Indian hedgehog and syndecans-3 coregulate chondrocyte proliferation and function during chick limb skeletogenesis



Hedgehog proteins exert critical roles in embryogenesis and require heparan sulfate proteoglycans (HS-PGs) for action. Indian hedgehog (Ihh) is produced by prehypertrophic chondrocytes in developing long bones and regulates chondrocyte proliferation and other events, but it is not known whether it requires HS-PGs for function. Because the HS-PG syndecan-3 is preferentially expressed by proliferating chondrocytes, we tested whether it mediates Ihh action. Primary chick chondrocyte cultures were treated with recombinant Ihh (rIhh-N) in absence or presence of heparinase I or syndecan-3 neutralizing antibodies. While rIhh-N stimulated proliferation in control cultures, it failed to do so in heparinase- or antibody-treated cultures. In reciprocal gain-of-function studies, chondrocytes were made to overexpress syndecan-3 by an RCAS viral vector. Cells became more responsive to rIhh-N, but even this response was counteracted by heparinase or antibody treatment. To complement the in vitro data, RCAS viral particles were microinjected in day 4–5 chick wing buds and effects of syndecan-3 misexpression were monitored over time. Syndecan-3 misexpression led to widespread chondrocyte proliferation and, interestingly, broader expression and distribution of Ihh. In addition, the syndecan-3 misexpressing skeletal elements were short, remained cartilaginous, lacked osteogenesis, and exhibited a markedly reduced expression of collagen X and osteopontin, products characteristic of hypertrophic chondrocytes and bone cells. The data are the first to indicate that Ihh action in chondrocyte proliferation involves syndecan-3 and to identify a specific member of the syndecan family as mediator of hedgehog function. Developmental Dynamics 229:607–617, 2004. © 2004 Wiley-Liss, Inc.


Ihh Indian hedgehog syn-3 syndecan-3 PtcPatchedrIhh-N recombinant N-terminal half of Ihh HS-PGs heparan sulfate proteoglycans BrdU 5′-bromodeoxyuridine PTHrP parathyroid hormone-related protein


Indian hedgehog (Ihh) is a member of a powerful family of secreted signaling proteins. Ihh is expressed by a restricted number of cell types and organs during embryogenesis, including prehypertrophic chondrocytes in growth plates of developing endochondral skeletal structures (Bitgood and McMahon, 1995; Koyama et al., 1996a; Vortkamp et al., 1996). Ihh expression by prehypertrophic chondrocytes has stimulated a great deal of interest and research activity owing to the finding that the growth plate is of critical importance for skeletal development and growth. Previous work on limb skeletogenesis from our laboratories indicated that one role of Ihh is to direct formation of an intramembranous bone collar around the growth plate (Koyama et al., 1996a; Nakamura et al., 1997), a step critical for long bone shaft morphogenesis. This role was confirmed by the phenotype of Ihh-null mice in which the bone collar does not form (St-Jacques et al., 1999). The growth plates in Ihh-null mouse embryos displayed another striking trait, namely the zone of chondrocyte proliferation preceding the prehypertrophic zone was considerably reduced. Reduced chondrocyte proliferation was subsequently demonstrated also in mouse embryos in which the hedgehog receptor Smoothened was conditionally deleted in cartilage, strengthening the conclusion that Ihh is an important regulator of chondrocyte proliferation in the growth plate (Long et al., 2001). Together, these and other studies have led to the important overall conclusion that Ihh has multiple roles and exerts influences in the proliferative zone as well as surrounding perichondrial tissues (Chung et al., 2001; Pacifici et al., 2002).

These putative but likely roles of Ihh in the growth plate raise an intriguing issue. Ihh is exclusively produced in the prehypertrophic zone. Hence, Ihh would need to undergo long-range diffusion from its site of synthesis to reach target cells in other zones and surrounding tissues. Indeed, we and others showed recently that Ihh is present not only in prehypertrophic zone, but also in proliferative and hypertrophic zones and the osteogenic inner layer of perichondrium (Gritli-Linde et al., 2001; Yin et al., 2002). While short-range movement of hedgehog proteins is believed to result from direct transfer between producing and target cells (McMahon, 2000), long-range diffusion is more complex and requires intervention of heparan sulfate proteoglycans (HS-PGs; Bellaiche et al., 1998; The et al., 1999; Rubin et al., 2002; Lum et al., 2003). HS-PGs would exploit the heparin-binding capacity of hedgehog proteins, counteract their tendency to remain near their site of synthesis because of their lipid tails, facilitate movement and diffusion, and increase their effective concentration for interaction with their receptor Patched at target sites (The et al., 1999; Ingham, 2001). There are, however, no data on whether and which HS-PG(s) may aid Ihh function in the growth plate.

