The osteocyte as a signaling cell

Abstract

Osteocytes, former osteoblasts encapsulated by mineralized bone matrix, are far from being passive and metabolically inactive bone cells. Instead, osteocytes are multifunctional and dynamic cells capable of integrating hormonal and mechanical signals and transmitting them to effector cells in bone and in distant tissues. Osteocytes are a major source of molecules that regulate bone homeostasis by integrating both mechanical cues and hormonal signals that coordinate the differentiation and function of osteoclasts and osteoblasts. Osteocyte function is altered in both rare and common bone diseases, suggesting that osteocyte dysfunction is directly involved in the pathophysiology of several disorders affecting the skeleton. Advances in osteocyte biology initiated the development of novel therapeutics interfering with osteocyte-secreted molecules. Moreover, osteocytes are targets and key distributors of biological signals mediating the beneficial effects of several bone therapeutics used in the clinic. Here we review the most recent discoveries in osteocyte biology demonstrating that osteocytes regulate bone homeostasis and bone marrow fat via paracrine signaling, influence body composition and energy metabolism via endocrine signaling, and contribute to the damaging effects of diabetes mellitus and hematologic and metastatic cancers in the skeleton.

CLINICAL HIGHLIGHTS

1. Osteocytes are the most abundant cells in bone and live deep in the mineralized bone matrix. Osteocytes have extensive long cytoplasmic extensions that form a complex network that enables direct communication among osteocytes and with other effector cells in the bone/bone marrow.

2. Osteocytes respond to mechanical and hormonal stimuli and, in turn, secrete molecules that affect cells of the bone/bone marrow niche by paracrine mechanisms and cells of distant organs by endocrine mechanisms.

3. Clinical evidence indicates that osteocyte viability and function are altered in common and rare bone diseases, supporting a pathophysiological role of osteocyte-derived molecules in skeletal diseases.

4. Cancer cells communicate with osteocytes to transform the bone marrow niche into an environment supportive of tumor cell proliferation and bone destruction. Cancer cells have proapoptotic effects on osteocytes and stimulate the production of osteocyte-derived osteoclastogenic cytokines and inhibitors of bone formation.

5. Therapeutic targeting of osteocytes and their derived factors (Sclerostin, RANKL, and FGF23) has proven to be an effective approach to treat common and rare bone diseases, as well as the deleterious effects of cancer in bone.

1. INTRODUCTION

Scientific advances in the last 25 years demonstrate the profound impact that osteocytes have on bone biology and pathophysiology. The vast advancement in the knowledge about osteocytes and their role in bone homeostasis started in the late 1990s, which was prompted by the discovery of rare skeletal diseases in humans linked to osteocyte dysfunction, the generation of osteocyte-like cell lines, and the development of genetic tools to manipulate in vivo gene expression in osteocytes. Today, it is recognized that osteocytes orchestrate osteoblast and osteoclast function, integrate and transmit hormonal and mechanical cues, regulate the differentiation and activity of bone/bone marrow cells via paracrine signaling, and influence the function of cells in distant tissues by secreting factors that act as endocrine signals. This boom in scientific research prompted the development and commercialization of therapies targeting osteocytes for the treatment of bone diseases, which prevent bone loss and/or increase bone mass and strength.

2. OSTEOCYTOGENESIS AND DEVELOPMENT OF THE OSTEOCYTIC GENE SIGNATURE

2.1. Osteocyte Differentiation from Osteoblasts

Osteocytes are the most fully differentiated cell type of the osteoblastic lineage. Driven by mechanisms still poorly understood, some osteoblasts on the bone surface become surrounded by the matrix proteins they produce and differentiate into osteocytes. Osteoblasts progressively decrease bone matrix production, undergo dramatic changes in morphology, and start expressing particular genes that constitute the osteocyte signature (1–3). Approximately 5–20% of osteoblasts undergo terminal differentiation and become osteocytes. The remainder either die by apoptosis or flatten and turn into cells, known as bone lining cells, that cover the quiescent surfaces of bone. During the process of osteocytogenesis, osteoblast morphology changes to the stellate morphology characteristic of osteocytes. This process is also characterized by an autophagy-mediated (4) reduction in endoplasmic reticulum area and number of mitochondria (5) and the development of cytoplasmic projections. These long cytoplasmic processes run through tunnels carved in the bone mineral known as canaliculi, allowing the osteocytes to physically interact with each other as well as to distribute molecules to adjacent osteocytes and other cells on the bone surface or in the bone marrow compartment (FIGURE 1). During osteoblast-to-osteocyte differentiation, major changes in transcription and translation lead to the expression in osteocytes of molecules that regulate bone homeostasis and phosphate metabolism (2). Different genes have been identified as responsible for the changes in cell morphology and the formation of the complex lacuno-canalicular network. Expression of Podoplanin/E11 is required by osteoid osteocytes (6) to initiate proper dendrite formation (7–9). The osteoblast-to-osteocyte transition is accompanied by increased expression of Dentin matrix protein 1 (DMP1) (10), which is essential for osteocyte maturation (11). The expression of matrix metalloproteinases (MMPs) also increases during osteoblast-osteocyte differentiation, enabling the formation and elongation of dendritic projections (12, 13). Connexin 43 (Cx43) allows communication between adjacent osteocytes and between osteocytes and the bone/bone marrow compartment, preserves osteocyte viability, and mediates osteocyte transduction of mechanical signals (14–16).

FIGURE 1.

Osteocyte morphology and lacuno-canalicular network. A: osteocytes are highly interconnected to each other and to cells on the bone surface. Canaliculi connect osteocytes to the endocortical bone surface and the bone marrow compartment. Electron microscopy imaging of an acid etched bone. B: osteocyte-derived factors circulate through the canalicular system to reach other osteocytes and bone cells on the bone surface or in bone marrow. Immunostaining of Sclerostin in paraffin-embedded bone tissue. C: MLO-Y4 osteocyte-like cells display a stellate morphology with dendritic processes resembling those found in primary osteocytes. White arrows point to individual osteocytes; yellow arrows point to canaliculi.

Osteocytogenesis is also characterized by increased expression of proteins that regulate phosphate metabolism and matrix mineralization (10). One of these regulators is fibroblast growth factor (FGF)23, which binds to FGF receptors and the coreceptor Klotho expressed in tubular cells in the kidney and regulates phosphate metabolism (17). Interestingly, FGF23 expression is in turn regulated by other factors produced by osteocytes, including Phosphate-regulating neutral endopeptidase X-linked (PHEX), DMP1, and Matrix extracellular phosphoglycoprotein (MEPE). In humans, inactivating mutations in PHEX lead to accumulation of FGF23 in the circulation and hypophosphatemia (18). Mice with deletion of PHEX exhibit osteomalacia and an irregular lacuno-canalicular system (19). Similarly, mutations and deletions in DMP1 cause autosomal recessive hypophosphatemic rickets (ARHR), which is associated with high FGF23 production (11, 20). MEPE is another product of osteocytes, and its genetic deletion in mice increases bone mineral density (21). Overall, the dysregulation in osteocytes of any of the genes listed above causes alterations in phosphate metabolism (17, 22–25).

Osteocytes are also a major source of receptor activator of nuclear factor-κΒ ligand (RANKL), the key regulator of osteoclastogenesis, and SOST/Sclerostin, a soluble Wnt signaling antagonist and potent suppressor of bone formation (26–30). The discovery of these molecules identified the long-sought mediators of osteocyte communication with the cells that execute the bone remodeling process. Several cells in the bone/bone marrow niche, such as cells of the osteoblastic lineage, adipocytes, B cells, and T cells, produce RANKL (31–33), yet genetic deletion of RANKL from osteocytes strikingly decreased the number of osteoclasts in bone and increased bone mass. These results contrast with the minimal effects of RANKL deletion in other bone cells, in particular osteoblasts (28–30), and identify osteocytes as a major source of RANKL in the adult skeleton. Regarding SOST/Sclerostin, Sclerostin is not expressed in the early stages of differentiation in the osteoblastic lineage. However, Sclerostin expression increases as osteocytes mature, with maximal levels being reached when the osteocytes become fully surrounded by mineralized bone (34–37). Novel evidence demonstrates that Sclerostin contributes to the regulation of osteocytic RANKL (38, 39), yet the mechanisms regulating osteocytic RANKL and SOST/Sclerostin gene expression in osteocytes are not completely known and likely involve a combination of several transcription factors and epigenetic modifications. Nevertheless, this evidence demonstrates that osteocyte-derived RANKL and SOST/Sclerostin contribute to bone homeostasis by regulating both osteoclastogenesis and osteoblast differentiation and activity (40, 41).

Recent mapping of the transcriptome of murine osteocytes revealed genes and molecular programs that control the intricate osteocytic cell network. The osteocyte signature comprises >1,000 genes differentially expressed in osteocytes compared to other cell types. The skeletal role of ∼80% of these genes is currently unknown. However, many of these genes have been identified as contributors to the formation of neuronal networks, which suggests that these genes may be important for the generation and maintenance of the osteocyte network. By evaluating numerous skeletal indexes in >700 mouse lines with genetic deletions, >26 genes within this osteocyte transcriptome signature were identified as critical regulators of bone structure and function. Furthermore, human homolog genes for some of these 26 genes cause monogenic skeletal dysplasias and are associated with polygenic diseases such as osteoporosis and osteoarthritis (42). These novel findings unravel the genetic identity of the osteocyte network and highlight the critical role of osteocytes in human bone diseases.

