12.2.1

Canonical and noncanonical Wnt signaling pathways

In the canonical Wnt pathway ( Fig. 12.1 ), in the absence of Wnts or if the binding to their receptors is repressed, β-catenin is associated with an intracellular complex (the destruction complex) containing Axin, adenomatous polyposis coli (APC), casein kinase I (CKI), and glycogen synthase kinase 3 (GSK3), and is degraded by ubiquitin-mediated proteolysis. Tcf/Lef transcription factors in the nucleus are bound to transcriptional corepressors (Groucho) thereby preventing gene transcription. Binding of the appropriate Wnt proteins to the Fzd-LRP5/6 coreceptor complex recruits and activates the cytoplasmic signaling protein Dishevelled (Dvl) to the cytoplasmic tail of Fzd. Dvl in turn recruits the axin–GSK3 complex, leading to LRP5/6 cytoplasmic tail phosphorylation and inhibition of the β-catenin destruction complex (axin–GSK3–Ck1–APC), stabilizing cytosolic β-catenin; β-catenin then accumulates in the cytoplasm and translocates to the nucleus where it interacts with Tcf/Lef transcription factors to activate transcription of Wnt downstream target genes .

Noncanonical Wnt signaling triggers its effects via alternative pathways through binding to Fzd or coreceptor complexes of Fzd/Ror2 or Ryk . Noncanonical pathways are complex and include Wnt/Ca ²⁺ , Wnt/PCP, Wnt/Rho-Rac, Wnt/G-protein-coupled receptors, Wnt/Ror, Wnt–RYK, and Wnt–mTOR pathways ( Fig. 12.1 ). These pathways are involved in developmental processes such as PCP and cell migration and recently some of them have been associated with the regulation of osteoblast differentiation and postnatal bone homeostasis . In the noncanonical Wnt signaling pathways, Wnt proteins signal through the Fzd receptors and activate Dvl leading to the activation of multiple distinct downstream effectors. Thus in these pathways, Dvl functions as a central player. On the other hand, noncanonical Wnt proteins can also signal through membrane receptors such Ror2 and Ryk, thereby activating downstream Wnt signals independently of Dvl .

12.2.2

The ligands: Wnt family of secreted proteins

The findings that Drosophila, Caenorhabditis elegans, Xenopus , and higher vertebrates express multiple Wnt genes indicate that these proteins have only partially redundant biological functions. Wnts are secreted cysteine-rich glycoproteins historically classified as either “canonical” or “noncanonical” depending on their ability/preference to activate β-catenin or β-catenin-independent signaling events, respectively. Although useful, this classification has been challenged by the evidence of cross talk between downstream effectors of the canonical and noncanonical cascades and by the findings that some Wnts can stimulate both pathways depending on the cellular context, the nature of receptors and coreceptors as well the presence of specific coactivators. For example, the “noncanonical” Wnt5a can induce Wnt/β-catenin signaling in the human 293 embryonic kidney cell line in the presence of overexpressed Fzd4 and LRP5 , while Wnt3a, the classical canonical Wnt ligand, can promote bone formation via Wnt/G-protein-coupled receptors . However, the molecular mechanism(s) by which Wnt proteins can activate one or the other signaling pathways remain to be elucidated.

12.2.3

Receptors and coreceptors

There are currently 10 known Fzd receptors. All are seven-transmembrane domain receptors with an extracellular N-terminus domain and an intracellular C-terminus domain. Fzd proteins contain a signal peptide, a cysteine-rich domain (CRD) in the N-terminal extracellular region (which interacts with both Wnts and other Wnt receptors), seven-transmembrane spanning domains, and a C-terminal PDZ domain-binding motif. Different Fzds regulate different intracellular signaling cascades . Importantly, the CDR domain is also found in tyrosine kinase receptors such as Ror2 (see later). Indeed, because Wnts bind the extracellular CRD of these receptors, Fzd receptors activate both canonical and noncanonical signaling, depending upon the nature of the coreceptors that are recruited by the ligand to the complex. Thus while the Wnt/Fzd/LRP complex activates β-catenin-dependent signaling, the Wnt/Fzd/Ror or Wnt/Fzd/RhoA–Rac1 activate noncanonical signaling cascades .

