Cartilage is a skeletal tissue that, together with the bone, forms the framework supporting the body as a whole.¹  It is a tissue of great biological importance, which is apparent from the fact that vertebrate life would be impossible without cartilage: different types of cartilage play major roles in the function of such crucial systems as the spine and the respiratory system. Among different types of cartilage, articular cartilage is the best known and perhaps the most studied type. Articular cartilage is the thin layer of connective tissue that covers the articulating ends of bones in synovial joints (e.g., knee, hip, shoulder, and many other movable joints in the body). The intense interest in articular cartilage is motivated by the critical role its degradation plays in arthritis and related joint diseases, which are the number one cause of disability in humans.² 

This chapter provides an essential (albeit incomplete) introduction to the multi-level and multi-scale properties of articular cartilage that give this tissue its extraordinary load-bearing characteristics. Section 1.2 describes the general relationship between cartilage and the joint, while the following sections focus on different aspects of articular cartilage that are important to understanding its biological properties and its magnetic resonance behaviour. Sections 1.3–1.6 introduce cartilage at cellular, extracellular, histological and biomechanical levels. Sections 1.7–1.9 discuss the spatial heterogeneity of cartilage properties over the whole joint, diseases of cartilage and joint, and research into arthritis and related joint diseases.

Cartilage is a specialised connective tissue present in animals, including humans, distinct from connective tissue proper.³  Cartilage tissue is stiff but also flexible. As a result, it is an integral part of many parts of the body where the supporting structures must accommodate limited movement.

Different varieties of cartilage occur at specific anatomical locations. Three types of cartilage are commonly distinguished morphologically.

1. Hyaline cartilage is named from the Latin word “hyalinus”, meaning “smooth”, “clear” or “glass-like”. It has a homogeneous appearance and a semi-solid consistency. Hyaline cartilage is the most abundant type of cartilage in the body and includes articular cartilage, which lines the articulating surfaces of bones within many movable joints. Other varieties of hyaline cartilage include costal cartilage connecting the anterior ends of the ribs to the sternum, nasal cartilage, and many laryngeal and tracheobronchial cartilages. In the foetus, hyaline cartilage forms most of the embryonic skeleton (“temporary cartilage”) before it is replaced by bone during ossification. Hyaline cartilage also forms epiphyseal growth plates, which enable rapid growth of long bones during childhood (see Section 1.3.1).^1,4 

2. Fibrocartilage is fasciculated and fibrous. It exhibits significant tensile strength and occurs in many areas subject to high mechanical stress. Fibrocartilage forms the annuli of intervertebral discs in the spine (anulus fibrosus), the menisci of the knee and certain other joints, and the plates connecting the opposing surfaces of bones within many immovable joints (e.g. the pubic symphysis). It is also present in some movable joints (e.g. glenoid and acetabular labra) and at the bone attachment sites of tendons and ligaments. Finally, fibrocartilage is found in articular discs, which facilitate independent movements in certain compound joints (e.g. the distal radioulnar joint in the wrist).^1,4,5 

3. Elastic cartilage is similar to hyaline cartilage, but exhibits greater elasticity. It is found in parts of the body where stretchability is required, most notably in the epiglottis – the elastic flap at the entrance to the larynx that acts as a valve allowing entry either into the trachea or the oesophagus. It is also present at the attachment of the vocal cords to the larynx and forms part of the structure of the external ear.^1,6 

Synovial joints are a certain type of joint that allow for a wide range of motion. They include the knee (Figure 1.1), hip, shoulder and elbow, as well as numerous smaller joints.¹  The adjective “synovial” refers to the presence of a synovial cavity – a space enclosed by a dense, fibrous articular capsule filled with synovial fluid.

A synovial joint behaves in many ways as a “complete organ”⁷  (or more precisely, an organ system). Its biomechanical function is carried out by its many components functioning in conjunction with each other:

1. Articular cartilage is a thin layer of connective tissue that covers the articulating surfaces of the bones within a synovial joint (see Figure 1.1).⁸  It protects the bone and processes mechanical load applied to the joint (see Section 1.6). The surface of the cartilage facing the synovial space is known as the articular surface. The articular surface is lubricated and exhibits very low friction.^9–13 

2. Synovial fluid is a filtrate of plasma that contains significant amounts of the polysaccharide hyaluronic acid (HA) and exhibits viscoelastic rheological properties. It contributes to the lubrication of the articular surface.^9–15 

3. The inner lining of the articular capsule is the synovial membrane, which is made up of loose connective tissue. The surface of the membrane is lined with squamous-to-cuboidal cells (synoviocytes) one to four layers deep. The cells are traditionally categorized into macrophage-like type A (which synthesize HA) and fibroblast-like type B (which synthesize various proteins). These two types are considered by some to be one cell population that alters its phenotype according to functional demands.¹⁴ 

4. Certain joints, most notably the knee, contain a meniscus – a fibrocartilaginous structure that fits between the articular surfaces of apposing bones and contributes to load processing.

5. Ligaments and muscles provide transmission of mechanical forces across the joint and reinforce the synovial cavity and the joint itself.

The articular capsule of a synovial joint is vascularized, but the interior of the joint (synovial cavity) is an avascular (as well as aneural) environment. This has two major implications for articular cartilage:

1. Articular cartilage is dependent upon the synovial fluid for the delivery of nutrients and the removal of metabolic wastes. Both these vital processes occur mainly via passive diffusion. The synovial lining lacks a basement membrane and merges with the underlying vascular connective tissue within the articular capsule; this allows for rapid exchange between blood and synovial fluid.¹⁴  Removal of particulate debris from the synovial cavity is performed by synoviocytes.¹ 

2. The absence of a direct blood supply renders the repair of diseased or damaged cartilage slow and inefficient. This is a major biological factor contributing to the prevalence of osteoarthritis, a degenerative disease of articular cartilage discussed in Sections 1.8 and 1.9.

In the major joints of large animals and adult humans, the thickness of articular cartilage ranges from approximately 2 mm (bovine knee articular cartilage) to approximately 4 mm (human knee). Cartilage in smaller human joints and in many laboratory animals is much thinner.¹⁶  Histologically, articular cartilage is commonly subdivided into four distinct and parallel zones based on the local orientation of the collagen fibrils. These zones are the superficial zone, the transitional zone, the radial zone, and the calcified zone (Figure 1.2).

Articular cartilage has a low cellularity, with the living cells (chondrocytes) occupying only approximately 2% of the volume of mature human cartilage.¹⁷  The bulk of the tissue is the extracellular matrix (ECM), which depends crucially on this small number of chondrocytes to synthesize and maintain articular cartilage over the lifetime. The loss of the cells will induce fatigue to the ECM and eventually a failure of the joint.