2.2. Modulation of Gene Expression Profiles in Osteocytes by Epigenetic Marks

The differentiation of mesenchymal stem cells (MSCs) toward osteoblasts, and ultimately osteocytes, is a process tightly controlled by epigenetic mechanisms (43, 44). Cytosine DNA methylation is a repressive epigenetic mark that negatively regulates gene expression (45). In addition, there are activating (mainly acetylation and phosphorylation) and repressive (methylation, ubiquitination, and sumoylation) histone epigenetic marks that contribute to the regulation of gene expression (46). Furthermore, production of microRNAs (miRNAs) and other small RNAs results in mRNA cleavage or translational repression, depending on the degree of complementarity (47). There are several examples of epigenetic regulation of gene programs leading to the mature phenotype of osteocytes. For instance, coordinated changes in DNA methylation and chromatin modifications determine the expression of several genes, including podoplanin/E11, Alkaline phosphatase (ALP), aromatase, Runt-related transcription factor 2 (RUNX2), RANKL, the osteogenic protein distal-less homeobox 5 (DLX-5), Activating transcription factor 4 (ATF4), Activator protein 1 (AP-1), Osterix, SMADs, SOST, and the estrogen receptor (ER) (43, 44, 48–52). In particular, epigenetic regulation of the SOST gene, whose expression in bone is restricted to osteocytes, is well studied (53). Binding of myocyte enhancer factor-2 (MEF2C) to evolutionarily conserved region 5 (ECR5) plays a critical role in the regulation of SOST expression (54–56). In addition, the loss of epigenetic DNA methyl marks at the SOST proximal promoter enables osteocytes to express SOST (57). Chromatin marks regulated by Sirtuin 1, a histone deacetylase, and HDAC5, a class IIa histone deacetylase, also directly regulate SOST expression in osteocytes (58–62).

In summary, osteoblasts undergo dramatic morphological changes and deep genetic and epigenetic reprogramming to differentiate into mature, matrix-embedded osteocytes. These changes enable osteocytes to become multifunctional cells capable of sensing mechanical loads, regulating bone remodeling, and producing endocrine regulators, yet the signals determining the final fate of osteoblasts (i.e., osteocytes, apoptosis, lining cells) in bone remain unclear. It is anticipated that cell lineage-tracing experiments will provide some light on the fate of osteoblasts. Similarly, the mechanisms regulating the changes in morphology and gene expression during osteoblast to osteocyte transition are yet to be determined. It is likely that epigenetic mechanisms play an important part in these processes, and this area of investigation demands further attention. Finally, it is expected that with the incorporation of single-cell sequencing approaches in the set of tools available to bone research, specific osteocyte gene signatures will be described for both physiological and disease conditions of the skeleton.

3. CHALLENGES OF WORKING WITH OSTEOCYTES IN VITRO AND IN VIVO

Isolating and culturing primary osteocytes, as well as reproducing the complexity of the osteocytic canalicular network in vitro, remain a major challenge for bone biologists. Several protocols to isolate primary osteocytes from bone have been developed, mainly based on controlled digestions of extracellular matrix proteins, followed by the separation of osteocytes from other bone and endothelial cells by virtue of genetic markers expressed by osteocytes (63–72). These approaches result in osteocytes deprived of their native environment, thus markedly changing their molecular signature and cellular phenotype. In addition, once placed in culture, authentic osteocytes frequently lose expression of osteocytic genes and dedifferentiate back to an osteoblastic phenotype (73). An alternative to primary cells is the use of osteocyte-like cell lines of rodent or human origin (60, 74–78) (TABLE 1). However, most of these cell lines do not exhibit all the features of authentic osteocytes, as reviewed recently (79). In this regard, a new osteocyte-like cell line named OmGFP66 has been generated recently. This cell line is capable of differentiating into highly dendritic osteocytes embedded in three-dimensional (3-D) lacunae (78). Yet, like other in vitro two-dimensional (2-D) and 3-D differentiation systems from osteoblast precursor cells, these cultures result in mixed populations of osteocytes, osteoblasts, and less differentiated cells, limiting their use for most purposes.

Table 1.

Osteocyte-like cell lines and their characteristics

Name	Origin	Characteristics	References
MLO-Y4	Mouse	Osteocyte	Kato et al. (74)
MLO-A5	Mouse	Late osteoblast-early osteocyte	Kato et al. (75)
IDG-SW3	Mouse	Preosteocyte	Woo et al. (76)
Ocy454	Mouse	Preosteocyte	Spatz et al. (60)
OMGFP66	Mouse	Preosteocyte	Wang et al. (78)
HOB-01-C1	Human	Preosteocyte	Bodine et al. (77)

The use of ex vivo bone organ cultures has become a convenient alternative to circumvent the limitations of in vitro cultures of isolated osteocytes (80). Osteocytes maintain their natural location, intact lacuna-canalicular network, and 3-D distribution in ex vivo bone organ cultures, making this approach better suited for the investigation of the mechanisms underlying bone growth and bone and cartilage matrix turnover, as well as the effects of cancer in bone (14–20). Ex vivo bone organ cultures also represent low-cost screening models with few ethical considerations and allow the study of both genetic and pharmacological manipulations of osteocytes and other bone cells in their natural environment. Detailed methodological procedures on how to establish ex vivo cultures for the study of osteocytes have been published recently (80).

Several transgenic Cre strains to conditionally manipulate gene expression in osteocytes in vivo are available (81). The design of these transgenic strains has been based on the regulatory regions of two genes primarily expressed by osteocytes. These are the DMP1 promoter, with 10-kb and 8-kb versions (14, 82), and the SOST promoter (28). DMP1 promoters are the most widely used Cre transgenic lines for conditional gene deletion in osteocytes in vivo. However, besides targeting osteocytes, studies with genetic reporters have shown that DMP1-related Cre strains have off-target recombination in mature osteoblasts potentially destined to be osteocytes, muscle, and other soft tissues (81). Similarly, the SOST-Cre strain targets osteocytes, not osteoblasts, but exhibits off-target recombination in hematopoietic stem cells (28). Furthermore, whether all osteocytes are targeted by the SOST-Cre driver has not been confirmed. It is likely that this driver targets a subpopulation of osteocytes more deeply embedded in the mineralized matrix, which, as shown before, express Sclerostin endogenously (36, 83, 84). Tamoxifen-inducible strains for DMP1-10Kb-Cre and SOST-Cre have also been generated recently (85, 86). Limitations with these inducible strains include some degree of “leaky” expression without tamoxifen, the need to carefully optimize tamoxifen doses, and off-target cell recombination (81). Thus many of these osteocyte-targeted Cre approaches are not as specific as intended, yet, despite these limitations, their careful use, inclusion of appropriate controls, performance of comparative studies with Cre lines targeting different cell populations, and confirmation that the deletion of the gene of interest is achieved in osteocytes and no other cells has markedly increased our knowledge about osteocyte functions and will continue to advance our understanding of the role of osteocytes in health and disease. Future research toward the generation of osteocyte-specific promoters without off-target recombination in other bone cells or tissues is warranted.

Because osteocytes reside deep in mineralized bone, imaging these cells in their environment is another challenging aspect of osteocyte research (87). Until recently, the majority of the imaging work on osteocytes was performed ex vivo, in both undecalcified and decalcified specimens along with conventional sectioning and imaging techniques. The development of confocal microscopy provided the opportunity for osteocyte biologists to image osteocytes in 3-D in their bone microenvironment (88). More recently, the use of micro-computed tomography (μCT) systems, scanning electron microscopy, and transmission microscopy has allowed researchers to image osteocyte lacunae noninvasively and in a 3-D fashion (87). In vivo live imaging is now possible thanks to the generation of transgenic mice with membrane-targeted fluorescent proteins in osteocytes (35, 89, 90). Moving forward, the development of multiphoton microscopy (91), together with tissue clearing techniques (92), is expected to revolutionize the way we image osteocytes and other bone cells.

In conclusion, the study of osteocyte biology is challenging because of the limited accessibility to mineral-embedded osteocytes and the lack of fully representative in vitro models. Researchers are overcoming these limitations with the use of ex vivo models and novel imaging techniques that allow the study of osteocytes in their native 3-D microenvironment. Although the use of the current Cre drivers to study osteocytes in vivo has tremendously advanced our understanding of osteocyte biology, future research is likely to focus on the generation of novel maneuvers to genetically modify osteocytes in a more specific manner.

4. OSTEOCYTE PROGRAMMED CELL DEATH: APOPTOSIS

Osteocyte density in bone is influenced by both the rate of osteoblast differentiation and the rate of apoptosis, as osteocytes are nondividing cells in vivo. Earlier studies of this area demonstrated that osteocyte viability is controlled by a combination of physical and chemical stimuli (93–95). Osteocytes can live for decades; however, these cells undergo programmed cell death by apoptosis like other cells do in bone. The first observations on osteocyte apoptosis were made in the context of estrogen withdrawal, revealing an association between apoptosis and estrogen signaling in osteocytes (96). This work stimulated a number of studies demonstrating that estrogen as well as selective estrogen receptor modulators (SERMs) preserve osteocyte viability and revealing multiple signaling pathways involved (97–106). The bone fragility syndromes ensuing with estrogen or androgen deficiency are characterized by increased osteocyte apoptosis, and similar observations have been made in animal models and human conditions of glucocorticoid (GC) excess, aging, and mechanical unloading (107–109). In contrast, physiological levels of mechanical stimulation (108, 110), estrogen/androgen replacement (100, 101), and treatment with bisphosphonates (111) have been shown to preserve osteocyte viability.

