This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Manolagas, S. C. Search for Related Content PubMed PubMed Citation Articles by Manolagas, S. C.

Birth and Death of Bone Cells: Basic Regulatory Mechanisms and Implications for the Pathogenesis and Treatment of Osteoporosis¹

Division of Endocrinology & Metabolism, Center for Osteoporosis & Metabolic Bone Diseases, University of Arkansas for Medical Sciences, and the Central Arkansas Veterans Healthcare System, Little Rock, Arkansas 72205, USA

	Abstract

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
The adult skeleton regenerates by temporary cellular structures that comprise teams of juxtaposed osteoclasts and osteoblasts and replace periodically old bone with new. A considerable body of evidence accumulated during the last decade has shown that the rate of genesis of these two highly specialized cell types, as well as the prevalence of their apoptosis, is essential for the maintenance of bone homeostasis; and that common metabolic bone disorders such as osteoporosis result largely from a derangement in the birth or death of these cells. The purpose of this article is 3-fold: 1) to review the role and the molecular mechanism of action of regulatory molecules, such as cytokines and hormones, in osteoclast and osteoblast birth and apoptosis; 2) to review the evidence for the contribution of changes in bone cell birth or death to the pathogenesis of the most common forms of osteoporosis; and 3) to highlight the implications of bone cell birth and death for a better understanding of the mechanism of action and efficacy of present and future pharmacotherapeutic agents for osteoporosis.

I. Introduction

II. Physiological Bone Regeneration

A. Remodeling by the basic multicellular unit (BMU)

III. Osteoblastogenesis and Osteoclastogenesis

A. Growth factors and their antagonists

B. Cytokines

C. Systemic hormones

D. Adhesion molecules

IV. Reciprocal Relationship Between Osteoblastogenesis and Adipogenesis

V. Serial and Parallel Models of Osteoblast and Osteoclast Development

VI. Function of the Mature Cells

A. Osteoblasts

B. Osteocytes

C. Lining cells

D. Osteoclasts

VII. Death of Bone Cells by Apoptosis

VIII. Regulation of Bone Cell Proliferation and Activity

IX. Pathogenesis of Osteoporosis

A. Sex steroid deficiency

B. Senescence

C. Glucocorticoid excess

X. Pharmacotherapeutic Implications of Osteoblast and Osteocyte Apoptosis

A. Intermittent PTH administration

B. Bisphosphonates and calcitonin

C. Novel pharmacotherapeutic strategies

XI. Summary and Conclusions

	I. Introduction

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
And Athena lavished a marvelous splendor on the prince so that all the people gazed in wonder as he came forward. The elders making way as he took his father’s seat. The first to speak was an old lord, Aegyptius, stooped with age, who knew the world by heart.

Homer, the Odyssey: translation by Robert Fagles

LOSS OF height (stooping), Dowager’s hump, and kyphosis are some of the most visible signs of old age in humans. The primary reason for these involutional changes is a progressive loss of bone mass that affects the axial (primarily trabecular) as well as the appendicular (primarily cortical) skeleton. Loss of bone mass, along with microarchitectural deterioration of the skeleton, leads to enhanced bone fragility and increased fractures—the bone disease known as osteoporosis (1 ). Both men and women start losing bone in their 40s. However, women experience a rapid phase of loss during the first 5–10 yr after menopause, due to the loss of estrogen (2 ). In men this phase is obscure, since there is only a slow and progressive decline in sex steroid production; hence, the loss of bone in men is linear and slower (3 ). In addition to losing bone faster at the early postmenopausal years, women also accumulate less skeletal mass than men during growth, particularly in puberty, resulting in smaller bones with thinner cortices and smaller diameter. Consequently, the incidence of bone fractures is 2- to 3-fold higher in women as compared with men (4 ).

In addition to sex steroid deficiency and the aging process itself, loss of bone mass is accentuated when several other conditions are present. The most common are chronic glucocorticoid excess (5 ), particularly its iatrogenic form, hyperthyroidism as well as inappropriately high T₄ replacement, alcoholism, prolonged immobilization, gastrointestinal disorders, hypercalciuria, some types of malignancy, and cigarette smoking (6 ).

Bone loss and eventually fractures are the hallmarks of osteoporosis, regardless of the underlying cause or causes. The bone loss associated with normal aging in women has been divided into two phases: one that is due to menopause and one that is due to aging and affects men as well (7 8 ). In elderly women these two phases eventually overlap, making it difficult to distinguish the effect of sex steroid deficiency from the effect of the aging process itself. The effect of the aging process itself is also frequently obscured because of secondary hyperparathyroidism (9 ), resulting from impaired calcium absorption from the intestine with advancing age (>75 yr old). The bone loss that is due to glucocorticoid excess shares several features with the bone loss due to senescence, but also has unique features of its own. Nonetheless, as is the case with the other types of bone loss, the heterogeneity of the underlying conditions, some of which (e.g., postmenopausal state, rheumatoid arthritis, etc.) independently contribute to skeletal deterioration, can distort the clinical and histological picture (10 ). Irrespective of the overlap, it is important to recognize that the pathogenetic mechanisms are quite distinct in the various forms of osteoporosis and that sex hormone deficiency and aging have independent effects.

During the last few years, there have been significant advances in our understanding of the pathogenetic mechanisms responsible for the bone loss associated with sex steroid deficiency, old age, and glucocorticoid excess. All these conditions do not cause loss of bone mass by turning on a completely new process. Instead, they cause a derangement in the normal process of bone regeneration. Therefore, to understand the pathogenesis of osteoporosis and rationalize its treatment, one must first appreciate the basic principles of physiological bone regeneration.

	II. Physiological Bone Regeneration

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
The skeleton is a highly specialized and dynamic organ that undergoes continuous regeneration. It consists of highly specialized cells, mineralized and unmineralized connective tissue matrix, and spaces that include the bone marrow cavity, vascular canals, canaliculi, and lacunae. During development and growth, the skeleton is sculpted to achieve its shape and size by the removal of bone from one site and deposition at a different one; this process is called modeling. Once the skeleton has reached maturity, regeneration continues in the form of a periodic replacement of old bone with new at the same location (11 ). This process is called remodeling and is responsible for the complete regeneration of the adult skeleton every 10 yr. The purpose of remodeling in the adult skeleton is not entirely clear, although in bones that are load bearing, remodeling most likely serves to repair fatigue damage and to prevent excessive aging and its consequences. Hence, the most likely purpose of bone remodeling is to prevent accumulation of old bone. Remodeling, with positive balance, does occur in the growing skeleton as well. Its purpose, quite different from those proposed for the adult skeleton, is to expand the marrow cavity while increasing trabecular thickness (12 ).

