Osteoblast dysfunctions in bone diseases: from cellular and molecular mechanisms to therapeutic strategies

Abstract

Several metabolic, genetic and oncogenic bone diseases are characterized by defective or excessive bone formation. These abnormalities are caused by dysfunctions in the commitment, differentiation or survival of cells of the osteoblast lineage. During the recent years, significant advances have been made in our understanding of the cellular and molecular mechanisms underlying the osteoblast dysfunctions in osteoporosis, skeletal dysplasias and primary bone tumors. This led to suggest novel therapeutic approaches to correct these abnormalities such as the modulation of WNT signaling, the pharmacological modulation of proteasome-mediated protein degradation, the induction of osteoprogenitor cell differentiation, the repression of cancer cell proliferation and the manipulation of epigenetic mechanisms. This article reviews our current understanding of the major cellular and molecular mechanisms inducing osteoblastic cell abnormalities in age-related bone loss, genetic skeletal dysplasias and primary bone tumors, and discusses emerging therapeutic strategies to counteract the osteoblast abnormalities in these disorders of bone formation.

Introduction

In the adult skeleton, the maintenance of the bone structure and strength is dependent on the balance between bone resorption and formation during the physiological process of bone remodeling. Bone remodeling begins with the recruitment and differentiation of bone resorbing osteoclasts, the resorption of a fraction of the calcified bone matrix followed by the replacement of the resorbed bone by new bone formed by osteoblasts. Both the quantity and the quality of bone structures are dependent on the balance between bone resorption and formation, and any defective bone formation relative to bone resorption causes bone loss and defective skeletal integrity [1, 2]. The excessive bone resorption in diseases such as osteoporosis or lytic bone tumors can be reduced by anti-resorptive drugs that target osteoclast differentiation or activity [3]. In contrast, bone disorders characterized by too much or disorganized bone formation (such as skeletal dysplasias or osteogenic bone tumors), or by reduced bone formation (caused by disuse or aging) are much more difficult to treat because of the limited number of safe and efficient drugs promoting osteoblast number and activity [4, 5]. In this context, the two important challenges in metabolic, genetic and oncogenic disorders of bone formation and their treatment are to identify the cellular and molecular mechanisms underlying these pathologies and to translate this knowledge into efficient therapy.

During the recent years, studies in human and mouse genetics greatly improved our knowledge of the mechanisms that control osteoblastogenesis and bone formation [6, 7]. Moreover, the identification of osteoblast dysfunctions in human skeletal diseases and mouse models led to propose novel targeted therapeutic approaches in these bone disorders. This article summarizes our current understanding of the cellular and molecular mechanisms underlying the osteoblast dysfunctions in selected metabolic, genetic and bone tumor pathologies, and focuses on emerging targeted therapeutic strategies that can attenuate the osteoblastic cell abnormalities in these bone disorders.

Mechanisms of osteoblastogenesis

Bone formation is a complex process [8] which starts with the recruitment of mesenchymal skeletal (stromal) cells (MSCs) located within the bone marrow stroma, or of pericytes in sinusoids in the bone marrow [9]. The commitment and differentiation of these cells to osteoblasts is dependent on the expression of the osteoblast transcriptional factors Runx2 and Osterix, followed by the expression of alkaline phosphatase, type 1 collagen and non-collagenous proteins, leading to matrix deposition and mineralization [10–12]. At the end of the bone formation period, some osteoblasts die by apoptosis or become flattened lining cells, whereas others are embedded in the bone matrix as osteocytes which control bone remodeling [13]. The activity of bone formation at the tissue level is dependent on both the number and function of differentiated osteoblasts [14, 15]. The number of osteoblasts depends on the osteogenic commitment from precursor cells and on the life-span of mature osteoblasts, whereas the amount of bone matrix produced by mature osteoblasts is dependent on the functional activity of each osteoblast [15, 16]. Both the number of functional osteoblasts and the activity of each osteoblast are controlled by transcriptional and epigenetic mechanisms [12, 17, 18] and are regulated by hormonal and local regulatory proteins [19], mechanical strain [20], cell–cell [21] and cell–matrix interactions [22]. Thus, any cellular or molecular abnormality in the recruitment, function or life-span of osteoblasts induced by metabolic, genetic or oncogenic diseases may impact bone formation at the tissue level, as detailed below.

