Functional anatomy of the musculoskeletal system

Extracellular matrix

The matrix is mostly comprised of collagen (and, in some cases, elastic) fibres, embedded in a highly hydrated proteoglycan gel ( Fig. 5.1 ). Large proteoglycan molecules have numerous side chains of glycosaminoglycans (GAGs), carbohydrates with remarkable water-binding properties. A preponderance of fixed negative charges on the surface of GAGs strongly attract polarized water molecules, causing wet cartilage to swell until restricted by tension in the collagen network, or by external loading. In this way, cartilage develops a compressive turgor that enables it to distribute loading evenly on to subchondral bone, rather like a water bed. Effectively, water is held in place by proteoglycans, which are themselves held in place by the collagen network. Cartilage proteoglycans are similar to those found in general, i.e. non-specialized, connective tissue. The most common GAG side chains in cartilage are chondroitin sulphate and keratan sulphate. The most common proteoglycan molecule, aggrecan, forms huge molecular aggregates with other proteoglycans and with hyaluronan (see Fig. 5.1 ). Other constituents of cartilage include dissolved salts, non-collagenous proteins and glycoproteins.

Collagens are described on p. 39 . The collagen network varies in different types of cartilage and with age. Collagen type I forms large fibres with a wavy ‘crimped’ structure: it is common in most fibrous tissues, but in cartilage it occurs only in the outer layers of the perichondrium, and in white fibrocartilage. More typical of cartilage is collagen type II, which forms very thin fibrils dispersed between the proteoglycan molecules so that they do not clump together to form larger fibres. Collagen type II fibrils are often less than 50 nm in diameter and are too small to be seen by light microscopy. Transmission electron microscopy reveals that they have a characteristic cross-banding (65 nm periodicity) and are interwoven to create a three-dimensional meshwork. The length of collagen fibres and fibrils in cartilage is unknown, but even relatively short fibrils can reinforce the matrix by interacting physically and chemically with each other, and with other matrix constituents including proteoglycans ( ), reflecting the fact that the word collagen means ‘glue-maker’. Collagen type II is found in hyaline cartilage, the notochord, the nucleus pulposus of an intervertebral disc, the vitreous body of the eye, and the primary corneal stroma.

Cartilage cells

The cells of cartilage are chondroblasts and chondrocytes. Chondroblasts are actively dividing cells, often flattened and irregular in shape, and are abundant in growing tissue where they synthesize the extracellular matrix ( Fig. 5.2 ). Small projections arising from the cell membrane ( Fig. 5.3 ) can form gap junctions with adjacent cells ( ), but these junctions may be lost when interstitial growth causes greater cell separation. As chondroblasts mature and lose the ability to divide, they develop into the larger but metabolically less active chondrocytes. These oval-shaped cells form sparse populations that maintain the extensive matrix of adult cartilage. The name ‘chondrocyte’ is commonly employed, as it is here, to denote all of the cartilage cells embedded in an extensive matrix. Chondrocytes are normally in close contact with their dense matrix (see Fig. 5.3 ); however, artefacts of tissue processing can sometimes give the illusion of an empty space or ‘lacuna’ surrounding each cell or group of cells in histological sections. One or more chondrocytes can form a chondron, which consists of the cells and their pericellular matrix ( Fig. 5.2B ), surrounded by a protective basket of collagen ( , ).

Cartilage cells synthesize and secrete all of the major components of their matrix, and their ultrastructure is typical of cells that are active in making and secreting proteins. The nucleus is round or oval, appears euchromatic and possesses one or more nucleoli. The cytoplasm is filled with rough endoplasmic reticulum, transport vesicles and Golgi complexes, and contains many mitochondria and frequent lysosomes, together with numerous glycogen granules, intermediate filaments (vimentin) and pigment granules. When these cells mature to the relatively inactive chondrocyte stage, the nucleus becomes heterochromatic, the nucleolus smaller and the protein synthetic machinery much reduced; the cells may also accumulate large lipid droplets.

Collagen is synthesized within the rough endoplasmic reticulum in the same way as it is in fibroblasts. Polypeptide chains are assembled into triple helices, and some carbohydrate is added. After transport to the Golgi apparatus, where further glycosylation occurs, the resulting procollagen molecules are secreted into the extracellular space. Terminal registration peptides are cleaved from their ends, forming tropocollagen molecules, and the final assembly into collagen fibrils takes place. Core proteins of the proteoglycan complexes are also synthesized in the rough endoplasmic reticulum and addition of GAG chains begins; the process is completed in the Golgi complex. Hyaluronan, which lacks a protein core, is synthesized by enzymes on the surface of the chondrocyte; it is not modified post-synthetically, and is extruded directly into the matrix without passing through the endoplasmic reticulum.

Matrix turnover is much slower in cartilage than in more metabolically active tissues. Collagen turnover is particularly slow, leaving it vulnerable to the slow process of non-enzymatic glycation, which makes the tissue yellow, stiff and vulnerable to injury ( ). Proteoglycans are turned over faster than collagen, with an estimated turnover time of 5 years for adult humans.

Cartilage is often described as avascular. Certainly, the ability of the matrix to deform under load makes it difficult for hollow blood vessels to persist in the tissue beyond early childhood, but a limited vascular supply is often found on the cartilage surface, from where it can revascularize the tissue following injury or degeneration ( ). Metabolite transport to cartilage cells is mostly by the process of diffusion down a concentration gradient from the cartilage surface, although fluid ‘pumping’ as a result of changing mechanical loading can contribute under certain circumstances. Metabolite transport severely limits cell density and metabolic rate in adult tissue, and this in turn restricts cartilage thickness to a few millimetres ( ). Cartilage cells situated further than this from a nutrient vessel do not survive, and their surrounding matrix typically becomes calcified. In the larger cartilages, and during the rapid growth of some fetal cartilages, vascular cartilage canals penetrate the tissue at intervals, providing an additional source of nutrients. In some cases these canals are temporary structures but others persist throughout life.

Hyaline cartilage

Hyaline (glassy) cartilage has a homogeneous, opalescent appearance, sometimes appearing bluish. It is firm and smooth to the touch and shows considerable deformability. The size, shape and arrangement of chondrocytes vary at different sites and with age. They are flat near the surface perichondrium, and rounded or angular deeper in the tissue ( Fig. 5.2A ). Groups of two or more cells frequently form a cell nest (isogenous cell group) surrounded by a basket of fine collagen fibrils (see Fig. 5.2B ). Within such a ‘chondron’, daughter cells of a common chondroblast often meet at a straight line. The pericellular matrix closest to the cells is typically lacking in collagen fibrils, but rich in proteoglycans that can exhibit basophilic and metachromatic staining. More distant and older interterritorial matrix appears paler and is mostly collagen type II (75% of dry weight) and proteoglycans (22%).

After adolescence, hyaline cartilage may become calcified, either as part of the normal process of bone development, or as an age-related, degenerative change. In costal cartilage, the matrix tends to fibrous striation, especially in old age when cellularity diminishes. The xiphoid process and the cartilages of the nose, larynx and trachea (excepting the elastic cartilaginous epiglottis and corniculate cartilages) resemble costal cartilage in microstructure. Hyaline cartilage is the prototypical form, but it varies more with age and location than either elastic or fibrocartilage. Its regenerative capacity following injury is poor.

Articular cartilage, which covers the articular surfaces of synovial joints, is a specialized hyaline cartilage that lacks a perichondrium ( Fig. 5.4 ). The synovial membrane overlaps and then merges into its structure circumferentially (see Fig. 5.32 ). The thickness of articular cartilage varies from 1 to 7 mm (typically 2 mm) in different joints, and decreases from middle to old age. Thickness does not increase in response to increased mechanical loading, at least in adults, although matrix composition and stiffness can adapt somewhat ( ). Central regions tend to be thickest on convex osseous surfaces, and thinnest on concave surfaces.

Articular cartilage provides an extremely smooth, firm yet deformable layer that increases the contact area between bones and thereby reduces contact stress (see Fig. 5.61 ). Microscopic undulations on the cartilage surface help to trap synovial fluid (see Fig. 5.60 ) and enable fluid-film lubrication to reduce friction and wear. Articular cartilage is generally too thin and stiff to be a good shock absorber, although shock absorption may be significant where there are multiple cartilage-covered surfaces, as in the carpus and tarsus.

Adult articular cartilage shows a structural zonation defined by its network of fine collagen type II fibrils ( Fig. 5.5 ). Embedded in the deep calcified zone, fibrils rise vertically through the radial zone towards the cartilage surface, where they appear to reorientate to run parallel to the surface in the tangential zone. These collagen arcades can be visualized using phase-contrast microscopy ( ) but it is likely that each arcade represents numerous discrete fibrils rather than a single fibre. Their three-dimensional orientation can be appreciated by repeatedly piercing the cartilage surface with a needle; this creates a series of permanent elongated splits in the surface, which can be stained by Indian ink. The resulting split line pattern ( ) reveals the predominant directions of collagen fibrils in the tangential zone, which may be related to internal lines of tension generated during joint movement.