Syndecans are a cell surface-associated subfamily of HS-PGs. The known four members are differentially expressed and regulate several developmental events and processes in which they act as receptors and coreceptors for factors, cytokines or matrix components such as fibronectin (Bernfield et al., 1999; Iozzo, 2001). We previously showed that syndecan-3 gene expression and distribution are restricted to the proliferative zone of the growth plate (Koyama et al., 1995; Shimazu et al., 1996; Kirsch et al., 2002), raising the possibility that syndecan-3 may participate in Ihh function in that zone and allow the factor to directly influence chondrocyte proliferation. The results presented here provide support for this possibility and are the first to identify a specific member of the syndecan family in hedgehog action.


In Vitro Analyses

We first carried out gain-of-function studies to determine whether syndecan-3 overexpression and exogenous Ihh treatment affected proliferation in chondrocyte cultures. Cell populations rich in resting and proliferating chondrocytes isolated from the caudal and peripheral cephalic portions of day 17 chick embryo sterna (Shimazu et al., 1996) were infected with RCAS(A) viral particles encoding chick syndecan-3 and cultured for 8 to 10 days until subconfluent. Companion control cells were mock-infected and grown in parallel. Approximately 70% of both caudal and cephalic cells had become infected by day 8–10 as revealed by immunostaining (Fig. 1B,D), while no cell in control uninfected cultures stained (Fig. 1A,C). Infected cells contained several-fold higher levels of syndecan-3 transcripts (Fig. 1E, lanes 3–4 and 7–8) and syndecan-3 itself (Fig. 1F, lanes 2 and 4) compared with their respective controls (Fig. 1E,F). To verify that the cell surface characteristics had been altered by syndecan-3 overexpression, uninfected and infected cells were tested for adhesion and spreading onto substrates coated with the syndecan-binding protein fibronectin. Control caudal and cephalic chondrocytes adhered quite poorly (Fig. 1G,H, dashed line) and failed to spread (Fig. 1I,K) during the 10- and 30-min test periods at any fibronectin dose used, while syndecan-3 overexpressing cells adhered (Fig. 1G,H, solid line) and spread well (Fig. 1J,L, arrowheads).

Figure 1.

Behavior of control and syndecan-3 overexpressing (RCAS Syn-3) chondrocytes. Caudal cells are in A, B, E (lanes 1–4), F (lanes 1–2), G, and I–J; cephalic cells are in C, D, E (lanes 5–8), F (lanes 3–4), H, K, and L. A–D: Immunocytochemistry showing dark-staining infected cells in B and D, and unstained control cells in A and C. E,F: Reverse transcriptase-polymerase chain reaction and immunoblots showing levels of syndecan-3 transcripts and macromolecule in control vs. overexpressing cells. G,H: Attachment assays with control and overexpressing cells onto substrates coated with increasing fibronectin doses. I–L: Spreading assays of control and overexpressing cells onto substrate coated with 5 μg/ml fibronectin. Scale bars = 50 μm in D (applies to A–D), in L (applies to I–L).

Next, day 8–10 semiconfluent caudal and cephalic peripheral cultures as above were switched to low serum-containing medium for 24 hr to induce mitotic quiescence. Cultures were then treated with recombinant chick Ihh (rIhh-N) for 24 hr, pulse-labeled with 5′-bromodeoxyuridine (BrdU) during the last 4 hr, and processed for BrdU incorporation. rIhh-N treatment increased proliferation in a dose-dependent manner in both uninfected (Fig. 2A,B, open bars) and syndecan-3 overexpressing cultures (Fig. 2A,B, solid bars); stimulation was maximal at 1 μg/ml rIhh-N and had increased 20 to 35% relative to respective untreated control cultures (Fig. 2A,B). Note that, in the absence of rIhh-N treatment, baseline proliferation rates were higher in syndecan-3 overexpressing than control cultures (Fig. 2A,B, bars on the far left), likely reflecting enhanced responses to endogenous or medium-derived growth factors. Cells infected with insertless viral particles behaved as control uninfected cells (not shown).

Figure 2.