Mechanisms of osteoblast dysfunctions in osteoporosis

Sex hormone deficiency at the menopause is associated with increased bone resorption, resulting in trabecular bone perforation, bone loss and increased risk of fractures [23]. Mechanistically, bone formation is insufficient to compensate for the increased bone resorbing activity. In addition to this mechanism, age-related bone loss occurs in both women and men as a slow process characterized by decreased osteoblast number and activity. This leads to thinning of bone trabeculae and cortical thickness, which contributes to bone fragility. Thus, the insufficient bone formation relative to bone resorption is an important mechanism mediating bone loss associated with aging and sex hormone deficiency [1]. Studies in humans and animal models allowed identifying multiple cellular mechanisms that contribute to the osteoblast dysfunctions in age-related bone loss. These mechanisms include the reduced recruitment and differentiation of osteoblast progenitor cells and the decreased life-span of mature osteoblasts, in addition to the reduced bone-forming capacity of individual osteoblasts [24–26] (Fig. 1a). Several extrinsic factors can cause these osteoblast dysfunctions in osteoporosis [27]. In addition, intrinsic senescence-related mechanisms, such as oxidative damage and epigenetic mechanisms affecting osteoblast genes, may contribute to the age-related defective osteoblastogenesis [26, 28] (Fig. 1a). Now that these pathogenic mechanisms have been identified in cells of the osteoblast lineage, the current aim is to find ways to target these processes and to develop effective therapeutic strategies to promote bone formation during skeletal aging [28, 29].

Fig. 1.

Mechanisms underlying the osteoblast dysfunctions in skeletal aging and potential therapeutic strategies to attenuate these osteoblast abnormalities. The defective bone formation associated with aging is characterized by preferential differentiation of mesenchymal stromal cells (MSCs) into adipocytes instead of pre-osteoblasts, reduced osteoblast differentiation and function and increased osteoblast and octeocyte death. These abnormalities result from alterations in local signaling factors, senescence mechanisms and epigenetic alterations of genes controlling osteoblastogenesis (a). Potential therapeutic strategies to reduce the osteoblast dysfunctions and improve bone formation in the aging skeleton include promoting anabolic signaling pathways, inhibiting negative regulators of bone formation, and attenuating senescence and epigenetic mechanisms in cells of the osteoblast lineage (b)

Current anabolic strategies in osteoporosis

Current therapeutic strategies to promote osteoblastogenesis in osteoporosis consist in increasing osteoblast number (by promoting the recruitment of osteoblast precursor cells or reducing osteoblast apoptosis), and increasing osteoblast activity [5, 30] (Fig. 1b). Up to now, however, bone anabolic molecules are limited [31] (Table 1). Intermittent administration of parathyroid hormone (iPTH) was shown to enhance osteoblast differentiation, function and life-span through various signaling pathways [32]. These effects result in increased bone microarchitecture and bone mass in osteopenic animals and in osteoporosis resulting from aging, sex hormone deficiency and glucocorticoid therapy [33]. Recent studies suggest that PTH-related Peptide (PTHrP) may be a physiological regulator of bone formation [34]. Recombinant hPTHrP1–34 and 1–84 were found to increase osteogenic cell differentiation in vitro [35] and intermittent treatment with a synthetic PTHrP 1–34 analog was recently shown to increase lumbar spine bone density in osteoporotic patients [36]. However, the clinical use of iPTH (or potentially iPTHrP) remains limited and other anabolic agents are needed for improving bone formation in osteopenic disorders [37]. In this context, several factors and signaling pathways have been considered for promoting osteoblastogenesis. This includes Insulin Growth Factor-1 (IGF-1), Fibroblast Growth Factor (FGF), Transforming Growth Factor-β (TGFβ), Bone Morphogenetic Proteins (BMPs), NOTCH and WNT signaling [38]. Some BMPs are known to promote the differentiation of mesenchymal osteoprogenitor cells into osteoblasts and were approved for clinical use in orthopedic surgery for non-union fractures [39]. However, only one BMP was reported to promote bone formation and bone mass in aged osteopenic ovariectomized rats [40], and the short half-life and off-effects of BMPs prevent a systemic use in osteoporotic patients. FGF, TGF-β and IGF-1 were shown to increase bone formation and bone mass in animals [41–44] but their non-specific effect is a major drawback [5]. Among the multiple pathways that control bone formation, NOTCH signaling was found to control osteoblastogenesis through increased osteoblast precursor cell proliferation [45, 46], but the activation of this pathway may contribute to the development of bone cancer [47]. Some pathways mediated by integrins, G protein-coupled receptors, receptor tyrosine kinases (RTKs), stretch-activated Ca(2+) channels, and WNT signaling are activated by mechanical stimulation, and their stimulation results in increased osteoblastic cell growth, differentiation and activity [20, 48, 49]. One of these pathways, WNT signaling, is a potential targetable pathway for bone anabolics. However, other potential therapeutic strategies are also emerging (Table 1).

