Nuclear translocation of β-catenin in osteoblasts. (A) Wnt3a-conditioned medium (Wnt3a) stimulated β-catenin translocation in MC3T3-E1 cells (arrows) compared with cells incubated with control conditioned medium (L-cells). Bar, 5 μm. (B) Dose-dependence of Wnt3a induced β-catenin translocation in MC3T3-E1 cells. No effect was observed in cells exposed to L-cell conditioned medium at all concentrations tested. (C) Dkk-1 inhibited the nuclear accumulation of β-catenin induced by 25% Wnt3a in MC3T3-E1 cells (mean ± s.e.m., n=12; ^**P<0.01, ^***P<0.001; n.s., not significant).

Mimicking Wnt signalling in osteoblasts stimulates β-catenin nuclear translocation and TCF/LEF-dependent transcriptional activation. (A) Inhibition of GSK3β with LiCl (20 mM) stimulated β-catenin nuclear translocation in SaOS-2 cells (arrows) compared with NaCl controls. Bar, 5 μm. (B) Inhibition of GSK3β with LiCl (20 mM) stimulated TOPFLASH TCF/LEF gene transcription in SaOS-2 cells compared with cells incubated with vehicle alone or those transfected with control reporter containing mutant TCF/LEF binding sites (FOPFLASH). Mean values are shown ± s.d., n=4. (C) Time course of LiCl-stimulated gene transcription in MG-63 cells. LiCl stimulated TCF/LEF reporter gene transcription from 5 hours post treatment (mean ± s.e.m., n=8; ^***P≤0.001).

β-catenin transactivates TCF/LEF gene transcription in osteoblasts. (A) Overexpression of β-catenin EGFP fusion protein significantly increased TCF/LEF-dependent transcription of a luciferase reporter gene (TOPFLASH) in MG-63 cells compared with cells transfected with control reporter containing mutant TCF/LEF binding sites (FOPFLASH). By contrast, overexpression of a β-catenin C-terminal mutant lacking the transactivation domain (β-catenin ΔC695-781) failed to activate transcription (values are mean ± s.d., n=4; ^***P≤0.001; n.s., not significant). (B) Both full-length β-catenin and β-catenin ΔC localised to the nucleus of osteoblasts (arrows). EGFP, Green; DAPI, blue. Bar, 10 μm.

(A) Schematic representation of the human RANKL promoter illustrating putative TCF/LEF transcription factor binding sites. MattInspector software was used to examine the human RANKL promoter (GenBank accession number AF333234) for the presence of the consensus TCF/LEF binding sequence ctttgww. Relative positions and nucleotide sequences of putative binding sites are shown (a-d: a, ctttgaa; b, ttcaaag; c, ctttgat; d, ctttgta). Positions indicated are relative to the translational start codon of the RANKL mRNA. Functional vitamin D₃ response elements (VDRE) are shown for comparison (Kitazawa et al., 2003). (B) Mimicking Wnt signalling in osteoblasts inhibited expression of receptor activator of NFκB ligand (RANKL). Inhibition of GSK3β with LiCl (20 mM) potently inhibited RANKL mRNA expression by MC3T3-E1 cells and protein expression by MG-63 osteoblastic cells compared with the NaCl control. (C) RT-PCR analysis illustrating the inhibitory effects of recombinant Wnt on RANKL mRNA expression by MC3T3-E1 cells. Exposure to Wnt3a (100 ng/ml) inhibited expression at levels comparable with LiCl (20 mM). Expression was unaffected by treatment with recombinant Wnt5a (500 ng/ml). (D) Overexpression of full-length β-catenin inhibited basal activity of the RANKL promoter in ST2 stromal cells. Inhibitory effects were dependent on the C-terminal transactivation domain of β-catenin deleted in β-catenin ΔC695-781. Values are mean ± s.e.m., n=8; ^***P≤0.001; n.s., not significant.

Wnt signalling inhibits osteoclastogenesis in mouse primary osteoblast/mononuclear spleen cell co-cultures. (A) Inhibition of GSK3β with LiCl (5-20 mM) inhibited the formation of TRAP-positive multinucleated cells compared with cells incubated with NaCl (5-20 mM) over 12 days in culture. (B) LiCl inhibited mononuclear cell fusion (arrows indicate characteristic cell fusion in control cultures; open circles, mononuclear cells; dashed arrow, rare fused cell in LiCl-treated cultures). TRAP stained osteoclasts predominate in NaCl-treated cultures but are absent in LiCl treated cultures. DIC, digital interference contrast. (C) Wnt3a conditioned medium (Wnt3a) inhibited the formation of TRAP-positive multi-nucleated cells compared with cells incubated with control conditioned medium (L-Cell) over 12 days in culture. Values are mean ± s.e.m., n=5; ^***P≤0.001; n.s., not significant.

Expression of LRP5 and LRP6 Wnt co-receptors by osteoblasts, osteoclast progenitors and mature osteoclasts. (A) RT-PCR demonstrating expression of LRP5 and LRP6 using cDNA generated from human osteoblastic cells (MG-63, SaOS-2 and TE85), primary human osteoblasts (HOB), mouse spleen and mature osteoclasts derived from human long-term bone marrow cultures. (B) Histogram illustrating the purity of CD14⁺ cells determined by flow cytometry following isolation from the mononuclear (MNC) fraction of human peripheral blood by magnetic activated cell sorting. RT-PCR demonstrating expression of LRP5 and LRP6 by CD14^- (depleted) and CD14⁺ (enriched) cells is shown on the right. The PCR product obtained using SaOS-2 cDNA is shown for comparison.

Effects of Wnt signalling on human osteoclast formation and activity. Wnt3a (10-100 ng/ml) had no effect on the formation of TRAP-positive multinucleated osteoclasts generated from peripheral blood mononuclear cells cultured in the presence of RANKL and M-CSF (open bars) and had no effect on their resorptive activity (closed bars). Data are mean ± s.e.m., n=5.

Evidence for endogenous Wnt signalling in mature osteoclasts. Fluorescent immunolocalisation of β-catenin in mature disaggregated rat osteoclasts. β-catenin predominately localised to nuclei with some peripheral staining (green). Multinucleation is disclosed by DAPI (blue) and peripheral actin ring formation, typical of active bone resorbing osteoclasts is stained with Rhodamine-phalloidin (red). Bar, 10 μm.