Proliferation in control and syndecan-3 overexpressing (RCAS Syn-3) chondrocytes during recombinant N-terminal half of Indian hedgehog (rIhh-N, IHH) treatment. Caudal cells are in A, C, E, and G; cephalic cells are in B, D, F, and H. A–D: Control cultures (open bars) and overexpressing cultures (filled bars) treated with different doses of rIhh-N (0–10 μg/ml) in absence (A,B) or presence (C,D) of 0.5 U/ml heparinase I (Hep'ase) for 24 hr, and pulsed with bromodeoxyuridine (BrdU) for the last 4 hr. BrdU incorporation in control uninfected untreated cultures was set at 100% and used to determine relative incorporation in all other cultures. Dashed line in C–D shows control incorporation. E,F: Control and overexpressing cultures treated with 1 μg/ml rIhh-N in absence or presence of 50 μg/ml syndecan-3 antibodies. G–H: Dose-dependent inhibition of proliferation in control and overexpressing cultures treated with 1 μg/ml rIhh-N in the presence of increasing amounts of syndecan-3 antibodies. Data are presented as averages ± standard deviation. Asterisks indicate significance of P < 0.01.

To show more directly that HS-PGs and syndecan-3 in particular are involved in chondrocyte mitotic behavior and responses, two sets of loss-of-function experiments were performed. Day 8–10 semiconfluent uninfected and syndecan-3 overexpressing cultures were treated with rIhh-N in the presence of 1 U/ml heparinase I (Fig. 2C,D), and proliferation rates were determined as above. All cultures failed to respond to rIhh-N and to mount a proliferative response (Fig. 2C,D) compared with cultures receiving no heparinase (Fig. 2A,B). Proliferation rates were 20 to 30% lower in heparinase-treated than untreated uninfected cultures (Fig. 2A–D, open bars); in syndecan-3 overexpressing cultures, rates were as much as 50% lower (Fig. 2A–D, solid bars). Effects of heparinase I were dose-dependent and maximal at 1 U/ml (not shown). Next, cultures were treated with rIhh-N in the presence of syndecan-3 neutralizing antibodies (Kirsch et al., 2002). Antibody treatment counteracted the stimulatory effects of rIhh-N on proliferation in both uninfected and syndecan-3–overexpressing cultures (Fig. 2E,F), particularly in cephalic cultures. As seen after heparinase I treatment, the antibodies were also able to reduce baseline proliferation rates (in the absence of exogenous rIhh-N) in both sets of cultures (Fig. 2E,F). Effects of neutralizing antibodies were dose-dependent and preimmune IgGs had no effect (Fig. 2G,H). Treatment with preimmune or immune antibodies had no obvious effects on the differentiated status of the chondrocytes, as assessed by gene expression analysis (not shown). Taken together, the above data show that Ihh and syndecan-3 regulate proliferation in cultured chondrocytes. Because similar responses were seen with caudal and cephalic cells, responses were not influenced by maturation status of cells before culturing nor anatomic origin.

In Vivo Studies

To extend the above findings, we determined the consequences of syndecan-3 misexpression on chondrocyte proliferation in vivo. Syndecan-3–encoding viral particles were microinjected in the vicinity of incipient cartilaginous humerus, radius, and/or ulna in wing buds of day 4–5 chick embryos in ovo; contralateral wing buds served as control (Yagami et al., 1999). Embryos were re-incubated for 2 to 3 days, injected with BrdU 2 hr before harvesting, and processed for detection of incorporated BrdU and syndecan-3 gene expression. Figure 3 shows the results of day 7 embryos in which radius (r) and ulna (u) had been infected. In control embryos, syndecan-3 gene expression and proliferating chondrocytes exhibited predictable patterns and location at that stage (Koyama et al., 1995; Kirsch et al., 2002) and were both restricted to the epiphyseal ends (Fig. 3A,C,E, arrows). In contrast, in virally infected specimens syndecan-3 expression was quite strong and widespread (Fig. 3B), and proliferating chondrocytes were present throughout radius and ulna, including the diaphyseal portion (Fig. 3D,F, arrowheads). Analysis of overall number of labeled cells showed that there were approximately twice as many proliferating chondrocytes in syndecan-3–misexpressing than control elements per microscopic field (20 ± 3 vs. 9 ± 3). Proliferation in surrounding perichondrial and mesenchymal tissues was not affected appreciably (Fig. 3C–F).

Figure 3.

Proliferation and syndecan-3 expression in limb skeletal elements. Longitudinal serial sections from control and syndecan-3 misexpressing day 7.5 wings were processed for detection of syndecan-3 transcripts (A,B) and bromodeoxyuridine (BrdU) -labeled proliferating cells (C–F). A,C,E: Control specimen showing that transcripts and proliferating chondrocytes are largely restricted to the epiphyseal ends (arrows) of radius (r) and ulna (u). h, humerus. B,D,F: Misexpressing specimen showing that syndecan-3 expression is widespread as are proliferating chondrocytes, which are present even in the diaphysis (arrowheads). An average of 20 ± 3 labeled cells/field were present in syndeca-3 misexpressing specimens compared with 9 ± 3 in control (P < 0.05). Scale bars = 200 μm in B (applies to A,B), 100 μm in F (applies to C–F).