Table 1.

Current therapeutic strategies and potential approaches targeting Wnt signaling to promote osteoblastogenesis and bone formation in osteopenic disorders

Therapeutic strategy	Osteoprogenitor cell number	Osteoblast differentiation and activity	Osteoblast/osteocyte survival
Current
BMP		+
iPTH	+	+	+
Potential
Anti-sclerostin Ab	+	+	+
Anti-DKK1, anti-sFRP1	+	+	+
Anti-N-cadherin	+	+	+

The reader is referred to the text for details and references

Emerging strategies to promote bone formation

Targeting WNT signaling

WNT signaling is a complex pathway that involves multiple ligands, receptors, extracellular and intracellular agonists and antagonists. The huge impact of WNT-β-catenin on bone homeostasis has been recently reviewed elsewhere [50]. Briefly, the original finding that mutations in Lrp5, a co-receptor of WNT signaling, cause abnormalities in bone mass [51] led to the discovery that WNT signaling is an important pathway controlling bone formation and bone mass. Mouse and human genetics as well as meta-analyses of genome-wide studies in humans showed that several WNT partners are involved in the control of bone formation and bone mass [7]. These findings support the concept that the WNT pathway could be targeted to modulate osteoblastogenesis in diseases where bone formation is compromised [52–54]. One current strategy is to target WNT signaling indirectly by antagonizing sclerostin, the product of the Sost gene, which is produced by mature osteocytes and antagonizes WNT binding to LRP4-6 co-receptors [55, 56] (Table 1). In line with this finding, blocking sclerostin using a monoclonal antibody was shown to increase bone formation, bone mass and strength in osteopenic animals [57]. This preclinical beneficial effect may translate into the clinics since an anti-sclerostin antibody was recently found to increase bone formation and bone mass in postmenopausal women [58]. Alternatively, other WNT signaling antagonists such as DKK-1 and sFRP1 [59] could be targeted since their ablation or inhibition results in increased bone formation and bone mass in osteopenic models [60, 61]. In this context, one potential strategy is to target miRNAs that repress sclerostin, DKK-1 or sFRP1 expression [62–64]. Another potential strategy is to target molecules that interact indirectly with WNT signaling. Notably, we showed that cadherin-2 (CDH2), a transmembrane homophilic protein that controls cell–cell adhesion, is a negative modulator of WNT signaling in osteoblasts. This effect results from both β-catenin sequestration at the cell membrane and CDH2 binding to LRP5 or LRP6, leading to reduced WNT-β-catenin signaling, decreased osteoblast differentiation and bone formation in mice [65, 66]. Consequently, blocking CDH2 and LRP5/6 interaction using a competitor peptide led to increased WNT-β catenin and enhanced periosteal bone formation in mice [67]. More importantly, we showed that peptide-mediated blockade of the CDH2-LRP5/6 interaction can increase WNT-β signaling, osteoblast differentiation, bone formation and bone mass in senescent osteopenic mice [68]. Others found that CDH2 impairs LRP6 availability to interact with the PTH receptor, and that ablation of CDH2 in osteoblasts can increase the bone anabolic response to iPTH in mice [69]. These findings support the concept that CDH2 in osteoblasts may be a potential pharmacologic target for enhancing the bone response to WNT-β signaling and iPTH in osteopenic disorders [70] (Table 1). Another example of a potential WNT signaling target is FHL2, a transcriptional cofactor which positively interacts with β-catenin. In mice, FHL2 deficiency results in osteopenia, whereas FHL2 transgenic mice exhibit increased bone mass [71, 72]. We showed that FHL2 positively controls the osteogenic differentiation of MSCs, bone formation and bone mass by promoting β-catenin nuclear translocation and modulating WNT ligand expression in mice [73, 74], suggesting another basis for therapeutic intervention. However, the ubiquitous role of WNT signaling as well as its potential implication in cancer development [75] may limit the use of direct WNT signaling activators in clinical settings.