4.1. The Role of Mechanical Signaling in the Regulation of Osteocyte Apoptosis

In vitro studies with osteocytic cells or authentic osteocytes showed that mechanical stimulation prevents apoptosis induced by proapoptotic stimuli, such as GC, tumor necrosis factor-alpha (TNF-α), or reactive oxygen species (ROS) (112, 113). Mechanistic studies revealed that mechanical forces sensed at the osteocyte membrane are transduced into antiapoptotic signals mediated by integrins, the Src and focal adhesion kinases (FAKs), as well as several cytoskeletal proteins. These membrane-associated events lead to the activation of intracellular survival kinases, including extracellular signal-regulated kinase (ERK) (112). Mechanical signals also activate the Wnt/β-catenin signaling pathway (106, 114–117), and there is bidirectional cross talk between survival kinase signaling and the Wnt pathway. For instance, preservation of osteocyte viability by ERK nuclear translocation and antiapoptotic signals induced by mechanical stretching or fluid flow are abrogated by Wnt antagonists, such as the stimulator of β-catenin degradation axis inhibition protein (Axin)2 and Dickkopf Wnt Signaling Pathway Inhibitor 1 (DKK1) (118). Conversely, the phosphorylation and accumulation of β-catenin and glycogen synthase kinase 3β (GSK3β) induced by mechanical cues are prevented by genetic inhibition of ERK or caveolin-1 signaling (118).

Osteocyte life span is also regulated by mechanical signals in vivo. Bones exposed to low or excessive mechanical strains exhibit an increased number of apoptotic osteocytes (93, 108, 119, 120). Under both circumstances, osteoclastic resorption is preceded by osteocyte apoptosis, with bone areas enriched in apoptotic osteocytes being subsequently resorbed by osteoclasts (108). This earlier evidence prompted us to propose that dying osteocytes are initiators of local bone resorption (121) (FIGURE 2). This hypothesis is further supported by the demonstration that genetic induction of osteocyte death stimulates the recruitment of osteoclasts and subsequent bone resorption and induces decreases in bone mass (122). Moreover, the skeleton of mice depleted of osteocytes does not exhibit the expected osteoclastogenic response to unloading (122), demonstrating that osteocytes initiate the cascade of events that lead to bone loss. One of the mechanisms by which dead osteocytes initiate local bone resorption involves increased expression of RANKL in osteocytes surrounding dead osteocytes (123–125). Importantly, an apoptosis inhibitor prevented the increased RANKL and the proresorptive action of either low or excessive mechanical strains (123, 124). Furthermore, mice lacking RANKL in osteocytes do not exhibit the typical increase in osteoclastogenesis and the loss of bone mass induced by unloading (29), demonstrating the requirement of osteocytic RANKL in the bone loss induced by mechanical disuse. Furthermore, treatment with a bisphosphonate that inhibits osteocyte apoptosis without directly affecting osteoclasts prevented the elevation in RANKL expression in osteocytes (and osteoblasts); however, it did not prevent unloading-induced bone loss, suggesting that, under these conditions, other cellular sources of RANKL are responsible for the resorption leading to the loss of bone (125). This suggests that there is a strong cause-effect association between apoptotic osteocytes and the production of RANKL by adjacent osteocytes. Yet, increased osteocytic RANKL does not always result in targeted local bone resorption, suggesting that other osteocyte-derived molecules are involved in osteoclast precursor recruitment to bone areas containing apoptotic osteocytes needing to be resorbed. One potential candidate is osteoprotegerin (OPG), a RANKL-soluble decoy also produced by osteocytes and osteoblasts (126). However, recent evidence showed that OPG deletion in osteocytes (or B lymphocytes) does not affect bone mass, whereas deletion of OPG in osteoblasts increases bone resorption and reduces bone mass, suggesting that under physiological conditions osteoblasts are a major source of OPG (127). Whether osteocytic OPG becomes relevant under pathological situations remains to be determined. Another osteocyte-derived factor involved in osteoclast precursor recruitment is high mobility group box 1 (HMGB1) protein (128), an osteoclast chemotactic cytokine released by apoptotic osteocytes known to regulate the expression of several cytokines controlling osteoclastogenesis, including RANKL, TNF-α, interleukin (IL)-6, and OPG. Moreover, the expression of both RANKL and vascular endothelial growth factor A (VEGF-A) is elevated in living osteocytes surrounding dead osteocytes in overloaded rat bones (129). These findings suggest that apoptotic osteocytes alter in neighboring cells the expression of angiogenesis-inducing molecules that facilitate the recruitment of osteoclast precursors from the circulation to mineralized bone areas that need replacement.

FIGURE 2.

Osteocyte apoptosis: key molecular mechanisms. A: scheme showing that proapoptotic stimuli activate death receptors in osteocytes and initiate caspase-3-mediated programmed cell death. Osteocytes adjacent to apoptotic osteocytes are thought to release proinflammatory and proosteoclastogenic cytokines that attract osteoclasts and promote local bone resorption. B: image showing TUNEL assay to detect degraded DNA in apoptotic osteocytes in paraffin-embedded bones infiltrated with myeloma cancer cells. C: image showing active caspase-3 immunostaining of paraffin-embedded human bone from an osteoporotic patient. Black arrows point to apoptotic osteocytes; red arrows point to healthy osteocytes. AGE, advanced glycation end-products; HMGB1, high mobility group box 1; ROS, reactive oxygen species; TNF, tumor necrosis factor-α; TRAIL, tumor necrosis factor-related apoptosis inducing ligand.

Taken together, these pieces of evidence identify osteocyte apoptosis as the initiating event for bone loss in skeletal disuse scenarios, as occurs with reduced physical activity during aging, bed rest, and motor paralyses (130) or under microgravity conditions during long-term spaceflights (131). Future research efforts are warranted to examine the potential of therapies targeting osteocytes to prevent or treat the bone loss induced by weightlessness.

4.2. Regulation of Osteocyte Apoptosis by Bisphosphonates and Sex Steroids

The loss of sex steroids increases osteocyte apoptosis, whereas estrogens and androgens preserve the viability of both osteocytes and osteoblasts (101, 106). The prosurvival effect of sex steroids is caused by rapid activation of phosphatidylinositol 3-kinase (PI3K) and Src/Shc/ERK, as well as by nongenotropic events downstream of the sex steroid receptors (101, 132). In vitro and in vivo studies show that bisphosphonates promote osteocyte/osteoblast survival via ERK activation and Cx43 signaling (15, 94, 111, 133). Thus, it is possible that the antiapoptotic properties of estrogens and bisphosphonates, and the consequent inhibition of resorption due to the lack of apoptotic osteocytes, contributes to the suppression of bone remodeling seen with these agents.

In conclusion, evidence gathered in the last 20 years demonstrates that untimely premature apoptosis of osteocytes is associated with bone fragility in conditions of dysregulated mechanical and hormonal stimulation. Osteocyte apoptosis triggers signaling mechanisms that drive bone resorption at particular skeletal sites, also known as “targeted bone remodeling.” However, the identity of only some of the potential molecular mediators of the communication between apoptotic osteocytes and osteoclasts and their precursors has been revealed. Future research is warranted to determine the prospective participation of accessory cells as well as additional molecular mechanisms underlying this phenomenon. It is expected that this new knowledge will provide tools to manipulate the apoptotic osteocyte/osteoclast axis toward preserving bone mass and strength.

5. OSTEOCYTE SENESCENCE

The study of osteocyte senescence and the potential of targeting senescent cells in bone as a therapeutic approach to treat bone loss has attracted much attention recently in the musculoskeletal research community. Cellular senescence is a complex process that results in irreversible cell cycle arrest and marked changes in gene expression leading to a senescence-associated secretory phenotype (SASP) (134). In aging cortical bone, ∼10% of osteocytes become senescent, a number that contrasts with <2% in young bones (135, 136). Khosla and colleagues (135, 137, 138) have shown that diabetes and radiation also induce accumulation of senescent osteocytes, which are characterized by high p16Ink4a expression and a development of a SASP. The specific function of senescent osteocytes is yet to be defined. However, emerging data link the accumulation of senescent osteocytes with bone loss and poor bone quality that occurs with aging, radiation, and diabetes (135, 137, 138). Consistent with this notion, SASP accumulation in vitro or in vivo stimulates osteoclastogenesis and suppresses osteoblast differentiation (135). Several pharmacological strategies to either eliminate senescent cells or inhibit the production of SASP have shown prevention of bone loss in preclinical animal models (135, 137, 138). Future studies are warranted to optimize the regimes and timing of therapeutics agents targeting senescent cells in bone.

It is intriguing that accumulation of senescent osteocytes has been described in aging, radiated bones, and diabetes, conditions in which osteocyte apoptosis also is increased (95, 135, 137, 138). To our knowledge, a potential relationship between osteocyte senescence and apoptosis has not been established. Future studies are warranted to determine the timing of senescence versus apoptosis, whether there is a cause-effect relationship between the two phenomena, as well as whether the same or neighboring osteocytes undergo senescence and then apoptosis (or vice versa). Answers to these crucial questions without a doubt will increase the chances of finding targets to regulate these events, establish their potential relationship, and clarify their role in bone pathophysiology.

