Organization of bone

Macroscopically, the skeleton is composed of flat bones (e.g. bones of the skull) and long bones (e.g. tibia), based on development and growth occurring through an intramembraneous process or endochondral and intramembraneous processes, respectively.¹ From a gross anatomical perspective, individual bones are made up of varying proportions of cortical (compact) and trabecular (cancellous) tissue, depending on location, and reflecting specific structural and functional differences. Compact bone is 80–90% mineralized tissue by volume, whereas trabecular bone is only 15–25% mineralized. Thus, trabecular bone has a much larger interface with soft tissue such as bone marrow, vascular, and connective tissue. In general terms, skeletal specialization is dictated by the relative abundance of cortical and trabecular tissue within specific bones; that is, cortical tissue accounts mainly for mechanical and protective functions while metabolic and cellular renewal functions are attributed mostly to trabecular tissue.

Microscopically, the hard tissue of bone is composed of a mineralized matrix and cellular elements that give rise to, maintain, and remodel bone. The bony matrix is composed of hydroxyapatite crystals associated with type-I collagen fibers and noncollagenous proteins. The predominant mineral composition of bone (95% of its mineral weight) is Ca₁₀(PO₄)₆(OH)₂, hydroxyapatite, with carbonate and other small impurities. Type I collagen accounts for about 90% of the total protein in bone with absorbed plasma proteins and proteins synthesized by bone-forming cells accounting for the remaining noncollagenous component. Bony matrix glycoproteins and proteoglycans serve to stabilize the mineral crystal. Before mineralization the organic matrix, called osteoid, is laid down by active osteoblasts and becomes mineralized soon after deposition. However, osteoid can accumulate under conditions such as rickets and chronic renal failure, when adequate calcium and/or phosphate ions become limiting.

The lamellar structure of adult bone forms from the alternating orientation of collagen fibers, allowing the greatest density of collagen per volume of bone tissue. Lamellae can be found in parallel sheets on the flat surfaces of periosteum and trabecular bone, or concentrically located around channels containing blood vessels in so-called Haversian systems. Deviation from this preferred arrangement of collagen is found in woven bone where fibers are oriented loosely and randomly. Woven or immature bone is found under conditions of rapid bone formation, as occurs with growth, fracture healing, tumor formation, and in some metabolic bone diseases.

Cellular elements

Osteocytes are bone-forming cells that become trapped in the bony matrix, and are produced and are found embedded within bone in small lacunae. These cells have many long processes, formed before matrix calcification, that interconnect with processes from bone-lining cells or other osteocytes and collectively make up a network of canaliculi. The periosteocytic space includes both the lacunae and this network, serving as a reservoir for calcium-containing extracellular fluid. The surface area of adult bone represented by periosteocytic space is 7–35 times the surface area of all lung capillaries and thus can be considered a sizeable bed for exchangeable bone calcium.

Despite their restricted environs, osteocytes continue to function as bone-forming cells at the surface of lacunae. In addition, they may serve as mechanosensors in the activation of local remodeling events under endocrine/paracrine and autocrine control. The ultimate fate of the osteocyte is probably death by phagocytosis during bone resorption.

Under the influence of fibroblast growth factors, bone morphogenetic proteins and Wnt proteins, bone marrow mescenchymal stem, or stromal cells differentiate into osteoblast precursors and finally bone-lining osteoblasts that produce osteoid. Osteoblasts respond to receptor-mediated signals such as parathyroid hormone, prostaglandins, estrogens, and vitamin D₃ to regulate synthesis of bony matrix. They also are intimately involved in bone remodeling, expressing both activating (e.g. colony stimulating factor-1 [CSF-1], receptor for activation of nuclear factor kappa B ligand [RANKL]) and inhibitory (e.g. osteoprotegerin, a decoy RANK receptor) factors of osteoclastogenesis (see Remodeling in the Spine below).

Osteoclasts are giant multinucleated cells whose precursors are bone marrow monocytes/macrophages derived from hematopoietic stem cells and whose sole function is bone resorption.² They are typically located in contact with a calcified bone surface within Howship’s lacunae. Their so-called zone of contact with bone is a ruffled membrane bounded circumferentially by an actin ring that serves to seal off the underlying bone as a discrete resorbing compartment. Specific attachment of osteoclasts to matrix components occurs through αvβ3, αvβ5, and α2β1 integrin receptors. The ruffled membrane contains both a proton ATPase and coupled chloride channel that acidifies the sealed-off bone resorbing compartment while maintaining electrical neutrality across the membrane. This is made possible by carbonic anhydrase within the cell and a HCO₃⁻/Cl⁻ exchanger on the basolateral membrane. RANK, the receptor for RANKL, and the macrophage-colony stimulating factor receptor are also located on the basolateral membrane, both being required for osteoclast differentiation.