Targeting NF-κB signaling

In addition to play an important role in osteoclastogenesis and inflammation [76], Nuclear factor kappaB (NF-κB) was recently found to control osteoblast differentiation and bone formation [77, 78]. In vitro, NF-κB activation inhibits osteoblast differentiation and activity through various mechanisms [78] including β-catenin degradation [79]. Consistently, specific pharmacological inhibition of IκB kinase (IKK)-NF-κB resulted in increased osteogenic capacity of MSCs and improved bone repair in a calvarial bone defect [79]. In mice, the inhibition of NF-κB in mature osteoblasts led to increased bone formation and bone mass with no change in bone resorption [80]. Similarly, the ablation of ReIA-mediated NF-κB signaling in mice resulted in increased osteoblast differentiation, bone formation and bone mass [81]. Collectively, these studies suggest that the inhibition of NF-κB signaling may be efficient to promote bone formation and repair in bone disorders. The recent findings that effective inhibition of NF-κB signaling in vivo can be achieved by noncanonical WNT4 [82] or by cell-mediated secretion of NF-kB inhibitors [83] may provide effective tools for targeting NF-κB signaling in bone and promoting bone formation in osteopenic disorders.

Targeting senescence mechanisms

Several senescence mechanisms were recently shown to negatively impact osteoblastogenesis and bone formation [24, 27]. These senescence mechanisms are not tissue specific and include telomere shortening, accumulation of oxidative damage, impaired DNA repair and altered epigenetic mechanisms. Interestingly, recent studies indicate that some of these mechanisms may be targeted to reduce osteoblast aging [27]. One promising target is the sirtuin deacetylase silent information regulator 1 (SIRT1), a transcription factor that counteracts cell senescence and decreases with aging [84]. SIRT1 upregulates osteoblast differentiation by downregulating the WNT antagonist gene Sost, by deacetylating β-catenin and by inhibiting NF-κB signaling [85–87]. Accordingly, Sirt1 overexpression [88] or pharmacological SIRT1 activation was found to protect against age-related bone loss [89]. Such strategy targeting a key senescence mechanism that affects both WNT and NF-κB signaling may be promising to promote bone formation and reduce bone loss associated with aging (Fig. 1a).

Targeting the ubiquitin-dependent proteolysis system

Another emerging approach to increase osteoblastogenesis is to target the ubiquitin-dependent proteolysis system (UPS), a physiological machinery that mediates the ubiquitination of proteins and their degradation by the proteasome [90]. In bone, UPS controls the degradation of key osteoblast modulators such as RTKs, BMP-2, RUNX2, β-catenin and GLI2 (a transcription factor regulating BMP-2 expression) [91]. Initial studies showed that proteasome inhibition led to increased osteoblast differentiation, trabecular bone formation and bone mass in ovariectomized mice [92–95]. This therapeutic approach is, however, limited by the non-specific effects of UPS inhibitors on overall protein degradation. A more selective approach may be to target key proteins involved in UPS-mediated protein degradation such as ubiquitin ligases that direct proteins to proteasomal degradation. Recent data indicate that some ubiquitin ligases may be targeted to promote bone formation [91]. One important ubiquitin ligase called SMAD ubiquitin regulatory factor 1 (SMURF1) controls the degradation of RUNX2 and SMAD1/5 which are critical signaling molecules involved in BMP signaling [96]. Consistently, silencing caseine-kinase 2 interacting protein-1 (Plekho1) that enhances SMURF1-mediated Smad1/5 degradation led to enhanced osteoblast differentiation, bone formation and bone mass in osteopenic rats [97]. Another E3 ubiquitin ligase, WWP1, interacts with RUNX2 and the zinc finger adapter protein SCHNURRI-3, leading to RUNX2 proteasome degradation. Schnurri-3 or Wwp1 deficiency in mice led to increased osteoblast activity and bone mass [98–100], suggesting that silencing these molecules may lead to promote bone formation. Other studies have highlighted the role of CBL ubiquitin ligases in osteoblastogenesis [91]. CBL proteins are scaffold proteins with multiple interaction domains that interact with a large number of proteins [101]. In bone-forming cells, CBL-b enhances RUNX2 stability and osteoblast-related genes [102], whereas c-CBL controls RTK ubiquitination and degradation [103, 104]. Recently, we showed that inhibiting the c-CBL-RTK interaction in human MSCs resulted in increased RTK expression and signaling causing enhanced osteoblast differentiation and survival [105]. Silencing c-CBL expression in MSCs also promoted osteoblast differentiation as a result of decreased ubiquitination of the transcription factor STAT5 that positively interacts with RUNX2 [106]. In vivo, the inhibition of c-CBL interaction with phosphatidylinositol-3′ kinase (PI3 K) led to increased PI3 K levels, bone formation and bone volume in mice [107]. Overall, these findings suggest that targeting selected ubiquitin ligases may lead to increased osteoblastogenesis, bone formation and bone mass in osteopenic disorders.