A. Remodeling by the basic multicellular unit (BMU)
Removal of bone (resorption) is the task of osteoclasts. Formation of new bone is the task of osteoblasts. Bone resorption and bone formation, however, are not separate, independently regulated processes. In the uninjured adult skeleton, all osteoclasts and osteoblasts belong to a unique temporary structure, known as a basic multicellular unit or BMU (13 ). Although during modeling one cannot distinguish anatomical units analogous to BMU per se, sculpting of the growing skeleton requires spatial and temporal orchestration of the destination of osteoblasts and osteoclasts, albeit with different rules and coordinates to those operating in the BMU of the remodeling skeleton. The BMU, approximately 1–2 mm long and 0.2–0.4 mm wide, comprises a team of osteoclasts in the front, a team of osteoblasts in the rear, a central vascular capillary, a nerve supply, and associated connective tissue (13 ). In healthy human adults, 3–4 million BMUs are initiated per year and about 1 million are operating at any moment (Table 1). Each BMU begins at a particular place and time (origination) and advances toward a target, which is a region of bone in need of replacement, and for a variable distance beyond its target (progression) and eventually comes to rest (termination) (10 ). In cortical bone, the BMU travels through the bone, excavating and replacing a tunnel. In cancellous bone, the BMU moves across the trabecular surface, excavating and replacing a trench. In both situations, the cellular components of the BMUs maintain a well orchestrated spatial and temporal relationship with each other. Osteoclasts adhere to bone and subsequently remove it by acidification and proteolytic digestion. As the BMU advances, osteoclasts leave the resorption site and osteoblasts move in to cover the excavated area and begin the process of new bone formation by secreting osteoid, which is eventually mineralized into new bone.

	III. Osteoblastogenesis and Osteoclastogenesis

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
Both osteoblasts and osteoclasts are derived from precursors originating in the bone marrow. The precursors of osteoblasts are multipotent mesenchymal stem cells, which also give rise to bone marrow stromal cells, chondrocytes, muscle cells, and adipocytes (14 15 16 ), whereas the precursors of osteoclasts are hematopoietic cells of the monocyte/macrophage lineage (17 18 ). Long before these cells could be cultured, the existence of multipotent mesenchymal stem cells was suspected (19 ), based on the evidence that fibroblastic colonies formed in cultures of adherent bone marrow cells can differentiate, under the appropriate stimuli, into each of the above mentioned cells; these progenitors were named colony forming unit fibroblasts (CFU-F). When CFU-F are cultured in the presence of ß-glycerophosphate and ascorbic acid, the majority of the colonies form a mineralized bone nodule; these bone-forming colonies are known as CFU-osteoblast (CFU-OB) (16 ). Osteoblast progenitors may originate not only from stromal mesenchymal progenitors of the marrow, but also pericytes — mesenchymal cells adherent to the endothelial layer of vessels (20 ). Whereas osteoclast precursors reach bone from the circulation, osteoblast precursors most likely reach bone by migration of progenitors from neighboring connective tissues.

The development and differentiation of osteoblasts and osteoclasts are controlled by growth factors and cytokines produced in the bone marrow microenvironment as well as adhesion molecules that mediate cell-cell and cell-matrix interactions. Several systemic hormones as well as mechanical signals also exert potent effects on osteoclast and osteoblast development and differentiation. Although many details remain to be established concerning the operation of this network, a few themes have emerged (21 ). First, several of the growth factors and cytokines control each other’s production in a cascade fashion and, in some instances, form positive and negative feedback loops. Second, there is extensive functional redundancy among them. Third, some of the same factors are capable of influencing the differentiation of both osteoblasts and osteoclasts. Fourth, systemic hormones influence the process of osteoblast and osteoclast formation via their ability to control the production and/or action of local mediators.

A. Growth factors and their antagonists
The only factors capable of initiating osteoblastogenesis from uncommitted progenitors are bone morphogenetic proteins (BMPs) (22 ). BMPs have been long implicated in skeletal development during embryonic life and fracture healing. More recently, it has become apparent that BMPs, and in particular BMP-2 and –4, also initiate the commitment of mesenchymal precursors of the adult bone marrow to the osteoblastic lineage (23 ). BMPs stimulate the transcription of the gene encoding an osteoblast-specific transcription factor, known as osteoblast specific factor 2 (Osf2) or core binding factor a1 (Cbfa1), hereafter referred to as Cbfa1 (24 ). In turn, Cbfa1 activates osteoblast-specific genes such as osteopontin, bone sialoprotein, type I collagen, and osteocalcin. The importance of Cbfa1 for osteoblasts has been highlighted by the evidence that knockout of the Cbfa1 gene in mice prevents osteoblast development (25 26 ). In addition to Cbfa1, BMP-4 induces a homeobox-containing gene, distal-less 5(Dlx5), which also seems to act as a transcription factor, probably as a heterodimer with another homeobox-containing protein (Msx2). Like Cbfa1, Dlx5 regulates the expression of osteoblast-specific genes such as osteocalcin and alkaline phosphatase, as well as mineralization (27 28 29 ). Other factors such as transforming growth factor ß (TGFß), platelet-derived growth factor (PDGF), insulin-like growth factors (IGFs), and members of the fibroblast growth factor (FGF) family can all stimulate osteoblast differentiation (30 31 ). However, whereas TGFß, PDGF, FGF, and IGFs are able to influence the replication and differentiation of committed osteoblast progenitors toward the osteoblastic lineage, they cannot induce osteoblast differentiation from uncommitted progenitor cells.

In addition to growth factors, bone cells produce proteins that modulate the activity of growth factors either by binding to them and thereby preventing interaction with their receptors, by competing for the same receptors, or by promoting the activity of a particular factor. For example, osteoblasts produce several IGF-binding proteins (IGFBPs). Of these, IGFBP-4 binds to IGF and blocks its action, whereas IGFBP-5 promotes the stimulatory effects of IGF on osteoblasts (30 ). During the last few years, several proteins able to antagonize BMP action have also been discovered. Of them, noggin, chordin, and cerberus were initially found in the Spemann organizer of the Xenopus embryo and shown to be essential for neuronal or head development (32 33 34 35 ). Noggin and chordin inhibit the action of BMPs by binding directly and with high affinity with the latter proteins (36 37 ). Such binding is highly specific for BMP-2 and 4, as noggin binds BMP-7 with very low affinity and does not bind TGFß or IGF-I. Addition of human recombinant noggin to bone marrow cell cultures from normal adult mice inhibits not only osteoblast, but also osteoclast, formation, and these effects can be reversed by exogenous BMP-2 (23 ). Consistent with this evidence, BMP-2 and -4 and BMP-2/4 receptor transcripts and proteins are found in bone marrow cultures and in bone marrow-derived stromal/osteoblastic cell lines, as well as in murine adult whole bone. Noggin expression has also been documented in all these cell preparations. These findings indicate that BMP-2 and -4 are expressed in the bone marrow in postnatal life and serve to maintain the continuous supply of osteoblasts.