Ossification in the spine

As in long bones, ossification in the spine occurs through an endochondral process.³ Mesenchymal cells give rise to prechondroblasts which then become progressively embedded within their cartilaginous matrix in lacunae. Now designated as chondrocytes, they continue to proliferate and then enlarge progressively to become hypertrophic, eventually undergoing apoptotic cell death. As cartilage cells die, and presumably under the influence of vascular endothelial growth factor (VEGF), the area is invaded by blood vessels. Osteoblasts move onto the cartilage surface and deposit bone in this area of primary spongiosa.

Osteoclasts or chondroclasts remove residual cartilage, which is then replaced by bone in the secondary spongiosa. This process is largely responsible for initial replacement of the embryonic cartilage template, fracture repair, and the growth/elongation of bones. Subsequent new bone formation is dictated by mechanical, humoral, or local signals and occurs by a remodeling process that continuously attempts to match bone architecture with function.

Cartilage models

Neural crest cells contribute to the branchial arch derivatives of the craniofacial skeleton, and the lateral plate mesoderm gives rise to the limbs. However, it is the paraxial mesoderm that gives rise to most of the axial skeleton, including the vertebral bodies, as well as the nonbranchial arch derivatives of the craniofacial skeleton. Mesenchymal cells condense to form regions with positional identity, which ultimately represent the future skeletal outline. The fate of mesenchymal cells within condensations dictates whether ossification occurs via an intramembranous or endochondral process. The former occurs as the result of osteoblast differentiation in the calvarium, maxilla, mandible, and subperiosteal bone-forming region of long bones. It is mediated by the transcription factors CBFA1/RUNX2 and osterix (OSX).^4–7 Endochondral ossification requires differentiation of mesenchymal cells into chondrocytes, creating cartilage models of the remaining skeleton that serve as templates for bony replacement. Chondrocyte differentiation is influenced by members of the SOX-family of transcription factors.^5–8

Cartilage models develop by interstitial and appositional growth until they take on the characteristic shapes and sizes of those intended bones. Patterning of the endochronal skeleton occurs even in the absence of eventual bone formation. Regulation of cartilage differentiation and growth in proper spatial sequence is essentially sufficient to complete a correctly shaped skeleton and this concept is supported by the phenotype of mice harboring two null alleles for the gene encoding CBFA1/RUNX2.^5,6 Despite the lack of osteoblast differentiation in these mice, a nearly intact skeleton is formed with accurately placed and correctly shaped skeletal elements, albeit consisting totally of cartilage.

Regulatory pathways

Transcriptional control mechanisms regulate osteoblastogenesis. Examples of early transcriptional regulators include the homeodomain proteins (e.g. Msx-2, Dlx-2, Dlx-5, BAPX1), steroid receptors, as well as the helix-loop-helix (HLH) proteins Id, Twist, and Dermo. The HLH proteins play important roles in the proliferation of osteoprogenitor cells, but repress osteoblast differentiation and must be down-regulated before a mature bone cell phenotype can be expressed.²¹ The activating protein cFOS is also expressed in osteoprogenitor cells, resulting in osteosarcomas if overproduced, but is not detected in mature osteoblasts.²² A similar theme is true for some of the homeodomain transcription factors, where mesenchymal precursors abundantly express them, but mature bone-forming cells do not.^23,24

In contrast, the runt homology domain proteins (e.g. Runx2/Cbfa1) and the zinc finger protein Osterix and others activate genes expressed in the mature osteoblast, such as osteocalcin and osteopontin (Fig. 38.1).^6,7,25–27 Null mouse mutants of Runx2 and Osterix result in inhibition of bone formation and perinatal mortality, demonstrating their requirement in osteogenesis.^6,7 Runx2 also plays important roles in the maturation of chondrocytes, cartilage mineralization, and endothelial cell migration necessary for vascular invasion.^28–30 Runx2 mediates its plethora of effects by interaction with diverse coregulatory proteins, including CBFβ1, Wnt pathway regulator LEF-1, and TGF-β/BMP responsive SMAD proteins, to name a few.^31–38 Temporal expression of Runx2 followed by Osterix is thought to be essential for the later stages of osteoblast differentiation.⁷

Fig. 38.1 Stem cell commitment and differentiation of osteoblasts. Cell types are indicated by boxes. Important regulatory factors are listed above arrows and transcription factors below arrows. Typical markers of later maturation stages are given above the block arrows. The dotted, double-sided arrow denotes possible transdifferentiation. Only representative examples of regulatory and other factors are shown (NCPs, noncollagenous proteins).