12.2.4

Other ligands: extracellular enhancers and antagonists

12.2.5

Intracellular effectors and regulators

The core of the canonical Wnt signaling pathway is the β-catenin destruction complex, which has a pivotal role in maintaining proper β-catenin levels in the cytoplasm and in the nucleus by promoting its ubiquitination and degradation. As proper levels of β-catenin are required for many biological events, the components of the destruction complex, GSK3–Axin–Ck1–APC, are important intracellular regulators of canonical Wnt signaling. Importantly, some of these intracellular regulators are also implicated in the regulation of noncanonical Wnt signaling and other signaling pathways, including BMP and FGF signaling.

GSK3, a serine/threonine kinase, is a negative regulator of several signaling pathways, including both canonical and noncanonical Wnt signaling. GSK3 phosphorylates β-catenin, thereby promoting its degradation through recruitment of beta-transducin repeat containing protein, ubiquitination and proteosomal degradation. Indeed, inhibition of GSK3-mediated β-catenin phosphorylation is the major mechanism that Wnt signals use to prevent β-catenin degradation . However, GSK3 also has a positive role in activating Wnt signaling, once it is recruited to the plasma membrane. Upon binding of Wnt ligands to Fzd/LRP5/6 receptor complexes, GSK3 and Axin move to the membrane where GSK3 phosphorylates LRP5/6 cytoplasmic tail, which in turn sequesters GSK3 thereby allowing β-catenin translocation to the nucleus and activation of gene transcription .

Axin1 and/or Axin2 act as a scaffold for all the components necessary for β-catenin degradation. In the presence of Wnt ligands, Axin together with other components of the destruction complex are recruited to LRP5/6 thereby activating downstream β-catenin signaling. In addition to bringing together the β-catenin destruction complex, Axin2 expression is a direct target of β-catenin, thus this scaffold protein functions as a negative feedback inhibitor of the canonical Wnt pathway .

APC is a tumor suppressor that is mutated in more than 80% of colon cancers. It has been established that the major function of APC is to antagonize canonical Wnt signaling. In addition to associating with β-catenin in the cytoplasm, APC also interacts with β-catenin in the nucleus where it prevents β-catenin from associating with Lef/Tcf transcription factors .

The CKI family of serine/threonine protein kinases is involved in many diverse and important cellular functions. Mammals have seven family members: α, β, γ1, γ2, γ3, δ, and ε. This family of kinases has been shown to regulate Wnt signaling both positively and negatively by interacting with and phosphorylating multiple components of this pathway. Targets for positive regulation of CKI include Dvl, Tcf3, and the coreceptor LRP6, whereas negative regulation occurs via APC and β-catenin where phosphorylation leads to complete or partial proteolysis of β-catenin and thereby inactivation of the pathway in the absence of ligand .

Another important intracellular regulator is Dvl, which has a central role in transmitting Wnt signaling from receptors to downstream effectors in both the canonical and noncanonical pathways . Three Dvl homologues (Dvl1, 2, and 3) have been identified in humans and mice. They contain three conserved domains: an N-terminal DIX (Dvl, Axin) domain, a central PDZ (postsynaptic density-95, discs large, zonula occludens-1) domain, and a C-terminal DEP (Dvl, Egl-10, Pleckstrin) domain. Studies in mouse models have demonstrated that Dvl is required for proper cardiac development, neural tube formation, and cochlear polarity . Dvl proteins can interact with a wide array of proteins and are thought to mediate protein–protein interactions and protein phosphorylation. Upon activation of the canonical Wnt pathway, Dvl is recruited by the receptor Fzd and prevents the destruction of cytosolic β-catenin by the proteosome. In the noncanonical Wnt pathways, Dvl is involved in the activation of several alternative pathways, including the Wnt/PCP signaling, calcium-dependent pathways, Wnt-Ror signaling, and Wnt/G protein signaling ( Fig. 12.1 ).

12.2.6

Ligand posttranslational modifiers

Wnts are known to be acylated. This posttranslational modification is essential for Wnt activity and is carried out by membrane-bound O-acyl transferases (MBOATs). Porcupine (Porcn), a Wnt–MBOAT, attaches palmitoleate and shorter cis unsaturated fatty acids onto Wnts . Palmiteoylation is required for (1) binding to Wntless, a protein that mediates Wnt intracellular transport; (2) secretion of Wnt ligands; and (3) binding to Frizzled receptors. Small-molecule inhibitors that block Porcn-mediated Wnt modification and Wnt signaling have been reported and have recently been shown to exhibit therapeutic efficacy in treating Wnt-driven cancers. However, it has been recently shown that in mice these porcupine inhibitors lead to decreased trabecular and cortical bone mass and strength due to both decreased bone formation and increased bone resorption On the other hand, a lipase that deacetylases Wnt proteins, Notum, has also been recently identified to negatively affect Wnt signaling . In particular, it has been shown that Notum enzymatically inhibits Wnt signaling activity by removing the palmitoleate moiety of Wnts, which contributes directly to Frizzled receptor binding . Deletion of the Notum gene in mice has been shown to increase cortical bone mass .

