Regulation of bone metabolism by Wnt signals

Abstract

Wnt ligands play a central role in the development and homeostasis of various organs through β-catenin-dependent and -independent signalling. The crucial roles of Wnt/β-catenin signals in bone mass have been established by a large number of studies since the discovery of a causal link between mutations in the low-density lipoprotein receptor-related protein 5 (Lrp5) gene and alternations in human bone mass. The activation of Wnt/β-catenin signalling induces the expression of osterix, a transcription factor, which promotes osteoblast differentiation. Furthermore, this signalling induces the expression of osteoprotegerin, an osteoclast inhibitory factor in osteoblast-lineage cells to prevent bone resorption. Recent studies have also shown that Wnt5a, a typical non-canonical Wnt ligand, enhanced osteoclast formation. In contrast, Wnt16 inhibited osteoclast formation through β-catenin-independent signalling. In this review, we discussed the current understanding of the Wnt signalling molecules involved in bone formation and resorption.

Bone remodelling is a continuous cycle of bone resorption and formation that maintains a constant bone mass and calcium homeostasis throughout the lives of vertebrates. This process is achieved by bone-resorbing osteoclasts, bone-forming osteoblasts and matrix-embedded osteocytes. An imbalance between bone resorption and formation leads to either osteopenia or high bone mass.

Osteoclasts are differentiated from monocyte–macrophage lineage cells (1). This differentiation is tightly regulated by osteoblast-lineage cells such as osteoblasts (2) and osteocytes (3, 4). Osteoblasts secrete two essential cytokines for osteoclast differentiation: receptor activator of the Nf-κb ligand (Rankl) and colony-stimulating factor-1 (Csf-1) (1). The expression of Rankl is induced by bone resorption factors such as 1α,25-dihydroxy vitamin D₃ [1α,25(OH)₂D₃] and parathyroid hormone (PTH) (1). Rankl binds to its receptor Rank in osteoclast precursors, which, in turn, activates several signalling pathways including Nf-κb and c-Fos (5). These signals induce the expression of nuclear factor-activated T cell cytoplasmic 1 (Nfatc1) (6) in osteoclast precursors, thereby inducing the expression of osteoclast-related genes such as cathepsin k and the dendritic cell (DC)-specific transmembrane protein (5, 7, 8). In contrast, osteoprotegerin (Opg), a decoy receptor of Rankl secreted from osteoblast-lineage cells, inhibits the binding of Rankl with Rank to negatively regulate osteoclastogenesis (9).

Analyses of loss-of-function and gain-of-function mutations in the Lrp5 gene, a co-receptor of Wnt/β-catenin signalling, revealed that Wnt/β-catenin signals regulated osteoblast differentiation in order to increase bone mass (10–12). Furthermore, experiments using conditional deletion of the β-catenin gene in skeletal progenitors such as perichondrial and periosteal cells showed that β-catenin was essential for osteoblast differentiation (13, 14). The early markers of the osteoblast lineage, such as alkaline phosphatase (Alpl) and Runx2, were shown to be still expressed in the perichondrial and periosteal cells of the β-catenin conditional knockout (cKO) mice (13). In contrast, the perichondrial and periosteal cells in these mutant mice failed to express osterix, the osteoblast commitment transcription factor. These findings indicated that Wnt/β-catenin signalling played a role in osteoblast differentiation, which, in turn, regulated bone formation.

In this review, we introduced recent findings on the molecules involved in Wnt signalling and discussed the roles of these molecules in bone formation and resorption.

Roles of Wnt Signals in Bone Formation

The roles of Lrp5 and β-catenin in bone formation have been established by a large number of studies (11). To date, 19 Wnt ligands have been identified in humans (15). In addition to the multiplicity of Wnt ligands, osteoblasts, osteocytes, chondrocytes and bone marrow cells have been reported to secrete various Wnt ligands (16). Therefore, difficulties have been associated with clarifying which types of cells are important as the source for bone accrual using gene knockout strategies against each Wnt ligand. However, this was solved by the discovery of Wntless (Wls) (17, 18).

Wls was identified as a novel transmembrane protein involved in Wnt signalling. The human Wls gene encodes a transcript for a 541 amino acid protein with eight membrane-spanning regions. The RNAi-mediated knockdown of Wls failed to activate β-catenin-dependent signals in Drosophila and human cells expressing Wingless (Wg) and Wnt3a, respectively (17). In contrast, the addition of exogenous Wg or Wnt3a into these cultures abrogated the effects of the knockdown of Wls on these signals. Furthermore, the knockdown of Wls markedly reduced the amount of Wg in culture media. These findings indicated that Wls was required for the secretion of Wnt ligands from Wnt-producing cells (Fig. 1).

Fig. 1.