Hormonal regulation of osteoblast differentiation includes steroid as well as polypeptide signaling molecules. Glucocorticoids enhance the differentiation of bone marrow stromal cells to osteoblasts in vitro, but in vivo have negative effects on bone that contribute to osteoporosis by inducing apoptotic cell death.^39–41 Vitamin D₃, sex steroids, adrenomedullin, and leptin all have osteogenic or anabolic effects on bone.^42–50 Retinoic acid contributes to skeletal development in the embryo and also plays a role in postnatal bone formation.^51–53 Among the polypeptide hormones, PTH and PTHrP have stimulatory effects on osteoprogenitor cells and regulatory effects on differentiation, both mediated by coupled G protein interactions with their receptors.^54–61

Osteoprogenitor cells

Osteoprogenitor cells can arise from stem cells in a variety of tissues. There are as yet no unique identifying markers for the mesenchymal stem cell (MSC) that gives rise to bone, fat, cartilage, and muscle. Bone marrow stroma contains highly proliferative cells that will form single colonies or colony-forming units-fibroblasts (CFU-Fs) and are thought to contain mesenchymal stem cells that can be distinguished from early hematopoietic precursors.^62–65 Stromal cells grown in vitro are heterogeneous with respect to the capacity for differentiation, and only a low percentage of all CFU-Fs have stem cell properties.^62,63,66,67 Further, only a small fraction of CFU-Fs are bone-forming cells.^68,69 Although a specific osteoprogenitor marker has not been identified, antibodies have been developed (e.g. STRO-1, SP-10, SH2, HOP-26) that recognize some subsets of precursor cells.^70–77

Mesenchymal differentiation may proceed by a multistep hierarchical process with greater cell lineage restriction to terminal cell types at each subsequent step.⁷⁸ This idea is supported by identification of bipotential adipocyte-osteoblast precursors, and has suggested that there may be an inverse relationship between adipocyte and osteoblast differentiation.^79,80 A related observation is the transdifferentiation of bone to fat cells, or fat to bone cells.^81–84 Depending on the local cellular environment, ‘committed’ MSC progeny may also dedifferentiate into another lineage.⁸⁵ Thus, plasticity may be a common phenomenon; however, at least some reports suggest that cell fusion events may confound this paradigm.^86–90 Figure 38.1 shows a simplified scheme for osteoblast differentiation.

MSCs similar to those found in bone marrow have been found in adult peripheral blood, fetal cord blood, fetal liver, tooth pulp, muscle satellite cells, and extramedullary adipose tissue and have osteogenic potential.^91–100 Multipotent adult progenitor cells (MAPCs) derived from bone marrow can contribute to most somatic cell types, including skeletal tissue.^101,102 Pericytes can also be induced toward the osteogenic lineage.¹⁰³

Osteogenic growth factors

Osteogenic growth factors have been described as signaling polypeptides that function in the bone microenvironment to regulate skeletal metabolism. These factors may be secreted by bone-forming cells, stromal cells, cells of the hematopoietic system, or by cells at distant sites whose products are released into the circulation and act systemically on bone.

Bone morphogenetic proteins (BMPs), members of the transforming growth factor β (TGF-β) superfamily, are produced by skeletal and extraskeletal tissues. BMP-2, −4, and −6 are secreted by osteoblasts and play an autocrine role by enhancing the differentiated function of bone-forming cells.¹⁰⁴ BMPs induce the differentiation of cells toward the osteoblast lineage, increase the terminally differentiated pool of osteoblasts, and induce endochondral ossification and cartilage formation in concert with Indian/Sonic hedgehog.^105–107 BMPs interact with three distinct serine/threonine kinase receptors which, after dimerization, can signal through at least eight different Smads that become phosphorylated and/or heterodimerize to regulate transcription.^108–110 Smads have activating and inhibitory functions, and also mediate signals initiated through TGF-β and activin receptors. BMPs and TGF-β signaling may also use Smad-independent pathways that rely on Ras or MAP kinase.¹¹¹ Suppression of BMP activity occurs at several levels, including nonsignaling pseudoreceptors, inhibitory Smads, Smurf-mediated ubiquitination and degradation of signaling Smads, intracellular Smad binding proteins, and extracellular BMP antagonists such as noggin, chordin, follistatin, and twisted gastrulation.¹¹²

TGF-β1, -β2, and -β3 are polypeptides synthesized by bone cells and function as mitogens for preosteoblasts, stimulatory factors for collagen synthesis, and inducers of osteoclast apoptosis.¹¹³ The effect of BMPs on osteoblast differentiation is inhibited by TGF-β, and transgenic mice overproducing TGF-β are, in fact, osteopenic.^114,115 Like BMPs, TGF-β signals through distinct serine-threonine kinase receptors that activate specific Smads to regulate transcription.^110,111