Targeting osteoblast progenitor cells

Mesenchymal cells present in the bone marrow stroma are able to differentiate into osteoblasts, adipocytes or chondrocytes under appropriate stimulation [108, 109]. Although MSCs have long been used as a source for bone regeneration in bone defects, the capacity of MSCs to differentiate and to generate new bone in vivo is limited [110, 111]. Several growth and differentiation factors were found to expand MSC proliferation or osteogenic differentiation in vitro [112] with limited impact on bone formation in vivo [113]. Novel targeted strategies have been recently developed to promote MSC osteogenic differentiation. Notably, targeting integrins, which are essential cell surface adhesion transmembrane molecules, proved to be effective to promote MSC osteogenic differentiation and bone formation [114]. We showed that lentiviral-mediated expression of α5β1 integrin, a cell surface receptor for fibronectin that is essential for osteoblast differentiation and survival [115, 116], promoted bone repair in two relevant critical-size bone defects in the mouse [117]. More importantly, an agonist of the α5 integrin subunit promoted osteoblast differentiation and survival, resulting in increased bone formation in mice [118]. These findings indicate that the α5 integrin subunit can be targeted to promote MSC osteogenic differentiation, bone formation and repair. Recent data indicate that the α4β1 integrin is another potential target for promoting MSC osteogenic differentiation. The transient expression of the α4 integrin subunit in MSCs was found to increase the homing of injected MSCs in the bone marrow, an effect that resulted in increased osteoblast formation [119]. Moreover, the injection of a chimeric molecule composed of a peptidomimetic which binds the α4β1 integrin coupled to the bone-seeking agent alendronate increased bone formation and bone mass in mice [120, 121]. Thus, direct targeting of α4β1 or α5β1 integrins appear to be promising therapies for promoting MSC osteogenic differentiation, bone formation and regeneration in age-related bone loss and repair [114].

One effective approach to promote MSC differentiation specifically at the bone surface is to silence genes that inhibit bone formation using small interference RNAs (siRNAs). Specifically, the systemic delivery of siRNA for Plekho1, a negative regulator of osteogenesis, using cationic liposomes attached to (AspSerSer)₆ that preferentially binds to bone-forming surfaces, resulted in increased bone formation and bone mass in osteopenic rats [97]. This may provide a cell-selective delivery system for gene knockdown and bone anabolic therapy. An alternative indirect method to target MSC osteogenic differentiation is the use of polymeric nanoparticles delivering siRNAs. This delivery system was recently used to target semaphorin 4D expression in osteoclasts, resulting in increased number of osteoblasts and bone volume in ovariectomized mice [122]. Thus, the RNA interference-based approach appears to be another promising strategy for targeting cells at the bone surface and promoting bone formation (Fig. 1b).

Fig. 2.

Osteoblast dysfunctions and potential targeted therapies in primary bone tumors. In primary bone tumors, genomic instability, aberrations in tumor suppressor genes, oncogenes and epigenetic deregulation of key transcription factors result in the transformation of osteoblast precursor cells into tumor cells which exhibit deregulated cell proliferation, defective terminal osteoblast differentiation and reduced cell apoptosis. These abnormalities can be specifically targeted to restore normal bone formation, as indicated

Osteoblast dysfunctions induced by genetic mutations in skeletal dysplasias

During the past years, the analysis of the genotype–phenotype relationship in skeletal dysplasias provided important information on the genetic control of bone formation [6, 123, 124]. This is best illustrated by the analysis of the skeletal phenotype in mice and humans induced by loss-of-function and gain-of-function mutations in the WNT-β-catenin signaling pathway [7, 125]. Studies of the osteoblast phenotype in genetic skeletal dysplasias unraveled some key pathophysiological mechanisms that may be targeted to correct the abnormalities in these bone diseases (Table 2), as detailed below.

Table 2.