Roles of sclerostin, Lrp4 and Wls in Wnt/β-catenin signals in osteoblasts. Wnt binds the receptor complex of Lrp5/6 and Frizzled and then activates Wnt/β-catenin signals. Osteoprotegerin is a target gene of this signalling pathway. Sclerostin secreted from osteocytes binds Lrp4 and Lrp5/6 to inhibit Wnt/β-catenin signals. Wls is involved in the secretion of Wnt ligands.

In order to clarify the importance of Wnt ligands secreted from osteoblasts in bone mass, osteoblast-specific Wls cKO mice were generated by crossing Wls-floxed mice with osteocalcin-Cre mice (Cre recombinase is expressed in mature osteoblasts), and their bone phenotypes were analysed (19). Wls cKO mice exhibited severe reductions in trabecular and cortical bone masses. A histomorphometric analysis revealed that the bone formation rate was significantly lower in Wls cKO mice, whereas the osteoblast surface was higher. Furthermore, the expression of osteoblast-related genes, such as Alpl and Sp7 (encoding osterix), and mineralization were impaired in Wls-deficient osteoblast cultures. These findings indicated that Wnt ligands secreted from osteoblasts were critical for the maturation and mineralization of osteoblasts.

The osteoclast surface per bone perimeter was higher in Wls cKO mice, similar to β-catenin cKO mice using Col (I) α1-Cre (20). Although the expression of Opg was decreased in β-catenin cKO mice, neither Opg nor Rankl expression was changed in Wls cKO mice (19). This finding suggested that Wnt ligands such as Wnt16 secreted from osteoblasts inhibited osteoclast formation in an Opg-independent manner. The inhibitory actions of Wnt 16 in osteoclastogenesis have been discussed in the next section of this review.

The importance of Lrp5/6 in bone formation has been underscored by studies on Wnt antagonists including sclerostin (encoded by the Sost gene) (21). Sclerostin has been implicated in the pathogenesis of sclerosteosis and van Buchem syndrome, which are both characterized by bone overgrowth (22, 23). Mutations in the SOST gene that disrupt its expression are associated with these diseases. Consistent with the bone phenotype observed in sclerosteosis, Sost-deficient mice exhibited a high bone mass with increased bone formation (24). Sclerostin, preferentially secreted from osteocytes, inhibited Wnt/β-catenin signalling by binding to Lrp5/6 (25).

The expression of Sost is regulated by its proximal promoter and the distal enhancer region ECR5, which is located within the 52-kb non-coding region. Deletion of the ECR5 region was identified in van Buchem syndrome (23). Mice in which this enhancer was deleted showed a high bone mass with the decreased expression of Sost (26). The ECR5 region has been shown to contain a Mef2 response element (27). Deletion of the Mef2c gene in osteoblast-lineage cells caused a high bone mass phenotype (26). These findings indicated that the expression of Sost was positively regulated by the ECR5 region to which Mef2c binds.

PTH reportedly decreased the expression of Sost (28). The intermittent administration of PTH induced anabolic effects on bone by stimulating bone formation. This anabolic action was abrogated in Sost-deficient or -overexpressing mice (24). Furthermore, mice constitutively expressing the active PTH receptor (encoded by Pth1r) in osteocytes showed a high bone mass with increased β-catenin signals (29). In contrast, mice lacking Pth1r in osteocytes showed a low bone mass with the increased expression of Sost and decreased β-catenin signals (30). The precise mechanism by which PTH inhibits the expression of Sost through Mef2 has not yet been elucidated in detail.

Lrp4 has been identified as a binding partner of sclerostin (31). The expression of Lrp4 was shown to facilitate the inhibitory effects of sclerostin on β-catenin signalling (Fig. 1). Notably, two mutations in the LRP4 gene (R1170W and W1186S) have been detected in patients with sclerosteosis (31). These mutant forms of LRP4 reduced the binding capacity and inhibitory effects of sclerostin. These findings indicated that Lrp4 bounds to sclerostin in order to facilitate the action of sclerostin. Previous studies examined the bone phenotypes of Lrp4-deficient mice (32, 33). Osteoblast-specific Lrp4 cKO mice (Lrp4^flox/flox; osteocalcin-cre mice) exhibited a high bone mass with increased bone formation. These findings strongly indicated that Lrp4 facilitated the inhibitory action of sclerostin in vivo.