Each zone of articular cartilage (see Fig. 5.5 ) has a distinct cell morphology and matrix composition. The tangential (or superficial) zone has relatively small, elongated cells orientated parallel to the surface. (Properties of the articular surface are described below under ‘Synovial joints’, p. 102 ). Deeper within the tangential zone, the collagen fibrils increase in diameter and density, and gradually merge with those of the transitional (or intermediate) zone. Here, the chondrocytes are large and rounded, and surrounded by collagen fibrils in a range of oblique orientations. Deeper still, in the radial zone, the cells are often disposed in vertical columns, interspersed with vertical collagen fibrils. The matrix in this zone contains collagen types IX and XI, as well as type II. An undulating band known as the tidemark indicates the start of the deepest zone, the zone of calcified cartilage, which has mechanical properties intermediate between cartilage and bone. This calcified zone is keyed into the subchondral bone by fine ridges and interdigitations, which serve to prevent shearing (gliding) movements between cartilage and bone. With age, articular cartilage thins by upward advancement of the tidemark, and gradual replacement of calcified cartilage by bone.

Cells of articular cartilage are capable of cell division, but mitosis is rarely observed in adult tissue and cartilage damage is not repaired. Superficial cells are lost progressively from normal young joint surfaces, to be replaced by cells from deeper layers. Age-related reductions in cell number and activity, and biochemical changes in the extracellular matrix, particularly affect the superficial zone of articular cartilage, increasing the risk of mechanical failure and of osteoarthritis ( ).

Articular cartilage derives nutrients by diffusion from vessels of the synovial membrane, synovial fluid and hypochondral vessels of an adjacent medullary cavity, some capillaries from which penetrate and occasionally traverse the calcified cartilage zone. The contributions from these sources are uncertain and may change with age. Small molecules freely traverse articular cartilage, with diffusion coefficients about half those in aqueous solution. Larger molecules have diffusion coefficients inversely related to their molecular size. The permeability of cartilage to large molecules is greatly affected by variations in its GAG (and hence water) content: a three-fold increase in GAGs increases the diffusion coefficient 100-fold.

Cartilaginous growth plates (see below under ‘Bone’, pp. 96–99 ) are also composed of hyaline cartilage, and there are similarities between active growth plates and growing articular cartilage on the epiphyses of long bones. In both cases, chondrocytes undergo a sequence of cell divisions and hypertrophy (with cells forming into columns) followed by cell death, and ossification by invading osteoblasts.

Fibrocartilage

Fibrocartilage is a dense, whitish tissue with a distinct fibrous texture. It forms the intervertebral discs of the spine and menisci of the knee, as well as smaller structures such as the glenoid and acetabular labra, and the lining of bony grooves for tendons. It is a versatile and tough material that combines considerable tensile strength with the ability to resist high compressive forces and to distribute them evenly on to underlying bone ( ). Histologically, fibrocartilage is intermediate between dense fibrous connective tissues such as tendon and ligaments, and hyaline cartilage. In some structures such as intervertebral discs, matrix composition and cell types vary from one location to another, reflecting varying mechanical properties.

Regions of fibrocartilage that are loaded predominantly in tension consist of large crimped fibres of collagen type I embedded in a hydrated proteoglycan gel. Cells are rounded in young tissue ( Fig. 5.6 ), but become elongated and fibroblast-like with age. They may be linked by gap junctions ( ). Regions of fibrocartilage that are loaded predominantly in compression appear more homogeneous, contain a high proportion of fine collagen type II fibrils in an abundant proteoglycan gel, and contain rounded, chondrocyte-like cells. Fibrocartilage could therefore be regarded as a mingling of two types of tissue rather than a separate type of cartilage. However, no other tissue combines high proportions of proteoglycans with collagen type I, suggesting that fibrocartilage should be regarded as a distinct class of connective tissue.

The articular surfaces of bones that ossify in mesenchymal membranes (e.g. squamous temporal, mandible and clavicle) are covered by white fibrocartilage. The deep layers, adjacent to hypochondral bone, resemble calcified regions of the radial zone of hyaline articular cartilage. The superficial zone contains dense parallel bundles of thick collagen fibres, interspersed with typical dense connective tissue fibroblasts and little ground substance. Fibre bundles in adjacent layers alternate in direction, as they do in the cornea. A transitional zone of irregular bundles of coarse collagen and active fibroblasts separates the superficial and deep layers. The fibroblasts are probably involved in elaboration of proteoglycans and collagen, and may also constitute a germinal zone for deeper cartilage. Fibre diameters and types may differ at different sites according to the functional load.

Elastic cartilage

Elastic cartilage occurs in the external ear, corniculate cartilages, epiglottis and apices of the arytenoids. Like hyaline cartilage, it contains typical chondrocytes, either singly or in small groups, surrounded by a matrix rich in type II collagen fibrils. However, the more distant interterritorial matrix is pervaded by very fine yellow elastic fibres ( Fig. 5.7 ) containing the protein elastin, which show no periodic banding structure under the electron microscope (as collagen fibrils do). A structure is termed ‘elastic’ if it returns to its original shape when loaded and then unloaded; elastic fibres (and cartilage) have the special property of being able to do this even after being subjected to deformations greater than 15%, which would damage collagen fibres. This characteristic is termed elastic recoil. Most sites in which elastic cartilage occurs have vibrational functions, such as laryngeal sound-wave production, or the collection and transmission of sound waves in the ear. Elastic cartilage is resistant to degeneration, and its capacity for limited regeneration following traumatic injury can be appreciated from the distorted repair of a cauliflower ear, as seen in participants of some contact sports.

Bone organic matrix

Approximately 10–20% of bone mass is water. A significant proportion (30–40%) of the remaining dry weight is made up of the organic component of the extracellular matrix. Approximately 30% of this organic matrix is collagen; the remainder includes various non-collagenous proteins, glycoproteins and carbohydrates. The proportions of these components vary with age, location and metabolic status.

Most of the collagen in bone is an ordered branching network of type I fibres ( Fig. 5.10 ). Although type I collagen fibres are found in most connective tissues, their molecular structure in bone is atypical: internal cross-linking between component fibrils is stronger and chemically more inert, and transverse spacings between collagen molecules within each fibril are larger, allowing more space for the deposition of minerals. A small amount of type V collagen is also present, probably to help regulate fibrillogenesis. Collagen fibres contribute greatly to the cohesive mechanical strength of bone, and also to its toughness (which is reflected in the energy required to break a bone).

Collagen is synthesized in bone by osteoblasts. Newly secreted molecules of tropocollagen lose part of their non-helical terminal regions, thus allowing them to polymerize in the extracellular matrix to form fibrils, which then associate to form fibres. These structures are stabilized by various cross-links, which increase in number and strength as the tissue matures. In primary bone, collagen fibres form a complex interwoven meshwork that incorporates other organic molecules; this ‘osteoid’ material is then mineralized to form woven (non-lamellar) bone. In time, primary bone is almost entirely replaced by regular laminar arrays of nearly parallel collagen fibres, which form the basis of lamellar bone ( ). Partially mineralized collagen networks can be seen within osteoid on the outer and internal surfaces of bone, and in the endosteal linings of vascular canals. Collagen fibres from the periosteum are incorporated in cortical bone (extrinsic fibres) and anchor this fibrocellular layer at its surface. Terminal collagen fibres of tendons and ligaments are incorporated deep into the matrix of cortical bone. They may be interrupted by new osteons during cortical drift (modelling) and turnover (remodelling), and remain as islands of interstitial lamellae or even trabeculae.

Bone organic matrix includes small amounts of various macromolecules attached to collagen fibres and surrounding bone crystals. They are secreted by osteoblasts and young osteocytes, and include osteonectin, osteocalcin, the bone proteoglycans biglycan and decorin, the bone sialoproteins osteopontin and thrombospondin, many growth factors including transforming growth factor beta (TGF-β), proteases and protease inhibitors, often in a latent form. The functions of some of these molecules are described with osteoblasts (see below).

Bone minerals

Approximately 60–70% of bone dry weight is made up of inorganic mineral salts in the form of microcrystalline hydroxyapatite (Ca ₁₀ (PO ₄ ) ₆ (OH) ₂ ). The microcrystals confer hardness and much of the rigidity of bone, and are the main reason why bone is easily seen on radiographs. (Bone must be 50% mineralized to be visible on radiographs produced with a standard X-ray unit.) Bone mineral also has an important carbonate content, and a lower Ca:P ratio than pure hydroxyapatite, together with a small amount of calcium phosphate. Bone crystals are extremely small (which gives them a high surface : volume ratio). They take the form of thin plates or leaf-like structures . The largest are 150 nm long × 80 nm wide × 5 nm thick, although most are half that size. Up to two-thirds of the mineral content of bone is thought to be located within collagen fibrils, where the crystals are packed closely together, with their long axes nearly parallel to the fibrils; crystal formation is probably initiated in the gaps between individual collagen molecules. Narrow spaces between the crystals contain water and organic macromolecules. The mineral substances of bone are mostly acid-soluble. If they are removed, using calcium chelators such as citrates or ethylene diamine tetra-acetic acid (EDTA), the bone retains its shape but becomes highly flexible.

The major ions in bone mineral include calcium, phosphate, hydroxyl and carbonate. Less numerous ions are citrate, magnesium, sodium, potassium, fluoride, chloride, iron, zinc, copper, aluminium, lead, strontium, silicon and boron, many of which are present only in trace quantities. Fluoride ions can substitute for hydroxyl ions, and carbonate can substitute for either hydroxyl or phosphate groups. ‘Group IIA cations’, such as radium, strontium and lead, all readily substitute for calcium and are therefore known as bone-seeking cations. Since they can be either radioactive or chemically toxic, their presence in bone, where they may be close to haemopoietic bone marrow, may cause illness and characteristic appearances on X-rays.