Insulin-like growth factors (IGFs) are produced in a variety of tissues and have been characterized as IGF-I and IGF-II, both sharing similar properties but the former being more potent. IGFs¹¹⁶ stimulate proliferation of the osteoblastic lineage, prevent apoptosis of mature osteoblasts, and enhance net type I collagen abundance by stimulation of transcription and inhibition of collagen protein degradation. IGF-I knockout mice have decreased bone formation while those overexpressing IGF-I have enhanced bone formation.^117,118 In humans, IGF-I has a generalized anabolic effect on musculoskeletal tissue with stimulation of bone remodeling.¹¹⁹ Parathyroid hormone (PTH) induces IGF-I synthesis in osteoblasts, and the anabolic effects of PTH on bone is diminished in the absence of IGF-I.¹¹⁷ Steroids decrease IGF-I expression and this may, in part, explain glucocorticoid-induced bone loss.¹²⁰

β-catenin in association with specific transcription factors can regulate BMP and TGF-β signaling.¹²¹ The so-called Wnts prevent the degradation of β-catenin. Low-density lipoprotein receptor-related protein 5 (LPR5) is a coreceptor for Wnt that is required for optimal signaling with resultant accumulation of β-catenin.¹²² BMP-stimulated osteoblasts and stromal cells express LRP5 and knockout mice lacking LRP5 are osteopenic.¹²³ In humans, inactivating mutations in LRP5 cause decreased bone mass, while other mutations in LRP5 cause a high bone mass phenotype.^124,125

Fibroblast growth factor (FGF) 1 and 2 stimulate osteoblast proliferation and in vivo are important in the maintenance of bone mass. FGF null mice display low numbers of osteoblasts and decreased bone formation.¹²⁶ Conversely, systemic FGF2 increases the preosteoblast pool that ultimately forms mature bone.¹²⁷ FGF receptor signaling is essential for chondrogenesis and normal skeletal and limb development, with an activating mutation in FGF receptor-3 causing achondroplasia associated with dwarfism and early closure of cranial sutures.¹²⁸ Platelet-derived growth factor (PDGF) has activities similar to those of FGF.¹²⁹

Synthesis of bony matrix

The extracellular matrix of bone is composed predominantly of type I collagen, noncollagenous proteins, and mineral in the form of hydroxyapatite. Osteoid also contains serum-derived proteins, including albumin and α₂-HS-glycoprotein. Bone-forming cells secrete proteoglycans, glycosylated proteins with and without cell attachment function, and γ-carboxylated (gla) proteins.

In early bone formation, a chondroitin sulfate proteoglycan (versican) and the glycosaminoglycan hyaluronan are secreted and may delineate areas of future bone deposition.¹³⁰ Versican is eventually replaced by two chondroitin sulfate proteoglycans, decorin and biglycan, each containing leucine-rich tandem repeat sequences. One possible function for these leucine-rich repeat proteins is their ability to bind and regulate TGF-β family members in the extracellular milieu.¹³⁰ Biglycan-deficient transgenic mice have underdeveloped trabecular bone.¹³¹

The most abundant noncollagenous glycoprotein secreted by bone cells is osteonectin, with putative functions in osteoblast proliferation and matrix mineralization. Other glycoproteins mediate cell attachment and have been designated members of the SIBLING family (e.g. fibronectin, osteopontin, bone sialoprotein).¹³² All members contain the cell consensus sequence RGD (Arg-Gly-Asn) that binds to integrin cell surface molecules. Although most members of the SIBLING family are found in other tissues, bone sialoprotein is specific to mineralized tissue.¹³³

The bone-specific gla-containing protein osteocalcin may function as an inhibitor of mineral deposition, as judged by osteocalcin-deficient mice who have increased bone mineral density.¹³⁴ Similarly, mice deficient in the matrix gla protein (MGP), found in many connective tissues, develop extracellular sites of calcification.¹³⁵

Bone mineral contains numerous impurities and is a carbonate-substituted hydroxyapatite with solubility properties that enable its constituent imperfect crystals to serve as a reservoir for calcium, phosphate, and magnesium ions.¹³⁵ While the mineral content provides mechanical rigidity and load-bearing strength, the organic matrix provides elasticity, flexibility, and microstructural organization largely secondary to the presence of type I collagen.¹³⁵

Extracellular matrix vesicles released from chondrocytes and osteoblasts accumulate calcium and phosphate ions, enzymes that degrade inhibitors of mineralization, and components of a nucleation core that induce apatite formation.¹³³ It is unclear if there is an association of vesicle mineral with mineral in the collagen matrix or if the matrix vesicle directly provides a calcium and phosphate reservoir that supports mineralization initiated in collagen fibrils. In either case, the first stable crystal formed is followed by the addition of ions and ion clusters that allows for the expansion of crystal structure with addition of secondary nucleation sites.¹³⁶ Removal of noncollagenous protein from matrix, particularly phosphoprotein, inhibits crystal nucleation and apatite formation.¹³⁷ Alkaline phosphatase may contribute to enhanced mineralization by increasing the local phosphate concentration, removing phosphate-containing inhibitors of apatite growth, or modifying phosphoproteins to control their function as potential nucleators.¹³³