B. Cytokines
Since the early stages of hematopoiesis and osteoclastogenesis proceed along identical pathways, it is not surprising that a large group of cytokines and colony-stimulating factors that are involved in hematopoiesis also affect osteoclast development (38 ). This group includes the interleukins IL-1, IL-3, IL-6, IL-11, leukemia inhibitory factor (LIF), oncostatin M (OSM), ciliary neurotropic factor (CNTF), tumor necrosis factor (TNF), granulocyte macrophage-colony stimulating factor (GM-CSF), M-CSF, and c-kit ligand. As opposed to the above mentioned cytokines that stimulate osteoclast development, IL-4, IL-10, IL-18, and interferon- inhibit osteoclast development. In the case of IL-18 the effect is mediated through GM-CSF (39 ).

IL-6 has attracted particular attention because of evidence that it plays a pathogenetic role in several disease states characterized by accelerated bone remodeling and excessive focal or systemic bone resorption (40 ). IL-6 is produced at high levels by cells of the stromal/osteoblastic lineage in response to stimulation by a variety of other cytokines and growth factors such as IL-1, TNF, TGFß, PDGF, and IGF-II (41 42 43 ). Binding of IL-6 or other members of the same cytokine family (IL-11, LIF, OSM) to cytokine-specific cell surface receptors (in the case of IL-6, the IL-6R) causes recruitment and homo- or heterodimerization of the signal transducing protein gp130, which is then tyrosine phosphorylated by members of the Janus family of tyrosine kinases (JAKs) (44 ). This event results in tyrosine phosphorylation of several downstream signaling molecules, including members of the signal transducers and activators of transcription (STAT) family of transcription factors (45 46 ). Phosphorylated STATs, in turn, undergo homo- and heterodimerization and translocate to the nucleus where they activate cytokine- responsive gene transcription (47 ). The -subunit of the IL-6 receptor also exists in a soluble form (sIL-6R), but unlike most soluble cytokine receptors, it functions as an agonist by binding to IL-6 and then interacting with membrane-associated gp130 to stimulate JAK/STAT signaling (44 ). On the other hand, the soluble form of gp130 blocks IL-6 action (48 ).

Alone or in concert with other agents, IL-6 stimulates osteoclastogenesis and promotes bone resorption. The cells that mediate the actions of the IL-6 type cytokines on osteoclast formation appear to be the stromal/osteoblastic cells, as stimulation of IL-6R expression on these cells allows them to support osteoclast formation in response to IL-6 (49 ). These findings indicate that the osteoclastogenic property of IL-6 depends not only on its ability to act directly on hematopoietic osteoclast progenitors, but also on the activation of gp130 signaling in the stromal/osteoblastic cells that provide essential support for osteoclast formation. STAT3 activation in stromal/osteoblastic cells is essential for gp130-mediated osteoclast formation (50 ). Despite the effects of IL-6 on osteoclastogenesis in experimental in vitro systems, IL-6 is not required for osteoclastogenesis in vivo under normal physiological conditions. In fact, osteoclast formation is unaffected in sex steroid-replete mice treated with a neutralizing anti-IL-6 antibody, or in IL-6-deficient mice (51 52 ). The most likely explanation for this is that the -subunit of the IL-6 receptor in bone is a limiting factor, and that both a change in the receptor and the cytokine are required for the IL-6- mediated increased osteoclastogenesis, seen in pathological states.

IL-6 type cytokines are capable of influencing the differentiation of osteoblasts as well. Thus, receptors for these cytokines are expressed on a variety of stromal/osteoblastic cells, and ligand binding induces progression toward a more mature osteoblast phenotype, characterized by increased alkaline phosphatase and osteocalcin expression, and a concomitant decrease in proliferation (53 54 ). Moreover, IL-6 type cytokines stimulate the development of osteoblasts from noncommitted embryonic fibroblasts obtained from 12-day-old murine fetuses (55 ). Consistent with the in vitro evidence, several in vivo studies have demonstrated increased bone formation in transgenic mice overexpressing OSM or LIF (56 57 ).

TGFß is another example of a factor affecting both bone formation and bone resorption (58 ). Thus, in addition to its ability to stimulate osteoblast differentiation, TGFß increases bone resorption by stimulating osteoclast formation. Injection of TGFß into the subcutaneous tissue that overlies the calvaria of adult mice causes increased bone resorption accompanied by the development of unusually large osteoclasts, as well as increased bone formation. The effects of TGFß might be mediated by other cytokines involved in osteoclastogenesis as TGFß can stimulate their production. Mice lacking the TGFß1 gene due to targeted disruption exhibit excessive production of inflammatory cells, suggesting that this growth factor normally operates to suppress hematopoiesis (59 ).

C. Systemic hormones
The two principal hormones of the calcium homeostatic system, namely PTH and l,25-dihydroxyvitamin D₃ [1,25-(OH)₂D₃], are potent stimulators of osteoclast formation (17 60 ). The ability of these hormones to stimulate osteoclast development and to regulate calcium absorption and excretion from the intestine and kidney, respectively, are the key elements of extracellular calcium homeostasis. Calcitonin, the third of the classical bone-regulating hormones, inhibits osteoclast development and activity and promotes osteoclast apoptosis. Although the antiresorptive properties of calcitonin have been exploited in the management of bone diseases with increased resorption, the role of this hormone in bone physiology in humans, if any, remains questionable (61 62 63 ). PTH, PTH-related peptide, and 1,25-(OH)₂D₃ stimulate the production of IL-6 and IL-11 by stromal/osteoblastic cells (49 64 65 66 ). Several other hormones, including estrogen, androgen, glucocorticoids, and T₄, exert potent regulatory influences on the development of osteoclasts and osteoblasts by regulating the production and/or action of several cytokines (21 64 67 68 69 ).