Osteoblast dysfunctions and potential therapeutic strategies in skeletal dysplasias induced by genetic mutations

	Osteoprogenitor cell replication	Osteoblast differentiation and function	Osteoblast/osteocyte death	Skeletal phenotype	Potential therapeutic strategy
Mutation
Gsα (+)	↑	↓	↑	Accelerated (local) bone formation	Wnt signaling inhibition
FGFR2 (+)	↔	↑	↑	Increased cranial ossification	FGFR2-PDGERα signaling inhibition
Twist (−)	↑	↑	↑	Increased cranial ossification	FGFR2 signaling inhibition
CFTR (−)	↔	↓	↔	Decreased bone formation	CFTR correction

(+) and (−) identify a gain- and loss-of-function, respectively, induced by the indicated genetic mutation in skeletal dysplasias

Gain-of-function mutations impacting bone formation

A typical example of gain-of-function mutation that affects osteoblastogenesis is provided by mutations in Gsα (R201C, R201H) causing fibrous dysplasia (FD), a rare disease characterized by immature bone deposition in dysplastic lesions [126]. The analysis of the osteoblast phenotype induced by Gsα mutations in FD provided insights into the role of Gsα in osteogenesis [127, 128]. The stimulatory α subunit of G protein (Gsα) binds and hydrolyses GTP, resulting in adenylate cyclase activation, cAMP production and protein kinase A (PKA) activation. In FD, Gsα mutations are expressed in osteoprogenitor cells in the bone marrow stroma in dysplastic lesions, causing excessive cAMP levels and alterations in the expression of c-Fos and c-Jun, resulting in increased proliferation of pre-osteoblastic cells, abnormal osteoblast maturation and deposition of an immature bone [126, 129, 130]. This phenotype can be reproduced in adult mice bearing the R201C mutation [131]. Conversely, selective silencing of the mutated Gsα allele in skeletal progenitor cells can revert the excessive cAMP production and the abnormal osteoblast differentiation in human FD cells [132]. A role for Gsα in bone formation is supported by the finding that activation of the Gsα-PKA signaling cascade mediates in part the anabolic action of iPTH in murine bone [133]. In contrast, the loss of Gsα led to the preferential commitment of bone marrow stromal cells to adipocytes [134]. Consistently, conditional deletion of Gsα in skeletal progenitors led to decreased osteoblast progenitor number, bone formation and bone mass in mice. This phenotype was associated with increased expression of the WNT antagonists, sclerostin and DKK-1, suggesting that Gsα interacts with WNT signaling to control cells of the osteoblast lineage [135]. In line with this finding, activating Gαs mutations in FD patients was found to potentiate WNT-β-catenin signaling and consistently, osteoblast differentiation abnormalities in FD can be attenuated by reducing β-catenin levels [136]. These findings suggest a potential therapeutic approach targeting WNT signaling to correct the osteoblast dysfunctions in FD patients with aberrant Gsα signaling (Table 2).

Another typical example of gain-of-function mutations affecting bone formation includes mutations in FGFR2 causing craniosynostosis (or premature cranial suture fusion) [137]. We initially showed that activating FGFR2 mutations in human osteoblasts promote osteoblast differentiation and osteogenesis [138, 139]. In mouse models, the phenotype induced by FGFR2 mutations is variable between the models [140–143] which reflects the complexity of FGF signaling in cells of the osteoblast lineage [144]. Mechanistically, the analysis of signaling pathways in human osteoblasts [43, 139] and mouse models [145, 146] revealed that MEK, PKC, SRC and WNT signaling pathways are implicated in the aberrant osteoblast phenotype induced by activating FGFR2 mutations. These findings led to propose selective therapeutic strategies in craniosynostosis [147, 148]. Specifically, the use of tyrosine kinase inhibitors was shown to attenuate FGFR signaling and to prevent the abnormal skeletal phenotype in mouse models of craniosynostosis [149, 150]. Recently, we and others showed that crosstalks between FGFR2 and platelet-derived growth factor receptor α (PDGFRα) signaling amplify the osteoblast abnormalities induced by FGFR2 activating mutations in vivo [151, 152]. Consistently, pharmacological inhibition of PDGFRα attenuated the abnormal bone formation induced by aberrant FGFR2 signaling in osteoblasts [153]. These findings suggest novel therapeutic strategies targeted to RTK signaling to attenuate the abnormal bone formation in this skeletal disorder (Table 2).