Growth of mineral crystals is primarily dictated by the collagen matrix template. Noncollagenous proteins that bind crystals can also influence the extent and direction of mineral deposition.¹³³ The presence of ions other than calcium and phosphate may influence bone apatite crystal growth, size, and solubility, and they include magnesium, strontium, cadmium, carbonate, and fluoride.¹³³

Osteoclast differentiation

Cells of the mononuclear/phagocytic lineage give rise to osteoclasts. Early precursors are committed to the myeloid series by the action of transcription factors PU-1 and MiTf. The lineage is further narrowed to monocytic cells by macrophage-colony stimulating factor (M-CSF).^2,138 Proliferation and survival of monocytic cells, with concomitant RANK receptor expression, is mediated by M-CSF. The presence of RANKL is then required for commitment to the osteoclast differentiation program. Osteoclastogenesis depends on the association of M-CSF and RANKL, found on stromal cells and osteoblasts, with their receptors on monocytes/macrophage cells.^2,138 Subsequent fusions of preosteoclasts result in the formation of giant multinucleated cells containing 4–20 nuclei and expressing characteristic markers of mature osteoclasts.

The osteoclast membrane has lost several macrophage markers and is lacking Fc and C₃ receptors, but retains phagocytic non-specific esterases, lysozyme synthetic function, and CSF-1 receptors.^2,138 Osteoclasts are distinguishable from macrophages by the abundance of RANK, vitronectin (integrin α_vβ3) receptors, and calcitonin receptors. Osteoclasts undergo an estrogen-promoted apoptotic cell death after a cycle of resorption.^2,138

Regulation of osteoclast activity

Resorption by osteoclasts may be stimulated by local or systemic factors that expand the pool of committed osteoclastic precursors or activate mature osteoclasts. Locally produced cytokines and hormones may exert more profound effects on normal bone resorption and remodeling than systemic hormones.

The key regulators of osteoclastic bone resorption are RANK ligand and its two confirmed receptors, RANK and osteoprotegerin (OPG). OPG is a member of the TNF receptor superfamily that lacks a transmembrane domain and, as such, is a secreted molecule. OPG regulates osteoclast activity by serving as a decoy receptor for RANKL.^138–141 The only other known ligand for OPG is TRAIL which, like RANKL, is also a type II membrane-bound TNF homolog. The significance of OPG–TRAIL interaction is unknown. Deficiency of OPG in mice causes increased osteoclast formation, activity, and survival that results in destruction of growth plates, loss of trabeculae, decreased bone strength, and reduced bone mineral density (BMD).^142,143 Parenteral administration of OPG causes a decrease in the number of active osteoclasts with a concomitant increase in BMD.¹⁴⁴ In vitro, OPG abrogates osteoclast formation mediated by M-CSF and RANKL, in the absence of osteoblasts or stromal cells.^139,140 OPG also inhibits osteoclastogenesis induced by a diverse group of osteotropic hormones.^138–141 Finally, there is a direct effect of OPG on isolated mature osteoclasts due, in part, to limiting osteoclast survival.

RANKL expression is obligatory for osteoclastic resorption and physiologic bone remodeling (Fig. 38.2). Membrane-bound RANKL is more potent than soluble forms and mature osteoblasts express RANKL constitutively.^145–147 Knockout mice lacking RANKL are devoid of osteoclasts and exhibit osteopetrosis with occlusion of the marrow cavity.¹⁴⁸ The bone-resorbing activity and survival of mature osteoclasts is greatly increased by exposure to recombinant RANKL in the absence of stromal cells or osteoblasts.^138–141

Fig. 38.2 Basic bone remodeling. Osteoclastogenesis depends on the association of macrophage colony-stimulating factor (M-CSF) and the receptor for activation of nuclear factor kappa B (RANK) ligand (RANKL), found on stromal cells and osteoblasts, with their receptors on monocytes/macrophage cells. This process is inhibited by osteoprotegerin (OPG). The differentiated osteoclast polarizes on the bone surface, forming a ruffled membrane that acidifies the extracellular microenvironment, mobilizes the mineral phase of bone and provides the milieu for organic matrix degradation. Bony dissolution releases various hormones and growth factors including members of the bone morphogenetic protein (BMP), transforming growth factor-β (TGF-β), insulin-like growth factor (IGF), and fibroblast growth factor (FGF) families. In a process that couples bone formation to resorption, these signaling molecules stimulate osteoblast differentiation and proliferation.