D. Adhesion molecules
In addition to autocrine, paracrine, and endocrine signals, cell-cell and cell-matrix interactions are also required for the development of osteoclasts and osteoblasts (70 71 72 ). Such interactions are mediated by proteins expressed on the surface of these cells and are responsible for contact between osteoclast precursors with stromal/osteoblastic cells and facilitation of the action of paracrine factors anchored to the surface of cells that are required for bone cell development. Adhesion molecules are also involved in the migration of osteoblast and osteoclast progenitors from the bone marrow to sites of bone remodeling as well as the cellular polarization of osteoclasts and the initiation and cessation of osteoclastic bone resorption. More important, for the purpose of this review, adhesion molecules play a role in the control of osteoblast and osteoclast development and apoptosis (73 74 75 76 77 ).

The list of adhesion molecules that influence bone cell development and function includes the integrins, particularly _vß₃ and ₂ß₁, selectins, and cadherins, as well as a family of transmembrane proteins containing a disintegrin and metalloprotease domain (ADAMS). Each of these proteins recognizes distinct ligands. For example, some integrins recognize a specific amino acid sequence (RGD) present in collagen, fibronectin, osteopontin, thrombospondin, bone sialoprotein, and vitronectin (78 ).

	IV. Reciprocal Relationship Between Osteoblastogenesis and Adipogenesis

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
The cells that comprise the bone marrow stroma can serve several diverse functions including support of hematopoiesis and osteoclastogenesis, fat accumulation, and bone formation (79 ). This functional adaptation is apparently accomplished by the plasticity of some of the stem cell progeny as exemplified by the ability of stromal cells to convert between the osteoblast and adipocyte phenotype. Thus, a stromal cell type known as the Westen-Bainton cell exhibits PTH receptors and high alkaline phosphatase activity and gives rise to osteoblasts during fetal development and in hyperparathyroidism. On the other hand, when marrow hematopoietic activity is reduced using chemotherapeutic agents, these cells convert into adipocytes and can support myeloid cell production (80 81 82 83 84 ). Further, adipocytes isolated by limiting dilution from cultures of rabbit bone marrow can form bone in diffusion chamber implants (85 ). Conversely, addition of certain fatty acids to cultures of osteoblastic cells causes them to differentiate into adipocyte-like cells (86 ).

It is likely that interconversion of stromal cells among phenotypes, as well as commitment to a particular lineage with suppression of alternative phenotypes, is governed by specific transcription factors. Indeed, Cbfa1 is required for commitment of mesenchymal progenitors to the osteoblast lineage. Mice that are deficient in this factor lack osteoblasts and mineralized bone matrix (26 ); and expression of Cbfa1 in fibroblastic cells induces transcription of osteoblast- specific genes (24 ). On the other hand, CCAAT/enhancer binding protein (C/EBP), C/EBPß, and C/EBP, as well as peroxisome proliferator activated receptor 1 (PPAR1) and PPAR2 orchestrate adipocyte differentiation (87 88 89 90 ). Introduction of C/EBP in fibroblastic cells induces adipocyte differentiation (91 92 ), and transfection of fibroblastic cells with PPAR2 and subsequent activation with an appropriate ligand causes the development of adipocytes (93 ).

Using clonal cell lines isolated from the murine bone marrow, it has been demonstrated that PPAR2 can convert stromal cells from a plastic osteoblastic phenotype that reversibly expresses adipocyte characteristics to terminally differentiated adipocytes. Moreover, PPAR2 suppresses the expression of Cbfa1 and thereby osteoblast-specific genes (94 ). Similar to the inhibitory effect of PPAR2 on the osteoblast phenotype, the combination of PPAR and C/EBP suppresses the muscle cell phenotype when transfected into G8 myoblastic cells (95 ). Taken together, these findings strongly suggest that PPAR2 plays a hierarchically dominant role in the determination of the fate of mesenchymal progenitors, due to its ability to inhibit the expression of other lineage-specific transcription factors. Studies with a clonal cell line (2T3) suggest that BMP-2 induces osteoblast or adipocyte differentiation in mesenchymal precursors, depending on whether the BMP receptor type IA or IB is activated. Therefore, BMP receptors may also play a critical role in both the specification and reciprocal differentiation of osteoblast and adipocyte progenitors (96 ).

	V. Serial and Parallel Models of Osteoblast and Osteoclast Development

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
Even though millions of small packets of bone are constantly remodeled, bone mass is preserved thanks to a remarkably tight balance between the amount of bone resorbed and formed during each cycle of remodeling. In any established BMU, bone resorption and formation are happening at the same time; new osteoblasts assemble only at sites where osteoclasts have recently completed resorption, a phenomenon referred to as coupling, and formation begins to occur while resorption advances. The end result is a new packet of bone, either a cylindrical osteon or Haversian system, or a plate-like hemiosteon, that has replaced the older bone that was removed (97 ).

As the BMU advances, cells are successively recruited at each new cross-sectional location. Osteoblasts do not arrive until the osteoclasts have moved on. However, during the longitudinal progression of the BMU as a whole, new osteoclasts and new osteoblasts are needed simultaneously, although not at the same location. Two models of osteoblast recruitment, a serial and a parallel (Fig. 1), can explain the distinction between the cross-sectional and longitudinal events during BMU progression (38 ). According to the serial model, factors released from resorbed bone or the local increase in mechanical strain resulting from bone resorption, stimulate osteoblast precursor cell proliferation and differentiation (98 99 100 ). According to the parallel model, osteoblast and osteoclast precursor proliferation and differentiation occur concurrently in response to whatever signal conveys the need for initiation of new BMUs, and whatever hormone prolongs their progression (10 38 ). With either model, new osteoblasts must be directed to the right location.

View larger version (26K): [in this window] [in a new window]   Figure 1. Serial and parallel models of osteoblast and osteoclast development. For explanation, please see text. PreOC, Preosteoclast; GF, growth factors released from the matrix of resorbed bone; preOB, osteoblast progenitors. The expressions of RANK ligand and RANK on preosteoblasts and preosteoclasts, respectively, are depicted to indicate their critical contribution in osteoclastogenesis and thereby the dependency of osteoclastogenesis on preosteoblastic cells.

In full agreement with the in vitro evidence for the dependency of osteoclast development on support by cells related to osteoblasts, mice lacking osteoblasts due to Cbfa1 deficiency also lack osteoclasts (26 ). In addition, marrow cells from SAMP6 mice, a strain with defective osteoblastogenesis, exhibit decreased osteoclastogenesis (105 ) and do not exhibit the expected increase in osteoclastogenesis nor do they lose bone after loss of sex steroids (106 ).