To determine whether the above changes had additional consequences on long bone development, embryos were examined at later stages by anatomy, histochemistry, and in situ hybridization. Figure 4B shows the results from day 10 embryos in which radius (r) but not ulna (u) had been infected. The uninfected ulna had normal appearance and displayed an ossifying diaphysis rich in blood-rich tissues (Fig. 4B) and mineralized alizarin-red staining bone tissue (Fig. 4D, arrow). Similar features were displayed by control radius and ulna in contralateral uninfected wing (Fig. 4A,C, arrows). On the other hand, the infected syndecan-3 misexpressing radius was entirely cartilaginous (Fig. 4B), and lacked alizarin red-staining endochondral bone and intramembranous collar (Fig. 4D, arrowhead) as well as a marrow cavity (cfr. Fig. 4F with E). Its most diaphyseal chondrocytes were larger in size than epiphyseal cells, indicating that they were advancing toward maturation and hypertrophy despite an overall delay in radius's development (Fig. 4F, arrowhead). No changes were seen in embryos infected with insertless viral particles (not shown).

Figure 4.

Anatomy and histology of wing skeletal elements. A,C,E: Control day 10 specimens exhibiting normal radius (r) and ulna (u), normal alizarin red bone tissue (arrows in C), and normal hypertrophy and marrow in E. B,D,F: Experimental specimens showing normal features in uninfected ulna but inhibition of osteogenesis in infected radius (arrowhead in D) and lack of full hypertrophy and marrow (arrowhead in F). Scale bar = 3 mm in B (applies to A,B), 200 μm in F (applies to C–F).

Gene expression patterns in control day 10 elements were typical (Fig. 5A–F). Syndecan-3 transcripts were abundant in the proliferative zone (Fig. 5B, arrowhead) which no longer occupied the epiphyseal end at this stage but was now located in the metaphysis below the articular cap (Fig. 5B, ac; Shimazu et al., 1996; Kirsch et al., 2002). Ihh gene expression was also strong and restricted to the prehypertrophic zone (Fig. 5C, arrowhead), while Patched (Ptc) expression was most obvious in the proliferative zone (Fig. 5D) along with syndecan-3 (Fig. 5B). Patched and syndecan-3 transcripts were also present in the surrounding perichondrium (Fig. 5B,D). Type X collagen and osteopontin RNAs were present primarily in diaphyseal hypertrophic and posthypertrophic zones, respectively (Fig. 5E,F, arrowheads), and type II collagen transcripts were abundant from articular cap to prehypertrophic zone (Fig. 5A). In syndecan-3 misexpressing element, however, syndecan-3 transcripts were very abundant and present in epiphysis, metaphysis, and diaphysis (Fig. 5I), particularly along the right side of the skeletal element where infection had occurred more efficiently (Fig. 5I, arrow). Of interest, gene expression of Ihh had been affected as well (Fig. 5J). Ihh transcripts were widespread and present even in the epiphysis (Fig. 5J, arrow), clearly overlapping with misexpressed syndecan-3 (Fig. 5I, arrow). Patched gene expression was also slightly stronger (Fig. 5K) than in control (Fig. 5D). On the other hand, type X collagen expression was severely reduced (Fig. 5L) and osteopontin RNAs were essentially undetectable (Fig. 5M), confirming that chondrocyte maturation and ossification had been markedly inhibited. Type II collagen expression remained strong from epiphysis to diaphysis (Fig. 5H), indicating that syndecan-3 misexpression affected maturation but not the differentiated state of the cells. Immunostaining confirmed that Ihh was more widely distributed in syndecan-3 misexpressing than control tissue (Fig. 5G,N), including the epiphysis (Fig. 5N, arrow), which was negative in control (Fig. 5G, arrow).

Figure 5.

Gene expression and Indian hedgehog (Ihh) distribution. Serial longitudinal sections from day 10 control (A–G) and syndecan-3 misexpressing (H–N) specimens were processed for in situ hybridization analysis of indicated genes (A–F,H–M) and immunostaining for Ihh (G,N). Arrowheads point to restricted expression of several genes. Arrows in I, J, and N point to epiphyseal misexpression and mislocalization of syndecan-3 and Ihh; arrow in G shows lack of epiphyseal Ihh in control. See text for further details. Scale bar = 150 μm in G (applies to A–N).