Articular cartilage is formed through condensation of cartilage progenitor cells (CPCs). During the initial growth of articular cartilage, the CPCs in the growth plate (at the end of the bone) undergo a series of tightly regulated differentiation and maturation processes¹⁸  and can be found to have four morphologically distinct layers (the resting, proliferative, prehypertrophic, and hypertrophic layers); the cells in these layers have differences in gene expression pattern and function.¹⁹  The resting layer is the farthest layer from the bone and consists of undifferentiated CPCs, which are fusiform, compactly positioned, have a low rate of proliferation, and are smaller and denser than mature chondrocytes, with larger nuclei and less cytoplasm.¹⁸  In the proliferative layer, CPCs gain a proliferative phenotype and have an increased proliferation rate that results in a flattened and oblate shape, and arrangement of multicellular clusters into longitudinal columns.²⁰  These highly organized columns may be directed by the cells in the resting layer, which have been postulated to produce a growth plate orientating factor.¹⁸  Both resting and proliferating cells can synthesize proteoglycans (e.g. versican) and collagens (e.g. type IIA), which have some distinctions while closely resembling those in mature articular cartilage.^21,22  Following the proliferation layer, cells pass through a transition layer (prehypertrophic layer), in which they lose their proliferative ability and gradually increase in size due to swelling. Finally, these cells become hypertrophic and undergo major phenotypic changes,²³  following a cascade of events including ECM mineralization, angiogenesis, and cellular apoptosis,²⁰  which eventually lead to the formation of bone.

Mature chondrocytes are the only cell type present in adult articular cartilage. Unlike the CPCs, mature chondrocytes normally maintain a stable phenotype.²⁴  Mature chondrocytes vary in numbers, sizes (approximately 10–15 µm in diameter), and shapes (oval to circular), depending on the tissue zone in which they are located (Figure 1.2).²⁵  There are a number of cellular gradients across the zones in cartilage. For example, the chondrocyte density per zone can increase from 7000 to 24 000 cells mm⁻³ along the tissue depth. Both cellular surface area per tissue volume and cellular volume per tissue volume have their maximum values in the upper radial zone.¹⁷ 

Morphologically, the mature chondrocytes in the superficial zone are flattened, in single-cell units, and oriented in parallel to the articular surface. Mature chondrocytes in the transitional zone are more rounded and often found in pairs, while mature chondrocytes in the radial zone are rounded or elliptic individually and aligned as a multi-cell group along the collagen fibrils.¹⁷  Mature chondrocytes in the calcified cartilage become hypertrophic and synthesize collagen type X.²⁶  Chondrocytes and their pericellular matrix in cartilage are organized as chondrons (“pericellular capsule”) (Figure 1.3), which consist of the chondrocyte and the pericellular molecular proteins, of which collagen type VI and IX are the major components. These collagens are only present in close proximity to the cells, and decrease to a very low level away from the cells, where chondroitin sulfate and keratan sulfate become rich.^27,28  The hollow spaces within the network that accommodate the chondrocytes are known as lacunae. Chondrons are the primary structural, functional, and metabolic units in articular cartilage.

Functionally, mature chondrocytes are highly specialized and responsible for synthesizing and remodeling the ECM (mainly comprised of collagen type II and aggrecan; see Section 1.4) which governs the functional and mechanical properties of articular cartilage.²⁹  To produce and maintain healthy cartilage, the mature chondrocytes display a specific pattern of gene expression to both biochemical cues (e.g. interleukins (IL), insulin-like growth factors, bone morphogenetic protein, transforming growth factor, and fibroblast growth factor) and mechanical influences (e.g. compression, shear, and hydrostatic pressure; see also chapter 15)^19,30  by various signal transduction mechanisms, such as integrins (e.g. α5β1 and α2β1)^31,32  and ion channels (e.g. Ca²⁺)³³  following Src and focal adhesion kinases, the Rho family of small GTPases, mitogen-activated protein kinase, phospholipase C and nuclear factor-κB pathways.^34,35  Normal chondrocytes isolated from the different zones of articular cartilage show differences in proliferation rates and proteoglycan/collagen synthesis, and in their responses to cytokines. Generally, chondrocytes from the deeper zones show higher rates of proliferation and synthesis of collagen and proteoglycan compared with those from the superficial and middle zones,^36,37  and those from the superficial zone are more sensitive to the catabolic effects of IL-1.³⁸  However, in injured or diseased cartilage, chondrocytes have morphologic alterations and altered patterns of gene expression,²⁹  which changes their ability to repair or maintain the ECM.^39–41 

Since mature chondrocytes are terminally differentiated and seem unable to regenerate and replace damaged cartilage, mesenchymal stem cells (MSCs) have been used to develop engineered cartilage tissue. MSCs can be isolated from both adult and fetal tissues (e.g. bone marrow, blood, amniotic fluid, lung, liver, spleen, and umbilical cord) and differentiated towards multiple phenotypes (e.g. chondrogenic, osteogenic, adipogenic, and neural).^42,43  Mature chondrocytes are the differentiated MSCs. A panel of surface markers has been developed to isolate and characterize MSCs, notably the expression of CD29, CD44, CD73, CD90, CD105, and CD166, and the absence of CD34 and CD45.^42,44 

Mesenchymal stem cells are present in normal and osteoarthritic human articular cartilage.⁴⁵  In cases of injury, tissue-specific MSCs in various adult tissues⁴⁶  can proliferate, differentiate, and regenerate the damaged tissue. However, the percentage of these cells decreases with age, concomitant with a lower capacity for proliferation and differentiation.^44,47  Fetal MSCs have an enhanced plasticity potential and proliferation propensity when compared with adult MSCs, since they have active telomerase and express pluripotency markers.⁴⁴  In addition, fetal MSCs offer very low immunogenic properties, which can be explained by the low expression of human leukocyte antigen (HLA)-I and the lack of HLA-II expression.^48,49  In addition, fetal MSCs isolated from different tissues exhibit heterogeneous multi-lineage differentiation potential.⁵⁰  For example, fetal femoral head MSCs showed a higher adipogenic differentiation potential than fetal spine MSCs. Moreover, since variations in the differentiation potential of MSCs isolated from second-trimester fetal tissues were observed, the plasticity of fetal MSCs may be, in part, determined by or dependent on the gestational age. For these reasons, fetal MSCs seem to be the best candidates for use in cartilage regeneration.