The molecular mechanism of the dependency of osteoclastogenesis on cells of the mesenchymal lineage has been elucidated during the last 2 yr with the discovery of three proteins involved in the TNF signaling pathway (reviewed in Ref. 107 ). Two of these proteins are membrane-bound cytokine-like molecules: the receptor activator of nuclear factor-B (NF-B) (RANK) and the RANK-ligand. Other names used in the literature for RANK are osteoprotegerin ligand (OPG-L) and TRANCE. RANK is expressed in hematopoietic osteoclast progenitors, while RANK-ligand is expressed in committed preosteoblastic cells and T lymphocytes (108 109 110 ). RANK-ligand binds to RANK with high affinity. This interaction is essential and, together with M-CSF, sufficient for osteoclastogenesis. 1,25-(OH)₂D₃, PTH, PTHrP, gp130 activating cytokines (e.g., IL-6, IL-11), and IL-1 induce the expression of the RANK-ligand in stromal/osteoblastic cells (50 107 ). Osteoprotegerin (OPG), the third of the three proteins, unlike the other two, is a secreted disulfide-linked dimeric glycoprotein. A hydrophobic leader peptide and three and one-half TNF receptor-like cysteine-rich pseudorepeats characterize the amino terminus of this protein. Unlike other members of the TNF receptor family, OPG does not posses a transmembrane domain. OPG has very potent inhibitory effects on osteoclastogenesis and bone resorption in vitro and in vivo (111 ). Consistent with an important role of OPG in the regulation of osteoclast formation, OPG transgenic mice develop osteopetrosis, whereas OPG knockout mice exhibit severe osteoporosis (112 ). The antiosteoclastogenic property of OPG is due to its ability to act as a decoy by binding to RANK-ligand and blocking the RANK-ligand/RANK interaction. In addition to skeletal metabolism, the RANK/RANK-ligand/OPG circuit may regulate several other biological systems. Indeed, OPG is produced by many tissues other than bone, including skin, liver, stomach, intestine, lung, heart, kidney, and placenta as well as hematopoietic and immune organs. Consistent with this, mice deficient in RANK-ligand completely lacked lymph nodes as well as osteoclasts (113 ). Moreover, OPG is also a receptor for the cytotoxic ligand TRAIL (TNF-related apoptosis-inducing ligand) to which it binds with high affinity and inhibits TRAIL-mediated apoptosis in lymphocytes (114 ) and also regulates antigen presentation and T cell activation (115 ).

Osteoblastic cells and T lymphocytes, the two cell types that express high levels of RANK-ligand, are also the two cell types that express high levels of the osteoblast-specific transcription factor Cbfa1 (24 ). More intriguingly, both the murine and human RANK-ligand genes contain two functional Cbfa1 sites, and mutation of these sites abrogates the transcriptional activity of the RANK-ligand gene promoter (116 ). Therefore, the cell-specific expression of RANK-ligand in cells of the stromal/osteoblastic lineage and concurrent differentiation of osteoblasts and osteoclasts might be dictated, at least in part, by interaction between an osteoblast-specific transcription factor and RANK-ligand. BMP 2 and -4 stimulate Cbfa1 expression. Based on these lines of evidence, it has been postulated that the molecular underpinnings of the control of the rate of bone regeneration and the concurrent production of osteoclasts and osteoblasts could well be a BMPCbfa1RANK-ligand gene expression cascade in cells of the bone marrow stromal/osteoblastic lineage (117 118 ). According to this hypothesis, BMPs may provide the tonic baseline control of both processes, and thereby the rate of bone remodeling, upon which other inputs (e.g., biomechanical, hormonal, etc.) operate.

Studies with transgenic and knockout animal models as well as models with spontaneous genetic mutations have identified at least three transcription factors that are required for osteoclast differentiation: PU-1, fos, and NF-b. A review of the precise role of these factors is beyond the scope of this article, but the reader is referred to a recent excellent review of the topic (119 ).

	VI. Function of the Mature Cells

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
A. Osteoblasts
The fully differentiated osteoblasts produce and secrete proteins that constitute the bone matrix (120 ). The matrix is subsequently mineralized under the control of the same cells. A major product of the bone-forming osteoblast is type I collagen. This polymeric protein is initially secreted in the form of a precursor, which contains peptide extensions at both the amino-terminal and carboxyl ends of the molecule. The propeptides are proteolytically removed. Further extracellular processing results in mature three-chained type I collagen molecules, which then assemble themselves into a collagen fibril. Individual collagen molecules become interconnected by the formation of pyridinoline cross-links, which are unique to bone. Bone-forming osteoblasts synthesize a number of other proteins that are incorporated into the bone matrix, including osteocalcin and osteonectin, which constitute 40% to 50% of the noncollagenous proteins of bone. Mice deficient in osteocalcin develop a phenotype marked by higher bone mass and improved bone quality, suggesting that osteocalcin functions normally to limit bone formation without compromising mineralization (121 ). Conversely, mice deficient in osteonectin exhibit decreased osteoclast and osteoblast numbers and bone remodeling and profound osteopenia, suggesting that, under normal conditions, this protein may play a role in the birth or survival of these cells (122 ). Other osteoblast-derived proteins include glycosaminoglycans, which are attached to one of two small core proteins: PGI (or biglycan) and decorin; the latter has been implicated in the regulation of collagen fibrillogenesis. A number of other minor proteins such as osteopontin, bone sialoprotein, fibronectin, vitronectin, and thrombospondin serve as attachment factors that interact with integrins.

In addition to being the cells that produce the osteoid matrix, mature osteoblasts are essential for its mineralization, the process of deposition of hydroxyapatite (123 124 ). Osteoblasts are thought to regulate the local concentrations of calcium and phosphate in such a way as to promote the formation of hydroxyapatite. In view of the highly ordered, well aligned, collagen fibrils complexed with the noncollagenous proteins formed by the osteoblast in lamellar bone, it is likely that mineralization proceeds in association with, and perhaps governed by, the heteropolymeric matrix fibrils themselves. Osteoblasts express relatively high amounts of alkaline phosphatase, which is anchored to the external surface of the plasma membrane. Alkaline phosphatase has been long thought to play a role in bone mineralization. Consistent with this, deficiency of alkaline phosphatase due to genetic defects leads to hypophosphatasia, a condition characterized by defective bone mineralization (125 ). However, the precise mechanism of mineralization and the exact role of alkaline phosphatase in this process remain unclear. Bone mineralization lags behind matrix production and, in remodeling sites in the adult bone, occurs at a distance of 8–10 µm from the osteoblast. Matrix synthesis determines the volume of bone but not its density. Mineralization of the matrix increases the density of bone by displacing water, but does not alter its volume.