The study provides the first evidence that Ihh action in chondrocyte proliferation is influenced by, and likely dependent on, syndecan-3. We find that exogenous Ihh and syndecan-3 overexpression both stimulate proliferation in vitro and that sydecan-3 misexpression causes widespread chondrocyte proliferation in vivo. The mitogenic effects of rIhh-N are all but prevented by treatment with heparinase I or syndecan-3 neutralizing antibodies, indicating that an intact complement of heparan sulfate chains and unburdened syndecan-3 molecules are required for action. We show also that there is extensive overlap in expression of syndecan-3 and Patched in the proliferative zone of growth plate. The data point to a close cooperation and mutual interdependence between Ihh and syndecan-3 in influencing and restricting chondrocyte proliferation. It is reasonable, thus, to conclude that syndecan-3 represents one of the HS-PGs in the growth plate which would allow Ihh produced by prehypertrophic chondrocytes to exert long-range roles and influence mitotic behavior in the preceding proliferative zone of the growth plate.

The proposition that Ihh acts directly on chondrocyte proliferation correlates quite well with recent observations by Long and collaborators involving genetic manipulations of mouse embryos (Long et al., 2001). On the other hand, our data and those of Long et al. appear to be at odds with previous studies indicating that Ihh acts indirectly on chondrocyte proliferation (Vortkamp et al., 1996). Ihh was suggested to induce expression of parathyroid hormone-related protein (PTHrP) in periarticular tissue in developing long bones; PTHrP would in turn diffuse into the epiphyseal zone and trigger proliferation of resident chondrocytes. These apparently contradictory views are actually reconcilable and suggest important implications. We showed previously and confirm here (Fig. 3) that chondrocyte proliferation is restricted to the most epiphyseal zone of long bone anlagen at early stages of development (Koyama et al., 1995). However, as joint formation and articular cartilage development progress, the zone of chondrocyte proliferation is relocated below the articular cap, thus becoming further away from periarticular tissues. The relocation of the proliferative zone is accompanied by an identical shift in syndecan-3 expression (Shimazu et al., 1996; Kirsch et al., 2002). Thus, it is possible that, at early stages of long bone development, chondrocyte proliferation is regulated both by PTHrP produced by periarticular tissue under Ihh influence and by Ihh itself produced by prehypertrophic chondrocytes and interacting with syndecan-3. At later stages of skeletogenesis and with formation of the mitotically quiescent articular cap (and eventually a secondary ossification center), chondrocyte proliferation may become more dependent on intrinsic growth plate pathways, including Ihh/syndecan-3 pathway, but less so on periarticularly-derived PTHrP (Karp et al., 2000; van der Eerden et al., 2000). It is important to point out here that, in addition to PTHrP and Ihh, other factors regulate chondrocyte proliferation. Fibroblast growth factors and their receptors have been shown to act as negative regulators (Colvin et al., 1996). Chondrocyte proliferation and skeletal growth are also influenced by factors such as insulin-like growth factors, growth hormone, and bone morphogenetic proteins (Liu and LeRoith, 1999; Butler and LeRoith, 2001; Minina et al., 2001). Clearly, chondrocyte proliferation and skeletal growth in general are the net result of multiple pathways with cooperating or antagonistic properties.

The complexities and multiplicity of mechanisms involved in chondrocyte proliferation may account for the lack of major skeletal defects in the recently reported syndecan-3 null mice (Kaksonen et al., 2002). This study focused on syndecan-3 roles in synaptic and memory function and, thus, did not examine the skeleton in detail. However, the null mice were found to be healthy and fertile and without obvious defects. It will be important then to carry out a direct analysis of the developing skeleton in these mice. If no defects in chondrocyte proliferation and skeletal development are found, the findings would suggest that other HS-PG(s) compensated for the lack of syndecan-3, including perlecan (Arikawa-Hirasawa et al., 1999). It may also be possible that there are differences between mouse and chick and that other syndecans, perlecan, and/or other HS-PGs actually exert primary roles in mammalian growth plates. At this regard, it is important to note that ablation of the perlecan gene in mice causes profound defects in growth plate and skeletogenesis, including deranged Ihh gene expression patterns (Arikawa-Hirasawa et al., 1999). As we show here, syndecan-3 misexpression inhibits chondrocyte maturation and blocks bone formation in the chick limb, despite widespread Ihh expression and distribution (Fig. 5). It appears then that normal patterns of HS-PG expression are rather important for growth plate function and skeletogenesis in both avian and mammalian embryos.