Mineralization of newly synthesized osteoid is a gradual process that slows over time; it typically reaches 70–80% in 3 weeks ( : data for rabbits). Immature woven bone mineralizes faster and so may be distinguished from adjacent lamellar bone by its higher degree of mineralization. In cortical bone, lamellae mostly take the form of cylindrical osteons (see Fig. 5.16 ). These structures mineralize from inside to out, so that the concentration of mineral is highest in the older, more peripheral, lamellae. Although new osteons are less mineralized than old ones, they may show one or more highly mineralized ‘arrest lines’ within their walls. Mineral distribution is most uniform in established, highly mineralized osteons. Overall, mineralization increases with age, even though bone mass decreases.

Osteoblasts

Osteoblasts are derived from osteoprogenitor (stem) cells of mesenchymal origin present in bone marrow and other connective tissues. They proliferate and differentiate into osteoblasts prior to bone formation, stimulated by bone morphogenetic proteins (BMPs). A layer of osteoblasts covers the forming surfaces of growing or remodelling bone ( Fig. 5.11 ). In relatively quiescent adult bone, they appear to be present mostly on endosteal rather than periosteal surfaces, but they also occur deep within compact bone wherever osteons are being remodelled. Osteoblasts are responsible for the synthesis, deposition and mineralization of the bone matrix, which they secrete. Once embedded in the matrix, they become osteocytes.

Osteoblasts are basophilic, roughly cuboidal mononuclear cells 15–30 μm across. They contain prominent bundles of actin, myosin and other cytoskeletal proteins associated with the maintenance of cell shape, attachment and motility. Their plasma membranes display many extensions, some of which contact neighbouring osteoblasts and embedded osteocytes at intercellular gap junctions. This arrangement facilitates coordination of the activities of groups of cells, e.g. in the formation of large domains of parallel collagen fibres.

Ultrastructurally, osteoblasts are typical protein-secreting cells. They synthesize and secrete collagens and a number of glycoproteins. Osteocalcin is required for bone mineralization, binds hydroxyapatite and calcium, and is used as a marker of new bone formation. Osteonectin is a phosphorylated glycoprotein that binds strongly to hydroxyapatite and collagen; it may play a role in initiating crystallization and may be a cell adhesion factor. RANKL is the cell surface ligand for RANK (receptor for activation of nuclear factor kappa B), an osteoclast progenitor receptor (see below). Osteoprotegerin is a soluble, high-affinity decoy ligand for RANKL, which restricts osteoclast differentiation. Biglycan and decorin are bone proteoglycans that attract water; decorin also binds the growth factor TGF-β. The bone sialoproteins, osteopontin and thrombospondin, mediate osteoclast adhesion to bone surfaces by binding to osteoclast integrins. In addition, osteoblasts secrete latent proteases and growth factors including BMPs and TGF-β (which is also secreted by osteoclasts and which may be a coupling factor for stimulating new bone formation at resorption sites).

Although extracellular fluid is generally supersaturated with respect to the basic calcium phosphates, mineralization does not occur in most tissues. In bone, osteoblasts secrete osteocalcin (which binds calcium at levels sufficient to concentrate the ion locally) and contain membrane-bound vesicles full of alkaline phosphatase (cleaves phosphate ions from various molecules to elevate concentrations locally) and pyrophosphatase (degrades inhibitory pyrophosphate in the extracellular fluid). The vesicles bud off from the osteoblast surface into newly formed osteoid, where they initiate hydroxyapatite crystal formation. Some alkaline phosphatase reaches the blood circulation, where it can be detected in conditions of rapid bone formation or turnover.

Osteoblasts also play a key role in the hormonal regulation of bone resorption. They express receptors for parathyroid hormone (PTH), 1,25-dihydroxy vitamin D ₃ (calcitriol) and other promoters of bone resorption. When activated, osteoblasts promote osteoclast differentiation via PTH-activated expression of cell surface RANKL, which binds to RANK on immature osteoclasts, establishes cell–cell contact and triggers contact-dependent osteoclast differentiation. In the presence of PTH, osteoblasts also downregulate secretion of osteoprotegerin, a soluble decoy ligand with higher affinity for RANKL. In conditions favouring bone deposition, secreted osteoprotegerin blocks RANKL binding to RANK, restricting the number of mature osteoclasts.

Bone-lining cells are flattened epithelial-like cells that cover the free surfaces of adult bone not undergoing active deposition or resorption. Generally considered to be quiescent osteoblasts or osteoprogenitor cells, they line the periosteal surface and the vascular canals within osteons, and form the outer boundary of the marrow tissue on the endosteal surface of marrow cavities.

Osteocytes

Osteocytes are the major cell type of mature bone and are distributed throughout its matrix, interconnected by numerous dendritic processes to form a complex cellular network ( Fig. 5.12 ). They are derived from osteoblasts that have become enclosed within their rigid matrix (see Fig. 5.11 ) and so have lost the ability to divide or to secrete new matrix. (The rigidity of mineralized bone matrix prevents interstitial growth, so that new bone must always be deposited on pre-existing surfaces.) Osteocytes retain contact with each other and with cells at the surfaces of bone (osteoblasts and bone-lining cells) throughout their lifespan.

Mature, relatively inactive osteocytes have an ellipsoid cell body with their longest axis (approximately 25 μm) parallel to the surrounding lamellae. The rather narrow rim of cytoplasm is faintly basophilic, contains relatively few organelles and surrounds an oval nucleus. Osteocytes in woven bone are larger and more irregular in shape ( Fig. 5.13 ). Numerous fine branching processes containing bundles of microfilaments and some smooth endoplasmic reticulum emerge from each cell body. At their distal tips, these processes form gap junctions with the processes of adjacent cells (osteocytes, osteoblasts and bone-lining cells) so that they are in electrical and metabolic continuity.

Extracellular fluid fills the small, variable spaces between osteocyte cell bodies and their rigid lacunae, which may be lined by a variable (0.2–2 μm) layer of unmineralized organic matrix. The same fluid fills the narrow channels or canaliculi that surround the long processes of the osteocytes. Approximately 0.25–0.5 μm wide, the canaliculi provide a route for the diffusion of nutrients, gases and waste products between osteocytes and blood vessels. Canaliculi do not usually extend through and beyond the reversal line surrounding each osteon and so do not communicate with neighbouring systems.

In well-vascularized bone, osteocytes are long-lived cells that actively maintain the bone matrix. The average lifespan of an osteocyte varies with the metabolic activity of the bone and the likelihood that it will be remodelled, but is measured in years. Old osteocytes may retract their processes from the canaliculi; when they die, their lacunae and canaliculi may become plugged with cell debris and minerals, which hinders diffusion through the bone. Dead osteocytes occur commonly in interstitial bone (between osteons) and in central regions of trabecular bone that escape surface remodelling. They are particularly noticeable by the second and third decades. Bones that experience little turnover, e.g. the auditory ossicles, are most likely to contain aged osteocytes and have low osteocyte viability. Osteocyte death leads to matrix resorption by osteoclast activity. Osteocytes themselves are often mineralized.

Osteoclasts

Osteoclasts are large (diameters of 40 μm or more) polymorphic cells containing up to 20 oval, closely packed nuclei (see Fig. 5.11 ). They lie in close contact with the bone surface in resorption bays (Howship's lacunae). Their cytoplasm contains numerous mitochondria and membrane-bound organelles, many of which are acid phosphatase-positive lysosomes. Rough endoplasmic reticulum is relatively sparse but the Golgi complex is extensive. The cytoplasm also contains numerous coated transport vesicles and microtubule arrays involved in vesicle transport between the Golgi stacks and the cell's ruffled border (the highly infolded region of plasma membrane of an active osteoclast at a site of bone resorption). A well-defined zone of actin filaments and associated proteins occurs beneath the ruffled border around the circumference of a resorption bay, in a region termed the sealing zone.

Functionally, osteoclasts are responsible for the local removal of bone during bone growth and remodelling ( Fig. 5.14 ). They dissolve bone minerals by proton release to create an acidic local environment, and they remove organic matrix by secreting lysosomal (cathepsin K) and non-lysosomal (e.g. collagenase) enzymes. Osteoclasts are stimulated to resorb bone by signals from local cells (including osteoblasts, macrophages and lymphocytes) and by blood-borne factors such as PTH and 1,25-dihydroxy vitamin D ₃ (calcitriol). Calcitonin, produced by C cells of the thyroid follicle, reduces osteoclast activity.

Osteoclasts differentiate from myeloid stem cells via macrophage-colony-forming units. Differentiation is primarily regulated by two cytokines: macrophage-colony stimulating factor, secreted by osteoblasts, and RANKL, expressed by osteoblasts (see above). The mononuclear precursors fuse to form terminally differentiated multinuclear osteoclasts ( ). Osteoclast differentiation inhibitors are potential therapeutic agents for bone loss-associated disorders, e.g. osteoporosis, rheumatoid arthritis, Paget's disease, periodontal disease and osteosarcoma.

Woven and lamellar bone

The mechanical properties of bone depend not only on matrix composition, as described above, but also on the manner in which the matrix constituents are organized. Woven bone and lamellar bone represent two quite distinct types of organization.