6. OSTEOCYTE PARACRINE FUNCTIONS

6.1. Regulation of Bone Formation by Osteocytes

Osteocytes sense mechanical forces and transmit bone anabolic signals to other bone effector cells through their canalicular system (139, 140) (FIGURE 3), making the osteocyte the main mechanosensor cell in the skeleton. Ablation of osteocytes by genetic means decreases bone mass and suppresses mechanically induced new bone formation (122). The key role of osteocytes in bone formation is further supported by the identification of genetic alterations in several components of the Wnt signaling pathway, which are accompanied by marked effects on bone mass (141). These clinical discoveries demonstrated that osteocytes are a major source of Wnt signaling antagonists, including DKK1 and Sclerostin (40, 142). The discovery of SOST, the gene encoding Sclerostin, and its impact on bone biology is perhaps one of the major advances in the last decades. Sclerostin was originally described in two rare genetic diseases, sclerosteosis and van Buchem disease, both characterized by high bone mass. Extensive in vitro and in vivo experimentation revealed that Sclerostin negatively regulates the interaction of Wnt ligands with low-density lipoprotein receptor-related proteins (LRPs) 5 and 6 and Frizzled receptors, ultimately preventing beta-catenin translocation to the nucleus (40, 141, 143). Furthermore, SOST knockout (KO) mice exhibit increased bone formation and bone mass, similar to the high bone mass phenotype that inactivating mutations in this gene induced in patients (141, 144) (FIGURE 4). In addition, consistent with the relevant role of osteocytes in mechanotransduction, downregulation of Sclerostin in osteocytes is required to achieve full bone anabolic response to mechanical loads (145, 146). Furthermore, Sclerostin expression is inhibited by parathyroid hormone (PTH), an FDA-approved anabolic agent for osteoporosis in the United States (147–152) (see sect. 8). Although deletion of DKK1 also results in a high-bone mass phenotype, the effects are minor compared with those observed with SOST inhibition. DKK1 inhibition increases SOST/Sclerostin expression in bone, which in turn counteracts the anabolic effects of DKK1 suppression (153). In fact, genetic deletion of both DKK1 and SOST leads to a robust anabolic, additive response greater than those seen when inhibiting each gene alone. Given the prominent role that SOST/Sclerostin and DKK1 have in the regulation of bone formation, several pharmaceutical approaches have been developed to neutralize their actions in the skeleton (discussed in sect. 5).

FIGURE 3.

Osteocyte paracrine signaling in bone. Osteocytes exert paracrine actions on osteoblasts and osteoclasts via cell-to-cell physical communication and/or secretion of molecules that stimulate or inhibit osteoblast or osteoclast differentiation. A: osteocytes produce the proosteoclastogenic cytokines receptor activator of nuclear factor-κΒ ligand (RANKL) and macrophage colony-stimulating factor (M-CSF) in both membrane-bound and soluble forms and the anti-osteoclastogenic factor osteoprotegerin (OPG). B: osteocyte-derived Sclerostin acts on mesenchymal stem cell precursors and promotes bone marrow adipogenesis. C: Wnt antagonists Dkk-1 and Sclerostin produced by osteocytes inhibit osteoblast differentiation and bone forming function. Furthermore, osteocytes express Notch receptors that can communicate physically with osteoblasts and potentially with osteoclasts.

FIGURE 4.

Osteocyte regulation of bone formation via Sclerostin production. A: Sclerostin immunostaining in paraffin-embedded human bone. Black arrow points to Sclerostin-positive osteocytes; red arrow points to Sclerostin-negative osteocytes. B: micro-computed tomography (μCT) image of vertebral cancellous bone from wild-type mice (WT) and mice with global deletion of SOST (SOST KO). C: toluidine blue-tartrate-resistant acid phosphatase (TRAP)-stained histological image of tibial cancellous bone from WT and SOST KO mice. D: dynamic histomorphometry in plastic-embedded bones from WT mice and mice overexpressing human SOST under the control of the DMP1-8kb promoter (Dmp1-8kb-hSOST). Green calcein labels and red alizarin labels are shown. E: toluidine blue-TRAP stained histological image of bones from C57BL/6 mice treated with IgG (control) or neutralizing anti-Sclerostin antibody (anti-Sclerostin-Ab). Yellow arrows point to osteoblasts on the bone surface.

Osteocytes are now considered the bone cells responsible for orchestrating the anabolic actions of the Wnt pathway. Importantly, Wnt signaling activation in osteocytes results in a robust increase in bone mineral density and bone volume (FIGURE 5). These anabolic bone effects are associated with increased osteoblast number and bone matrix production (39). These results contrast with previous work showing that Wnt/β-catenin signaling activation in earlier stages of the osteoblastic lineage (preosteoblasts or osteoblasts) increases OPG and blocks bone resorption with no alteration of physiological bone apposition (154–156).

FIGURE 5.

The osteocyte as the anabolic Wnt signaling cell in bone. Mice with genetic constitutive activation of canonical Wnt/β-catenin signaling in osteocytes (daβcat^Ot) have been used to determine the cell responsible for the bone anabolic effects of canonical Wnt signaling. A: daβcat^Ot mice are smaller in size compared with wild-type (WT) control littermates and exhibit denser bones. B: daβcat^Ot mice exhibit exuberant bone formation that expands into the bone marrow cavity of the femoral midshaft. C: dynamic histomorphometry in a plastic-embedded bone from a daβcat^Ot mouse. Green calcein labels and red alizarin labels are shown.

Osteocytes also can affect osteoblasts through other secreted factors (157) and physical interactions (158). It is possible that the Notch pathway, which regulates direct physical communication between adjacent cells, mediates the physical communication between osteocytes and osteoblasts (159, 160) (see sect. 9).

6.2. Osteocytes as Regulators of Bone Resorption

Osteocytes regulate osteoclast formation and function under both normal and disease conditions (FIGURE 3). Although not the only source, osteocytes are major suppliers of the proosteoclastogenic cytokine RANKL, which regulates the formation and activity of osteoclasts. Genetic deletion of RANKL (driven by Dmp1-Cre) in osteocytes, and some mature osteoblasts, markedly affects bone resorption, and adult mice lacking RANKL in osteocytes exhibit an osteopetrotic phenotype due to lack of mature osteoclasts (29, 30). Supporting the idea that osteocyte-derived RANKL is critical for adult bone remodeling, genetic deletion of RANKL with SOST-Cre, which targets osteocytes but not osteoblasts, resulted in a very similar high-bone mass phenotype (28). More recent studies examined the specific contribution of soluble RANKL resulting from cleavage of the membrane-bound RANKL. Intriguingly, deletion of soluble RANKL results in normal bone mass during growth, whereas adult mice display more cancellous bone due to fewer osteoclasts (161). The similarities between the phenotypes of adult mice with genetic deletion of RANKL/soluble RANKL in osteocytes suggest that soluble RANKL is an important mechanism by which osteocytes modulate bone resorption, yet deletion of soluble RANKL did not mitigate the bone loss induced by estrogen deficiency. In contrast, soluble RANKL appears to accelerate tumor metastasis to bone (162). Nonetheless, the majority of osteocytes are not physically in contact with cells on the bone surface or in bone marrow or near blood vessels. Therefore, how an osteocyte’s membrane-bound RANKL reaches osteoclast progenitors to induce their differentiation to mature, bone-resorbing osteoclasts demands further investigation. Finally, intriguing data presented in a recent report suggest the existence of reverse RANKL signaling in osteoblasts [i.e., receptor activator of NF-κB (RANK) secreted by osteoclasts binds to RANKL in osteoblasts], a mechanism proposed to contribute to the coupling of bone resorption and bone formation in local bone areas (163). Yet, whether reverse RANKL signaling occurs in osteocytes and its consequences remain to be determined.

Osteocytes also secrete OPG, a soluble decoy for RANKL, impeding the binding to its RANK on osteoclasts. OPG production in both osteocytes and osteoblasts is regulated by the Wnt/β-catenin pathway. Supporting this notion, mice lacking β-catenin in osteocytes are osteoporotic because of decreased OPG, increased osteoclast numbers, and bone resorption (126). However, as discussed above, recent evidence suggests that production of OPG by osteoblasts, instead of osteocytes, is responsible for the suppression of osteoclast formation and bone resorption under physiological conditions. Finally, osteocytes also appear to be a source of macrophage colony-stimulating factor (M-CSF), a cytokine essential for osteoclast differentiation (164).

6.3. Osteocytes and Bone Marrow Adipogenesis

Bone marrow adipose tissue (BMAT) is an active and dynamic fat depot confined in the bone marrow cavity that secretes adipokines involved in the regulation of glucose and lipid metabolism and bone homeostasis (165). The exact function and regulation of BMAT are largely unknown. The potential role of osteocytes in BMAT has been reviewed recently (166). Emerging evidence suggests that the Wnt antagonists SOST/Sclerostin and DKK-1 produced by osteocytes could play a role in the regulation of BMAT as well as peripheral fat depots (discussed in sect. 5; FIGURE 3). Supporting this notion, in vitro work shows that recombinant Sclerostin can induce adipocyte differentiation (167, 168). Furthermore, both genetic and pharmacological inhibition of Sclerostin decrease BMAT (168). Mechanistically, this proadipogenic effect is due to inhibition of the Wnt signaling pathway followed by increases in Peroxisome proliferator-activated receptor γ (PPARγ) gene expression (168). Furthermore, pharmacological inhibition of Sclerostin with a neutralizing antibody significantly reduced the increase in BMAT by rosiglitazone, an antidiabetic drug frequently used in the clinic (169).

In summary, the evidence discussed above demonstrates that osteocytes are essential for physiological, normal bone remodeling. Some, but not all, of the mechanisms used by osteocytes to regulate osteoblast and osteoclast function have been discovered during the last decade, leading to the development of new therapeutic approaches to combat low-bone mass conditions. However, it is important to note that bone remodeling is a complex process likely regulated by the coordinated action of several bone cells and factors. The new findings suggesting that osteocytes also influence bone marrow adiposity open the possibility that osteocytes regulate other cell populations in the bone marrow, a topic that demands further investigation.