Elegant recent studies have identified the portion of hedgehog proteins rich in basic amino acids that interact with the negatively charged heparan sulfate chains (Rubin et al., 2002). However, it remains unclear how, after binding to HS-PGs, the proteins would be able to interact with their signaling receptor Patched and trigger cellular responses. This situation is reminiscent of other factors that require HS-PGs for effective and efficient interactions with their signaling receptors, most notably members of the fibroblast growth factor family (Rapraeger, 1995). In both cases, HS-PGs may serve to “capture” the factors and “present” them to the signaling receptors. By virtue of their considerable length, the heparan sulfate chains should also be able to bind multiple ligands and, thus, significantly increase the effective concentration of ligand on the cell surface. These possibilities offer an explanation for the inhibitory effect of heparinase I treatment on rIhh-N–induced chondrocyte proliferation. But how can we account for the equally inhibitory effects of antibody treatment? One explanation can be found in our recent study showing that syndecan-3 is present in dimeric and oligomeric forms on the surface of chondrocytes (Kirsch et al., 2002). Syndecan-3–syndecan-3 interactions had been invoked in previous studies to account for biological properties of the molecule (Asundi and Carey, 1995; Bernfield et al., 1999). We provided evidence that syndecan-3 dimers/oligomers are actually present on the chondrocyte cell surface and their assembly probably involves parallel pairing of β-sheet segments within the ectodomains of adjacent core proteins. It is possible then that antibody treatment may have interfered with syndecan-3 function by preventing dimerization and oligomerization and/or restricting lateral mobility and functional dynamism of the molecule.

The first insights into a role of HS-PGs in hedgehog function came from studies of a Drosophila phenotype caused by mutations in tout-velu, a gene that turned out to be homologous to vertebrate EXT-1 and EXT-2 genes involved in heparan sulfate synthesis (Bellaiche et al., 1998; Lind et al., 1998). As pointed out above, HS-PG roles in hedgehog function have been confirmed and extended by several subsequent studies. As far as we know, however, ours is the first study to identify a specific syndecan in hedgehog function. The HS-PG family comprises four syndecans, six glypicans (anchored to the cell surface by means of a phosphatidylinositol linkage), and extracellular HS-PGs such as perlecan and agrin (Bernfield et al., 1999; Iozzo, 2001). The growth plate and adjacent tissues express other HS-PGs in addition to syndecan-3, including perlecan (Handler et al., 1997), in keeping with the fact that EXT genes are widely expressed within and around it (Stickens et al., 2000). It is, therefore, conceivable that different HS-PGs or specific combinations of HS-PGs are involved in each of the roles Ihh is believed to have in the growth plate. For instance, the role Ihh appears to play in intramembranous bone collar formation may be mediated by syndecan-2 and syndecan-3, which are both expressed in perichondrium (David et al., 1993; Koyama et al., 1996b). Ihh presence in the hypertrophic zone and its roles in endochondral ossification may involve perlecan, which is strongly expressed in that zone and whose absence leads to ossification defects (Arikawa-Hirasawa et al., 1999). A major challenge for the future then is to understand how expression of different HS-PG family members, including syndecan-3, is restricted to specific portions within and around the growth plates of developing skeletal elements and how different HS-PGs, and combinations of, participate in multiple Ihh functions.


Cell Cultures and Viral Infection

Chondrocytes isolated from the caudal and peripheral cephalic portions of day 17 chick embryo sterna were cultured in complete medium consisting of high-glucose DMEM, 10% fetal bovine serum (FBS), 2 mM L-glutamine and 1% penicillin-streptomycin (Shimazu et al., 1996). When indicated, freshly isolated chondrocytes were infected with concentrated viral preparations, maintained for 5 to 7 days in primary culture and then subcultured into secondary cultures at 3 × 106 cells/100-mm dish by trypsinization. To favor adhesion to the substrate, secondary cultures received 4U/ml of testicular hyaluronidase (Sigma) in complete medium for the first 2 days and were grown for a total of 8 to 10 days until subconfluent. More than 85% of cells in day 8–10 secondary cultures were routinely infected, as revealed by immunocytochemistry of viral antigens (see below). Basal media, trypsin, and salines were from Gibco, and defined FBS was from Hyclone Labs. Data reported here were obtained in a minimum of three independent experiments; statistical significance was calculated by Student's t-test and is indicated by an asterisk in specific data sets.