While chondrocytes synthesize the ECM of cartilage, the ECM provides the structural and physiological environment for the living chondrocytes. The ECM is also responsible for the remarkable biomechanical properties of articular cartilage discussed in Section 1.6 and elsewhere in this book. The chemical composition of the articular cartilage ECM differs between young and adult cartilage, with the compositional changes occurring primarily during fetal development and skeletal maturation.⁵¹  The major components of the ECM of adult human articular cartilage are collagen (15−20% of the wet weight (w.w.) of cartilage), proteoglycans (3−10% w.w.), and water (65–75% w.w.).¹⁴ 

Collagen is a structural protein and the main solid component of the cartilage ECM (about 1/2 to 2/3 of the dry weight of articular cartilage). More than 50 different types of collagen and collagen-like proteins occur in vertebrates, and some 28 types are found in humans,⁵²  but the type of principal importance in adult articular cartilage is collagen type II.⁵³  The molecular building block of collagens is the tropocollagen molecule, which contains three amino acid chains. The amino acid sequence in each chain consists of repeating blocks glycine–proline−Y or glycine−X−hydroxyproline, where X and Y can be any amino acid. The unusual glycine- and (hydroxy)proline-rich amino acid composition is responsible for the chains forming a “coiled-coil” assembly;⁵⁴  each chain twists into a left-handed helix, and the three twisted chains are then twisted together in a right-handed triple helix⁵⁵  (see Chapter 4, Figure 4.10). The resulting molecule has the geometry of a rod ∼300 nm long and 1.5 nm in diameter (Figure 1.4a).

Collagen II is a fibrillar protein¹: the coiled-coil molecules are assembled, via a combination of covalent cross-linking and non-covalent interactions, into fibrils of a typical diameter 50–80 nm.^55–59  The molecules forming a fibril are staggered by a distance of 67 nm (see Figure 1.4a), which gives collagen its characteristic striated appearance in high-resolution electron microscopy images and Bragg reflections in small-angle X-ray scattering. The fibrils form a cross-linked network that serves as the structural scaffold of the ECM of articular cartilage. The typical fibril–fibril separation within the network is a few hundred nanometers. Multiple collagen fibrils bundle into collagen fibers in tissues.

The collagen network of the articular cartilage’s ECM is synthesized by the chondrocytes during skeletal growth. The synthesis virtually ceases in the adult articular cartilage. In the absence of an injury, the turnover time of collagen II in human femoral head cartilage has been estimated as several hundred years. The rate of collagen synthesis can increase by up to an order of magnitude following an injury;⁵³  however, this is almost never sufficient to repair the damage.

Other types of collagen present in articular cartilage ECM include types III, VI, IX, X, XI, XII, and XIV. These types occur in smaller quantities than type II and play specialized structural roles (e.g. in chondrons). Types IX and XI occur in significant quantities in young articular cartilage during skeletal development, but their content is much lower in mature cartilage.⁵³  Collagen type I is the main ECM structural protein in ligaments and tendons and a major component of fibrocartilage.¹ 

Glycosaminoglycans (GAGs) are macromolecular polysaccharides with a linear chain structure. Along with collagen, they play a key role in determining the biomechanical properties of articular cartilage. Major GAGs in articular cartilage include chondroitin 4-sulphate, chondroitin 6-sulphate, dermatan sulphate and keratan sulphate.¹  The polymeric GAG molecules are composed of repeating disaccharide blocks: N-acetylgalactosamine and glucose in the case of chondroitin, and N-acetylgalactosamine and d-galactose in the case of keratan.^60,61  Importantly, the sulphate and carboxyl groups of GAGs carry negative electric charges, which are the main source of repulsive electrostatic interactions in the cartilage ECM.⁶² 

Glycosaminoglycans are also present in synovial fluid, primarily in the form of HA composed of repeating units of glucuronic acid and N-acetylglucosamine.⁶³  In healthy joints, HA has a polydisperse mass distribution with a typical molecular weight (MW) of 10⁷ Da.^15,63  Hyaluronic acid is present in healthy synovial fluid at a concentration ∼3 mg mL⁻¹ and is the major determinant of the fluid’s rheological properties. HA has also been demonstrated to have an anti-inflammatory¹⁵  and an analgesic effect,⁶³  and possibly contributes to the reconstitution of the superficial cartilage within the joint.¹⁵ 

Proteoglycans are macromolecules that contain a protein core and covalently attached sulfated GAGs (Figure 1.4b).⁶⁴  The core protein typically has a MW ∼250 000 Da and can have in excess of 100 GAG molecules laterally bound to it in a branched, bottlebrush-like arrangement.¹  Cartilage proteoglycans include aggrecan, decorin, biglycan, versican, PG-Lb, and fibromodulin. Aggrecan occurs mostly in articular cartilage and is a copolymer of the cartilage-specific proteoglycan core protein and chondroitin sulfate.⁶⁵  A single aggrecan molecule has a MW ∼ 1–3 × 10⁶ Da. Numerous aggrecan molecules non-covalently bind to a HA chain via a link protein, forming giant aggregates with molecular weights up to 10⁹ Da.⁶⁶ 

Due to their huge molecular weight, PG–HA aggregates have very low mobility in cartilage. They are packed to only a fraction of their fully extended aqueous volume and entrapped between the collagen fibrils.⁶⁷  Their negatively charged GAG chains are the source of fixed electric charge in the cartilage ECM. The intermolecular electrostatic repulsive interactions between GAG groups contribute to the stiffness and elasticity of cartilage tissue. Proteoglycans are also the major determinant of cartilage hydration: owing to their polyanionic nature, proteoglycans bind large amounts of water, in effect acting as giant macromolecular “sponges” that osmotically retain water in the cartilage tissue.

Proteoglycans in adult articular cartilage undergo continuous removal and replacement. The proteoglycan turnover is significantly faster than that of collagen, with the turnover time estimated to be 1–5 years.⁶⁸  This process is regulated by chondrocytes. On one hand, chondrocytes synthesize proteoglycans. On the other hand, proteoglycans are continuously broken down via fragmentation of the protein core by proteolytic enzymes or free radicals and subsequent hydrolysis of the GAG chains in the ECM. In healthy cartilage, the synthesis and replacement of proteoglycans is a steady-state process, with the rates of breakdown and biosynthesis determining the turnover rate and the concentration of proteoglycans in the tissue.⁶⁹ 

Articular cartilage is highly hydrated: water accounts for 65–75% of the tissue weight. Most of the cartilage water resides in, and is considered a part of, the cartilage ECM; i.e. the ECM comprises “solid components” (collagen and proteoglycans) as well as a liquid component (water). The physico-chemical state of the ECM water is a crucial determinant of the physical and physiological properties of articular cartilage, such as its elasticity, its ability to process mechanical load and the transport of nutrients and metabolites through the tissue. The characteristics and properties of water are also of utmost importance to cartilage nuclear magnetic resonance (NMR) and magnetic resonance imaging (MRI) because most such measurements rely on the signal of the proton (the nucleus of the lightest isotope of hydrogen).