In woven (or bundle) bone, the collagen fibres and bone crystals are irregularly arranged. The diameters of the fibres vary, so that fine and coarse fibres intermingle, producing the appearance of the warp and weft of a woven fabric. Woven bone is typical of young fetal bones, but is also seen in adults during excessively rapid bone remodelling and during fracture repair ( Fig. 5.15 ). It is formed by highly active osteoblasts during development, and is stimulated in the adult by fracture, growth factors or prostaglandin E ₂ .

Lamellar bone, which makes up almost all of an adult skeleton, is more organized and is produced more slowly. The precise arrangement of lamellae (bone layers) varies from site to site. In trabeculae and the outer (periosteal) and inner (endosteal) surfaces of cortical bone, a few lamellae form continuous circumferential layers that are more or less parallel to the bony surfaces. However, in more central regions of cortical bone, the lamellae are arranged in concentric cylinders around neurovascular channels called Haversian canals ( Fig. 5.16 ). This interconnecting, three-dimensional, laminated construction increases the toughness of lamellar bone because the interfaces between lamellae are effective in stopping the growth of cracks; more energy is therefore required to propagate cracks that are sufficiently extensive to fracture the bone.

Each lamella consists of a sheet of mineralized matrix containing collagen fibres of similar orientation locally, running in branching bundles 2–3 μm thick and often extending the full width of a lamella. The orientation of collagen fibres and crystals differs between 0° and 90° in adjacent lamellae, as may be demonstrated by polarized light microscopy (see Fig. 5.19B ). At the borders of lamellae, packing of collagen fibres into bundles is less perfect and intermediate and random orientations of collagen predominate.

Cortical bone

The cylindrical structural units that comprise most cortical bone are termed Haversian systems or osteons ( Fig. 5.17 ). Osteons usually lie parallel with each other ( Fig. 5.18 ); in long bones, they lie parallel with the long axis of the bone. Adjacent osteons may encroach on one another because they are usually formed at different times, during successive periods of bone remodelling. Irregular gaps between osteons are filled with interstitial lamellae ( Fig. 5.17A ), which are the fragmentary remains of older osteons and circumferential lamellae. Osteons may be spiral or they may branch, and some end blindly. They are round or ellipsoidal in cross-section. The main direction of collagen fibres within osteons varies: in the shaft of long bones, fibres are more longitudinal at sites that are subjected mainly to tension, and more oblique at sites subjected mostly to compression. Peripheral lamellae of osteons contain more transverse fibres.

It has been estimated that there are 21 million osteons in a typical adult skeleton. Their diameter varies from 100 to 400 μm, and they usually contain 5–20 lamellae. Each osteon is permeated by the canaliculi of its resident osteocytes, which form pathways for the diffusion of metabolites between osteocytes and blood vessels. The maximum diameter of an osteon ensures that no osteocyte is more than 200 μm from a blood vessel, a distance that may be a limiting factor in their survival.

The central Haversian canals of osteons vary in size, with a mean diameter of 50 μm; those near the marrow cavity are somewhat larger. Each canal contains one or two capillaries lined by fenestrated endothelium and surrounded by a basal lamina, which also encloses typical pericytes. They usually contain a few unmyelinated and occasional myelinated axons. The bony surfaces of osteonic canals are perforated by the openings of osteocyte canaliculi and are lined by collagen fibres.

Haversian canals communicate with each other and directly or indirectly with the marrow cavity via vascular (nutrient) channels called Volkmann's canals, which run obliquely or at right-angles to the long axes of the osteons (see Fig. 5.18 ). The majority of these channels appear to branch and anastomose, but some join large vascular connections with vessels in the periosteum and the medullary cavity.

Osteons are distinguished from their neighbours by a cement line that contains little or no collagen, and is strongly basophilic because it has a high content of glycoproteins and proteoglycans. Cement lines are also known as reversal lines because they mark the limit of bone erosion prior to the formation of a new osteon. Canaliculi occasionally pass through cement lines, and so provide a route for exchange between interstitial bone lamellae and vascular channels within osteons. Basophilic resting lines can occur in the absence of erosion; they indicate where bony growth has been interrupted and then resumed.

Trabecular bone

The organization of trabecular bone (also known as cancellous or spongy bone) is basically lamellar, as shown most clearly under polarized light ( Fig. 5.19 ). Trabeculae take the form of branching bars and curved plates of varying width, length and thickness (50−400 μm) (see Fig. 5.9 ). They are covered in endosteal tissue because they are adjacent to marrow cavities. Thick trabeculae and regions close to compact bone may contain small osteons, but blood vessels do not otherwise lie within trabeculae; osteocytes therefore rely on canalicular diffusion from adjacent medullary vessels. In young bone, calcified cartilage may occur in the cores of trabeculae, but this is generally replaced by bone during subsequent remodelling.

Periosteum, endosteum and bone marrow

The outer surface of bone is covered by a condensed collagenous layer, the periosteum. The inner surface is lined by a thinner, more cellular endosteum. Osteoprogenitor cells, osteoblasts, osteoclasts and other cells important in the turnover and homeostasis of bone tissue lie in these layers.

The periosteum is tethered to underlying bone by thick collagen fibres (Sharpey's fibres), which penetrate deep into the outer cortical bone tissue. It is absent from articular surfaces, and from the points of insertion of tendons and ligaments (entheses) (see Fig. 5.51 ). The periosteum is highly active during fetal development, when it generates osteoblasts for the appositional growth of bone. These cells form a layer, 2–3 cells deep, between the fibrous periosteum and new woven bone matrix. Osteoprogenitor cells within the mature periosteum are indistinguishable morphologically from fibroblasts. Periosteum is important in the repair of fractures; where it is absent (e.g. within the joint capsule of the femoral neck) fractures are slow to heal.

Quiescent osteoblasts and osteoprogenitor cells act as the principal reservoir of new bone-forming cells for remodelling or repair on the endosteal surfaces of resting adult bone. Bone endosteum is likely to be important in calcium homeostasis because it provides a total surface area of approximately 7.5 m ² . It is formed by flattened osteoblast precursor cells and reticular (type III collagen) fibres, and lines all the internal cavities of bone, including the Haversian canals. It overlies the endosteal circumferential lamellae and encloses the medullary cavity.

Vascular supply

The osseous circulation supplies bone tissue, marrow, perichondrium, epiphysial cartilages in young bones, and, in part, articular cartilages. The vascular supply of a long bone depends on several points of inflow that feed complex and regionally variable sinusoidal networks within the bone. The sinusoids drain to venous channels that leave through all surfaces that are not covered by articular cartilage. The flow of blood through cortical bone in the shafts of long bones is mainly centrifugal ( Fig. 5.20 ).

One or two main diaphysial nutrient arteries enter the shaft obliquely through nutrient foramina, which lead into nutrient canals. Their sites of entry and angulation are almost constant and characteristically directed away from the dominant growing epiphysis. Nutrient arteries do not branch in their canals but divide into ascending and descending branches in the medullary cavity; these approach the epiphyses, dividing repeatedly into smaller helical branches close to the endosteal surface. The endosteal vessels are vulnerable during surgical operations, such as intramedullary nailing, which involve passing metal implants into the medullary canal. Near the epiphyses, diaphysial vessels are joined by terminal branches of numerous metaphysial and epiphysial arteries (see Fig. 5.20 ). The former are direct branches of neighbouring systemic vessels; the latter come from peri-articular vascular arcades formed on non-articular bone surfaces. Numerous vascular foramina penetrate bones near their ends, often at fairly specific sites; some are occupied by arteries but most contain thin-walled veins. Within bone, the arteries are unusual in consisting of endothelium with only a thin layer of supportive connective tissue. The epiphysial and metaphysial arterial supply is richer than the diaphysial supply.

Medullary arteries in the shaft give off centripetal branches, which feed a hexagonal mesh of medullary sinusoids that drain into a wide, thin-walled central venous sinus. They also possess cortical branches, which pass through endosteal canals to feed fenestrated capillaries in Haversian systems. The central sinus drains into veins that retrace the paths of nutrient arteries, sometimes piercing the shaft elsewhere as independent emissary veins. Cortical capillaries follow the pattern of Haversian canals, and are mainly longitudinal with oblique connections via Volkmann's canals (see Fig. 5.18 ). At bone surfaces, cortical capillaries make capillary and venous connections with periosteal plexuses (see Fig. 5.20 ) formed by arteries from neighbouring muscles that contribute vascular arcades with longitudinal links to the fibrous periosteum. A capillary network permeates the deeper, osteogenic periosteum from this external plexus. At muscular attachments, periosteal and muscular plexuses are confluent and the cortical capillaries then drain into interfascicular venules.

In addition to the centrifugal supply of cortical bone, there is an appreciable centripetal arterial flow to outer cortical zones from periosteal vessels. The large nutrient arteries of epiphyses form many intraosseous anastomoses, their branches passing towards the articular surfaces within the trabecular spaces of the bone. Near the articular cartilages, these form serial anastomotic arcades (e.g. there are three or four in the femoral head), which give off end-arterial loops. The latter often pierce the thin hypochondral compact bone to enter, and sometimes traverse, the calcified zone of articular cartilage, before returning to the epiphysial venous sinusoids.