B. Osteocytes
Some osteoblasts are eventually buried within lacunae of mineralized matrix. These cells are termed osteocytes and are characterized by a striking stellate morphology, reminiscent of the dendritic network of the nervous system (126 127 ). Osteocytes are the most abundant cell type in bone: there are 10 times as many osteocytes as osteoblasts. Osteocytes are regularly spaced throughout the mineralized matrix and communicate with each other and with cells on the bone surface via multiple extensions of their plasma membrane that run along the canaliculi; osteoblasts, in turn, communicate with cells of the bone marrow stroma which extend cellular projections onto endothelial cells inside the sinusoids. Thus, a syncytium extends from the entombed osteocytes all the way to the vessel wall (128 ) (Fig. 2). As a consequence, the strategic location of osteocytes makes them excellent candidates for mechanosensory cells able to detect the need for bone augmentation or reduction during functional adaptation of the skeleton, and the need for repair of microdamage, and in both cases to transmit signals leading to the appropriate response; albeit this remains hypothetical (129 ). Osteocytes evidently sense changes in interstitial fluid flow through canaliculi (produced by mechanical forces) and detect changes in the levels of hormones, such as estrogen and glucocorticoids, that influence their survival and that circulate in the same fluid (130 131 132 ). Therefore, disruption of the osteocyte network is likely to increase bone fragility.

View larger version (61K): [in this window] [in a new window]   Figure 2. Functional syncytium comprising osteocytes, osteoblasts, bone marrow stromal cells, and endothelial cells. [Adapted from G. Marotti and reproduced with the permission of the Editor of Journal of Clinical Investigation 104:1363–1374, 1999 (219 ).

D. Osteoclasts
Mature osteoclasts are usually large (50 to 100 µm diameter) multinucleated cells with abundant mitochondria, numerous lysosomes, and free ribosomes. Their most remarkable morphological feature is the ruffled border, a complex system of finger-shaped projections of the membrane, the function of which is to mediate the resorption of the calcified bone matrix (17 123 ). This structure is completely surrounded by another specialized area, called the clear zone. The cytoplasm in the clear zone area has a uniform appearance and contains bundles of actin-like filaments. The clear zone delineates the area of attachment of the osteoclast to the bone surface and seals off a distinct area of the bone surface that lies immediately underneath the osteoclast and which eventually will be excavated. The ability of the clear zone to seal off this area of bone surface allows the formation of a microenvironment suitable for the operation of the resorptive apparatus.

The mineral component of the matrix is dissolved in the acidic environment of the resorption site, which is created by the action of an ATP-driven proton pump (the so-called vacuolar H+-ATPase) located in the ruffled border membrane. The protein components of the matrix, mainly collagen, are degraded by matrix metalloproteinases, and ca-thepsins K, B, and L are secreted by the osteoclast into the area of bone resorption (134 ). The degraded bone matrix components are endocytosed along the ruffled border within the resorption lacunae and then transcytosed to the membrane area opposite the bone, where they are released (135 136 ). Another feature of osteoclasts is the presence of high amounts of the phosphohydrolase enzyme, tartrate-resistant acid phosphatase (TRAPase). This feature is commonly used for the detection of osteoclasts in bone specimens (137 ). Mice deficient in TRAPase exhibit a mild osteopetrotic phenotype (due to an intrinsic defect of osteoclastic resorptive activity) and defective mineralization of the cartilage in developing bones (138 ).

	VII. Death of Bone Cells by Apoptosis

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
The average lifespan of human osteoclasts is about 2 weeks, while the average lifespan of osteoblasts is 3 months (Table 1). After osteoclasts have eroded to a particular distance, either from the central axis in cortical bone or to a particular depth from the surface in cancellous bone, they die and are quickly removed by phagocytes (139 ). The majority (65%) of the osteoblasts that originally assembled at the remodeling site also die (140 ). The remaining are converted to lining cells that cover quiescent bone surfaces or are entombed within the mineralized matrix as osteocytes (Fig. 3A). Both osteoclasts and osteoblasts die by apoptosis, or programmed cell death, a process common to several regenerating tissues (141 ). As in other tissues, bone cells undergoing apoptosis are recognized by condensation of chromatin, the degradation of DNA into oligonucleosome-sized fragments, and the formation of plasma and nuclear membrane blebs (Fig. 4). Eventually the cell breaks apart to form so-called apoptotic bodies. Osteoblast apoptosis explains the fact that 50–70% of the osteoblasts initially present at the remodeling site of human bone cannot be accounted for after enumeration of lining cells and osteocytes (142 ). Moreover, the frequency of osteoblast apoptosis in vivo is such that changes in its timing and extent could have a significant impact in the number of osteoblasts present at the site of bone formation (130 ). Osteocytes are long-lived but not immortal cells; some die by apoptosis (132 143 144 ). Osteocyte apoptosis could be of importance to the origination and/or progression of the BMU. Indeed, osteocytes are the only cells in bone that can sense the need for remodeling at a specific time and place. Moreover, osteocytes are in direct physical contact with lining cells on the bone surface, and targeting of osteoclast precursors to a specific location on bone depends on a "homing" signal given by lining cells (133 ).

View larger version (31K): [in this window] [in a new window]   Figure 3. Osteoblast apoptosis and its implications. A, The average life span of a matrix forming osteoblast (200 h in the mouse) is indicated by the continuous line. The process of apoptosis represents only a small fraction of this time period. The alternative two fates of osteoblasts are to become lining cells or osteocytes. The fraction of osteoblasts that undergo apoptosis in vivo (fApoptosis) can be estimated from a bone biopsy specimen. The duration of the apoptosis phase that can be observed in the specimen (tApop) depends on the sensitivity of the detection method. For example, in the case of the TUNEL technique (without CuSO4 enhancement), the TUNEL-labeled phase of apoptosis is estimated to be approximately 2 h. In a steady state, the fraction of cells at a particular stage is the same as the corresponding fraction of time spent in that stage. Assuming an apoptosis detection time of 2 h and a 200-h life span, a prevalence of TUNEL positive osteoblasts in the biopsy of 0.005 indicates that half of the osteoblasts die by apoptosis. B, A change in the timing and extent of osteoblast apoptosis (fApoptosis) from 50% to zero should increase the number of osteoblasts present at the site of bone formation and thereby the work output, i.e., the amount of bone formed by a given team of matrix-forming osteoblasts. It will also lead to an increase in the density of osteocyte apoptosis, as illustrated by the example shown.