7. OSTEOCYTE ENDOCRINE FUNCTIONS

7.1. Osteocytes and the Regulation of Whole Body Composition and Metabolism

Osteocyte-derived factors have recently emerged as endocrine signals involved in the cross talk between bone and distant organs regulating peripheral fat and glucose metabolism (170). Genetically induced osteocyte death in mice results in severe lymphocytopenia and peripheral white adipose tissue mass loss (171). Moreover, growing evidence suggests that osteocyte-derived Sclerostin plays a role in the regulation of peripheral fat. In vivo, mice with genetic overproduction of Sclerostin have elevated peripheral fat mass and impaired glucose metabolism. In contrast, mice with global deletion of SOST exhibit low peripheral fat mass compared with control mice (172, 173). Similarly, the elevation in Sclerostin that results from the genetic deletion of LRP4 in osteocytes or the osteocytic activation of β-catenin is accompanied by elevated body fat and peripheral white adipose fat and compromised glucose metabolism (174). Moreover, pharmacological inhibition of Sclerostin with neutralizing antibodies mitigates the high-fat diet-induced body fat accumulation and impairment of glucose metabolism (172). Although clinical data in this area are scarce, circulating Sclerostin is increased in type 2 diabetes patients, and serum Sclerostin levels positively correlate with body and fat mass in healthy adults and diabetic patients (175–177). Similarly, LRP5 mutations blocking the binding of Sclerostin to Wnt coreceptors result in altered fat distribution (178). It is important to note that there are some controversial results. For instance, the increase in Sclerostin that follows the genetic deletion of the stimulatory subunit of G protein G_sα in osteocytes results in loss of gonadal and inguinal white adipose tissue mass (179). This recent body of literature supports the notion that Sclerostin is an endocrine signaling factor that mediates the cross talk between bone and peripheral fat depots and the pancreas. Further investigations are required to determine the clinical relevance of these results, as well as the cellular and molecular mechanisms underlying Sclerostin’s actions on adipocyte biology.

7.2. Regulation of Phosphate Metabolism by Osteocytes: FGF23

Osteocytes express higher levels of genes related to mineralization and phosphate metabolism than osteoblasts, including PHEX, DMP1, and MEPE (2). Osteocytes also produce FGF23, which as discussed above plays a crucial role in the regulation of inorganic phosphate (Pi) homeostasis by inhibiting its renal reabsorption (11, 17, 180–182).

FGF23 secretion is stimulated by high dietary Pi and the active form of vitamin D₃, 1,25-dihydroxy-vitamin D₃ (1,25-D₃) (183–186). FGF23 is a recognized biomarker of altered Pi metabolism as it increases before detectable changes in PTH in early chronic kidney disease (CKD) (187). FGF23 production is also regulated by PTH. Indeed, elevation in both FGF23 and PTH is seen in the hyperphosphatemia exhibited by CKD patients (188, 189), and the increase in FGF23 in a rat model of kidney failure is prevented by parathyroidectomy (190).

FGF23 is also elevated by hyperparathyroidism, as seen in both patients and mouse models (191, 192). Jansen-type metaphyseal chondrodysplasia, caused by missense mutations resulting in constitutive active PTHR1 receptor signaling, is a rare disorder characterized by abnormal bone formation in the metaphysis of long bones. Patients with these mutations have high serum circulating FGF23 levels, regardless of exhibiting normal 1,25-D₃ and low Pi levels (193). Moreover, FGF23 expression is increased in osteocytes in mice overexpressing the same mutation in the PTH receptor found in Jansen patients (DMP1-caPTHR1^Ot transgenic mice) in DMP1-8kb-expressing cells. Despite circulating FGF23 being elevated in DMP1-8kb-caPTHR1 mice, plasma Pi and renal Pi reabsorption are normal. In addition, the FGF23-FGFR1-KLOTHO signaling axis is active in osteocytes, and osteocytes from DMP1-caPTHR1 mice exhibit increased expression of FGFR1 and Polypeptide N-Acetylgalactosaminyltransferase 3 (GALNT3), downstream targets of FGF23 signaling. Thus, PTH receptor signaling has the ability to integrate the endocrine and paracrine functions of osteocytes by regulating FGF23 through cAMP- and Wnt-dependent mechanisms (194).

The exact mechanisms by which osteocytes sense blood Pi and modulate FGF23 production are still unclear. In a recent article, glycerol-3-phosphate was identified as a potential signaling factor regulating FGF23 in bone (195). These results support the existence of a kidney-to-bone reverse signaling axis by which the kidney, the main target for FGF23 actions, releases glycerol-3-phosphate to regulate osteocyte FGF23 production in bone. Future studies to confirm this hypothesis are warranted.

7.3. Osteocytes and Skeletal Muscle

Growing evidence indicates that there is cross talk between bone and muscle, and recent findings point to osteocytes as potential mediators of this communication, yet the work conducted in this area is very limited. Studies using conditioned media from osteocyte-like cells demonstrated that factors released by osteocytes induce acceleration of myogenesis of C2C12 myoblasts (196, 197). Mechanistic studies identified prostaglandin E₂ and Wnt3a released by osteocytes as responsible for the enhanced myogenic differentiation of C2C12 cells in vitro (196, 198). Conditional deletion of Membrane Bound Transcription Factor Peptidase, Site 1 (MBTPS1) in osteocytes leads to an age-related increase in contractile force in adult slow twitch-dominant soleus muscles (197). Osteocytes also express a significant number of cytokines capable of inhibiting myogenesis (199). Conversely, the exercise-induced muscle factor β-aminoisobutyric acid can act as a bone-protective factor and prevents osteocyte cell death induced by ROS in vitro (200). Additional studies are needed to fully understand the role of osteocytes in the cross talk between bone and muscle and to determine the potential of targeting osteocytes to treat bone pathologies accompanied by muscle dysfunction.

In conclusion, emerging data suggest that osteocytes can regulate the function of distant organs through the secretion of endocrine factors. In addition to the well-accepted endocrine role of osteocyte-derived FGF23 in the kidney, evidence from animal models shows that osteocytes can also influence whole body fat, pancreas, and muscle function. However, a causal relationship between these new osteocyte-derived endocrine factors and human metabolic diseases is yet to be established. Future interventional clinical studies are needed to confirm these potential novel endocrine functions of osteocytes.

8. OSTEOCYTES AS MEDIATORS OF HORMONAL SIGNALING

It has long been hypothesized that osteocytes sense changes in mechanical strain and initiate an adaptive skeletal response by coordinating the activity of osteoclasts and osteoblasts. However, that osteocytes are also major target cells in bone and integrators of hormonal signaling was less expected. Evidence accumulated during the last few years confirms that osteocytes are crucial drivers of the skeletal changes induced by systemic hormones with recognized effects on bone.

8.1. Effects of Glucocorticoids on Osteocytes

Osteoblastic cells, and in particular osteocytes, are profoundly affected by GC (FIGURE 6). The role of GCs in different bone cell types has been elucidated by extensive research taking advantage of the enzymatic system of 11β-hydroxysteroid dehydrogenase (11β-HSD) type 1 and type 2, which regulates the local levels of active GC metabolites by a prereceptor mechanism. 11β-HSD2 overexpression in cells early in the osteoblastic differentiation stage revealed that GC signaling is essential for proper bone mass acquisition during growth (201). In addition, elevated prevalence of osteoblast and osteocyte apoptosis is a hallmark of GC excess in the adult skeleton (202, 203). Osteoblast apoptosis explains at least in part the reduced osteoblast number and low bone formation induced by the hormones. In addition, osteocyte apoptosis is likely to disrupt the osteocyte network contributing to GC-induced bone fragility. Indeed, blocking GC action by overexpressing 11β-HSD2 in osteoblasts and osteocytes prevents GC-induced apoptosis, preserving osteoblast function and bone formation (204, 205). GC-induced osteoblast and osteocyte apoptosis depends on the glucocorticoid receptor (GR) and is replicated in vitro, demonstrating that indeed apoptosis is due to direct hormonal effects on osteoblasts and osteocytes (206, 207). Moreover, despite the bone loss seen in osteocalcin gene 2 (OG2)-11β-HSD2 transgenic mice treated with GC, osteocyte apoptosis is prevented and bone strength is preserved, suggesting a direct connection between osteocyte apoptosis and bone fragility (208).

FIGURE 6.

Osteocytes as mediators of mechanical and hormonal signals in bone. A: osteocytes sense mechanical forces in bone and transduce them into biological cues to effector cells. Mechanical forces decrease Sclerostin, increase Wnt signaling, and decrease osteoprotegerin (OPG) expression in osteocytes to enable bone gain. ER, estrogen receptor; PTH1R, parathyroid hormone (PTH) receptor 1. B: glucocorticoid direct actions on osteocytes increase apoptosis (via Pyk2 activation) and Sclerostin expression, leading to decreased Wnt signaling and OPG expression and overall contributing to the bone loss induced by glucocorticoids. C: intermittent administration of PTH (iPTH) requires osteocytic PTH1R signaling to induce full bone anabolism. Mechanistically, PTH decreases SOST expression, activates Wnt signaling, and increases osteoblast number and function. Additionally, iPTH increases receptor activator of nuclear factor-κΒ ligand (RANKL) expression and promotes Notch signaling to regulate PTH-induced bone resorption.

Besides regulating cis- or trans-interactions between the ligand-bound receptor with DNA and the consequent regulation of gene transcription (209, 210), GCs modulate in bone cells the activity of several kinases, including c-Jun NH₂-terminal kinase (JNK), Proline-rich tyrosine kinase 2 (PYK2), and ERKs (211–216). In particular, GCs activate in osteoblastic cells PYK2, a kinase related to the focal adhesion kinase FAK. These kinases regulate cell survival by opposing mechanisms. Thus, FAK prevents and PYK2 promotes cell attachment-induced apoptosis (named anoikis) (112, 207, 209, 217–228). Recent evidence showed that genetic deletion or pharmacological inhibition of PYK2 prevents GC-induced bone loss by overriding GC effects on anoikis in osteocytes, osteoblasts, and osteoclasts (207, 229). GC extended osteoclast life span, induced osteoblast/osteocyte apoptosis, increased bone resorption, promoted microarchitectural deterioration, and reduced bone strength in control mice, whereas genetic deletion or pharmacological inhibition of PYK2 prevented all these effects. These findings support the notion that inhibition of PYK2 and anoikis signaling prevents the damaging skeletal effects of GC by promoting osteoblast/osteocyte survival and reducing osteoclast life span and bone resorption activity. Thus, targeting PYK2 is a promising new therapeutic approach for the treatment of GC-induced bone loss.

