto
M.
,
Kanatani
N.
,
Yoshida
C.
,
Liu
Y.
,
Enomoto-Iwamoto
M.
,
Ohmori
T.
,
Enomoto
H.
,
Nakata
K.
,
Takada
K.
, et al. 
. (
2001
).
Skeletal malformations caused by overexpression of Cbfa1 or its dominant negative form in chondrocytes
.
J. Cell Biol.
153
,
87
-
100
.
Vortkamp
A.
,
Lee
K.
,
Lanske
B.
,
Segre
G. V.
,
Kronenberg
H. M.
,
Tabin
C. J.
(
1996
).
Regulation of rate of cartilage differentiation by Indian hedgehog and PTH-related protein
.
Science
273
,
613
-
622
.
Wagner
T. U.
,
Renn
J.
,
Riemensperger
T.
,
Volff
J. N.
,
Koster
R. W.
,
Goerlich
R.
,
Schartl
M.
,
Winkler
C.
(
2003
).
The teleost fish medaka (Oryzias latipes) as genetic model to study gravity dependent bone homeostasis in vivo
.
Adv. Space Res.
32
,
1459
-
1465
.
Walker
M. B.
,
Kimmel
C. B.
(
2007
).
A two-color acid-free cartilage and bone stain for zebrafish larvae
.
Biotech. Histochem.
82
,
23
-
28
.
Wang
Y.
,
Wan
C.
,
Gilbert
S. R.
,
Clemens
T. L.
(
2007
).
Oxygen sensing and osteogenesis
.
Ann. N. Y. Acad. Sci.
1117
,
1
-
11
.
Witten
P. E.
,
Hansen
A.
,
Hall
B. K.
(
2001
).
Features of mono- and multinucleated bone resorbing cells of the zebrafish Danio rerio and their contribution to skeletal development, remodeling, and growth
.
J. Morphol.
250
,
197
-
207
.
Wu
X.
,
Walker
J.
,
Zhang
J.
,
Ding
S.
,
Schultz
P. G.
(
2004
).
Purmorphamine induces osteogenesis by activation of the hedgehog signaling pathway
.
Chem. Biol.
11
,
1229
-
1238
.
Yasuda
H.
,
Shima
N.
,
Nakagawa
N.
,
Yamaguchi
K.
,
Kinosaki
M.
,
Mochizuki
S.
,
Tomoyasu
A.
,
Yano
K.
,
Goto
M.
,
Murakami
A.
, et al. 
. (
1998
).
Osteoclast differentiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to TRANCE/RANKL
.
Proc. Natl. Acad. Sci. USA
95
,
3597
-
3602
.
Yelick
P. C.
,
Schilling
T. F.
(
2002
).
Molecular dissection of craniofacial development using zebrafish
.
Crit. Rev. Oral Biol. Med.
13
,
308
-
322
.
Yelick
P. C.
,
Driever
W.
,
Neuhauss
S.
,
Stashenko
P.
(
1996
).
Craniofacial cartilage development in zebrafish
.
Ann. N. Y. Acad. Sci.
785
,
360
-
361
.
Zelzer
E.
,
Olsen
B. R.
(
2005
).
Multiple roles of vascular endothelial growth factor (VEGF) in skeletal development, growth, and repair
.
Curr. Top. Dev. Biol.
65
,
169
-
187
.
Zelzer
E.
,
Mamluk
R.
,
Ferrara
N.
,
Johnson
R. S.
,
Schipani
E.
,
Olsen
B. R.
(
2004
).
VEGFA is necessary for chondrocyte survival during bone development
.
Development
131
,
2161
-
2171
.

Supplementary information