In immature long bones, the supply is similar but the epiphysis is a discrete vascular zone. Epiphysial and metaphysial arteries enter on both sides of the growth cartilage and rarely, if ever, anastomose. Growth cartilages are probably supplied from both sources, and from an anastomotic collar in the adjoining periosteum. Occasionally, cartilage canals are incorporated into a growth plate. Metaphysial bone is nourished by terminal branches of metaphysial arteries and by primary nutrient arteries of the shaft, which form terminal blind-ended sprouts or sinusoidal loops in the zone of advancing ossification. Young periosteum is more vascular; its vessels communicate more freely with those of the shaft than their adult counterparts and give off more metaphysial branches.

Large, irregular bones such as the scapula and innominate not only receive a periosteal supply but are also often supplied by large nutrient arteries that penetrate directly into their cancellous bone, the two systems anastomosing freely. Short bones receive numerous fine vessels that supply their compact and cancellous bone and medullary cavities from the periosteum. Arteries enter vertebrae close to the base of their transverse processes (see Fig. 46.20 ). Each vertebral medullary cavity drains to two large basivertebral veins, which converge to a foramen on the posterior surface of the vertebral body (see Fig. 46.21 ). Flatter cranial bones are supplied by numerous periosteal or mucoperiosteal vessels. Large, thin-walled veins run tortuously in cancellous bone. Lymphatic vessels accompany periosteal plexuses but have not been convincingly demonstrated in bone. (For further reading on the role of the vasculature in bone development, function and regeneration, see ).

Innervation

Nerves are most numerous in periosteum, the articular extremities of long bones, vertebrae and the larger flat bones. The nutrient blood vessels are richly innervated by postganglionic sympathetic axons. Periosteum is innervated by myelinated axons with encapsulated nerve endings and by unmyelinated axons with free endings typical of nociceptors. Many of these afferents are mechanically sensitive and respond rapidly to mechanical distortion, some can be activated by multiple stimuli and may be polymodal, and some may be stretch receptors.

Fine myelinated (Aδ) and unmyelinated (C) axons with free fibre endings accompany nutrient vessels into bone and marrow, and lie in the perivascular spaces of Haversian canals. Animal studies have revealed that many of these sensory neurones have a nociceptive phenotype, e.g. they express substance P, alpha calcitonin gene-related peptide (αCGRP), tropomyosin receptor kinase A (TrkA), and transient receptor potential cation channel subfamily V member 1 (TRPV1) and respond to noxious chemical and mechanical stimuli such as raised intraosseous pressure ( ). Osteoblasts and osteoclasts possess receptors for several of the neuropeptides found in these nerves; there is increasing experimental evidence that neuropeptide-mediated signalling plays a role in modulating bone cell metabolism ( ). Pathological changes that cause the release of inflammatory mediators in bone marrow and/or raised intraosseous pressure or mechanical compression or distortion of bone and periosteum can all trigger bone pain by activating nociceptors in bone.

Intramembranous ossification

Intramembranous ossification is the direct formation of bone (membrane bone) within highly vascular sheets or ‘membranes’ of condensed primitive mesenchyme. At centres of ossification, mesenchymal stem cells differentiate into osteoprogenitor cells, which proliferate around the branches of a capillary network, forming incomplete layers of osteoblasts in contact with the primitive bone matrix. The cells are polarized, and secrete osteoid only from the surface that faces away from the blood vessels. The earliest crystals appear in association with extracellular matrix vesicles produced by the osteoblasts. Crystal formation subsequently extends into collagen fibrils in the surrounding matrix, producing an early labyrinth of woven bone, the primary spongiosa. As layers of calcifying matrix are added to the early trabeculae, osteoblasts become enclosed within primitive lacunae. These new osteocytes retain intercellular contact by means of their fine cytoplasmic processes (dendrites) and, as these elongate, matrix condenses around them to form canaliculi.

As matrix secretion and calcification proceed, trabeculae thicken and vascular spaces become narrower. Where bone remains trabecular, the process slows and the spaces between trabeculae become occupied by haemopoietic tissue. Where compact bone is forming, trabeculae continue to thicken and vascular spaces continue to narrow. Meanwhile, the collagen fibres of the matrix, secreted on the walls of the narrowing spaces between trabeculae, become organized as parallel, longitudinal or spiral bundles, and the cells they enclose occupy concentric sequential rows. These irregular, interconnected masses of compact bone each have a central canal and are called primary osteons (primary Haversian systems). They are later eroded, together with the intervening woven bone, and replaced by generations of mature (secondary) osteons.

While these changes are occurring, mesenchyme condenses on the outer surface to form a fibrovascular periosteum. Bone is laid down increasingly by new osteoblasts, which differentiate from osteoprogenitor cells in the deeper layers of the periosteum. Modelling of the growing bone is achieved by varying rates of resorption and deposition at different sites.

Endochondral ossification

The hyaline cartilage model that forms during embryogenesis is a miniature template of the bone that will subsequently develop. It becomes surrounded by a condensed, vascular mesenchyme or perichondrium, which resembles the mesenchymal ‘membrane’ in which intramembranous ossification occurs. Its deeper layers contain osteoprogenitor cells.

The first appearance of a centre of primary ossification ( Fig. 5.21 ) occurs when chondroblasts deep in the centre of the primitive shaft enlarge greatly, and their cytoplasm becomes vacuolated and accumulates glycogen. The intervening matrix is compressed into thin, often perforated, septa. The cells degenerate and may die, leaving enlarged and sometimes confluent lacunae (primary areolae) whose thin walls become calcified during the final stages ( Fig. 5.22 ). Type X collagen is produced in the hypertrophic zone of cartilage. Matrix vesicles originating from chondrocytes in the proliferation zone are most evident in the intercolumnar regions, where they appear to initiate crystal formation. At the same time, cells in the deep layer of perichondrium around the centre of the cartilage model differentiate into osteoblasts and form a peripheral layer of bone. Initially, this periosteal collar, formed by intramembranous ossification within the perichondrium, is a thin-walled tube that encloses and supports the central shaft (see Figs 5.21 – 5.22 ). As it increases in diameter, it also extends towards both ends of the shaft.

The periosteal collar, which overlies the calcified cartilaginous walls of degenerate chondrocyte lacunae, is invaded from the deep layers of the periosteum (formerly perichondrium) by osteogenic buds. These are blind-ended capillary sprouts that are accompanied by osteoprogenitor cells and osteoclasts. The latter excavate newly formed bone to reach adjacent calcified cartilage, where they continue to erode the walls of primary chondrocyte lacunae ( Figs 5.23 – 5.24 ). This process leads to their fusion into larger, irregular communicating spaces, secondary areolae, which fill with embryonic medullary tissue (vascular mesenchyme, osteoblasts and osteoclasts, haemopoietic and marrow stromal cells, etc.). Osteoblasts attach themselves to the delicate residual walls of calcified cartilage and lay down osteoid, which rapidly becomes confluent, forming a continuous lining of bone. Further layers of bone are added, enclosing young osteocytes in lacunae and narrowing the perivascular spaces. Bone deposition on the more central calcified cartilage ceases as the formation of subperiosteal bone continues.

Osteoclastic erosion of the early bone spicules then creates a primitive medullary cavity in which only a few trabeculae, composed of bone with central cores of calcified cartilage (see Fig. 5.23 ), remain to support the developing marrow tissues. These trabeculae soon become remodelled and replaced by more mature bone or by marrow. Meanwhile, new, adjacent, cartilaginous regions undergo similar changes. Since these are most advanced centrally, and the epiphyses remain cartilaginous, the intermediate zones exhibit a temporospatial sequence of changes when viewed in longitudinal section ( Fig. 5.25 ). This region of dynamic change from cartilage to bone persists until longitudinal growth of the bone ceases.

Expansion of the cartilaginous extremity (usually an epiphysis; see Fig. 5.20 ) keeps pace with the growth of the rest of the bone by both appositional and interstitial growth. The growth zone expands in all dimensions. Lateral growth of a developing long bone is caused by occasional transverse mitosis in its chondrocytes, and by appositional growth as a result of matrix deposition by cells from the perichondrial collar or ring at this level. The future growth plate therefore expands in concert with the shaft and adjacent future epiphysis. A zone of relatively quiescent chondrocytes (the resting zone) lies on the side of the plate closest to the epiphysis. An actively mitotic zone of cells faces towards the shaft of the bone; the more frequent divisions in the long axis of the bone soon create numerous longitudinal columns (palisades) of disc-shaped chondrocytes, each in a flattened lacuna (see Fig. 5.25 ). Proliferation and column formation occur in this zone of cartilage growth (the proliferative zone), and its continued longitudinal interstitial expansion provides the basic mode of elongation of a bone.

The columns of cells show increasing maturity towards the centre of the shaft, as their chondrocytes increase in size and accumulate glycogen. In the hypertrophic zone, energy metabolism is depressed at the level of the mineralizing front (see Fig. 5.23 ). The lacunae are now separated by transverse and longitudinal walls, and the latter are impregnated with apatite crystals in what has become the zone of calcified cartilage (or ossification zone; see Fig. 5.22 ). The calcified partitions enter the zone of bone formation and are invaded by vascular mesenchyme containing osteoblasts, osteoclasts, etc. The partitions, especially the transverse ones, are then partly eroded while osteoid deposition, bone formation and osteocyte enclosure occur on the surfaces of the longitudinal walls. Lysis of calcified partitions is mediated by osteoclast action, aided by cells associated with the terminal buds of vascular sinusoids that occupy, and come into close contact with, each incomplete columnar trabecular framework.