Another mechanism by which GC impacts bone homeostasis is through the accumulation of ROS in bone (230), associated with endoplasmic reticulum (ER) stress that is alleviated by phosphorylation of eukaryotic translation initiation factor 2α (eIF2α) (231, 232). Supporting the relevant role of ROS/ER stress as a mediator of the effects of GC on bone, the eIF2α dephosphorylation inhibitors salubrinal and guanabenz are capable of blocking ROS-induced ER stress (233, 234), as well as blocking the increase in osteoblast and osteocyte apoptosis in vitro and the inhibition of osteoblast differentiation induced by GC (235). In addition, salubrinal inhibited osteoblast and osteocyte apoptosis in vivo and blunted the GC-induced decrease in bone mass and bone formation. These findings provide an additional pathway to interfere with the deleterious actions of GC on bone.

GCs also have profound effects on the gene expression profile of osteocytes. SOST/Sclerostin expression is increased by GC, as well as the expression of several Wnt target genes linked to anabolism/survival (i.e., Cyclin D1) and anticatabolism (i.e., OPG) (236). Consistent with a driving role of SOST/Sclerostin in GC’s skeletal actions, mice lacking the gene SOST are protected from the loss of bone mass, microarchitectural deterioration, and decrease in structural and material strength induced by GC. Remarkably, the preservation of bone integrity in SOST KO mice is due to prevention of the increase in bone resorption induced by GC but not to the restoration of bone formation or the inhibition of osteoblast/osteocyte apoptosis (236). These studies indicate that, in the frame of GC excess, inhibition of SOST/Sclerostin and activation of Wnt/β-catenin signaling opposes bone catabolism and maintains bone integrity. In addition to SOST, GCs decrease the expression of several proteases required for perilacunar remodeling [MMP13, MMP14, tartrate-resistant acid phosphatase (TRAP)] (237). This suppression of perilacunar remodeling causes disruption of the osteocyte lacuno-canalicular network and contributes to the subchondral bone degeneration seen in GC-induced osteonecrosis.

8.2. Osteocytes as Mediators of PTH Signaling in Bone

PTH is an important mediator of extracellular calcium and phosphate levels and has profound effects on the skeleton. Chronic PTH elevation, as occurs in hyperparathyroidism, results in bone loss due to stimulation of bone resorption. In contrast, intermittent PTH elevation, as occurs with daily injections of PTH 1–34, stimulates bone formation more than resorption, an approach used in the clinic to treat osteoporosis (238–240). Mechanistically, the bone anabolic effects of PTH have been linked to its ability to stimulate proliferation of preosteoblasts, to promote osteoblast survival, to transform lining cells into matrix-synthesizing osteoblasts, or to a combination of these effects, downstream of the PTH receptor 1 (PTH1R) (238, 241). Basic and clinical work over the last three decades demonstrates that osteocytes are critical target cells for the anabolic actions of PTH in the skeleton (242) (FIGURE 6). Genetic activation of the PTH1R in osteocytes in mice is sufficient to elicit the effects of PTH on the skeleton (240, 243–245), and, conversely, genetic deletion of the PTH1R in osteocytes leads to a defective bone anabolic response to PTH in mice (85, 194, 246–249).

Parfitt postulated in the 1970s (250) that osteocytes must control the rapid release of calcium, whereas osteoclasts and resorption determine the later phase of calcium release in response to PTH. That osteocytes are involved in the regulation of calcium homeostasis by the hormone is supported by more recent findings showing that mice lacking the PTH1R in osteocytes (tamoxifen-induced DMP1-10kb-Cre) are unable to maintain normal circulating levels of calcium when fed a low-calcium diet (85). During lactation, PTH-related peptide (PTHrP) is produced to stimulate the release of calcium from the skeleton to the blood and enable milk production by the mammary glands. This phenomenon is accompanied by increased osteocyte lacunar size as a result of perilacunar bone remodeling providing the additional source of calcium and contributes to the overall bone loss exhibited during lactation (251–254). Perilacunar remodeling, or “osteocytic osteolysis,” is mediated by the induction of an “osteoclast-like” gene expression profile in osteocytes, including genes that promote acidification and degrade and/or dissolve the extracellular matrix and mineral (251–253). Importantly, conditional deletion of PTH1R in osteocytes mitigates the increase in osteocyte lacunar size and the bone loss seen with lactation in mice, demonstrating the involvement of direct actions of PTH1R signaling in these cells (252).

In contrast to the requirement of the PTH receptor in osteocytes for bone anabolism induced by PTH injections, chronic elevation of PTH in mice lacking the PTH1R in osteocytes fed a low-calcium diet still increased bone resorption and induced the expected bone loss (246). However, T cell depletion or conditional deletion of PTH1R in T cells in mice prevented the bone loss induced by chronic elevation of PTH (255–257). Thus, complementary cellular mechanisms facilitate the skeletal actions of PTH.

The molecular mechanisms downstream from PTH1R signaling involved in the bone effects of PTH remain unclear. It is well documented that salt-inducible kinases (59, 62, 258) mediate PTH decreases in SOST/Sclerostin expression, an effect lost in mice with conditional deletion of PTH1R in osteocytes (147, 148, 246). However, DMP1-8kb-SOST mice, in which SOST is overexpressed in osteocytes and thus cannot be downregulated by PTH, displayed the expected increase in bone formation and Wnt/β-catenin signaling induced by daily PTH injections, demonstrating that SOST downregulation is dispensable for bone anabolism induced by PTH (246). In addition, PTH also increased bone mass and induced new bone formation in global SOST KO mice (259). Furthermore, direct effects of PTH on osteocytes regulate the production of RANKL and promote bone resorption (260). In fact, genetic deletion of the PTH-responsive distal control region (DCR) in the RANKL gene (DCR−/− mice) or pharmacological inhibition of MMP14, a metalloproteinase involved in the generation of soluble RANKL from membrane-bound RANKL, attenuated the exacerbated bone resorption displayed by DMP1-caPTH1R^Ot mice (245, 260). Altogether, these studies suggest that osteocyte RANKL, and soluble RANKL in particular, are important components of the bone catabolic response to PTH signaling.

In conclusion, today it is recognized that osteocytes are critical target cells for the action of hormones that affect the skeleton, in particular GC and PTH. Extensive evidence from several research laboratories demonstrates that GCs regulate osteocyte survival and induce critical changes in osteocytic gene expression that impact bone remodeling. Regarding PTH, expression of the PTH1R in osteocytes is required for full bone anabolism by the hormone, and PTH controls bone formation and resorption by regulating the expression of several osteocytic genes, including RANKL, SOST/Sclerostin, DKK1, and MMP14. Future research is warranted to reveal the whole spectrum of mechanisms and molecular mediators downstream of signaling in osteocytes induced by GC and PTH, with the ultimate goal of identifying new targets and developing improved bone therapeutics.

9. OSTEOCYTE PATHOPHYSIOLOGY

9.1. The Effects of Aging on Osteocytes

One of the functions of osteocytes is to sense bone microdamage and initiate repair (1, 261). With aging, there is a decline in osteocyte density and an increase in number of empty osteocyte lacunae, an index of premature osteocyte death (262, 263). These changes in osteocytes are associated with accumulation of microdamage in old bones. It is possible that the reduction in osteocyte number is a direct consequence of apoptosis of osteoblasts, which is also increased. Furthermore, osteocyte apoptosis might be due to reduced skeletal loading, elevated levels of ROS in the bone/bone marrow microenvironment (264), and/or increased endogenous GC signaling with age (265). This loss of osteocytes could partially explain the discrepancy between bone quantity and quality that occurs during aging.

Another potential cause of osteocyte death is Cx43 expression, which is reduced with age (266). Cx43 is required to preserve osteocyte viability and regulates the expression of pro- and antiosteoclastogenic cytokines in osteocytes (14, 267). In vivo and in vitro observations show that osteocytes lacking Cx43 are more prone to undergo apoptosis and exhibit an increased RANKL-to-OPG ratio (14, 268). Furthermore, long bones from old rodents and humans, as well as from mice lacking Cx43 in osteocytes, display altered cortical bone geometry resulting from concomitant increased periosteal bone formation and exacerbated endocortical bone resorption (14, 109, 268–270).

9.2. Diabetes Mellitus and Osteocytes

Diabetes mellitus (DM) is associated with elevated bone fracture risk and impaired fracture healing, resulting from alterations in bone structure and decreased strength that compromises bone quality (271–273). Type 1 DM (T1D) patients have reduced bone mass, whereas type 2 DM patients display normal or even high bone mass. In both conditions, however, the bone disease is characterized by decreased bone formation and increased resorption. Yet, the mechanisms leading to DM-induced bone disease are not clear. Using a preclinical mouse model of DM, recent work demonstrated that osteocyte functions are altered in DM. Mice with T1D induced by streptozotocin display low bone mass, decreased structural and material properties, higher osteoclast number and bone resorption, and reduced osteoblast number and bone formation. In addition, T1D mice have increased osteocyte apoptosis and upregulation of Sclerostin, DKK1, and RANKL/OPG expression in bone (274, 275). The changes in formation and resorption rates induced by DM, as well as the bone loss, were rectified by peptides derived from PTH-related peptide (PTHrP)(1–37) and (107–111), which activate PTH1R signaling (274). Comparable effects in DM models have been shown with similar doses of PTH(1–34) (275), and more recent studies demonstrated that abaloparatide, a PTHrP analog that, like PTH, binds to the PTH1R and is an FDA-approved bone anabolic agent, was superior to PTH in increasing or restoring bone mass in control or DM mice, respectively (276, 277). However, both agents corrected the impaired bone structural and material properties, citrate content, and increased Wnt antagonist expression in mice with T1D. Whether the actions of abaloparatide and PTH are due to effects mediated by osteocytes remains to be established.