The signaling receptor for RANKL is RANK, a transmembrane protein that is sufficient to induce, upon RANKL binding, all downstream signals for osteoblast differentiation as well as later activation of mature osteoclasts. Mutations in the RANK gene that cause constitutive activation have been detected in familial expansile osteolysis, resulting in osteolytic lesions and osteopenia.^149,150 Engineered mice lacking RANK have severe osteopetrosis secondary to the absence of osteoclasts.^151,152

M-CSF or CSF-1 is required for normal osteoclast formation in the neonatal period and in rodent models of osteopetrosis there is impaired CSF-1 production. CSF-1 is secreted by stromal cells and cells of the osteoclast lineage express CSF-1 receptors. Interleukin (IL)-1α and β are potent stimulators of osteoclasts that work at all stages of formation and activation whose effects are mediated through RANKL.^141,153 IL-1 has been implicated in conditions of increased bone turnover, including osteoporosis, bone loss seen in some malignancies, and in rheumatoid arthritis.¹⁵⁴ Lymphotoxin and TNF-α are functionally related to IL-1 and their effects on bone are synergistic with IL-1.¹⁵⁵

The list of other locally-acting factors is growing. Those that stimulate osteoclast formation and activity include IL-6, IL-15, IL-17, vitamin A, and TGF-α. Inhibitors of osteoclast formation and activity include interferon-γ (IFN-γ), TGF-β, IL-18, inorganic phosphate, and high extracellular calcium. In humans, exogenous TGF-β inhibits proliferation and differentiation of osteoclast precursors and enhances osteoclast apoptosis, presumably by increasing OPG expression in osteoblasts and bone marrow stromal cells. However, endogenous TGF-β may actually stimulate osteoclastogenesis.^156,157 The role of other factors such as prostaglandins and leukotrienes is less clear.

The major systemic hormones that influence osteoclast activity are PTH, vitamin D₃, calcitonin, glucocorticoids, and the sex hormones. PTH has dual effects on bone mediated through its regulation of Cbfa1, which is obligatory for bone formation and also influences RANKL expression on osteoblasts.¹⁵⁸ Intermittent administration of PTH stimulates bone formation while continuous administration stimulates osteoclastic bone resorption. PTH-related protein has identical skeletal effects to those of PTH. The D vitamin 1,25 (OH)₂D₃ stimulates osteoclastic bone resorption by stimulating differentiation and fusion of osteoclast progenitors.¹⁵⁹ These effects of 1,25 dihydroxyvitamin D are mediated indirectly by osteoblasts through RANKL signaling. Calcitonin, made by the parafollicular cells of the thyroid gland, is a potent inhibitor of osteoclastic bone resorption whose effects are transient owing to downregulation of its receptor.¹⁶⁰ Glucocorticoids have inhibitory effects on formation of osteoclasts in vitro and on bone resorption in organ culture, but in vivo are associated with increased bone resorption. The latter effect may, in part, be due to glucocorticoid inhibition of calcium absorption from the gut and secondary hyperparathyroidism. For up to 10 years after the menopause, lack of estrogen is associated with enhanced osteoclastic bone resorption. Estrogens may have a direct effect on osteoclasts, but they may also suppress peripheral blood monocyte production of cytokines such as IL-1, TNF-α, and IL-6 that stimulate osteoclast formation.^{154,161–163}

Coupling of bone formation and resorption

Remodeling of bone occurs in focal and discrete microscopic sites throughout the skeleton. Complete remodeling at each site takes 3–6 months, and is probably longer in cancellous than in cortical bone. Cancellous bone is more abundant in the vertebral column, comprising more than 66% of total bone in the lumbar spine. This is in contrast to the intertrochanteric portion of the femur, where bone is 50% cortical and 50% cancellous, and in the mid-radius where greater than 95% of bone is cortical bone. At sites of bone turnover an activation phase is followed by a resorption phase that is followed by a bone formation phase. Although anatomically different from each other, cortical and cancellous bone follow the same general remodeling sequence. However, turnover events in cancellous bone are probably influenced more by factors produced by adjacent bone marrow cells, whereas these events in cortical bone, distant from marrow-derived cytokines, are likely influenced more by systemic hormones such as PTH and 1,25 dihydroxyvitamin D₃.¹⁶⁴

In the initial step of the remodeling sequence, osteoclasts are activated in focal sites by mechanisms that are not well understood and the rate of bone turnover is dependent on the activation frequency of osteoclasts. Interactions between integrins on the osteoclast cell membrane and bony matrix proteins may initiate the activation phase. After a resorptive phase lasting up to 10 days, repair of resulting defects proceeds by attraction of osteoblasts and subsequent bone formation. The mechanisms by which bone formation is coupled to previous osteoclastic resorption are still unclear, but several possibilities can be supported by existing evidence.