Proteoglycans constitute the principal factor responsible for cartilage hydration. The polyanionic nature of proteoglycans combined with their hydrophilicity as carbohydrates enable them to bind large amounts of hydration water. The large size of proteoglycan aggregates means that they are effectively immobilized within the cartilage ECM and do not “leak” from the swollen tissue. Proteoglycans osmotically attract water into articular cartilage and create a sufficiently large osmotic pressure for the tissue to retain significant amount of water even under mechanical compression (see Section 1.6).

A crucial element of the osmotic homeostasis in cartilage is the molecular state of the ECM water. In the first approximation, water is distributed between a pool of bulk-like interstitial water and a (smaller) pool of water associated with the ECM macromolecules, mainly proteoglycans and collagen. These two pools are referred to as “free” and “bound” water, respectively. The distribution is dynamic rather than static: there is rapid (sub-millisecond time-scale) chemical exchange between the two pools, which results in the two populations existing in a state of dynamic chemical equilibrium:

In eqn (1.1), A and B represent the “free” and “bound” water populations, and k₊ and k₋ are the forward and reverse exchange rate constants, respectively. The equilibrium condition requires that the rates of the forward and the reverse reactions be equal; from this, the relative sizes of the two populations can be calculated:

where p_A and p_B are the molar fractions of the free and bound water, respectively, and p_A + p_B ≡ 1. At any given time, relatively few water molecules are actually bound to an ECM macromolecule: p_B ≪ p_A. However, rapid chemical exchange ensures an effective averaging of the chemical potential between the two populations; the presence of the “bound” pool therefore lowers the chemical potential for all the water within the ECM. This enables the proteoglycans to osmotically attract significantly more water into the cartilage tissue than can be present within the immediate proteoglycan hydration shell.

The dynamic equilibrium given by eqn (1.1) also describes the state of other aqueous species present within the ECM, such as the sodium ion (Na⁺) or artificially introduced electrically charged MRI contrast agents (GdDTPA²⁻). Due to the negative charge of proteoglycans, the spatial distribution of electrically charged solutes within cartilage is sensitive to the distribution of proteoglycans. This can be used for non-invasive mapping of the proteoglycan content within cartilage, as described in chapter 6.

In reality, the two-state equilibrium model given by eqn (1.1) is a significant simplification.⁷⁰  Detailed studies of the molecular state of water in hydrogels indicate that even in the presence of a single type of water-binding macromolecule, the model needs to include more than one “bound” state in order to account for certain experimental observations.⁷¹  The situation in cartilage is further complicated by the presence of two different kinds of major macromolecules (collagen and proteoglycan), each possessing chemically distinct water-binding sites (see also chapter 18). It is also complicated by the presence of intrahelical and intrafibrillar water, which exhibits the molecular hydrodynamic properties different to those of the water on the fibril surface.^72–75  Finally, the single-state model of the “free” water is a simplification: the fact that k₊ ≪ k₋ (despite the forward reaction being energetically favorable) implies that k₊ is, in fact, a diffusion-controlled effective rate constant and not the elementary rate constant of the binding step alone. Despite these complexities, the two-state model given by eqn (1.1) is adequate for the description of many physical and NMR properties of cartilage water, as seen in chapters 3 and 7.

The equilibrium illustrated by eqn (1.1) is also instrumental for the understanding of anisotropic NMR properties of water and aqueous solutes. Transient association with aligned collagen fibrils can impart a residual alignment order to the aqueous species. This can be manifested in the NMR experiment as anisotropic spin relaxation rates (chapters 4, 5, and 12), residual dipolar or quadrupolar coupling constants (chapters 8 and 9), or anisotropic diffusivities (chapters 7, 14, and 22).

Besides the three major components, the cartilage ECM contains numerous minor components. Cell adhesion proteins and glycoproteins facilitate the attachment of chondrocytes to the ECM.¹  Chondronectin (MW ∼150 000 Da) is a glycoprotein important to the adhesion of young chondrocytes (chondroblasts) to collagen II fibres.⁷⁶  Anchorin (MW ∼34 000 Da) plays a similar role with mature chondrocytes. Various lipids occur both on the articular surface and throughout the ECM.¹  Articular surface phospholipids are surface-active molecules oriented with their hydrophilic heads towards the cartilage and hydrophobic tails towards the synovium. Articular surface phospholipids have been postulated to contribute to the static lubrication of the articular surface.¹¹ 

The most recognized feature of articular cartilage histology is its zonal division, where cartilage is sub-divided into four distinct zones: the superficial zone, transitional zone, radial zone, and calcified zone (Figure 1.2). However, we must point out that the histological zones in the non-calcified cartilage merely represent a conceptual “discretization” of the continuous functions in the tissue.⁷⁷  This is because in high-resolution imaging, one does not find any abrupt “line” or “border” separating any two histological zones in non-calcified cartilage. Instead, the collagen fibrils change their structural organization gradually over a certain distance away from the articular surface in the tissue. These changes of the fibril orientation could be faster or slower within the tissue, but are never abrupt. Nevertheless, the concept of “discrete tissue zones” is a very useful one, since it conveniently draws our attention to a particular depth of tissue where the morphology is distinct from other regions.

The spatial arrangements of chondrocytes and the ECM in adult articular cartilage vary in different zones (Figure 1.2b).^25,78  In the superficial zone, thin collagen fibrils are arranged parallel to the surface to which the chondrocytes are elongated with their long axis. In this zone, the proteoglycan content is at its lowest level and the water content at its highest level. The collagen fibers in this surface zone have the finest diameters. In the transitional zone, collagen fibers are formed with larger diameters and in less apparent organization within the sparsely distributed round chondrocytes. In the radial zone, chondrocytes are surrounded by organized type VI collagen fibers to form a chondron, in a columnar fashion that is in parallel to the collagen fibers in this part of the ECM. The diameters of the collagen fibers in the radial zone are also bigger than in the superficial zone. Between chondrons is the intercellular matrix which contains the highest concentration of proteoglycans and the lowest water content. In the deepest layer, cartilaginous matrix is encrusted with apatitic salts. Histological staining with hematoxylin and eosin often shows a wavy bluish line, called the tidemark (Figure 1.2a), which represents the mineralization front of the calcified zone that separates the radial zone of the non-calcified articular cartilage from its calcified zone.⁷⁹  (For a more detailed description of histology dyes in cartilage research, see Chapter 22.)