Osteocytes might be involved in the metabolic changes induced by DM. Regulation of body composition and glucose metabolism in mice fed a high-fat diet have been associated with osteocytic expression of peroxisome proliferator-activated receptor γ (PPARγ) (278). Conditional deletion of osteocytic PPARγ results in mice with less fat mass and increased insulin sensitivity and energy expenditure. In addition, mice lacking PPARγ in osteocytes retain glycemic control when fed with a high-fat diet, supporting the notion that osteocytes regulate glucose homeostasis via a PPARγ-dependent mechanism. Furthermore, PPARγ signaling in osteocytes positively regulates Sclerostin expression (279). These results support the hypothesis that osteocyte signaling mediates the bone-deleterious effects of DM and open the possibility of targeting osteocytes to treat the deleterious effects of DM on the bone.

9.3. Role of Osteocytes in Cancer That Grows in Bone

Bone is a preferred site for the metastasis of cancer cells, and cancer cells communicate with host bone cells to generate a niche conducive for the progression of cancer in bone (280). Advances in the last decade revealed some of the mechanistic underpinnings by which osteocytes contribute to tumor growth, skeletal destruction, and bone pain in cancer (281, 282) (FIGURE 7). Increased osteocyte apoptosis is found in bones from multiple myeloma (MM) patients (283), particularly in bone areas colonized by MM cells (284). Osteocyte apoptosis is triggered by MM-induced activation of Notch signaling in osteocytes, which along with TNF-α secreted by MM cells, induces caspase-3-mediated apoptosis (284). Autophagy might be also involved in the stimulation of osteocyte apoptosis by MM cells (285). The increase in osteocyte apoptosis is accompanied by accumulation of osteoclasts in bone areas adjacent to dead osteocytes (283), supporting the notion that local bone resorption and the development of lytic lesions can be initiated by apoptotic osteocytes. Supporting this hypothesis, MM cells also increase RANKL and decrease OPG expression in osteocytes and enhance the ability of osteocytes to attract osteoclast precursors in vitro (283, 284).

FIGURE 7.

Osteocytes generate a microenvironment supportive for the growth of cancer cells and exacerbated bone resorption. A: cocultures of osteocytes and myeloma-green fluorescent protein (GFP) cancer cells enable the study of communication by cell-to-cell direct cell contact and by the exchange of soluble factors. Under these conditions, myeloma cells establish physical contact with the body and dendritic processes of osteocytes. B: osteocytes transfected with nuclear GFP cocultured with myeloma cancer cells. Fragmented (yellow arrow) and normal (white arrow) nuclei in MLO-Y4 osteocytes cocultured with MM1.S myeloma cells are displayed. C: murine intratibial injection of myeloma cells engrafts and produces multiple myeloma (MM) tumors in the bone marrow tumors that induce osteolytic lesions (red arrows) similar to those seen in myeloma patients. D: pharmacological blockade of the osteocyte-derived factor Sclerostin with a neutralizing antibody (anti-Sclerostin Ab) prevents the progression of the myeloma-induced bone disease in mice bearing myeloma tumors. Three-dimensional micro-computed tomography (μCT) reconstructions of C57BL/KaLwRijHsd bones bearing 5TGM1 myeloma cells are shown.

Osteocytes, through the production of Sclerostin, play a prominent role in the suppression of osteoblast differentiation and new bone formation in MM (284, 286). Sclerostin production by osteocytes is increased in MM-colonized bones (284). In addition, the production of Sclerostin in bones in breast cancer models is associated with increased breast cancer cell migration and invasion, as well as with the development of bone osteolysis (287). Blockade of SOST/Sclerostin’s actions in MM promotes osteoblast formation and stimulates new bone formation without affecting tumor growth, and thus could become an important therapeutic target to combat the bone disease that accompanies MM (287–290) (see sect. 10). Of note, Sclerostin inhibition reduced the proliferation of metastatic breast cancer tumor cells (291). These studies suggest that Sclerostin’s actions on tumor cells are cancer type dependent, and thus use of anti-Sclerostin antibodies requires individual evaluation. Finally, MM cells also increase the expression of FGF23 and VEGFA in osteocytes (292, 293), which in turn elevate heparanase expression and contribute to angiogenesis, respectively.

Finally, osteocytes also alter the biology of malignant cells, enhancing cancer cell invasiveness and migration capacity and stimulating tumor growth (284, 294–297). Also, novel intriguing findings suggest the existence of cross talk between osteocytes and bone sensory nerves, which might contribute to the pain induced by cancer cell growth in bone (298).

9.4. Osteocytes and Rare Diseases

Analysis of osteocytes and their signaling factors in rare skeletal diseases has unequivocally facilitated the identification of new targets and the development of new therapeutic approaches to combat both rare and common skeletal diseases. Various genetic skeletal rare bone diseases are associated with dysfunctional osteocytes, as recently reviewed (299). As discussed in sect. 2, sclerosteosis and van Buchem disease are characterized by mutations in the osteocyte-derived factor Sclerostin, leading to low or lack of SOST expression and the development of a skeletal dysplasia characterized by high bone mass (300–305). These findings led to the generation of antibodies neutralizing Sclerostin skeletal actions and successful FDA approval for its use in the treatment of postmenopausal bone loss. Mutations in the PHEX and DMP1 genes cause hypophosphatemia and ARHR, respectively, diseases characterized by high levels of osteocyte-derived FGF23 and decreased bone mass (11, 20, 306). Antibodies against FGF23 were generated and recently approved for the treatment of children with X-linked hypophosphatemia (see sect. 10) (307).

Osteogenesis imperfecta is another rare skeletal disease with low bone mass and frequent fractures linked to mutations in COL1A1 and COL1A2 genes (308). Morello and colleagues reported that the osteocyte transcriptome is dysregulated in this disease (309) and that loss of osteocytic RANKL is sufficient to increase cancellous bone mass in murine models of osteogenesis imperfecta (310). Future studies are needed to fully understand the specific role of these mutations in osteocytes versus the effects that environmental clues sent by a diseased bone microenvironment may have on them.

In summary, the function and viability of osteocytes are altered in several disorders that compromise the integrity of bone, yet whether the changes in osteocyte behavior are the origin or just a consequence of the skeletal complications demands further investigation. Given the potential role of osteocytes as regulators of distant organs, the contribution of these cells to metabolic diseases and other nonbone disorders (i.e., sarcopenia) should also be investigated.

10. OSTEOCYTES AS TARGETS AND MEDIATORS OF THERAPEUTIC INTERVENTIONS

10.1. Bisphosphonates

The potent antiresorptive bisphosphonates have been widely used to treat bone diseases for decades and to preserve bone mass by inhibiting osteoclast activity and bone resorption. However, the antifracture properties of bisphosphonates cannot be fully accounted for by the effect on bone mass. Studies in the late 1990s provided an explanation for this apparent discrepancy by demonstrating that bisphosphonates preserve osteocyte (and osteoblast) viability (94) (FIGURE 8). All bisphosphonates inhibit apoptosis of osteoblastic cells, even analogs that are inactive on osteoclasts, supporting that these drugs act through different mechanisms in osteoclasts versus osteocytes/osteoblasts (311). Moreover, an analog that does not inhibit resorption still was able to prevent apoptosis of osteocytes and osteoblasts induced by GC or by unloading (125, 312). Furthermore, this analog protected from GC-induced loss of bone mass and strength. Preservation of osteoblast function and the osteocyte network by promoting cell viability in the absence of antiresorptive effects might be beneficial in the treatment of bone fragility diseases in which decreased bone resorption is not desired.

FIGURE 8.

Osteocytes and osteocyte-derived factors as therapeutic targets. A better understanding of osteocyte biology prompted the generation of therapeutics targeting osteocytes and osteocyte-secreted signaling molecules. Antibodies designed to neutralize Sclerostin (clinically tested and FDA approved) and Dkk1 (clinically tested but not FDA approved) function were developed to boost bone formation in osteoporotic patients. Similarly, a neutralizing antibody against receptor activator of nuclear factor-κΒ ligand (RANKL) (clinically tested and FDA approved) was generated to stop bone loss in osteoporotic and cancer patients. More recently, a neutralizing antibody against fibroblast growth factor (FGF)23 (clinically tested and FDA approved) has been approved for the treatment of X-linked hypophosphatemia and tumor-induced osteomalacia. Furthermore, we know now that bisphosphonates (clinically tested and FDA approved), proteasome inhibitors (clinically tested and FDA approved), and pan-inhibitors of the Notch pathway (clinically tested but not FDA approved; bone-targeting molecule (BT)-gamma secretase inhibitor (GSI) experimental] preserve osteocyte viability. Finally, pharmacological inhibition of Pyk2 signaling with PF-431396 (experimental, not FDA approved) decreases glucocorticoid induced osteocyte apoptosis and promotes osteoclast cell death. NFKB, nuclear factor-κΒ; OPG, osteoprotegerin; PTH, parathyroid hormone.

The antiapoptotic action of bisphosphonates is strictly dependent on the expression of Cx43, the most abundant member of this gap junction family of proteins in bone cells (267). Remarkably, the survival effect of bisphosphonates does not require the formation of gap junction channels (15, 111). Instead, bisphosphonates open Cx43 hemichannels, which in turn activate a pathway that inhibits apoptosis by phosphorylating p90^RSK kinase and its substrates, C/EBPβ and BAD.