Continuing cell division in the growth zone adds to the epiphysial ends of cell columns, and the bone grows in length as this sequence of changes proceeds away from the diaphysial centre. The bone also grows in diameter as further subperiosteal bone deposition occurs near the epiphyses, and its medullary cavity enlarges transversely and longitudinally. Internal erosion and remodelling of the newly formed bone tissue continues.

Growth continues in this way for many months or years in different bones, but eventually one or more secondary centres of ossification usually appear in the cartilaginous extremities. Initially, these epiphysial centres (or the ends of bones that lack epiphyses) do not display cell columns. Instead, isogenous cell groups hypertrophy, with matrix calcification, and are then invaded by osteogenic vascular mesenchyme, sometimes from cartilage canals. Bone is formed on calcified cartilage, as described above. As an epiphysis enlarges, its cartilaginous periphery (perichondrium) also forms a zone of proliferation in which cell columns are organized radially; hypertrophy, calcification, erosion and ossification occur at increasing depths from the surface. The early osseous epiphysis is thus surrounded by a superficial growth cartilage, and the growth plate adjacent to the metaphysis soon becomes the most active region.

As a bone reaches maturity, epiphysial and metaphysial ossification processes gradually encroach upon the cartilaginous growth plate from either side; when they meet, bony fusion of the epiphysis occurs and longitudinal growth of the bone ceases. The events that take place during fusion are broadly as follows. As growth ceases, the cartilaginous plate becomes quiescent and gradually thins; proliferation, palisading and hypertrophy of chondrocytes stop, and the cells form short, irregular, conical masses. Patchy calcification is accompanied by resorption of calcified cartilage and some of the adjacent metaphysial bone, forming resorption channels that are invaded by vascular mesenchyme.

Some endothelial sprouts pierce the thin plate of cartilage, and the metaphysial and epiphysial vessels unite. Ossification around these vessels spreads into the intervening zones and results in fusion of epiphysis and metaphysis. This bone is visible in radiographs as a radiodense epiphysial line (a term that is also used to describe the level of the perichondrial collar or ring around the growth cartilage of immature bones, or the surface junction between epiphysis and metaphysis in a mature bone). In smaller epiphyses, which unite earlier, there is usually one initial eccentric area of fusion, and thinning of the residual cartilaginous plate. The original sites of fusion are subsequently resorbed and replaced by new bone. Medullary tissue extends into the whole cartilaginous plate until union is complete and no epiphysial ‘scar’ persists. In larger epiphyses, which unite later, similar processes also involve multiple perforations in growth plates, and islands of epiphysial bone often persist as epiphysial scars. Calcified cartilage coated by bone forms the epiphysial scar, and is also found below articular cartilage. It has been called metaplastic bone, a term also applied to sites of attachments of tendons, ligaments and other dense connective tissues to bone.

The cartilaginous surfaces of epiphyses that form synovial joints remain unossified, but the typical sequence of cartilaginous zones persists in them throughout life. A similar developmental sequence occurs at synchondroses, except that the proliferative rates of chondrocytes and the replacement of cartilage by bone are similar, although not identical, on either side of the synchondrosis.

Postnatal growth and maintenance of bone

Modelling, by which is meant changes in general shape, occurs in all growing bones. The process has been studied mainly in cranial and long bones with expanded extremities.

During calvarial growth, a bone such as the parietal thickens and expands, but also decreases in curvature. Accretion continues at its edges by proliferation of osteoprogenitor cells at sutures; periosteal bone is mainly added externally and eroded internally, but not at uniform rates or at all times. The rate of formation increases with radial distance from the centre of ossification (in this case, the future parietal eminence). Bone formation may also occur endocranially as well as ectocranially, so changing the curvature of the bone. The relative positions of the original centres of ossification change in three dimensions as the skull bones thicken and grow at the sutures and as the calvaria (vault) of the skull expands to accommodate the growth of the brain. Development of the outer and inner cortical plates is accompanied by internal development of trabeculae and marrow spaces.

Long bones increase in length mainly by endochondral ossification at the epiphysial growth plates. Simultaneous increase in width occurs by subperiosteal deposition and endosteal erosion. Growth at different locations can occur at different rates, or even be replaced by resorption, resulting in a change in the shape of a bone. This explains how, for example, the tibia changes its cross-sectional shape from tubular to triangular. Similarly, the waisted contours of metaphyses are preserved by differential rates of periosteal erosion and endosteal deposition, as metaphysial bone becomes diaphysial in position. The junction between a field of resorption and one of deposition on the surface of a growing bone is called a surface reversal line. The relative position of such a line may remain stable over long periods of growth and shape change.

Lamellar bone forms and is remodelled at variable rates throughout adult life (see below).

Normal development and maintenance of bone requires adequate intake and absorption of calcium, phosphorus and vitamins A, C and D, and a balance between growth hormone (GH, somatotropin), thyroid hormones, oestrogens and androgens. Other biological influences include prostaglandins and glucocorticoids. Vigorous mechanical loading is important for the maintenance of adequate bone mass. Prolonged deficiency in any of these factors can lead to loss of bone tissue (osteopaenia); if bone loss is severe (osteoporosis), it can lead to fracture and deformity.

Vitamin D influences intestinal transport of calcium and phosphate, and therefore affects circulatory calcium levels. In adults, prolonged deficiency (with or without low intake) produces bones that contain regions of deformable, uncalcified osteoid (osteomalacia). During growth, vitamin D deficiency can lead to severe disturbance of growth cartilages and ossification, such as reductions of regular columnar organization in growth plates, and failure of cartilage calcification even though chondrocytes proliferate. Growth plates also become thicker and irregular, as exemplified in classic rickets or juvenile osteomalacia. In rickets, the uncalcified or poorly calcified cartilaginous trabeculae are only partially eroded; osteoblasts secrete layers of osteoid but these fail to ossify in the metaphysial region, and ultimately gravity deforms these softened bones.

Vitamin C is essential for the adequate synthesis of collagen and matrix proteoglycans in connective tissues. When vitamin C is deficient, growth plates become thin, ossification almost stops, and metaphysial trabeculae and cortical bone are reduced in thickness, causing fragility and delayed healing of fractures.

Vitamin A is necessary for normal growth and for a correct balance between deposition and removal of bone. Deficiency retards growth as a result of the failure of internal erosion and remodelling, particularly in the cranial base. Foramina are narrowed, sometimes causing pressure atrophy of the nerves that pass through them. The cranial cavity and spinal canal may fail to expand at the same rate as the developing central nervous system, impairing nervous function. Conversely, excess vitamin A stimulates vascular erosion of growth cartilages, which become thin or totally lost, and longitudinal growth ceases. Retinoic acid, a vitamin A derivative, is involved in pattern formation in limb buds and in the differentiation of osteoblasts.

Balanced endocrine functions are also essential to normal bone maturation, and disturbances in this balance may have profound effects. In addition to its role in calcium metabolism, excess parathyroid hormone (primary hyperparathyroidism) stimulates unchecked osteoclastic erosion of bone, particularly subperiosteally and later endosteally (osteitis fibrosa cystica). Growth hormone (GH) is required for normal interstitial proliferation in growth cartilages, ensuring normal increase in stature. Termination of normal growth is imperfectly understood, but may involve a fall in hormone production or in the sensitivity of chondroblasts to insulin-like growth factors regulated by GH. Reduction of GH production in the young leads to quiescence and thinning of growth plates and hence pituitary dwarfism. Conversely, continued hypersecretion in the immature leads to gigantism, and in the adult results in thickening of bones by subperiosteal deposition; the mandible, hands and feet are the most affected, a condition known as acromegaly.

While continued longitudinal growth of bones depends on adequate levels of GH, effective remodelling to achieve a mature shape also requires the action of thyroid hormones. Moreover, growth and skeletal maturity are closely related to endocrine activities of the ovaries, testes and suprarenal cortices. High oestrogen levels increase deposition of endosteal and trabecular bone; conversely, osteoporosis in postmenopausal women reflects reduced ovarian function. In men, fluctuations in the rate of growth, and the timing of skeletal maturation, depend on circulating levels of suprarenal and testicular androgens. In hypogonadism, growth-plate fusion is delayed and the limbs therefore elongate excessively; conversely, in hypergonadism, premature fusion of the epiphyses results in diminished stature.

Bone remodelling

Stiff materials (including bone) are vulnerable to the accumulation of microdamage during repeated loading. In metals this can result in crack propagation and ‘fatigue failure’. Bone reduces the risk of such failure by periodically renewing itself, one small region of tissue at a time. This process is referred to as ‘remodelling’ because the volume and orientation of newly replaced matrix are not necessarily the same as the old; instead, bone takes this opportunity to adapt its mass and architecture to prevailing mechanical demands. Remodelling affects the local balance between resorption and deposition of bone. Its primary purpose is to renew bone rather than increase its mass, and the process continues throughout life, replacing approximately 10% of bone each year in adults ( ).