Non-canonical Wnt5a is also involved in osteoblast differentiation. We previously demonstrated that Wnt5a^+/– mice exhibited a low bone mass with impaired osteoblast and osteoclast differentiation (34). Ex vivo experiments revealed that the formation of mineralized nodules was lower in Wnt5a^–/– osteoblast cultures. Furthermore, the expression of Alpl, osterix and osteocalcin was also lower in Wnt5a^–/– osteoblasts in culture, whereas that of adipocyte marker genes including aP2 was higher. These findings suggested that Wnt5a promoted osteoblast differentiation and prevented adipocyte differentiation. In order to further clarify the effects of Wnt5a on osteoblast differentiation, we examined the expression of co-receptors for Wnt signalling in Wnt5a^–/– osteoblast cultures (35). The expression of Lrp5/6 was lower in Wnt5a^–/– osteoblasts. Although Wnt5a bound to its receptor, receptor tyrosine kinase orphan receptor 2 (Ror2), the expression of Lrp5/6 in Ror2^–/– osteoblasts remained unchanged. Notably, the expression of Axin2, a target gene of Wnt/β-catenin signals, and T-cell factor (Tcf)/lymphoid enhancer factor (Lef) activity were lower in Wnt5a^–/– osteoblasts. The overexpression of Lrp5 rescued impaired Tcf/Lef activity and mineralization in Wnt5a^–/– osteoblast cultures. These findings indicated that Wnt5a enhanced the expression of Lrp5/6, thereby promoting osteoblast differentiation.

Roles of Wnt Signals in Bone Resorption

The roles of Wnt/β-catenin signals in osteoclastogenesis have already been established. Mice expressing a constitutively active form of β-catenin in osteoblasts (CA β-catenin mice) exhibited a high bone mass phenotype with impaired osteoclastogenesis (20). Bone formation in these mice remained unchanged. Furthermore, Opg was highly expressed in the osteoblasts of CA β-catenin mice in order to inhibit bone resorption.

Osteoclasts have been formed in co-cultures of osteoblasts and bone marrow cells in the presence of 1α,25(OH)₂D₃. Wnt3a, a typical canonical Wnt ligand, was previously shown to inhibit 1α,25(OH)₂D₃-induced osteoclast formation in these co-cultures, but not Rankl-induced osteoclast formation in osteoclast precursor cultures (36). These findings suggest that Wnt3a induces the expression of Opg in osteoblasts. The expression of Wnt16 mRNA was recently reported to be the highest in cortical bone among the various tissues tested, including trabecular bone (37). Wnt16 induced the expression of Opg through the activation of Wnt/β-catenin signalling in MC3T3-E1 cells, an osteoblastic cell line. 1α,25(OH)₂D₃-induced osteoclast formation was enhanced in co-cultures using Wnt16^–/– osteoblasts. These findings suggested that Wnt16 secreted from osteoblasts negatively regulated osteoclastogenesis. Furthermore, the effects of Wnt16 on Rankl-induced osteoclast formation in precursor cultures without osteoblasts were examined (37). This topic has been described in the latter part of this section.

We more recently confirmed that Wnt3a inhibited 1α,25(OH)₂D₃-induced osteoclast formation in co-cultures prepared from wild-type mice but not in those from Opg^–/– mice (38). This finding suggested that Wnt3a inhibited osteoclast formation through the up-regulated Opg expression in osteoblasts. Thus, Wnt/β-catenin signalling in osteoblasts induces the expression of Opg in order to inhibit osteoclast differentiation.

Non-canonical Wnt ligands such as Wnt5a activate β-catenin-independent signalling pathways such as the Wnt/Ca⁺⁺ (39) and Wnt/planer cell polarity pathways (40). Wnt5a has been shown to bind to Ror1 and Ror2, in order to regulate cell polarity, migration and invasiveness (41, 42). We previously reported that Wnt5a enhanced the expression of Rank in osteoclast precursors, which, in turn, promoted Rankl-induced osteoclast formation (34). We found that calvaria-derived osteoblasts strongly expressed Wnt5a, and that osteoclast precursors strongly expressed Ror2, but not Ror1. We generated osteoclast precursor-specific Ror2 cKO mice (Ror2 cKO) by crossing Ror2-floxed with Rank-Cre mice. Ror2 cKO mice exhibited a high bone mass with impaired osteoclastogenesis. We found that the expression of Rank mRNA, but not Csf-1 receptor, was lower in tibiae from Ror2 cKO mice. Mechanistically, Ror2 signals phosphorylated c-Jun N-terminal kinase, thereby activating c-Jun. Activated c-Jun was recruited to the specificity protein 1-binding sequences on the proximal Rank promoter, which, in turn, promoted the transcription of Rank mRNA.

As described above, Wnt16 reportedly inhibited Rankl-induced osteoclast formation in a Wnt/β-catenin signalling-independent manner (37) (Fig. 2). In contrast to the effects of Wnt16 on osteoblasts, Wnt16 failed to induce the cytosolic accumulation of β-catenin or expression of Axin2 in osteoclast precursors. Furthermore, Wnt16 suppressed Rankl-induced osteoclastogenesis in human and mouse osteoclast precursors in cultures. Wnt16 also inhibited Rankl-induced Nf-κb signals and suppressed the expression of Nfatc1 in these precursors. We confirmed that the treatment of bone marrow macrophages as osteoclast precursors with Wnt16 inhibited osteoclast formation and that Wnt16 inhibited 1α,25(OH)₂D₃-induced osteoclast formation in co-cultures prepared from Opg^–/– mice (38). These findings suggested that Wnt16 directly acted on osteoclast precursors, thereby inhibiting osteoclastogenesis in an Opg-independent manner.