Virus Preparation

Skin fibroblasts were isolated from SPAFAS virus-free 11-day-old chick embryos and cultured in medium 199 containing 10% FBS. Cells were transfected with RCAS(A) plasmid encoding full-length chick syndecan-3, using FuGENE 6 transfection reagent according to the manufacturer's protocol (Rech Diagnostic, Inc.). Fibroblasts were grown for a total of approximately 10 days with two passages at which point 100% of the cells were infected. Recombinant viral particles present in the medium were concentrated by ultracentrifugation at 25,000 rpm for 3 hr and used to infect freshly isolated chondrocytes (Yagami et al., 1999). Insert-less control viral particles were produced, isolated, and used in the same manner. To verify that chondrocytes had become infected, second-passaged monolayer cultures were trypsinized, resuspend in 1× phosphate buffered saline (PBS) and plated onto glass coverslips precoated with 0.5% poly-L-lysine (Sigma) at 1 × 106 cells/2 ml for 10 min. Attached cells were fixed with 70% ethanol for 15 min at room temperature. After rinsing with PBS, coverslips were incubated with PBS containing 10% nonimmune goat serum to block nonspecific protein binding for 10 min, and reacted with 1:300 dilution of rabbit antibodies to viral coat protein p27 (Spafas) in PBS for 30 min at room temperature. After three rinses, coverslips were reacted with streptavidin–peroxidase conjugated secondary antibodies for 10 min; after rinsing again, they were stained with ABC Histostain substrate–chromogen mixture (Zymed labs).

Syndecan-3 Antibodies

Preparation and characterization of rabbit antibodies to syndecan-3 were described previously (Kirsch et al., 2002). Briefly, antibodies were raised against a recombinant extracellular region of syndecan-3 (amino acids 215-313) prepared by using pGEX expression vector. That region was chosen because it shows no homology to other syndecans and is thought to be important for function (Bernfield et al., 1999). Rabbits were injected with 200 μg of purified syndecan-3-glutathione S-transferase fusion protein four times (Cocalico, Inc.). Total IgGs were purified by protein A–Sepharose affinity chromatography. Preimmune IgGs were isolated from the same rabbits before immunization and were used as controls.

Syndecan-3 Immunoblotting

To determine expression and overexpression of syndecan-3, chondrocyte cultures were processed for sequential immunoprecipitation and immunoblotting procedures. Second-passage control and RCAS-infected semiconfluent monolayer cultures were rinsed with ice-cold PBS and lysed in ice-cold lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 1% NP-40, 10 mM NaF, 1 μg/ml leupeptin, 2 μg/ml aprotinin, and 1 mM phenyl methyl sulfonyl fluoride, pH 7.4) on ice. After sheering DNA by passage through a syringe, aliquots of homogenates corresponding to 100 μg of total cellular proteins were precleared by incubation with 0.25 μg of rabbit IgGs and protein G–agarose beads for 30 min at 4°C. Beads were pelleted by centrifugation at 2,500 rpm for 5 min at 4°C, and the cleared supernatants (cell lysates) were transferred to fresh 1.5 ml microcentrifuge tubes at 4°C. Lysates were mixed with 10 μl of syndecan-3 antibodies (approximately 5 μg) for 1 hr at 4°C; 20 μl of protein G–agarose beads were added and mixture was incubated at 4°C on a rotator overnight. Beads were recovered by centrifugation, rinsed three times with lysis buffer, and resuspended in sample buffer (50 mM Tris-HCl, pH 6.8, 2% sodium dodecyl sulfate [SDS] 0.1% bromophenol blue, and 10% glycerol). Proteins were size-separated on SDS-polyacrylamide gels and transferred to nylon membranes (Immobilon-P, Millipore Co.). Membranes were blocked with 2% nonfat dry milk in TBS overnight at 4°C and incubated with a 1:200 dilution of syndecan-3 antibodies in TBS, 0.1% Tween, and 5% nonfat milk overnight at 4°C. After rinsing, membranes were reacted with a 1:5,000 dilution of peroxidase-conjugated anti-rabbit IgGs and processed for detection by chemiluminescence (ECL, Amersham).

Cell Proliferation, Adhesion, and Spreading Assays

Second-passage chondrocytes were seeded in 96 microwell plates at 2 × 104 cells/well in complete medium for 2 to 3 days until subconfluent. Cells were switched to DMEM containing 0.5% FBS for 24 hr and were then treated or cotreated with different doses (0–10 μg/ml) of rIhh-N in the absence or presence of indicated doses of heparinase I (Sigma) and syndecan-3 antibodies for 24 hr. Cultures were pulse-labeled with BrdU during the last 2 hr of culture, and processed with an enzyme-linked immunosorbent assay BrdU assay kit (Boehringer Mannheim Corp). Each time point was measured in triplicate, and each assay was repeated in three independent experiments. rIhh-N (amino acids 24-198) was prepared in baculovirus-infected insect cells, and bioactivity was verified by induction of supernumerary digits in wing buds and Patched expression in cultured chondrocytes.