The calcified zone plays an integral role in securing the hyaline cartilage into the bone matrix, by cementing the collagen fibrils to the subchondral bone. The anchoring fibrils are aligned perpendicularly to the subchondral bone from which the fibrils are extended into the articular surface in the shape of arcades known as Benninghoff’s arcades.⁸⁰  This arcade-like architecture takes months or even years to form and may not appear in neonatal/immature and regenerated cartilage. This architecture plays an essential role for the zonal changes and distribution of chondrocytes, ECM molecules and biomechanical properties, even though their shapes and directions can vary widely from one anatomic location to another depending on the distribution of local stress.⁸¹  Water content varies throughout the depth of cartilage, decreasing in concentration from ∼80% at the surface to 65% in the radial zone. During the early phases of osteoarthritis, water content may increase to more than 90% before disintegration of the tissue occurs. A small percentage of this water is contained in the intracellular space, ∼30% (bound water) is associated with the intrafibrillar space within the collagen, and the remainder (free water) is contained in the molecular pore space of the ECM.⁸¹ 

In addition to the fact that these histological zones merely represent a conceptual “discretization” of the continuous functions in articular cartilage,⁷⁷  each histological zone itself also has its own distinct secondary structures. For example, the superficial zone was found to have several distinctly different sub-layers, with the topmost layer being called the lamina splendens.²⁵  This lamina splendens is a thin (several micrometers in thickness) cell-free layer of tightly packed and tangentially oriented collagen fibers,^25,82  which can be viewed as a filter or membrane that prevents the entrance or escape of molecules larger than 6 nm.^83,84  Consequently, proteoglycan aggregates in healthy cartilage are too large to pass through the small pores in the surface layer of articular cartilage, preventing the “leakage” of the charge density. An injury or lesion of this membrane would surely signal an onset of the cartilage disease.

The distribution of collagens and proteoglycans is not homogeneous throughout the depth of articular cartilage. The superficial zone is relatively poor in proteoglycans, but rich in collagen fibers, which align along the articular surface. In the transitional zone, collagen fibers have a less ordered organization with increased distribution of proteoglycans. In the radial zone, proteoglycan has the highest concentration but more variable in its distribution around cells. In the pericellular matrix of each chondrocyte, the concentration of proteoglycan can be two-fold higher than that in the matrix distant from the cells. These proteoglycan macromolecules are rich in polysaccharide side-chains with negatively charged carboxyl and sulfate groups that give articular cartilage a hydrophilic nature.

The spatial distribution of GAGs in cartilage can be visualized by using a number of histochemical assays (as well as MRI;⁸⁵  see chapter 6). Sulfated GAG content (both free GAGs and GAGs that are part of the proteoglycan matrix) can be measured using a 1,9-dimethylmethylene blue assay.⁶⁹  This assay is sensitive to interference from proteoglycan core proteins and non-GAG polyanions such as DNA and HA. These sources of interference can be minimized through the digestion of the core protein with endopeptidases and controlling the pH and osmolarity of the buffer, respectively. The safranin-O staining assay^86,87  is also a sensitive and easy method to quantify the GAG concentration. The assay involves the preparation of microtomed slices (10–20 µm) of articular cartilage, which are then dehydrated with ethanol and stained with 0.1% (w/v) safranin-O. Safranin-O is a cationic dye that binds to both sulfate and carboxyl groups and precipitates GAG. The precipitate is stable and tolerates interference from many other chemical species. The absorbance of the sample at λ = 475 nm is proportional to the local concentration of GAG. Reflectance measurement can also be used, subject to the qualification that reflectance obeys the non-linear Kubelka−Munk law and therefore requires linearization in order to quantify GAG concentration.⁸⁶  Further discussion of histochemical assays of cartilage can be found in chapter 22.

In addition to the depth-dependent macromolecules in cartilage (proteoglycans and collagen fibers), many other molecular concentrations in cartilage are depth-dependent. A recent study quantitatively analyzed extractable proteins in human lateral tibial cartilage, by first applying a non-targeted mass spectrometry approach (iTRAQ: isobaric tags for relative and absolute quantitation) and subsequently by analyzing protein distribution using a targeted multiple reaction monitoring mass spectrometry. The unique distribution patterns of 70 ECM proteins were identified in the lateral tibial plateau, revealing groups of proteins with a preferential distribution to the superficial, intermediate or deep regions of articular cartilage.⁸⁸  It is likely that most, if not all measurable quantities in articular cartilage have some types of depth-dependency.

Since the primary function of a synovial joint is load bearing and facilitation of mechanical motion, the biomechanical properties of articular cartilage are the ultimate measure of its health.⁵¹  Several characteristics describe the essential biomechanical function of articular cartilage:

1. Through its low friction, articular cartilage facilitates flexional and rotational movement within the joint.

2. Articular cartilage prevents direct contact between bones and other parts of the joint, thus protecting the relatively brittle bone from frictional wear and the softer tissues of the joints from abrasion.

3. Through deformation, articular cartilage distributes the forces exerted on the bone to a greater contact area. Known as load processing, this reduces contact stresses between the bones and protects the bones from fatigue.

Because of the unique molecular and morphological structures of articular cartilage, the biomechanical properties of cartilage in joints are both poroviscoelastic and depth-dependent.

In describing the intrinsic swelling nature of cartilage, Ogston proposed the idea that polysaccharide content is a primary component in maintaining the “inflation” of cartilage.^89,90  As described in Section 1.4, polysaccharides are the primary molecule that comprises the proteoglycan subunits and the heavily sulfated GAG chains of the bottlebrush shaped proteoglycan aggregate (Figure 1.4b). The sulfated parts of the GAG molecules are highly negatively charged, and through electrostatic repulsion cause the proteoglycan bottlebrush structure to expand and straighten into a rigid form that takes up a large molecular domain with a highly negative fixed charge density (FCD).⁹¹  The fixed charges are subject to the rule of electroneutrality, which states that a charge (e.g. ion) cannot be without a counter-charge for an extended period, and causes counter-ions and co-ions, namely Na⁺ and Cl⁻, to be pulled into the tissue. The counter- and co-ions cloud the negative fixed charges and become electrostatically bound to the proteoglycan molecule (Figure 1.5). This causes a greater ionic concentration in the ECM with respect to the external environment, causing water influx of the tissue, which generates an outward osmotic stress known as the Donnan osmotic swelling pressure.⁹²  The Donnan swelling pressure has been experimentally found to be 0.05–0.35 MPa.⁶² 