12.3.1

LRPs plays a key role in bone

The first study to show a link between Wnt signaling and bone mass reported that inactivating mutations in LRP5 in human cause osteoporosis pseudoglioma syndrome (OPPG, OMIM259770), an autosomal recessive disorder characterized by severe early-onset osteoporosis . Shortly after these observations, two groups independently reported a specific point mutation (G171V) in LRP5 resulting in high bone mass (HBM) . Most interestingly, this mutation decreased the LRP5 binding affinity of two endogenous Wnt inhibitors, Dkk1 and sclerostin, resulting in increased canonical Wnt signaling. Further support for a key role of LRP5 in skeletal homeostasis was provided by several GWAS, which showed association between LRP5 and bone mass and fracture risk in diverse human populations .

Studies in animal models have confirmed that loss- and gain-of-function of LRP5 recapitulate the OPPG and the HBM phenotype. Thus while mice with germline deletion of LRP5 develop low bone mass due to impaired osteoblast proliferation and bone formation rate, transgenic mice expressing the G171V mutation in immature osteoblasts develop an HBM phenotype because of increased bone formation and decreased osteoblast and osteocyte apoptosis . However, no changes in osteoclastogenesis and bone resorption were observed in these mice. Thus overall these studies suggested that LRP5, expressed in osteoblasts, regulates bone formation via proliferation and apoptosis of cells of the osteoblast lineage.

These findings were challenged when Yadav et al. reported that neither conditional deletion of LRP5 in osteoblast progenitors or mature osteoblasts nor conditional knock-in of the G171V mutation into mature osteoblasts resulted in changes in bone formation . These authors suggested that it is LRP5 expressed in the duodenum that, via serotonin secretion, regulates bone formation. In support of this hypothesis were the findings that OPPG patients have high levels of circulating serotonin, whereas patients with HBM have lower serotonin levels . These findings were very surprising in view of the phenotype of transgenic mice expressing the G171V mutation in LRP5 targeted to osteoblasts . Independent of whether gut-derived serotonin affects bone formation or not, a very thorough recent study strongly supports the earlier findings that LRP5 functions locally in bone to regulate bone mass . In these studies, conditional deletion or knock-in of the HBM mutation of LRP5 in osteocytes led to a marked decrease or increase in bone mass, respectively, while LRP5 deletion in the intestinal stem cells did not affect bone mass. In addition, knock-in of the HBM mutation in LRP5 in limb mesenchymal progenitors resulted in increased bone mass only in the appendicular skeleton, firmly establishing the local nature of these events, and excluding a significant role of gut-derived factors. Although the reason(s) for this inconsistency is still unknown, the results from Cui et al. unequivocally establish that the HBM mutation exerts its effects in cells of the osteoblast lineage, and most probably osteocytes.

Other members of the LDLRs family have also been shown to be involved in the regulation of bone mass. Mutations in LRP6 have been linked to changes in bone mass and haploinsufficiency of LRP6 in LRP5 null mice leads to more severe osteopenia, indicating that these receptors are partially redundant . In addition, it has recently been shown that LRP4 acts as a coreceptor for Dkk1 and sclerostin in vitro and that LRP4-deficient mice develop shortened femurs, reduced cortical femoral perimeter, reduced total femur bone mineral content, and BMD .

12.3.2

β-Catenin and bone

Since LRP5/6 are coreceptors for β-catenin-dependent Wnt signaling, the role of these receptors in the regulation of bone mass requires consideration of how β-catenin affects bone mass. Gain-and-loss-of-function studies in mice have shown an essential role for β-catenin in osteoblast differentiation and in the regulation of bone mass and emphasized that the ultimate response to activation of the canonical Wnt signaling is cell stage-dependent.