Coupling may be mediated by local humoral factors acting on osteoblasts or osteoblast precursors (see Fig. 38.2). These mediators may serve as osteoblast-stimulating factors, such as IGF-I or TGB-β, that are released from the bone matrix with osteoclastic resorption.^165–167 A factor that stimulates resorption may also stimulate osteoblasts, but at a much slower rate, so as to time the initiation of bone formation to occur after osteoclast activity has ceased.¹⁶⁸ Coupling may instead involve transcription factors, such as Runx2/Cbfa1, that are both obligatory for bone formation but also play a role in the regulation of osteoclast formation.¹⁶⁹ Alternatively, systemic factors such as leptin may play a role in the regulation of the bone formation component of the coupling process by presumptive hypothalamus-mediated β₂-adrenergic stimulation of osteoblasts and their precursors.¹⁷⁰ Lastly, coupling may be a more simple process whereby osteoblasts or their precursors in proximity to resorption sites repopulate bony defects after osteoclasts leave or die after their last resorptive cycle.

Regulation of the resorption phase

The resorption phase begins with osteoclast activation and is followed by osteoclast formation, polarization, ruffled membrane demarcation of future resorbing sites, resorption, and apoptosis. Osteoclast programmed cell death may be precipitated by resorption inhibitors such as estrogen and TGF-β and often occurs at reversal sites, locations where osteoclastic resorption has ceased and preosteoblastic migration and differentiation are soon to proceed. The regulation of the resorption phase is under complex control as described above under Regulation of Osteoclast Activity.

Regulation of the formation phase

The formation phase of bone remodeling occurs through a series of events beginning with osteoblast precursor chemotaxis and proliferation. These events are followed by osteoblast differentiation, mineralized bone formation, and cessation of osteoblast activity. Chemotaxis of osteoblast precursors is likely mediated by soluble factors produced locally as well as factors released from the dissolution of bone by osteoclastic resorption. Likely candidates include TGF-β, PDGF, and digested fragments of type I collagen and osteocalcin.^171–176

Proliferation of osteoblast precursors is also likely to be stimulated by local factors released during the resorptive process including IGFs, FGFs, TGF-βs, and PDGF. Differentiation of osteoblastic precursors into mature bone-forming cells is likely to be induced by bone-derived growth factors such as IGF-I and BMP2. The role of TGF-β may thus be to trigger the process of bone formation by first attracting and then expanding the pool of osteogenic precursors, after which its disappearance or inactivation allows these precursors to differentiate into osteoblasts. The cessation of osteoblast activity may also be mediated by TGF-β, since TGF-β decreases osteoblast activity and it is expressed by osteoblasts as they become terminally differentiated.¹⁷⁷ Osteoblast life span and factors that regulate it (see below under Osteoblast Senescence) may be critically important to terminating the formation phase of bone turnover, particularly with aging.

Bone remodeling and disease

In diseases which tend to increase osteoclast activation, such as primary hyperparathyroidism, hyperthyroidism, and Paget’s disease, there is also a roughly balanced increase in the amount of new bone formation, although new bone is not necessarily architecturally equivalent to bone formed under normal physiologic conditions. In diseases that cause osteolytic lesions of bone, namely multiple myeloma and certain malignant solid tumors, osteoblasts do not completely repair defects made by osteoclastic resorption. In the former case there appears to be a defect(s) in osteoblast differentiation while in the later case malignant cells produce local factors that stimulate osteoclast activity.^178,179 Osteoblastic lesions occur at sites where no previous bone resorption has occurred, and can be found with cancers of the prostate or breast, or with administration of pharmacologic doses of fluoride.

With aging, there appears to be an uncoupling of bone resorption and formation that begins after about 35 years of age, depending on the skeletal site. There is a resulting decrease in trabecular wall thickness owing to the inability of osteoblasts to repair osteoclastic resorptive defects, presumably secondary to decreased osteoblast activity or a decrease in the number of osteoblasts available to initiate bone formation.¹⁸⁰

Age-related changes in biomechanical properties

The strength of cancellous bone declines 8–10% per decade in comparison to cortical bone strength which declines only 2–5% per decade. This translates into a much higher risk of fracture in bones of the vertebral column. Inherent fragility of bone increases with age, allowing damage to accumulate more readily with repeated loading.¹⁸¹ Damaged bone is less able to sustain further deformation and less able to absorb energy with any impact, predisposing it to failure.¹⁸² Fracture risk in older adults is greater than predicted by loss of bone mass alone, suggesting that age-related changes in bone quality or architecture are important.

Ultramicroscopic tears in bone or fatigue damage results in fracture whenever damage occurs faster than the remodeling apparatus can repair it (slower bone turnover) or when remodeling is defective (uncoupling of bone formation to resorption).¹⁸³ Bones vertically loaded, such as the vertebral bodies, depend on horizontal, cross-hatching trabeculae, for support. Loss of trabecular connectivity is thought to be a major contributor to the female preponderance of vertebral bone loss and susceptibility to vertebral fractures.