Collagen fibrils have an extremely high tensile modulus, individually found to be 850 MPa,⁹³  but can reach values of several GPa,^94,95  depending on their environment. The collagen fibrils provide the required structural resistance to swelling pressures, but although strong in tension they provide little resistance to compression due to the high length-to-width ratio.⁹⁶  The Donnan osmotic swelling pressure induces a “pre-stress” in the collagen network,⁹⁷  which generates hydrostatic pressure (see chapter 2).⁹⁰  As described in the following paragraphs, osmotic and hydrostatic pressure provides most of the load support when cartilage is subject to compression.⁹⁸ 

When a load is applied to the cartilage surface, the force is distributed across the contact area, generating pressure (stress) and resulting in tissue compression.^99,100  The modeling of the compressive behavior of cartilage has been the subject of considerable scientific debate.^101–103  The response of cartilage to applied mechanical load has been variously described as elastic, viscoelastic, poroelastic, poroviscoelastic, porohyperelastic, and poroviscohyperelastic.¹⁰⁴  Much of the controversy originated in the fact that the response is strongly dependent upon the strain rate. Different strain-rate regimes can be approximated by different mechanical models, and conversely, for any given regime, several models may yield very similar numerical results.^104,105  Certain features of the compressive response of cartilage are well described by the viscoelastic model, e.g., the time-dependent strain under a step applied load, or the time-dependent stress under a step strain.¹⁰⁶  But its general description requires elements of the poroelastic model,^107–109  which is based on the Biot theory of consolidation.¹¹⁰ 

The poroelastic model of articular cartilage can be qualitatively encapsulated as follows. Because water is incompressible, compression can only occur as a result of an outflow of water from the tissue. The solid matrix of the ECM is a porous–permeable material that allows the flow of interstitial fluid. Although cartilage ECM has a high water content, varying from 80% to 65% from the articular surface to the subchondral bone,⁷⁹  it has low permeability and is therefore highly resistant to fluid flow. The frictional interaction between the pore walls and interstitial fluid generates large shear forces that dampen the outflow of water through the tissue.^111–113  The rate of the outflow can be described by Darcy’s law.¹⁰⁷  The compression rate of the tissue is therefore limited by the migration of the fluid through the ECM, which in turn is controlled by the porosity of the “solid” network and the viscosity of water.¹⁰¹  Upon application of constant load, the strain slowly increases and eventually plateaus. This is accompanied by load sharing between the fluid and the solid components of the ECM. The load is initially borne by the fluid component in the form of the hydrostatic excess pore pressure (HEPP). The HEPP initially increases, reaches a maximum, and then slowly decays as the load is transferred to the solid components of the ECM in the form of electrostatic and osmotic interactions. This compressive response is known as “consolidation” and can be characterized by two plots: strain vs. time and HEPP vs. time (see Figure 1.6).¹¹⁴ 

The viscous dampening of the flow of interstitial fluid is crucial to both the physiology and biomechanics of articular cartilage. Biomechanically, it provides load-carrying capacity through the slow transient response to load. Physiologically, the dampening provides protection for the chondrocytes and ECM, and even represents a major factor in controlling the metabolic behavior of chondrocytes and ECM maintenance.^115,116 

The poroelastic model provides a highly realistic physicochemical picture of cartilage compression. However, under many commonly used loading conditions (e.g. step applied load), the compressive behavior of cartilage can be empirically described as viscoelastic, such that there is a solid (elastic) and a fluid (viscous) phase in the biomechanical response to deformation and loading. Similar to the poroelastic model, viscoelastic response to a step stress or strain is time dependent and not instantaneous. In general, viscoelasticity is theoretically modeled using a spring–dashpot system to represent the elastic (solid) and fluid (interstitial water) phases, respectively. Several works have attempted to theoretically model the biomechanical behavior of cartilage using the viscoelastic model.^117–120  In 1980, Mow et al. developed a biphasic model from mixture theory, which assumed a porous-permeable solid component of the cartilage ECM and a viscous fluid. Similar to the pure poroelastic model, the compressive response of cartilage in the biphasic model was dependent upon the viscous fluid efflux from the porous–permeable solid phase of the ECM.¹²¹  It has been shown that, under certain assumptions, the biphasic model is equivalent to the poroelastic Biot consolidation model.¹²²  An extension of the biphasic model was the triphasic model, developed by Lai et al., which additionally allowed quantitative treatment of the FCD. The triphasic model considered a three-phase ECM consisting of a porous–permeable solid phase, fluid phase, and an additional ionic phase (counter-ion and co-ion species).¹²³ 

The two fundamental behaviors of biphasic/poroelastic materials are creep and stress–relaxation (Figure 1.6). The creep response (Figure 1.6a) involves a step-like load applied to the tissue and a resultant time-dependent (transient) deformation. As the tissue compresses, water is rapidly forced out of the ECM. Tissue volume and the molecular volume of aggrecan domains decrease, while the FCD and the concentration of counter-ion and co-ion species increase, causing increasing intermolecular repulsion and interstitial swelling stress. The tissue deformation and water exudation slow down as the repulsive and swelling stresses slowly reach an equilibrium with the external load.^91,96  Water exudation ceases when equilibrium is reached, which may take several hours depending on tissue thickness (the time to equilibrium varies inversely to the square of the thickness).¹²¹  Since creep is tissue response-mediated by interstitial fluid flow, tissue permeability can be calculated.¹²⁴ 

In a stress–relaxation experiment (Figure 1.6b), a displacement (strain) is applied to the cartilage and maintained at a constant value. As the tissue is compressed, water is forced out from the tissue. The reduction of water content leads to a time-dependent decrease in the interstitial stress as the tissue relaxes. The interstitial mechanisms involved in stress–relaxation are subtly different to those in the creep response, in that the relaxation process involves interstitial molecular rearrangement. The reorganization occurs since the fibrils closest to the compressed surface are more compacted than the deeper fibrils,⁹⁶  which generates a gradient pressure within the tissue that eventually equalizes. An interesting feature of stress–relaxation is that the peak stress is highly dependent upon the rate of compression (Figure 1.7). Since the shear stress from the solid–fluid interaction limits the fluid flow, a higher displacement rate creates a higher interstitial pressure peak.¹²⁵  If loaded slowly enough, the tissue will reach equilibrium without experiencing a spike in interstitial stress. The differences in peak loading stress govern stress redistribution between the ECM and the subchondral bone.¹⁰⁵  Under physiological loading conditions, peak interstitial stresses can be large, as much as 20 MPa in the human hip,¹²⁶  due to the small contact area of cartilaginous surfaces and the associated shear forces. For perspective, the pressure of 20 MPa is approximately 100 times the air pressure in the tires supporting an automobile, although automobile tires utilize a significantly larger surface area. Experimentally, it has been shown that interstitial water pressure can contribute to more than 90% of the initial load-carrying capacity at the cartilage surface.¹²⁷ 