View larger version (121K): [in this window] [in a new window]   Figure 4. Two osteoblasts undergoing apoptosis in a section of murine cancellous bone (TUNEL staining with toluidine blue counterstain, x630). Apoptotic osteoblasts (shown in brown) are adjacent to an intact osteoblast (shown in blue), on the surface of a trabecula occupying the right lower portion of the picture in which two intact (blue stained) osteocytes are also seen. Apoptotic osteoblasts display nuclear condensation and fragmentation. [Photomicrograph provided by Robert S. Weinstein, M.D., University of Arkansas for Medical Sciences.]

	VIII. Regulation of Bone Cell Proliferation and Activity

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
A large body of literature suggests that growth factors, cytokines, hormones, and drugs regulate the proliferation of committed cells or the biosynthetic and functional activity of the differentiated osteoblasts and osteoclasts. Hence, in addition to cell number, alterations in the functional activity of individual cells, i.e., cell vigor, may contribute to changes in the rate of bone resorption and formation. However, because of the inherent difficulty in demonstrating changes in individual cell vigor in vivo, the vast majority of such literature and its conclusions rely heavily, if not exclusively, on in vitro experimentation. A detailed discussion of this work is beyond the scope of this review, and the reader is referred to other articles (30 147 148 149 150 151 ). Nonetheless, some general aspects merit discussion here, as they are important for putting the significance of birth rate and apoptosis into a broader perspective.

In regenerating tissues, the initial commitment of a stem cell progeny is followed by amplification with several or many rounds of cell division. In most tissues, division of stem cells is infrequent, and almost all of the divisions that produce the final population of differentiated cells occur in the so-called transit compartment (152 ). This notion is obscured by the frequent practice of showing linear diagrams representing the transition from one cell type to another and ignoring completely the amplification during the transition. In general, terminally differentiated cells do not divide, and an osteoblast, defined as a cell making bone matrix, and osteoclast, as a cell resorbing bone, are in this category. Therefore, the in vivo relevance of much of the in vitro evidence on the regulation of osteoblastic or osteoclastic cell proliferation, using established cell lines or primary cultures of isolated cells, must be largely confined to changes in this transit compartment. Irrespective of whether a given regulatory factor, be that a cytokine or a hormone, influences the initial commitment of a stem cell, or the subsequent amplification of its progeny, or both, the end result is a change in the rate of production and therefore the number of cells available for the execution of the biological task. For the sake of simplicity, the terms birth, rate of birth, osteoblastogenesis, or osteoclastogenesis, as used in this article, implicitly combine commitment and amplification.

The in vivo relevance of numerous reports of in vitro observations of changes in the biosynthetic activity of osteoblastic cells, e.g., changes in the level of expression of osteocalcin or alkaline phosphatase in response to a given agent, is also a matter of conjecture. Most likely, given the nature of commonly used in vitro cell models which, by and large, represent preterminally differentiated cells, such observations might be more relevant to postcommitment differentiation events, than to altered activity of the fully differentiated cell. Moreover, even if some agents can alter the function of terminally differentiated cells in short-term cultures in vitro, heretofore there is no evidence that short-term change in the rate of collagen production or bone matrix digestion, for example, are ultimately translated into differences in the amounts of bone matrix formed or resorbed.

In the bone literature there is considerable ambiguity when using the term "activation." It is important that one distinguishes between activation as a switch from an off state to an on state, and activation as modulation of activity of an already active cell. Morphological evidence summarized elsewhere does not support the notion that completely inactive osteoclasts are waiting for a stimulus to make them active (117 ). However, this evidence does not address the issue of whether there are variations in the rate of bone matrix dissolution by individual osteoclasts after they started work.

Studies with the widely used bone slice pit bioassay have shown that several regulatory factors can cause a decrease or increase in the resorptive ability of individual osteoclasts (153 154 155 ). However, it is not clear to what extent these observations reflect a change in cell vigor as opposed to a change in the precariously short lifespan of osteoclasts in vitro. In estrogen deficiency, individual osteoclasts are seemingly more "active" as they dig deeper resorption cavities often leading to trabecular perforation, but it has been convincingly argued that this is due to delayed apoptosis (133 ). In Paget’s disease osteoclasts are certainly far more aggressive than normal, perhaps as a result of their uniquely large size and number of nuclei (156 ). Today, the most compelling evidence in support of the notion that the vigor of individual osteoclasts may not always be maximal comes from in vitro as well as in vivo experiments with soluble RANK-ligand (157 ). Specifically, it has been found that RANK-ligand acts on mature rat osteoclasts in vitro to stimulate more frequent cycles of resorption and induce rearrangements of the cytoskeleton. Moreover, intravenous administration of RANK-ligand to mice elevates the circulating concentration of ionized calcium within 1 h. RANK-ligand has potent antiapoptotic effects on cultured osteoclasts (Ref. 158 and William Boyle, personal communication). Therefore, definitive conclusions regarding RANK-ligand’s ability to modulate osteoclast vigor will have to await dissection of the contribution of its effects on osteoclast survival in vivo. Similar to osteoclasts, the bone-forming ability of osteoblasts may not be always maximal, as suggested by the evidence that PTH can rapidly enhance bone formation when administered by subcutaneous injections to rats (159 ).

To conclude this section, it is intuitive that the amount of bone resorbed or formed by a team of osteoclasts and osteoblasts should be a function of the total cell number as well as the vigor of individual cells. However, whereas cell number can be quantified on bone sections from animals and humans with conventional histomorphometric techniques, quantification of individual cell vigor cannot. This situation makes it difficult to judge at present whether cell vigor is or is not a critical component in the pathogenesis of abnormal skeletal regeneration in common acquired metabolic bone diseases such as postmenopausal, senile, or steroid-induced osteoporosis. For this reason and space limitations, changes in bone cell vigor or other potential mechanisms of osteoporosis resulting from changes in extraskeletal tissues, for example altered calcium absorption or excretion, will not be discussed in the following section. This also reflects the author’s intention to focus on the dynamics of bone cell number, rather than a dismissal of other mechanisms.

	IX. Pathogenesis Of Osteoporosis

Top Abstract I. Introduction II. Physiological Bone... III. Osteoblastogenesis and... IV. Reciprocal Relationship... V. Serial and Parallel... VI. Function of the... VII. Death of Bone... VIII. Regulation of Bone... IX. Pathogenesis Of Osteoporosis X. Pharmacotherapeutic... XI. Summary and Conclusions References

 
From the brief discussion of the principles of physiological bone regeneration and the role of osteoblasts and osteoclasts in the process, it is obvious that the rate of supply of new osteoblasts and osteoclasts and the timing of the death of these cells by apoptosis are critical determinants of the initiation of new BMUs and/or extension or shortening of the lifetime of existing ones. Recent advances in our understanding of the pathogenesis of the various forms of osteoporosis have confirmed this truism by revealing that over- or undersupply of these cells relative to the need for remodeling are the fundamental problems in all these conditions (160 ) (Table 2).