Loss-of-function mutations affecting bone formation

Some typical loss-of-function genetic mutations also cause abnormalities in bone formation. Notably, mutations in Twist1, a basic helix–loop–helix (bHLH) factor involved in mesodermal differentiation, induce the Saethre–Chotzen syndrome (SCS) which is characterized by premature fusion of coronal sutures [154]. TWIST1 heterodimerizes with bHLH E proteins that bind E-boxes present in the promoter of target genes. In SCS, Twist haploinsufficiency causes TWIST1 protein degradation which abolishes TWIST1 binding activity [155]. However, in the developing mouse, TWIST1 can inhibit the functional activity of RUNX2 independently of the bHLH domain [156]. In human SCS osteoblasts, Twist haploinsufficiency increased collagen expression and osteogenic capacity independently of Runx2 expression [157, 158]. This phenotype resulted from activated PI3 K/AKT-dependent osteoprogenitor cell growth and increased number of collagen-forming osteoblasts [159]. Another potential pathophysiological mechanism involved in SCS is the interaction between Twist1 and FGFR2 signaling [142, 160–162]. We showed that molecular or pharmacological inhibition of FGFR2 signaling can correct the abnormal osteoblast phenotype induced by Twist1 silencing in human MSCs [162]. Consistently, inhibition of FGF signaling was shown to prevent the premature cranial suture fusion in Twist1 heterozygotic mice [161]. Collectively, these studies indicate that targeting FGF/FGFR signaling may be a valuable option to prevent the abnormal osteoblastic cell replication and function induced by Twist1 genetic mutations [139] (Table 2).

Another example of a loss-of-function genetic mutation causing abnormalities in osteoblast function is the F508del mutation in the cystic fibrosis transmembrane conductance regulator (CFTR). Patients with cystic fibrosis (CF) who exhibit this prominent CFTR mutation display low bone density and decreased bone formation [163, 164] but the underlying mechanism was unknown. We recently showed that the F508del CFTR mutation impacts osteoblast activity, bone formation and bone mass in a mouse model, independently of other etiological factors, suggesting a link between this CFTR mutation and bone formation in this disease [165]. Pharmacological correction of the mutant CFTR-dependent channel led to improve bone formation, bone mass and microarchitecture in F508del mice [166], suggesting a potential therapeutic strategy to reduce bone loss in this disorder (Table 2). Further investigation of the molecular mechanisms involved in the defective osteoblast function in CF is needed for developing a specific therapeutic strategy and improving bone formation in this genetic disorder affecting the skeleton.

Osteoblast dysfunctions in primary bone tumors

During the recent years, the analyses of the cellular and molecular dysfunctions in primary bone tumors led to propose novel targeted therapies capable of counteracting bone tumor development and metastasis. Osteosarcoma, the main primary bone tumor, results from genetic or epigenetic changes that cause uncontrolled osteoprogenitor cell proliferation, halted terminal differentiation and reduced apoptosis (Fig. 2). One current concept underlying tumor development is that osteosarcoma cells retain a high proliferative capacity and lose their potential to differentiate as the consequence of deregulations of various processes [167]. These dysfunctions may result from aberrations in tumor suppressor genes, such as p53 that is implicated in cell survival [168], or Rb that controls terminal osteoblast differentiation [169]. Overexpression of oncogenes such as Fos, Jun [170] or Met [171] also causes deregulated osteosarcoma cell proliferation. In addition, osteosarcoma development is often associated with RUNX2 deregulation [172], Osterix downregulation [173] and Sox2 overexpression [174], resulting in blockade of terminal osteoblast differentiation. Deregulation of growth factors and receptor RTKs may also contribute to the aberrant cell growth and differentiation in primary bone tumors [175]. Notably, primary bone tumorigenesis is often associated with increased expression of TGF-β1 [176], IGF-1 [177], EGFR [178, 179] and PDGFR [179, 180]. In line with these findings, we recently reported that ErbB3, a member of the EGF family which heterodimerises with ErbB2, is overexpressed in human osteosarcoma cell lines and tumors. Consistent with this finding, ErbB3 silencing reduced osteosarcoma cell proliferation and decreased tumor growth in mice, suggesting that ErbB3 can be targeted to reduce bone tumorigenesis [181]. One potential mechanism for the increased RTK signaling in bone tumors may be the reduction of RTK degradation by the ubiquitin–proteasome system [91]. In support of this concept, proteasome inhibition was found to suppress cell growth and to induce apoptosis in osteosarcoma [182]. Recently, we showed that the expression of the ubiquitin ligase c-CBL, which negatively controls RTK levels and signaling, is downregulated in osteosarcoma cells. Consistently, increasing c-Cbl expression reduced EGFR and PDGFRα levels and inhibited osteosarcoma cell growth and tumorigenesis in mice [179]. Taken together, these studies suggest an effective therapeutic approach targeting the proteasome to inhibit primary bone tumor development [91] (Fig. 2).