The close association between proteoglycan aggregates, interwoven collagen fibrils and interstitial fluid provides the compressive resilience to cartilage through negative electrostatic repulsion forces.^78,128  Since the density of the FCD increases with depth from the surface,¹²⁹  and the fluid (water) content and permeability decrease with tissue depth,^62,130,131  the magnitude of compressive resistance in articular cartilage is highly depth-dependent. Using fluorescence-labeled chondrocyte nuclei as intrinsic markers, Schinagl et al.¹³²  measured the intra-tissue deformation within the cartilage matrix, and found that the compressive modulus increased significantly with depth from the articular surface, ranging from 0.079 MPa in the superficial zone to 2.10 MPa in the deepest zone. This depth dependence of articular cartilage is the reason that, when articular cartilage is compressed externally (either in vitro or in vivo), the surface regions of the tissue undergo compression before the deeper zones do (Figure 1.8).^133–135  A consequence of this depth-dependent compression is that an external loading on an intact joint damages the surface cells to a greater extent than the deep cells. (Further discussion of tissue−cell mechanical interactions in cartilage can be found in chapter 15.) A change in the depth dependence of GAG concentration, which is related to the FCD, has been suggested to alter the depth dependence of the biomechanical behavior.¹³⁵ 

The properties of articular cartilage discussed in the last several sections are generic in nature, i.e. no consideration is given to the extensive variations in these properties of articular cartilage from different sampling sites on a single joint, even if the sites are just a few millimeters apart. For example, the thickness of the superficial zone was ∼200 µm at the summit of the medial femoral condyle and increased to ∼600 µm at the periphery.²⁵  These site-to-site variations within a single joint are likely governed by the mechanical demands of a given joint in a given animal species. In addition to variations within a single joint, cartilage from different types of joints in the same animal may vary because of the different load-bearing patterns of the different joints. Likewise, cartilage from the same kind of joint in different animal species might also vary as a consequence of variations in their mechanics (e.g. human shoulder versus canine shoulder).¹³⁶ 

Across different animal species, a linear relationship of body mass to cartilage thickness was found, with femoral condyle cartilage thickness ranging from 90 µm in mouse, to 2000 µm in humans, to 3000 µm in an Asian elephant with the medial condyle being on average thicker than the lateral condyle.¹⁶  This study also found a non-dependency of GAG to the body mass of different species but a negative relationship of DNA (chondrocytes) to body mass. Since different species have a different distribution of applied stress in their joints, the properties of articular cartilage should be considered different between species.¹³⁷ 

The development of a synovial joint can be divided into two stages. Initially, early limb buds develop from the somatopleural mesoderm of the embryo. The contiguous mesenchyme condensations in the early limb buds undergo chondrification,¹³⁸  and then separate by transverse bands of relatively flattened cells, which is known as an interzone.¹³⁹  Previous studies suggest that the expression of Hox and Cux1 determine the position of the joints^140,141  and Wnt14 probably demarcates the limits of the joint interzone.¹⁴²  The interzone is presumed to divide into three layers (two outer chondrogenic layers and an intermediate layer). The band of CD44-expressing cells could only present in the intermediate layer. Subsequently, the interzone cavitates to form the joint cavity, followed by the morphogenesis of an interlocking structure.¹⁴³  The presence of hyaluronan-binding protein CD44 on the interzonal cells represents the first sign of cavitation, and the massive upregulation of HA, mainly depended on movement (mechanical stimuli), would promote cell separation by saturation of HA-binding proteins, accounting for the loss of tensile properties in the ECM.¹⁴⁴  Moreover, lytic enzymes increase in the interzone, involved in the breakdown of matrix components, and promote the cavitation process.¹⁴⁵  The periphery of the interzone, destined to become synovium, becomes vascularized, but the center remains avascular. Following cavitation, the skeletal elements undergo morphogenesis (Figure 1.9).¹⁴⁶  In the rest of this section, any joint mentioned is a mature joint in an adult.

The adult knee (tibiofemoral) joint is probably the most studied joint in osteoarthritis research, both because of the high incidence of osteoarthritis associated with this joint and because of its easy surgical access. As shown in Figure 1.1, the knee joint is a complex mechanical organ, the smooth movement of which is supported and controlled by multiple structural components (femur, tibia, patella, ligaments, and tendons). Different surfaces of cartilage within the knee have different modes of contact and motion (e.g. rolling, sliding, and twisting) during physiological loading (e.g. walking), likely tailored by different reactions to an applied load.¹⁴⁷ 

Figure 1.10 shows the top view of a canine tibia, where the topographical surface variation is clearly visible over the entire tibial plateau. The cross-sectional slices from anterior to posterior directions (the coronal planes) reveal clearly the significant topographical (site to site) variations of both lateral and medial tibial plateaus, which likely provides effective load dissipation.¹⁴⁸  There are large variations in the cartilage thickness from lateral to central direction on the lateral plateau and central to medial on the medial plateau. In canine tibia, cartilage around the central region (not covered by the meniscus) is thicker than the periphery (covered by the meniscus). Cartilage thickness also varies from the anterior to posterior direction. Topographical variations have been found in many properties of tibial and condyle cartilage, including thickness, shear modulus, birefringence, water content, sulphated-GAG content, and collagen content.^25,149–156  An example of multidisciplinary correlation among the tibial parameters can be found in chapter 17.^155–157  As osteoarthritis progresses, the differences between healthy and osteoarthritis tissues could become increasingly site-specific and load-dependent. Even with the progress of age, there are site-specific changes in water and collagen content in asymptomatic cartilage.¹⁵⁸ 

Cartilage from the adult shoulder joint (a ball and socket joint between the scapula and the humeral head) has also been studied in many biomedical studies, which in animal models is probably due to its availability as a spare tissue. Compared with knee cartilage, humeral cartilage has much simpler contours, which has be used as a classical three-layer model of articular cartilage in a number of studies.¹⁵⁹  However, various topographical variations are also known in humeral cartilage, including morphological, biochemical, and biomechanical properties.^{147,160–167}  These variations are also age- and sex-dependent,^160,168,169  which show the complexity of the issue.