Deletion of β-catenin in early skeletal progenitor cells leads to impaired osteoblast differentiation and ectopic chondrogenesis in the mouse embryo demonstrating a role for β-catenin in osteoblast proliferation and differentiation and in the balance between osteoblasts and chondrocytes ( Fig. 12.3 ). On the other hand, mice expressing constitutively active β-catenin in early mesenchymal cells display an absence of appendicular and skull bones, suggesting that increased levels of β-catenin during embryonic development negatively affect skeletal development . Moreover, activation of the canonical Wnt pathway regulates the expression of OPG, the decoy RANK receptor that inhibits osteoclast differentiation by interacting with RANKL. Disruption of β-catenin in mature osteoblasts and in osteocytes in mice affects bone resorption rather than bone formation because of elevated expression of RANKL and decreased expression of OPG . Confirming these findings, constitutive activation of β-catenin in osteoblasts and osteocytes results in decreased osteoclast numbers . Thus canonical Wnt signaling controls osteoblasts at different levels: commitment, proliferation/apoptosis, and function. In addition, β-catenin can directly regulate osteoclastogenesis . Using different Cre-mice to delete or constitutively express β-catenin in osteoclast progenitors, macrophage precursors, and mature osteoclasts, these studies have suggested that constitutive activation of β-catenin blocks osteoclast differentiation and β-catenin deletion impairs osteoclast proliferation, both causing osteopetrosis. On the other hand, 50% deletion of β-catenin leads to osteoporosis due to increased osteoclast differentiation. Thus deregulation of β-catenin has a biphasic effect on osteoclasts, inhibiting osteoclastogenesis when activation is either low or high.

Because the phenotypes of the β-catenin and LRP5 null mice are different, it is important to mention that, on the one hand, β-catenin activation may not be the only signaling event downstream of Wnt receptor complexes and, on the other hand, β-catenin may also function independently of the Wnt signaling pathway to regulate osteoblast proliferation and differentiation and bone formation. For instance, β-catenin interacts with N-cadherin, which has been shown to regulate osteoblast proliferation and survival, and mechanical strain in osteoblasts and osteocytes can activate β-catenin directly to promote bone formation .

12.3.3

Wnts and bone

Many Wnts have been shown to impact bone biology. Early experiments in chickens and in mice have demonstrated that ligands such as Wnt2b, Wnt3a, Wnt4, Wnt5a, Wnt8c, and Wnt7a are expressed early during limb development and are crucial for proper limb outgrowth and patterning along the proximal–distal and dorsoventral axes . Studies in mice have shown that several Wnt proteins are required for chondrocyte and osteoblast differentiation during limb development as well as for bone homeostasis. Importantly, several GWAS showed association of some of the Wnt proteins, including, Wnt4, Wnt5b, and Wnt16 with BMD and/or fracture risk . Thus loss- and gain-of-function studies have demonstrated that ligands such as Wnt1, Wnt3a, Wnt4, Wnt5a, Wnt7b, Wnt10b, and Wnt16 affect osteoblastogenesis, osteoclastogenesis, and adipogenesis thereby contributing to bone mass via β-catenin-dependent or β-catenin-independent pathways . In particular, the first functional characterization of Wnt16 has confirmed its critical role in the regulation of cortical bone mass and bone strength in mice, confirming Wnt16’s strong association with cortical bone fragility in GWAS . These studies also supported the emerging hypothesis that the homeostasis of the trabecular and cortical bone compartments is, at least in part, differentially regulated. Most interestingly, Wnt16 appears to be a strong antiresorptive soluble factor acting on both osteoblasts and osteoclast precursors.