WNT signaling may also play a role in osteosarcoma tumorigenesis. Notably, osteosarcoma cells often exhibit overexpression of WNT ligands and receptors, downregulation of WNT inhibitors and increased β-catenin-mediated activity [75, 183–185]. This suggests that inhibiting WNT signaling may reduce osteosarcoma tumorigenesis [186]. Indeed, inhibition of WNT signaling decreased bone tumorigenesis in animal models [184, 187] and increased osteosarcoma cell sensitivity to chemotoxic drugs [188]. The inhibition of WNT signaling by silencing FHL2, a Wnt-β-catenin activator, also resulted in decreased cell growth and tumorigenesis in murine osteosarcoma [73]. However, downregulation of WNT signaling in osteosarcoma cells may block WNT-mediated osteoblast differentiation and thereby promote cell self-renewal [174], suggesting that WNT signaling in osteosarcoma may have pro-tumoral or anti-tumoral effects depending on the state of cell progression.

An alternative strategy to inhibit bone tumorigenesis may be to target the epigenetic mechanisms that are involved in the deregulation of tumor cell proliferation and survival in osteosarcoma [189–191]. In this context, the inhibition of histone deacetylase (HDAC) was shown to inhibit osteosarcoma cell growth and to induce terminal osteoblast differentiation [192]. Current evidence indicate that miRNAs play a role in the pathogenesis of bone tumors, suggesting that miRNAs could be targeted to reduce osteosarcoma tumorigenesis [191, 193, 194]. One promising method to deliver miRNAs or HDAC inhibitors locally in bone tumors may be the use of polymeric nanoparticles associated with bone-binding molecule (bisphosphonate), thus allowing bone homing and delivery of anticancer drugs [195]. A better understanding of the role of epigenetic mechanisms in gene deregulation in primary bone tumors is likely to result in selective therapeutic strategies to efficiently reduce primary bone tumor growth and to sensitize tumor cells to chemotherapy (Fig. 2).

Conclusion and perspectives

During the recent years, the phenotypic and functional analyses of human bone pathologies and mouse models allowed identifying some important mechanisms that are implicated in metabolic, genetic and oncogenic skeletal disorders. This led to propose novel therapeutic approaches such as the modulation of WNT signaling, the pharmacological modulation of proteasome-mediated protein degradation, the inhibition of senescence mechanisms, the induction of osteoblastic differentiation in osteoprogenitor cells and the inhibition of cell proliferation in osteosarcoma cells (Figs. 1, 2). The next step will be to improve the specificity and efficiency of these therapeutic approaches in preclinical studies, and to translate these strategies into effective therapies for correcting the abnormal osteoblastogenesis in human skeletal disorders.

In the future, a number of important issues need to be addressed in osteoblast biology and pathology. First, we need to identify specific markers of cells of the osteoblast lineage at different stages of differentiation to determine which cell(s) could be best targeted to correct the defective osteoblastogenesis in bone disorders. In this context, the lineage tracing methods that are currently available [196, 197] may help identifying the nature of cells in the bone marrow or periosteum to be targeted for selective therapeutic effects. Second, there is mounting evidence that dysfunctions in epigenetic mechanisms contribute to both age-related bone loss [198, 199] and primary bone tumors [191, 193, 194]. It will be of major interest to determine if the specific modulation of epigenetic mechanisms can attenuate the abnormal osteoblastic cell fate and function in disorders of bone formation. Finally, recent evidence indicate that cells of the osteoblast lineage control both hematopoietic stem cell differentiation [200, 201] and oncogenesis [202–205]. It will be thus essential to determine whether and how cells of the osteoblast lineage may be selectively targeted to improve or to correct bone marrow cell fate. It is hoped that such research in osteoblast biology and pathology will translate into efficient and safe therapeutics not only for correcting the abnormal osteoblastogenesis in human metabolic and genetic skeletal disorders, but also for preventing the development of oncogenic bone disorders.