10.2. Neutralizing Antibodies Against Sclerostin

Osteocyte-derived Sclerostin has rapidly become an important therapeutic target for the treatment of bone diseases characterized by low bone mass (313–316) (FIGURE 8). The neutralizing anti-Sclerostin antibody is considered a “dual” therapeutic agent, as it increases osteoblast number and stimulates bone formation but also inhibits osteoclast differentiation and suppresses bone resorption. In contrast to other therapeutic strategies in which bone formation and resorption are decreased (bisphosphonates) or increased (PTH1R ligands) at the same time, the bone gain induced by the anti-Sclerostin antibody results from opposing effects on bone formation (increased) and bone resorption (decreased). However, anti-Sclerostin discontinuation results in BMD loss, which eventually returns to pretreatment levels. In addition, the increased risk of cardiovascular events when the anti-Sclerostin antibody was compared with a bisphosphonate in clinical trials has raised safety issues (317, 318). Despite this observation, anti-Sclerostin was approved by the FDA in 2019 for the treatment of postmenopausal women at high risk of fracture.

Additionally, as discussed in sect. 9, preclinical animal studies showed that anti-Sclerostin antibody induces new bone formation in bones infiltrated with cancer cells without affecting tumor growth (287–290). Future studies are warranted to explore in clinical trials the potential benefits of anti-Sclerostin therapy in patients with cancer-related skeletal complications.

10.3. Neutralizing Antibodies Against RANKL

Taking into consideration that osteocytes are one of the major producers of RANKL in adult bone, it is possible that the antiresorptive properties of neutralizing antibodies against RANKL on bone are due to blockade of RANKL produced by osteocytes (FIGURE 8). Denosumab is a fully human monoclonal antibody against RANKL developed as a therapeutic agent to pharmacologically block bone resorption (319, 320). Clinical and preclinical work demonstrated that denosumab blocks RANKL signaling downstream of RANK in osteoclasts, resulting in inhibition of osteoclast formation and activity and decreased bone resorption. Denosumab is approved by the FDA for the treatment of postmenopausal osteoporosis, GC-induced osteoporosis, and the skeletal complications of metastatic prostate and breast cancer patients (321). A clinical issue that has recently attracted much attention is the bone loss that occurs after denosumab discontinuation, which can be mitigated by bisphosphonates (322). Croucher and McDonald have shown via intravital imaging of osteoclasts in vivo how removal of OPG-Fc, which mimics the effects of denosumab, increases fusion and fission of surviving osteoclasts, leading to the formation of active bone-resorbing osteoclasts potentially responsible for the initial rapid bone loss that ensues after denosumab withdrawal (323). In addition to the current patient populations approved to receive denosumab, a recent clinical study showed that denosumab prevents skeletal-related events in MM patients (324). Although it is clear that RANKL from osteocytes contributes to adult bone remodeling, whether the membrane-bound form, the soluble form, or both derived from osteocytes contribute to cancer-induced bone disease remains unclear and investigation is warranted.

10.4. Neutralizing Antibodies Against FGF23

Osteocytes produce FGF23 and regulate phosphate reabsorption in the kidney in an endocrine manner (24). Earlier this year, the FDA approved the use of burosumab, a neutralizing antibody against FGF23, for the treatment of X-linked hypophosphatemia (XLH) as well as tumor-induced osteomalacia (307) (FIGURE 8). Interestingly, the levels of FGF23, coming from either osteocytes or cancer cells, are elevated in prostate cancer and MM (293, 325–327). Whether targeting FGF23 in cancer would lead to anticancer effects and improvement of bone health is yet to be determined.

10.5. Proteasome Inhibitors and Osteocytes

Proteasome inhibitors (PIs) are drugs capable of blocking the action of the proteasome, a key cellular complex that degrades unneeded or damaged proteins. PIs are considered the backbone of MM treatment because of their ability to arrest tumor growth (328). Although preferentially used to stop the growth of tumors, both preclinical and clinical data demonstrate that PIs (329) also have effects on bone cells (330). Treatment with PIs inhibits RANKL expression in osteoblasts and NF-κB-dependent osteoclast differentiation, thus decreasing bone resorption (331). Furthermore, PIs have transient effects on osteoblasts that increase bone formation. (331). Treatment of MM patients with PIs maintained osteocyte viability (285) and decreased the serum levels of Sclerostin by 50% compared to control subjects (332) (FIGURE 8). However, whether these effects are due to direct actions on osteocytes remains to be determined. In the frame of GC excess, PIs might represent a therapeutic opportunity, as the hormones upregulate in bone (and muscle) the expression of atrogenes, E3 ubiquitin ligases that prime protein for proteasomal degradation (333).

10.6. Targeting Notch Signaling in Osteocytes

The effects of Notch signaling in osteocytes are complex. A series of studies by Canalis and collaborators showed that activation of Notch signaling in osteocytes by genetic overexpression of the Notch receptor intracellular domain (NICD)1 results in increased cancellous bone volume and fewer osteoclasts (160). Surprisingly, deletion of Notch receptors 1 and 2, which should inhibit Notch signaling, also increases trabecular bone mass by decreasing osteoclast number (334). The cause of some of these unexpected skeletal phenotypes might be interference with Notch signaling during bone development. Activation of Notch signaling in osteocytes in adult mice, which avoids potential developmental issues, demonstrated that Notch activation is sufficient to stimulate new bone formation and prevent the bone loss induced by age and ovariectomy (335). Additionally, activation of Notch in osteocytes in vitro, either by culturing osteocytes on Delta-like canonical Notch ligand 1 (DLL1)-coated plates or by coculturing them with MM cells, induces osteocyte apoptosis (284) (FIGURE 8). Moreover, pharmacological inhibition of the Notch pathway with gamma secretase inhibitors (GSIs), which block the cleavage of the NICD, decreases bone resorption (336, 337). However, this approach is clinically limited by severe side effects in the gut (338). To bypass gut toxicity, we generated a novel bone-targeted Notch inhibitor and showed that bone-targeted inhibition of Notch decreases tumor growth and bone destruction, without inducing gut toxicity (336). Importantly, bone-targeted inhibition of Notch does not affect physiological bone formation and seems to preserve the bone gain induced by daily PTH injections (339). An alternative strategy to avoid toxicity is the use of pharmacological targeting of individual Notch components. In particular, neutralizing antibodies against specific Notch receptors or ligands Delta‐1 and Jagged-1 are currently being studied in preclinical models (340, 341). It also has been shown that pharmacological inhibition of Jagged-1 blocks osteoblast mineralization (335). Additional efforts are needed to identify the role of specific components of the Notch pathway in osteocytes in bone homeostasis.

In summary, meticulous studies of osteocyte biology during the last two decades guided the development of new therapeutic approaches targeting osteocytic signaling pathways and derived factors to improve skeletal health. Although not discussed in this review, the role of osteocytes, both as contributors and/or as potential targets, in craniofacial disease and periodontal disease is an important and emerging area of research (342–344). Full characterization of osteocyte dysfunction in common and rare skeletal diseases could provide clues for the discovery of new targets to treat bone-related diseases.

11. CONCLUSIONS AND FUTURE DIRECTIONS

Advances during the last two decades demonstrated that osteocytes are multifunctional signaling cells that regulate osteoblasts and osteoclasts in a paracrine manner to maintain skeletal homeostasis. The knowledge gained about the underlying mechanisms provided the rationale for the development of new approaches to treat skeletal diseases targeting osteocytes. It is also now recognized that osteocytes are affected by mechanical and hormonal signals and by bone therapeutic agents. In addition, osteocytes contribute to the skeletal complications of diabetes and cancer as well as rare bone diseases. More recently, much attention has been attracted to the potential role of osteocytes as endocrine signaling cells in the regulation of body fat and glucose metabolism, beyond the well-recognized role of these cells in the regulation of phosphate metabolism through actions in the kidney. Future studies are needed to establish the contribution of osteocyte-derived molecules to whole body fat and their potential as therapeutic target cells for metabolic diseases.

Despite the tremendous advances in our knowledge of osteocyte biology, key areas of osteocyte biology remain underexplored. For instance, how osteocyte life span and gene expression signature are regulated is not completely understood. Unbiased genomewide approaches such as single-cell RNA sequencing (RNA-seq), proteomics, and epigenomics should advance our knowledge of the molecular changes preceding osteocyte dysfunction and thus guide the development of novel therapeutic targets. Furthermore, a better understanding of osteocyte energetic demands and utilization of glucose, fatty acids, and amino acids is needed. Moreover, although several paracrine and endocrine signaling mechanisms driven by osteocytes have been elucidated, autocrine signaling in osteocytes is underexplored.

In conclusion, osteocytes contribute to the physiological and pathophysiological regulation of bone metabolism. A complete comprehension of how osteocytes regulate the formation/function of osteoblast, osteoclasts, bone marrow adipocytes, and potentially other cells in remote organs is key to improving current therapeutic approaches and identifying novel approaches to target osteocytes, with the ultimate goal of finding superior and safer treatments to combat bone diseases.

GRANTS

This work was supported by the National Institutes of Health (R01-AR059357, R01-CA209882, R01-AT008754, and O2S1-AT008754 to T.B. and R37-CA251763 to J.D.-C.); an Arkansas Research Alliance Scholar award to T.B.; the Arkansas COBRE program NIGMS P20GM125503, a Scholar Award by the American Society of Hematology Scholar Award, and a Brian D. Novis Award by the International Myeloma Foundation (to J.D.-C.); the Winthrop P. Rockefeller Cancer Institute at UAMS (to T.B. and J.D.-C.); and the Department of Veterans Affairs (I01 BX002104 and IK6BX004596 to T.B.).

DISCLAIMERS

The contents do not represent the views of the US Department of Veterans Affairs or the United States Government.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.D.-C. prepared figures; J.D.-C. and T.B. drafted manuscript; J.D.-C. and T.B. edited and revised manuscript; J.D.-C. and T.B. approved final version of manuscript.