Fig. 2.

Roles of Wnt5a, Wnt16,and Wnt4 in osteoclastogenesis. Wnt5a acts on osteoclast precursors and enhances the expression of Rank, thereby promoting Rankl-induced osteoclastogenesis. Wnt16 inhibits Rankl-induced Nf-κb and Nfatc1 signals, which, in turn, prevents osteoclastogenesis. Wnt4 also inhibits osteoclastogenesis through the expression of osteoprotegerin (Opg). Wnt5a abrogates the inhibitory effects of Wnt16 on Rankl-induced osteoclastogenesis.

Furthermore, Wnt4 inhibited osteoclast formation in a Wnt/β-catenin signalling-independent manner (43). Wnt4 transgenic (Wnt4-Tg) mice, in which Wnt4 was expressed in osteoblasts, were generated, and their bone phenotype was analysed. Wnt4-Tg mice were protected from estrogen deficiency-induced and tumor necrosis factor α (Tnfα)-induced bone loss. The number of osteoclasts was significantly increased in estrogen-deficient and Tnfα transgenic mice. The overexpression of Wnt4 in osteoblasts of these models also decreased the number of osteoclasts.

The treatment of osteoclast precursors with Wnt4 inhibited the Rankl-induced phosphorylation of transforming growth factor β-activated kinase 1, which suppressed the activation of Nf-κb signals (43). Consequently, Wnt4 suppressed Rankl-induced Nfatc1 expression in these cells. Wnt4 failed to activate Wnt/β-catenin signals in osteoclast precursors. Therefore, we tested the inhibitory effects of Wnt4 on the osteoclast formation in culture (38). Although Wnt4 failed to inhibit Rankl-induced osteoclast formation in osteoclast precursor cultures, it inhibited 1α,25(OH)₂D₃-induced osteoclast formation in co-cultures prepared from wild-type mice but not in those from Opg^–/– mice. This finding suggested that Wnt4 inhibited osteoclast formation through the expression of Opg (Fig. 2). Because Wnt4 reportedly failed to activate Wnt/β-catenin signalling in C2C12 cells or mesenchymal stem cells (44), it may change the expression of endogenous Wnt ligands and co-receptors such as Lrp5/6 in osteoblasts. We previously demonstrated that non-canonical Wnt5a enhanced the expression of Lrp5/6 in osteoblasts (35), thereby promoting Wnt/β-catenin signalling through endogenous Wnt10b.

Osteoblasts express Wnt16, an inhibitory Wnt for osteoclastogenesis, as well as Wnt5a, which promotes it. We subsequently examined the effects of Wnt16 on Wnt5a-induced Rank expression in osteoclast precursors using Rank-enhanced green fluorescent protein (EGFP) mice, in which EGFP was expressed in Rank-expressing cells (38). The treatment of osteoclast precursors with Wnt16 failed to enhance the expression of Rank or inhibit the Wnt5a-enhanced expression of Rank in the precursors. We also determined whether Wnt5a abrogated the inhibitory effects of Wnt16 on Rankl-induced osteoclast formation in cultures. Wnt5a partially abrogated the inhibitory effects of Wnt16 on Rankl-induced osteoclast formation. These findings suggested that Wnt16 mainly acted on osteoclast precursors expressing Rank and then inhibited osteoclast formation. The inhibitory effects of Wnt16 on osteoclast formation may be abrogated under Wnt5a-rich conditions such as arthritis (34, 45). Thus, Wnt5a and Wnt16 tightly regulate osteoclast formation in a manner that is dependent on various conditions.

Conclusion

The discovery of Wls led to the roles of Wnt ligands secreted from osteoblasts being clarified in bone formation. Wnt5a secreted from osteoblasts also facilitates Wnt/β-catenin signals through the promotion of Lrp5/6 expression. In addition, Wnt5a enhances Rankl-induced osteoclast formation through Ror2-mediated signals. In contrast, osteoblast-derived Wnt16 inhibits osteoclast formation by inhibiting Rank signals. Thus, Wnt ligands secreted from osteoblasts tightly regulate osteoblast and osteoclast differentiation. A more detailed understanding of Wnt signals may lead to the development of new therapeutic drugs for osteoporosis.

Funding

This work was supported in part by Grants-in-Aid for Scientific Research (KAKEN) [25221310 (N.T.) and 25293423 (Y.K.)] from the Ministry of Education, Cultures, Sports, Science and Technology of Japan.

Conflict of Interest

None declared.