Internal remodelling continuously supplies young osteons with labile calcium reserves, and provides a malleable bony architecture that is responsive to changing patterns of stress. A bone-remodelling unit consists of an advancing cutting cone and a closing cone. Activated osteoclasts form a cutting cone that excavates a cylindrical tunnel of bone (resorption canal) and advances ahead of a central growing blood vessel at a rate of 50 μm/day. A cutting cone is typically 2 mm long and takes 1–3 months to form; a similar period is required to create the new (secondary) osteon by completing the closing cone ( Fig. 5.26 ). Osteoblasts follow the osteoclasts, filling in the space created with new osteoid, starting at the peripheral surface or walls of the tunnel. Successive layers of bone are deposited on the surface of the previous layer as cohorts of osteoblasts become embedded (as osteocytes) in the matrix they secrete, until the most central lamella is close to the blood vessel at the axis of the cylinder. The ‘closing cone’ (see Fig. 5.26 ) may contain 4000 osteoblasts per mm ² . In this way, the walls of resorption canals are lined with new lamellar matrix and the vascular channels are progressively narrowed. A hypermineralized basophilic cement (or growth-reversal) line marks the edge of a new osteon, indicating the border between the resorptive activity of the cutting cone and the bony matrix not remodelled by this activity. Remnants of the circumferential lamellae of old osteons form interstitial lamellae between newer osteons (see Fig. 5.17A ).

The remodelling unit in cancellous bone, equivalent to the secondary osteon of compact bone, is the basic (or bony) structural unit; it has an average thickness of 40–70 μm and an average length of 100 μm, but may be more extensive and irregular in shape. Separate structural units can sometimes be visualized in microradiographs because of differences in their age and extent of mineralization ( ).

Adult bone shape and mass are partly determined by genetic inheritance ( ). However, the pattern and extent of remodelling are largely dictated by the mechanical loading applied to the bone. Bone resorption occurs when muscle or gravitational forces are reduced, as occurs in bed rest, or in zero gravity conditions in space ( ). Reduced activity in old age is another major cause of bone loss. The rate of remodelling decreases with age, which means that numbers of osteons and osteon fragments can be used to estimate the age of skeletal material at death. Conversely, increased sporting or occupational loading of the skeleton can cause bone hypertrophy, as exemplified by the 35% increase in cortical thickness in the racket arm of elite tennis players ( ). Bones appear to respond to the maximum deformation they experience (see Fig. 5.68 ), rather than to cumulative load. Bone subjected to constant pressure can actually resorb, a response that underpins much orthodontic treatment, because teeth can be made to migrate slowly through alveolar bone by the application of steady lateral or medial pressure.

Growth of individual bones

Ossification centres appear over a long period during bone growth: many in embryonic life, some in prenatal life, and others well into the postnatal growing period. Ossification centres are initially microscopic but soon become macroscopic, which means that their growth can then be followed by radiological and other scanning techniques.

Some bones, including carpal, tarsal, lacrimal, nasal and zygomatic bones, inferior nasal conchae and auditory ossicles, ossify from a single centre, which may appear between the eighth intrauterine week and the tenth year: a wide interval for studying growth or estimating age. Most bones ossify from several centres, one of which appears in the centre of the future bone in late embryonic or early fetal life (seventh week to fourth month). Ossification progresses from the centres towards the ends, which are still cartilaginous at birth ( Fig. 5.27 ). These terminal regions ossify from separate centres, which are sometimes multiple, and which appear between birth and the late teens; they are therefore secondary to the earlier primary centre from which much of the bone ossifies. This is the pattern in long bones, as well as in some shorter bones such as the metacarpals and metatarsals, and in the ribs and clavicles.

At birth, a bone such as the tibia is typically ossified throughout its diaphysis from a primary centre that appears in the seventh intrauterine week, whereas its cartilaginous epiphyses ossify from secondary centres. As the epiphyses enlarge, almost all the cartilage is replaced by bone, except for a specialized layer of articular (hyaline) cartilage that persists at the joint surface, and a thicker zone between the diaphysis and epiphysis. Persistence of this epiphysial growth plate, or growth cartilage, allows increase in bone length until the usual dimensions are reached, by which time the epiphysial plate has ossified. The bone has then reached maturity. Coalescence of the epiphysis and diaphysis is fusion, the amalgamation of separate osseous units into one.

Many long bones have epiphyses at both their proximal and their distal extremities, whereas metacarpals, metatarsals and phalanges have only one epiphysis. Typical ribs have epiphyses for the head and articular tubercle and another for the non-articular area. The costal cartilages represent the unossified hyaline cartilage of the developing rib and therefore do not display epiphyses. Epiphysial ossification is sometimes complex, e.g. the proximal end of the humerus is wholly cartilaginous at birth, subsequently developing three centres during childhood that coalesce into a single mass before they fuse with the diaphysis. Only one of these centres forms an articular surface, and the others form the greater and lesser tubercles, to which muscles are attached. Similar composite epiphyses occur at the distal end of the humerus and in the femur, ribs and vertebrae.

Many cranial bones ossify from multiple centres. The sphenoid, temporal and occipital bones are almost certainly composites of multiple elements in their evolutionary history;: some show evidence of fusion between membrane and cartilage bones that unite during growth.

If bone growth rate were uniform, ossification centres would appear in a strict descending order of bone size. However, disparate rates of ossification occur at different sites and do not appear to be related to bone size. The appearance of primary centres for bones of such different sizes as the phalanges and femora are separated by, at most, a week of embryonic life. Those for carpal and tarsal bones show some correlation between size and order of ossification, from largest (calcaneus in the fifth fetal month) to smallest (pisiform in the ninth to twelfth postnatal year). In individual bones, succession of centres is related to the volume of bone that each centre produces. The largest epiphyses, e.g. the adjacent ends of the femur and tibia, are the earliest to begin to ossify (immediately before or after birth) and are of forensic interest. At epiphysial plates, the rate of growth is initially equal at both ends of those bones that possess two epiphyses. However, experimental observations in other species have revealed that one epiphysis usually grows faster than the other after birth. Since the faster-growing end also usually fuses later with the diaphysis, its contribution to length is greater. Though faster rate has not been measured directly in human bones, later fusion has been documented radiologically. The more active end of a long limb bone is often termed the growing end but this is a misnomer.

The rate of increase in stature, which is rapid in infancy and again at puberty, demonstrates that rates of growth at epiphyses vary. The spurt at puberty, or slightly before, decreases as epiphyses fuse in post-adolescent years.

Growth cartilages do not grow uniformly at all points, which presumably accounts for changes such as the alteration in angle between the humeral shaft and its neck. The junctions between epiphysis and diaphysis at growth plates are not uniformly flat on either surface. Osseous surfaces usually become reciprocally curved by differential growth, and the epiphysis forms a shallow cup over the convex end of the shaft, with cartilage intervening – an arrangement that may resist shearing forces at this relatively weak region. Reciprocity of bone surfaces is augmented by small nodules and ridges, as can be seen when the surfaces are stripped of cartilage. These adaptations emphasize the formation of many immature bones from several elements held together by epiphysial cartilages. Most human bones exhibit these complex junctions, at which bone is bonded to bone through cartilage, throughout the active years of childhood and adolescence.

Forces at growth cartilages are largely compressive but with an element of shear. Interference with epiphysial growth may occur as a result of trauma but more frequently follows disease; the resulting changes in trabecular patterns of bone are visible radiographically as dense transverse lines of arrested growth (Harris's growth lines). Several such lines may appear in the limb bones of children afflicted by successive illnesses.

Variation in skeletal development occurs between individuals, sexes and possibly also races. The timing rather than the sequence of events varies, and females antedate males in all groups studied. Differences that are perhaps insignificant before birth may be as great as two years in adolescence.

Suture

Sutures are restricted to the skull (see Chapter 34 for descriptions of individual sutures). In a suture, the two bones are separated by a layer of membrane-derived connective tissue. The sutural aspect of each bone is covered by a layer of osteogenic cells (cambial layer) overlaid by a capsular lamella of fibrous tissue that is continuous with the periosteum on both the endo- and ectocranial surfaces. The region between the capsular coverings contains loose fibrous connective tissue and decreases with age, so that the osteogenic surfaces become apposed. On completion of growth, many sutures synostose and are obliterated. Synostosis occurs normally as the skull ages; it can begin in the early twenties and continues into advanced age. A schindylesis is a specialized suture in which a ridged bone fits into a groove on a neighbouring element, e.g. where the cleft between the alae of the vomer receives the rostrum of the sphenoid (see Fig. 5.28 ).

Gomphosis

A gomphosis is a peg-and-socket junction between a tooth and its socket, where the two components are maintained in intimate contact by the collagen of the periodontium connecting the dental cement to the alveolar bone. Strictly speaking, a gomphosis is not an articulation between two skeletal structures.

Syndesmosis

A syndesmosis is a truly fibrous connection between bones. It may be represented by an interosseous ligament (e.g. the interosseous membrane between the radial and ulnar shafts), a slender fibrous cord, or a denser fibrous membrane (e.g. the posterior region of the sacroiliac joint: see Fig. 5.28 ).

Primary cartilaginous joints

Primary cartilaginous joints or synchondroses occur where advancing centres of ossification remain separated by an area of hyaline (but non-articular) cartilage ( Fig. 5.29A ). They are present in all postcranial bones that form from more than one centre of ossification. Since hyaline cartilage retains the capability to ossify with age, synchondroses tend to synostose when growth is complete. Primary cartilaginous joints are almost exclusively associated with growth plates (see above).