In line with the fact that loss of sex steroids increases the rate of bone remodeling, in addition to up-regulating osteoclastogenesis, loss of sex steroids increases the number of osteoblast progenitors in the murine bone marrow. These changes are temporally associated with increased bone formation and parallel the increased osteoclastogenesis and bone resorption (175 ). As IL-6 type cytokines can stimulate osteoblast development and differentiation (54 55 146 ), increased sensitivity to IL-6 and other members of this cytokine’s family may account also for the increased osteoblast formation that follows the loss of gonadal function. In view of the fact that mesenchymal cell differentiation and osteoclastogenesis are tightly linked, stimulation of mesenchymal cell differentiation toward the osteoblastic lineage after sex steroid loss may be the first event that ensues after the hormonal change, and increased osteoclastogenesis and bone loss might be downstream consequences of this change (106 ).

In addition to IL-6, estrogen also suppress TNF and M-CSF (176 177 ), and estrogen loss may increase the sensitivity of osteoclasts to IL-1 by increasing the ratio of the IL-1RI over the IL-1 receptor antagonist (IL-RII) (178 ). As in the case of IL-6, the effects of estrogen on TNF and M-CSF are mediated via protein-protein interactions between the estrogen receptor and other transcription factors. In agreement with the evidence that IL-1 and TNF play a role in the bone loss caused by loss of estrogen, administration of IL-1RA and/or a TNF soluble receptor ameliorates the bone loss caused by ovariectomy in rats and mice (179 180 181 ). Because of the interdependent nature of the production of IL-1, IL-6, and TNF, a significant increase in one of them may amplify, in a cascade fashion, the effect of the others (161 ). Interestingly, recent in vitro studies with human osteoblastic cells indicate that OPG production is stimulated by estrogen, suggesting that this cytokine may also play an important role in the antiosteoclastogenic (antiresorptive) action of estrogen on bone (182 ).

Increased remodeling, resulting from up-regulation of osteoblastogenesis and osteoclastogenesis, alone can cause a transient acceleration of bone mineral loss because bone resorption is faster than bone formation, and new bone is less dense than older bone. However, in addition to increased bone remodeling, loss of sex steroids leads to a qualitative abnormality: osteoclasts erode deeper than normal cavities (133 183 ). In this manner, sex steroid deficiency leads to the removal of some cancellous elements entirely; the remainder are more widely separated and less well connected. An equivalent amount of cancellous bone distributed as widely separated, disconnected, thick trabeculae is biomechanically less competent than when arranged as more numerous, connected, thin trabeculae. Concurrent loss of cortical bone occurs by enlargement and coalescence of subendocortical spaces, a process due to deeper penetration of endocortical osteoclasts.

This deeper erosion can be now explained by evidence that estrogen acts on mature osteoclasts to promote their apoptosis; consequently, loss of estrogen leads to prolongation of the lifespan of osteoclasts (133 ). Specifically, estrogen promotes osteoclast apoptosis in vitro and in vivo by 2- to 3-fold, an effect seemingly mediated by TGFß (139 ). In direct contrast to their proapoptotic effects on osteoclasts, estrogen (as well as androgen) exerts antiapoptotic effects on osteoblasts and osteocytes; consequently, loss of estrogen or androgen leads to shorter lifespan of osteoblasts and osteocytes (184 ). Extension of the working life of the bone-resorbing cells and simultaneous shortening of the working life of the bone-forming cells, can explain the imbalance between bone resorption and formation that ensues after loss of sex steroids. Furthermore, the increase in osteocyte apoptosis could further weaken the skeleton by impairment of the osteocyte-canalicular mechanosensory network. The increase in bone remodeling that occurs with estrogen deficiency would partly replace some of the nonviable osteocytes in cancellous bone, but cortical apoptotic osteocytes might accumulate because of their anatomic isolation from scavenger cells and the need for extensive degradation to small molecules to dispose of the osteocytes through the narrow canaliculi. Hence, the accumulation of apoptotic osteocytes caused by loss of estrogen, or glucocorticoid excess (130 ), could increase bone fragility even before significant loss of bone mass, because of the impaired detection of microdamage and repair of substandard bone.

In conclusion, the increased rate of bone remodeling in estrogen deficiency may be due to increased production of both osteoclasts and osteoblasts, and the imbalance between bone resorption and formation is due to an extension of the working lifespan of the osteoclast and shortening of the working lifespan of the osteoblast. Moreover, a delay of osteoclast apoptosis seems responsible for the deeper resorption cavities and thereby the trabecular perforation associated with estrogen deficiency.

Clinical observations of decreased bone mass in a male with mutant estrogen receptor (185 ), and increased bone mass after treatment with estrogen in two males with P-450 aromatase deficiency (186 187 ), have raised the possibility that estrogen derived by peripheral aromatization of androgens is critical for the maintenance of bone mass in men as well as in women (188 ). However, in all three cases, the decreased bone mass in young males with estrogen deficiency in the face of androgen sufficiency could be due to failure in achieving peak bone mass from defects occurring during development or growth, not to loss of bone mass, as it is the case with common forms of osteoporosis. In addition, individuals with complete androgen insensitivity, due to mutations in the androgen receptor gene on the X chromosome and increased testosterone and estrogen production, have decreased bone mass, in spite of the elevated estrogen levels (189 ). Moreover, androgens, including nonaromatizable ones, have identical effects to those of estrogen on the biosynthetic activity and the birth as well as the death of bone cells in vitro and in vivo, at least in rodents (67 106 190 ). It is therefore more likely that both estrogen and androgen are important for the maintenance of bone mass in the adult male skeleton.

B. Senescence
The amount of bone formed during each remodeling cycle decreases with age in both sexes. This is indicated by a consistent histological feature of the osteopenia that occurs during aging, namely a decrease in wall thickness, especially in trabecular bone (191 192 193 ). Wall thickness is a measure of the amount of bone formed in a remodeling packet of cells and is determined by the number or activity of osteoblasts at the remodeling site.

Studies measuring bone turnover by histomorphometry (194 ), or indirectly by circulating markers (195 196 197 ), have suggested that in aging women, even in extreme old age, bone turnover is most likely increased by secondary hyperparathyroidism or by the continuing effect of estrogen deficiency. Increased turnover and reduced wall thickness are not inco