12.3.4

Secreted Wnt antagonists and bone: sFRPs, Dkks, and Sclerostin

As mentioned earlier, gain-of-function mutations in LRP5 decrease the ability of antagonists such as Dkk1 and sclerostin to bind to LRP5, thereby leading to HBM. Sclerostin is the protein encoded by the Sost gene located on the human chromosome 17q12-q21. Mutations in the Sost gene cause sclerosteosis (MIM269500) and van Buchem disease (MIM239100) and polymorphisms of the Sost promoter have been linked to osteoporosis . Genetically engineered mouse models designed to mimic the mutations seen in sclerosteosis and van Buchem’s patients reproduce the HBM phenotypes seen in these patients, while transgenic expression of sclerostin in mice leads to osteopenia . Sclerostin is secreted by osteocytes and it is believed to reach the bone surface through the network of osteocyte canaliculi and to regulate osteoblast and osteoclast differentiation and function . However, recent studies in which Sost was deleted in mesenchymal progenitor cells, mature osteoblasts or osteocytes have shown that, although Sost deletion in each cell types leads to a significant increase in trabecular bone mass, only Sost deletion in mesenchymal progenitor cells (Prx1-Cre) fully recapitulated the high trabecular and cortical bone mass phenotype seen in the appendicular skeleton of the global KO as well as the B-cell phenotype. Importantly, despite normal expression of Sost in the axial skeleton of Prx1-Cre deleted mice, these mice developed HBM in the vertebrae, while the sclerostin released in circulation by the axial skeleton did not affect bone mass in the appendicular skeleton. Thus these findings add new insights into sclerostin and the cells secreting it and suggest that the only Sost expression affecting bone mass is derived from descendants of Prx1 -positive limb bud progenitors . As discussed later, sclerostin expression is regulated by mechanical loading and parathyroid hormone (PTH) and it is thought that the anabolic effects of both mechanical stimulation and PTH, are due, at least in part, to suppression of sclerostin expression and secondary enhancement of canonical Wnt signaling. In addition to expressing sclerostin, osteocytes also express high levels of RANKL , suggesting that sclerostin might function in osteocytes to support osteoclastogenesis by repressing the Wnt-induced increase in OPG .

Dkk1 also binds to LRP5/6 and suppresses canonical Wnt signaling, inhibiting bone formation. Dkk1 null mice die shortly after birth with severe developmental abnormalities . However, Dkk1 haploinsufficient mice are viable and develop increased bone formation and bone mass without changes in bone resorption . Recent studies have shown that Dkk1 deletion in cells of the osteoblastic lineage results in HBM with increased bone formation, despite increased levels of sclerostin and that osteoprogenitors but not mature osteoblasts or osteocytes are the main source of systemic Dkk1 . Consistent with these findings and confirming a negative role for Dkk1 in the regulation of bone mass, transgenic overexpression of Dkk1 in osteoblasts significantly decreases osteoblast number and bone formation rate and tends to increase bone resorption . Although Dkk1 mutations affecting bone mass in humans have not been discovered, high levels of Dkk1 have been implicated in the development of osteolytic lesions in metastatic bone disease and multiple myeloma . Similar to Dkk1, Dkk2 blocks canonical Wnt signaling and inhibits osteoblast differentiation pathways in osteoarthritic osteoblasts . However, it has been reported that Dkk2 activates β-catenin in Xenopus and that mice lacking Dkk2 develop osteopenia , suggesting that in contrast to the inhibitory functions of Dkk1 on bone mass, Dkk2 can stimulate bone formation, possibly by favoring the last stages of osteoblast maturation.

As mentioned previously, in contrast to sclerostin and Dkk1, sFRPs and Wif-1 bind to Wnt ligands preventing their functional association with both LRP5/6 and Fzd receptors. Although, sFRPs are all expressed in bone , they play quite distinct roles in skeletal development and homeostasis. Mice with germline deletion of sFRP1 develop modest increases in trabecular bone mass, mineral apposition rate, and total volume, but not in cortical BMD in adult mice , whereas transgenic sFRP1 overexpression decreased bone density . Mechanistically, sFRP1 deletion reduces osteoblast and osteocyte apoptosis in vivo, and cell proliferation and differentiation in vitro . It has also been suggested that sFRP1 binds RANKL thereby impairing osteoblast-induced osteoclastogenesis . In addition to sFRP1, other members of the family are involved in skeletal development and/or bone homeostasis. sFRP2 inactivation results in brachydactyly, without other obvious skeletal changes , while sFRP3 has been associated with the development of osteolysis and heterotopic ossification . sFRP4 is expressed in cells of the osteoblast lineage and it inhibits osteoblast proliferation by suppressing Wnt3a activity in vitro . Studies in mice have shown that sFRP4 expression is markedly increased in osteopenic accelerated aging SAMP6 mice and that overexpression/deletion of sFRP4 results in specific skeletal phenotypes . Directly relevant to osteoporosis in humans, several independent GWAS have linked sFRP4 to BMD, including cortical sites . Recent studies have established that SFRP4 loss-of-function mutations in humans are linked to Pyle’s disease , a rare genetic skeletal disease first described in 1931 by Edwin Pyle as a “case of unusual bone development” . The clinical hallmarks of Pyle’s disease are genu valgum and long bones characterized by wide and expanded metaphyses with increased trabecular bone, while cortical bone is thinner than normal. As a result, the bones are fragile and fracture easily. Affected individuals also present with mild cranial sclerosis, thin calvarium with expanded diploe and impaired formation of cranial sinuses. In addition, some patients present with a smaller head circumference (>2 SD smaller) or with an abnormal shape, characterized by prominence of the frontal bones and prognathism. Dental problems are also common in Pyle’s disease, including delayed eruption of permanent teeth, cavities, and malocclusion . Loss-of-function studies in mice have demonstrated that sFRP4 deletion results in a skeletal phenotype similar to that observed in Pyle’s disease and uncovered that, while activation of canonical Wnt signaling leads to increased trabecular bone mass, activation of the noncanonical Wnt/Jnk signaling cascade, together with increased BMP signaling and sclerostin levels, leads to decreased cortical thickness and homeostasis . sFRP4 has also been described as a phosphate regulator both in vivo and in vitro . However, ablation of sFRP4 in mice does not result in changes in serum or urine phosphate levels, thus dampening the potential role of sFRP4 in phosphate homeostasis .