Secondary cartilaginous joints

Secondary cartilaginous joints, or symphyses, are largely defined by the presence of an intervening pad or disc of fibrocartilage interposed between the articular (hyaline) cartilage that covers the ends of two articulating bones ( Fig. 5.29B ). Intervertebral discs are prime examples. The pad or disc varies from a few millimetres to over a centimetre in thickness, and the whole region is generally bound by strong, tightly adherent, dense connective tissues. Collagenous ligaments extend from the periostea of the articulating bones across the symphysis. The ligaments blend with the hyaline and fibrocartilaginous perichondria but do not form a complete capsule. They contain plexuses of afferent nerve terminals, which also penetrate the periphery of the fibrocartilage. The combined strength of the ligaments and fibrocartilage can exceed that of the associated bones. A symphysis is designed to withstand a range of stresses (compression, tension, shear, bending and torsion) but the range of movement is generally limited, both by the physical nature of the articulation and by adjacent bones. Tears are usually the result of sudden stresses that occur when the body is in an inappropriate posture.

All symphyses occur in the midline (mandibular, manubriosternal, pubic and intervertebral) and all except the mandibular symphysis occur in the postcranial skeleton and resist synostosis. The mandibular symphysis (symphysis menti) is histologically different from the other symphyses; however, the widespread use of this descriptive term ensures that it remains, perhaps inappropriately, within this category.

The concept that synchondroses are temporary and concerned with growth, whereas symphyses are permanent and concerned with movement, is an oversimplification and only partly correct. Both types of joint must be strong, both are sites at which growth occurs, and both contribute either directly or indirectly to the total movement patterns of the parts involved. Movements that occur at a symphysis often depend on more than the mechanical properties of the fibrocartilaginous pad or disc, e.g. movements between vertebrae depend not only on the deformability of the intervertebral disc but also on the morphology of the apophysial joints and the properties of associated ligaments ( ).

The prominent role of synchondroses in skeletal growth is widely recognized, whereas growth of symphyses has received less attention. Symphysial growth may, for convenience, be considered from two interrelated aspects: namely, intrinsic growth of the fibrocartilaginous disc, and growth of the hyaline cartilaginous plates into which endochondral ossification progresses.

Articular surfaces

Articular cartilage comprises a specialized type of hyaline cartilage, reflecting its origin as part of the cartilaginous ‘model’ of bone in embryonic life. Exceptions include the sternoclavicular, acromioclavicular and temporomandibular joints, where articulating surfaces are covered by dense fibrous tissue containing isolated groups of chondrocytes with little proteoglycan in their surrounding matrix, presumably reflecting their formation by intramembranous ossification.

The deformability of articular cartilage enables opposing cartilage surfaces to flatten slightly at their area of contact, increasing contact area and decreasing contact stress (see Fig. 5.57 ). This load-distributing property of articular cartilage depends on the congruence of opposing joint surfaces (see Fig. 5.61 ). Slight undulations in the surface trap synovial fluid so that fluid-film lubrication is possible under most circumstances; effectively, the bones ‘aqua-plane’ on each other ( Fig. 5.60 ). This ensures very low friction and, consequently, low wear of the cartilage.

The most superficial layer of articular cartilage, directly adjacent to the synovial fluid, is acellular and approximately 3 μm thick. It contains fine collagen fibrils running parallel to the surface, and functions as an elastic and protective ‘skin’ for the underlying layers. It can appear to recoil under tension when damaged.

The acellular surface layer is coated with a large glycoprotein, lubricin, which projects from the surface so that a hydrophobic region of the molecule lies in the joint space, where it repels its counterpart on the opposing articular surface. In this way, lubricin acts in the manner of a lubricant such as grease to reduce friction and wear of the surfaces. This ‘boundary lubrication’ mechanism becomes important when the fluid film has been squeezed out, e.g. after sustained forceful loading of the joint, and loss of lubricin can lead to cartilage degeneration ( ). Transmission electron microscopy shows this lubricant layer as an interrupted electron-dense surface coat 0.03–0.1 μm thick. Synovial fluid and membranous debris, the product of chondrocytic necrosis, may contribute to this surface coat, which is transient in nature. A ‘lamina splendens’ is sometimes identified as a bright line at the free surface of articular cartilage when oblique sections are examined by negative phase contrast microscopy: it may be a microscopical artefact at the border between regions of different refractive index, rather than an anatomically distinct surface layer. Deeper zones of articular cartilage are described on p. 87 .

With advancing age, undulations on the articular surfaces deepen and develop minute, ragged projections, perhaps as a consequence of wear and tear. These changes occur extremely slowly in healthy joints, but can be accelerated in pathologically ‘dry’ joints, or where synovial fluid viscosity is altered, or following injury.

Fibrous capsule

A fibrous capsule completely encloses each synovial joint except where it is interrupted by synovial protrusions (see descriptions of individual joints for details). It is composed of interlacing bundles of parallel fibres of collagen type I, and is attached continuously round the ends of the articulating bones. In small bones this attachment is usually near the periphery of the articular surfaces, but in long bones it varies considerably, and part or all of the attachment may be a significant distance from the articular surface. The joint capsule is perforated by vessels and nerves, and may contain apertures through which synovial membrane protrudes as bursae.

A fibrous capsule usually exhibits local thickenings of parallel bundles of collagen fibres, called capsular (intrinsic) ligaments, and named by their attachments. Some capsules are reinforced or replaced by tendons of nearby muscles, or expansions from them. Accessory ligaments are distinct structures that may be located inside or outside the joint capsule. All ligaments, although stiff in tension, are pliant in bending. They can rebound elastically from being stretched by up to 10–15%, and are protected from injury by reflex contraction of appropriate muscles. They do little to resist normal movements but become taut at the end of each normal range of movement.

Synovial membrane

Synovial membrane lines the fibrous joint capsule and exposed osseous surfaces, intracapsular ligaments, bursae and tendon sheaths ( Fig. 5.32 ). It does not cover intra-articular discs or menisci, and stops at the margins of articular cartilages in a transitional zone that occupies the peripheral few millimetres of cartilage. Where a tendon is attached to bone inside a synovial joint, an extension of the synovial membrane usually accompanies it beyond the capsule. Some extracapsular tendons are separated from the capsule by a synovial bursa that is continuous with the interior of the joint. All of these protrusions are potential routes for the spread of infection into joints.

Synovial membrane secretes and absorbs a fluid that lubricates the movement between the articulating surfaces.

Pink, smooth and shining, the internal synovial surface displays a few small synovial villi that increase in size and number with age. Folds and fringes of membrane may also project into a joint cavity; some are sufficiently constant to be named, e.g. the alar folds and ligamentum mucosum of the knee. Synovial villi are more numerous near articular margins and on the surfaces of folds and fringes, and become prominent in some pathological states.

Accumulations of adipose tissue (articular fat pads) occur within the synovial membrane in many joints. These pads, together with synovial folds and fringes, are deformable cushions that occupy potential spaces and irregularities in joints that are not wholly filled by synovial fluid. During movement they accommodate to the changing shape and volume of the irregularities, a function they share with intra-articular discs and menisci. They also increase the area of synovial membrane, and may help to spread synovial fluid over the articular surfaces.

Synovial membrane consists of a highly cellular intimal layer resting on a fibrous and vascular subintimal layer (subsynovial tissue). The subintima is often composed of loose, irregular connective tissue, but also contains organized collagen and elastin fibres lying parallel to the membrane surface, interspersed with occasional fibroblasts, macrophages, mast cells and fat cells. The elastic component may prevent formation of redundant folds during joint movement. Subintimal adipose cells form compact lobules surrounded by highly vascular fibroelastic interlobular septa that provide firmness and compressive turgor.

The intimal layer consists of pleomorphic synovial cells embedded in a granular, amorphous matrix. In normal human joints, synovial cells form an interlacing, discontinuous layer, 1–3 cells and 20–40 μm thick, between the subintima and the joint cavity. They are not separated from the subintima by a basal lamina, and are distinguished from subintimal cells only because they associate to form a superficial layer. In many locations, but particularly over loose subintimal tissue, areas are commonly found that are free from synovial cells. The synovial cells over fibrous subintimal tissue may be flattened and closely packed, forming endothelium-like sheets. Human synovial cells are generally elliptical, with numerous cytoplasmic processes. Neighbouring cells are often separated by gaps, but their processes may interdigitate where they lie close together. There is considerable regional variation in cell morphology and numbers.

There are at least two morphologically distinct populations of synovial cells or synoviocytes: type A macrophage-like synoviocytes (MLS) and type B fibroblast-like synoviocytes (FLS) ( ). The phenotype of type A synoviocytes is similar to that of other tissue resident macrophage populations: morphologically they are characterized by surface ruffles or lamellipodia, plasma membrane invaginations and associated pinocytotic vesicles, a prominent Golgi apparatus but little rough endoplasmic reticulum. They probably synthesize and release lytic enzymes and phagocytose joint debris from synovial fluid. Type B synoviocytes, which predominate, resemble fibroblasts and have abundant rough endoplasmic reticulum but fewer vacuoles and vesicles, and a less ruffled plasma membrane than type A synoviocytes. They probably synthesize some of the hyaluronan of synovial fluid, the boundary lubricant lubricin, and inhibitors of the degradative enzymes synthesized by type A cells, limiting their potential to damage joint tissues. Synoviocytes do not divide actively in normal synovial membranes, but may do so in response to trauma and haemarthrosis. Under such conditions, type B synoviocytes divide in situ , while the population of type A cells increases by immigration of bone marrow-derived precursors. FLS and MLS play essential roles in the joint pathology of rheumatoid arthritis ( ).