For adhesion assays, bovine fibronectin (Gibco) was diluted with 0.1% BSA in PBS to final concentrations of 0.3 to 5.0 μg/ml. Immunologic 96-well plates were precoated with fibronectin solutions at 4°C overnight, blocked with 6% BSA in PBS for 1 hr at room temperature, and washed with PBS three times. Second-passaged chondrocytes were plated at 3 × 104 cells/ well in 50 μl of serum-free DMEM for up to 1 hr. Wells were then flooded with additional DMEM and placed bottom up for 15 min at room temperature. After discarding the floating cells, attached cells were fixed with 4% paraformaldehyde, stained with methylene blue, and quantified by dye extraction with measurement of absorbance at 620 nm. For spreading assays, second-passaged cells were re-plated in 35-mm dishes (5 × 105 cells/dish) in complete medium. After 1 hr, cells were examined by microscopy.

Syndecan-3 Misexpression and BrdU Labeling In Vivo

Protocols to achieve virally driven gene misexpression were as described (Yagami et al., 1999). Concentrated syndecan-3-encoding RCAS viral particles prepared as above or a small fragment of confluent fibroblast cell layer producing the virus were microinjected or implanted by microsurgical procedures in the vicinity of incipient cartilaginous humerus, radius, and/or ulna in wing buds of day 4.5–5 chick embryos in ovo; contralateral wing buds served as controls. Similar wing buds in companion embryos were injected or implanted with insert-less viral particles or fibroblasts producing insertless virus. Eggs were returned to the incubator and allowed to develop further. At indicated time points, embryos were fixed in 4% paraformaldehyde and sections were processed for (1) staining with hematoxylin and eosin, (2) histochemical detection of mineral by staining with 0.5% alizarin red S solution, pH 4.0, for 5 min at room temperature, or (3) in situ hybridization. When indicated, embryos were microinjected and labeled for 2 hr with BrdU according to manufacturer's instructions (Boehringer-Mannheim), and serial limb sections were processed for cell proliferation by immunostaining with anti-BrdU antibodies. Proliferation indexes were determined by counting labeled cells in three microscopic fields over the epiphysis, metaphysis, and diaphysis in slides from independent experiments; results are expressed as averages ± standard deviation. Comparable results of syndecan-3 misexpression on chondrocyte proliferation and long bone development were obtained in 4 of 6 independent experiments.

RT-PCR, Northern Blots, and In Situ Hybridization

These procedures have all been detailed in previous studies (Koyama et al., 1995; Nakamura et al., 1997; Yagami et al., 1999). Briefly, RNA was isolated from chondrocyte cultures with TRIZOL (Life Technologies, Inc.) following manufacturer's protocol. For RT-PCR, 5 μg of total RNA was reverse-transcribed by using random primer synthesis kits (Gibco). Amplification was performed with syndecan-3–specific primers for 30 or 35 cycles using denaturation at 94°C for 2 min, annealing at 55°C for 1 min, and extension at 72°C for 1 min. Polymerase chain reaction (PCR) products were resolved by electrophoresis on 1% agarose gels. For Northern blots, 20 μg of total RNA were denatured by glyoxalation, electrophoresed on 1% agarose gels, and blotted. Blots were stained with 0.04% methylene blue to verify transfer efficiency and were hybridized for 16 hr to 32P-labeled DNA probes (2.5 × 106 dpm/ml) in hybridization solution containing 50% formamide, 6× standard saline citrate (SSC), 1% SDS, 200 μg/ml sheared denatured salmon sperm DNA, and 10× Denhardt's reagent at 42°C. Blots were rinsed at a final high-stringency of 0.1× SSC and 0.5% SDS and exposed to Kodak X-ray films at −70°C. For in situ hybridization, 5-μm-thick paraffin sections were pretreated with 1 μg/ml proteinase K for 1 min, post-fixed in 4% paraformaldehyde buffer for 10 min, washed twice in PBS containing 2 mg/ml glycine for 10 min/wash, and treated for 15 min with a freshly prepared solution of 0.25% acetic anhydride in triethanolamine buffer. Sections were hybridized with 35S-labeled antisense or sense riboprobes (approximately 1 × 106 DPM/section) at 50°C for 16 hr, washed three times with 2× SSC containing 50% formamide at 50°C for 20 min/wash, treated with 20 μg/ml RNaseA for 30 min at 37°C, coated with Kodak NTB3 emulsion diluted 1:1 with water, exposed for 7–10 days, and developed with Kodak D19. After staining with hematoxylin and eosin, slides were analyzed with a Nikon microscope by using bright- and darkfield optics. The cDNA clones encoding portions of chick syndecan-3, Ihh, type X collagen, Patched and osteopontin were described previously (Koyama et al., 1995, 1996a; Yagami et al., 1999).


We thank Dr. Robert Kosher (University of Connecticut) for providing a full-length cDNA clone of chick syndecan-3, and Dr. William R. Abrams for help with informatics. M.P. was funded by the NIH.