The decline in collagen content with age is associated with a reduction in bone strength and stiffness, two parameters generally used to define bone health. Fewer reducible collagen cross-links per unit of total collagen are associated with increase bone fragility in osteoporotic women.^184,185 Periosteal apposition of bone with age serves as a compensatory mechanism for reduced bone volume in men, but not in women, at both the femoral neck and for the vertebral bodies.^186–188

Bone loss: secondary hyperparathyroidism, menopause, and somatopause

The uncoupling of bone formation to resorption leads to remodeling imbalances that ultimately result in age-related bone loss. With aging, bone formation is affected by the reduction of osteoblast differentiation, activity, and life span that is potentiated by secondary hyperparathyroidism and, in women, by estrogen deprivation and increased osteoclast activity with menopause.¹⁸⁹ Secondary hyperparathyroidism is caused by decreased intestinal and renal calcium resorption, precipitated by vitamin D deficiency in the housebound (lack of sunlight) and undernourished older adult, and by age-onset changes in renal function. Both processes are affected by estrogen deficiency. However, secondary hyperparathyroidism is ameliorated by a diet high in calcium.

Impaired osteoblast function is thought to be a major contributing factor to remodeling imbalances that occur with age. Decline in growth hormone production, decrease in physical activity, and osteoblast senescence (see below) impart age-related deleterious effects on osteoblast function. Some of these effects may be mediated by low levels of circulating and/or local IGF-I. Changes in the levels of other locally acting growth factors and cytokines that also impair osteoblast function may be mediated by estrogen withdrawal. The consequences of impaired osteoblast function are a relative increase in osteoclast activity and uncompensated bone resorption made worse by the increase in activation frequency (of remodeling events) with estrogen loss. Random remodeling errors that accumulate over time may also play a role, albeit small, in the fidelity of aging bone.

Osteoblast senescence

Telomeres shorten with age in most human tissues, including bone, and because telomere shortening is a cause of cellular replicative senescence in cultured cells, including osteoblasts and MSCs,^190–192 it is hypothesized that telomere shortening contributes to the aging of bone. Conversely, after forced ectopic expression of telomerase in human MSCs, proliferative capacity is extended in vitro and the capacity for bone formation is enhanced in vivo.^190–192 Telomerase not only extends the life span of osteogenic precursors, but accelerates the osteogenic differentiation of MSCs.¹⁹³ These observations provide strong evidence that telomerase status in MSCs is likely a critical component of bone formation.

Functional deficits in osteoblasts that occur with cellular senescence play a major role in the uncoupling of bone formation and resorption, resulting in a net loss of bone tissue.¹⁹⁴ Thus, recruitment of osteoblast precursors (MSCs) and osteoblast differentiation become a critical component in maintaining the balance between these two opposing processes. If the number and activity of osteoblasts responsible for synthesizing new bone matrix are substantially reduced, then the diminished recruitment of osteogenic precursors to replace senescent osteoblasts may potentially explain many aspects of age-related bone loss. The osteogenic potential of murine MSCs have been reported to decline with increasing donor age.¹⁹⁵ There are conflicting reports regarding the effects of age on human MSCs with the weight of evidence in favor of modest declines in several measures associated with osteogenic potential, particularly after the age of 40.^196–203 However, there are at least two well-conceived studies that suggest no significant age-related differences.^201,202 Thus, decrements in osteogenic stem cell potential with age may only partially account for changes in bone integrity associated with osteoporosis and poor fracture repair.

It has also been suggested that reductions in bone density seen with age may be the result of osteoblast replicative senescence, either directly (decreased proliferation and apoptosis) or secondary to other changes that accompany replicative senescence (e.g. failure to express an osteoblast phenotype).²⁰⁴ It is unclear if these alterations are solely a consequence of decreased osteoblast responsiveness to extracellular signals, but reduced expression of osteoblast markers such as alkaline phosphatase, osteocalcin, and type I collagen has been shown in response to various hormones and growth factors including 1,25 (OH)₂ vitamin D₃ [1,25 (OH)₂ D₃], insulin-like growth factor-I (IGF-I), parathyroid hormone (PTH), and prostaglandin E₂ (PGE₂).^205–212 Many of these changes have been documented in primary osteoblast cultures from old donors as well as in osteoblasts aged in vitro by serial passage.

If the osteogenic potential of bone marrow stromal cells is sufficiently intact, but the number and activity of osteoblasts responsible for synthesizing new bone matrix are substantially reduced, then the diminished recruitment of osteogenic precursors to replace senescent osteoblasts may explain many aspects of age-related bone loss.