12.3.5

Noncanonical Wnt signaling and bone

Unlike canonical Wnt signaling, the role of noncanonical Wnt signaling in bone remains to be fully elucidated, albeit a number of recent papers have indicated that it also plays a role in skeletal development and homeostasis. The Wnt7b–Gα _q/11 –PKCδ noncanonical cascade has been shown to be required for proper osteoblast differentiation and bone formation through the regulation of Runx2 expression . Oishi et al. reported that Wnt5a binds to Ror2 and activates JNK phosphorylation, suggesting that during embryonic development Ror2 functions as a receptor for Wnt5a to activate noncanonical signaling. Indeed, Wnt5a or Ror2 null mice display a similar phenotype characterized by dwarfism, facial abnormalities, and shortened limbs . In humans, Ror2 mutations cause brachydactyly type B, a dominant skeletal disorder characterized by hypoplasia/aplasia of distal phalanges , and Robinow syndrome, a recessive condition characterized by short stature and limb bone shortening . Confirming these findings, Wnt5a is required not only for osteoblastogenesis but also for proper osteoclastogenesis via a Ror2–JNK signaling, suggesting a role for noncanonical Wnt signaling in the regulation of postnatal bone homeostasis . In addition, recent work has shown that Sfrp4-dependent suppression of the noncanonical Wnt5a/Ror2/Jnk cascade in osteoclasts, but not that of β-catenin signaling, regulates osteoclastogenesis and protects cortical bone from excessive endocortical resorption . Furthermore, Takada et al. have shown that Wnt5a heterozygous mice display an osteopenic phenotype, with decreased trabecular bone mass and increased bone marrow adipogenesis. Both of these effects were shown to be due to CAMKII–TAK1–TAB2–NLK-dependent activation of PPARγ transcriptional activity .

Thus collectively these studies suggest that, similar to canonical Wnt signaling, activation of noncanonical Wnt signaling can affect osteoblastogenesis, bone marrow adipogenesis, and osteoclastogenesis ultimately regulating bone mass ( Fig. 12.3 ).

12.4.1

Wnt and bone morphogenetic protein signaling

BMP and Wnt signaling are essential signals implicated in multiple events during skeletal development and bone formation. BMPs share high homology with the decapentaplegic (dpp) protein in Drosophila , and with the Xenopus Vg-1, which are proteins involved in early embryo pattern formation . BMPs were originally identified as molecules derived from bone, which were capable of inducing new bone formation ectopically when subcutaneously injected in animal models . In vitro, BMP-2 has been reported to induce, the expression of many osteoblast-specific genes in different cell types and analysis of mutations of mammalian BMP signaling have confirmed the critical role for members of this family both in skeletal development and postnatal bone . In addition, human disorders such as chondrodysplasia and fibrodysplasia ossificans progressiva confirm that BMP signaling plays an important role in regulating bone formation.

Although, the role of BMP signaling in promoting bone formation is clear, preclinical studies in animals have shown that inhibition of BMP signaling has an anabolic effect on bone and that activation of BMP signaling by deleting Noggin results in osteopenia . This can be explained by the findings that BMP signaling represses Wnt canonical signaling through induction of sclerostin and subsequently bone formation. Indeed, accumulated evidences suggest that the BMP and Wnt signaling pathways regulate one another both synergistically and antagonistically in a context and age-dependent fashion . In vitro studies have demonstrated that canonical Wnt signaling induces BMP expression and that BMPs induce Wnt expression and in vivo studies have confirmed the interaction between these two signaling pathways. It has been shown that BMP signaling inhibits Wnt signaling by increasing Sost expression and decreases bone mass by supporting osteoclastogenesis through the RANKL–OPG pathway .

12.4.2

Wnt and parathyroid hormone signaling

It is well established that PTH controls the rate of bone remodeling in both humans and animal models. PTH, which exerts its action via the PTH/PTH-related peptide type 1 receptor (PTH1R), functions both as an anabolic and catabolic effector in bone depending upon whether its action is continuous, as in hyperparathyroidism, or intermittent when used in therapeutic intervention for osteoporosis. Intermittent administration of PTH in rodents induces osteoblast differentiation, prolongs the life span of mature osteoblasts, and favors bone formation. In contrast, continuous administration of PTH, which also increases osteoblast number and bone formation, increases RANKL production by these cells thereby increasing bone resorption as well .

In the last few years, several studies have demonstrated that PTH signaling interacts with and regulates canonical Wnt signaling through multiple mechanisms, supporting the hypothesis that the anabolic effect of PTH in bone is mediated, at least in part, via activation of the canonical Wnt signaling pathway in osteocytes . Most importantly, PTH represses the expression of Sost in vitro and in vivo by inhibiting Mef2c (myocyte enhancer factor 2), a transcription factor required for activation of Sost expression . PTH also inhibits the expression of other Wnt signaling agonists, such as Dkk1 and Wif-1 . In addition, similar to the findings in Sost null mice, sFRP1 affects, in a gene dosage–dependent fashion, the PTH-induced increase in trabecular bone, indicating that activation of Wnt signaling contributes to the anabolic action of PTH . Mice lacking PTH1R in osteocytes develop osteopenia with increased Sost expression and decreased canonical Wnt signaling , whereas osteocyte-specific constitutive activation of the PTHR1 leads to increased canonical Wnt signaling, increased cortical bone area, and a higher rate of periosteal and endocortical bone formation . Deletion of the Wnt receptor LRP5 in these mice reduces the increase in bone mass and PTH-induced bone gain is blunted in Sost-deficient or overexpressing mice, although with different effects on bone formation and resorption . Thus these findings demonstrate that PTH signaling regulates bone homeostasis in large part through osteocytes and emphasize the role of Wnt signaling in PTH responses.

12.4.3

Wnt and Hippo signaling

An emerging signaling cascade involved in cell fate, tissue development and homeostasis, the Hippo/YAP–TAZ signaling, has been recently implicated in bone biology . The Hippo pathway is a critical regulator of cell proliferation and organ size, and uncontrolled activity of this signaling cascade leads to tissue overgrowth, mainly due to increased cell proliferation . In mammals the Hippo pathway is a kinase cascade comprising the Mst1/2 and Lats1/2 kinases that end up phosphorylating and inhibiting YAP and TAZ in the cytoplasm. In this cascade, YAP and TAZ shuffle between the cytoplasm and the nucleus where they mediate the transcription of genes involved in cell proliferation, apoptosis, and differentiation. As such, their nuclear accumulation is the key determinant of their function. It has been shown that TAZ is a coactivator of Runx2 activity and several studies have now reported that deletion of TAZ and/or YAP in mature osteoblasts leads to impaired osteoblast differentiation and decreased bone mass in mice. Confirming these findings, Zebrafish embryos with decreased TAZ levels have defective bone formation, and mice overexpressing TAZ in osteoblasts display increased bone mass. The regulation of the Hippo cascade by upstream cues remains captivating in that no specific ligands have yet been identified. Multiple layers of interaction between WNT signaling and TAZ and YAP have been described . However, the mechanisms of the WNT–Hippo interaction or of how TAZ and YAP function independent of Hippo are not clear, especially in skeletal homeostasis. It has been proposed that cytoplasmic interactions between β-catenin and TAZ/YAP as well as TAZ and DVL inhibit canonical WNT signaling . Azzolin et al. have proposed a WNT/YAP/TAZ axis in which both TAZ and YAP are part of the destruction complex and WNT signaling activation leads to their translocation into the nucleus with subsequent activation of YAP/TAZ biological responses . Although these studies clearly show that YAP/TAZ intersects with canonical WNT signaling, these interactions have mainly been examined in the light of their ability to influence cell fate decisions, while their effects in the context of skeletal homeostasis remain unclear.