A peculiar graphical pattern can be formed on the intact surfaces of many joints, by using a sharp pin to prick the surface of articular cartilage and coating the cartilage surface with a layer of India ink.^170–172  After the ink is wiped off, it can be seen that a pattern of elongation lines has been formed over the joint surface (Figure 1.11). This pattern clearly represents some type of surface anisotropy of articular cartilage over the joint, which could be the anisotropy of molecular structure as well as biomechanical nature. Some have considered the pattern to reflect the directions of the surface fibers,¹⁷³  which could represent the stress patterns on the surface of any joint.^174,175 . (Incidentally, the collagen fibers in symphysis fibrocartilage also tend to follow the stress lines.⁶ ) However, some studies revealed little evidence that the split-line direction correlated strongly with any preferred alignment of fibrils.¹⁷⁶  Although it is an invasive approach, the split-line pattern and its modification due to lesion or during growth could potentially be explored in in vitro studies of the surface structure of a joint.

Osteoarthritis is a progressively degenerative joint disease with a massive socioeconomic burden, as it is the leading cause of disability.^2,177  Osteoarthritis is thought to be initiated when an imbalance occurs between chondrocyte-controlled anabolic and catabolic processes. This imbalance is characterized by ECM degradation, tissue loss, joint space narrowing, subchondral bone sclerosis, and osteophyte formation.¹⁷⁸  Classical symptoms of a later stage osteoarthritis in humans include joint pain, deformity, restricted motion, and dysfunction.

Osteoarthritis can be classified into two types: primary and secondary. There is little difference in their clinical symptoms, pathogenesis, and treatment, but there are differences in their triggers. Primary osteoarthritis, which is also known as “wear and tear” or idiopathic osteoarthritis, has no known cause and occurs mostly in seniors. Repetitive use of the joints over five or six decades may result in degeneration of articular cartilage in a large percentage of the senior population. Secondary osteoarthritis usually has an identifiable cause, such as trauma (e.g. sports injuries or vehicle accidents), obesity, genetic and developmental joint abnormalities, or hormone disorders (e.g. diabetes or gout). Secondary osteoarthritis tends to strike at an earlier age. It should be noted that some “primary” osteoarthritis patients might forget one or several sub-acute traumas or injuries many years earlier, which had actually contributed to the initial degradation of his or her cartilage and joint.

The molecular mechanisms underlying the pathogenesis of osteoarthritis are still not fully understood and there is no widespread agreement. Some recent studies relate osteoarthritis pathogenesis to the re-initiation of the transient chondrocyte phenotype (as seen in terminal differentiated growth-plate chondrocytes) and the upregulation of collagenase matrix metalloproteinase (MMP)-13.¹⁷⁹  Epigenetics has been implicated in osteoarthritis, with three known mechanisms, including DNA methylation, histone modifications and non-coding RNAs presenting a capacity to control the chondrocyte phenotype. MMP-13-promoted methylation in osteoarthritic cartilage, in part, may drive the chondrocyte hypertrophy.¹⁸⁰  Furthermore, the breakdown of the ECM in cartilage is thought to be mediated by a number of MMPs (e.g. MMP-3 and MMP-13) and aggrecanases (adisintegrin and metalloproteinase with thrombospondinmotifs (ADAMTS)-4 and ADAMTS-5).¹⁸¹ 

Although osteoarthritis is not classified as an inflammatory joint disease, several inflammatory components (e.g. cytokines, immunoglobulins, and prostaglandins (PGs)) have been implicated in the progressive erosion of cartilage.¹⁸²  Interestingly, chondrocytes from osteoarthritis patients express higher levels of IL-1 receptor type I than normal controls.¹⁸³  Thereby, osteoarthritic chondrocytes are highly sensitive to the effects of IL-1β, which can induce the expression of matrix-degrading enzymes such as MMPs.¹⁸⁴  Studies have shown that osteoarthritis cartilage spontaneously releases cyclo-oxygenase-2-dependent prostaglandins, especially (PG)E2, 50-fold higher than in normal cartilage and 18-fold higher than in normal cartilage stimulated by cytokines.¹⁸⁵  The quantity of PGE2 could promote cartilage damage by exacerbating joint inflammation, and inducing IL-1b expression. In addition, nitric oxide (NO) can be spontaneously produced by osteoarthritis-affected cartilage to cause catabolic effects, including inhibiting PGE2 synthesis as well as activating MMPs.¹⁸⁶ 

Along with the progressive loss of articular cartilage, osteoarthritis is characterized by increased subchondral bone sclerosis with thickening of cortical plate and formation of osteophytes,¹⁸⁷  which at some stages can be diagnosed as bone bruises and edema via medical imaging. Like cartilage, the subchondral bone of osteoarthritis patients releases high levels of alkaline phosphatase, osteocalcin, osteopontin, IL-6, IL-8, and progressive ankylosis protein homolog, PG, and insulin growth factor-1.¹⁸⁸  These biochemical factors not only contribute to the anabolic activity of subchondral bone cells, but also directly interact with chondrocytes, leading to a shift towards increased cartilage catabolism, chondrocytes hypertrophy and ECM calcification through increased vascularization and microcracks.¹⁸⁹  For example, hepatocyte growth factor secreted from subchondral bone can diffuse to the intermediate and deep layers of articular cartilage and induce the expression of collagenase-3 or MMP-13.¹⁹⁰  In addition, the high turnover rate of osteoarthritis-affected subchondral bone would lead to a higher proportion of immature bone (including increased narrowing of fibril diameters, reduced level of pyrrole crosslinks and decreased mineralization), all contributing to reduced mechanical properties and finally resulting in cartilage destruction upon repetitive loading.¹⁹¹  Therefore, the subchondral bone both affects the response to mechanical loading of articular cartilage and influences the progression of joint degeneration.

The current gold standard for diagnosing and measuring clinical efficacy in osteoarthritis is radiographic joint space narrowing.¹⁹²  However, it may not be a reliable measure since it could be subject to the pre-imaging loading of the patient (e.g. walking or sitting). For imaging biomarkers, with the exception of plain radiography, MRI is recommended to detect all the joint tissues, including quantitative cartilage morphometry, cartilage defects, and bone marrow lesions on semi-quantitative analysis, bone shape/attrition, and subchondral bone area¹⁹³  (see chapters 22–24). The elevated presence of biochemical biomarkers measured from the serum, urine, and synovia of osteoarthritis patients is often accompanied by cartilage degradation and subchondral bone turnover, such as cartilage oligomeric matrix protein, c-terminal telopeptide of type II collagen, helical fragments (Helix-II and Coll 2-1, Coll 2-1 NO2), amino-terminal type II procollagen propeptide, carboxy-terminal type II procollagen propeptide, chondroitin sulfate 846 epitope (CS 846), HA, n-terminal type I collagen telopeptides, c-terminal type I collagen (CTX I) or serum CTX I, amino-terminal procollagen propeptide of type I collagen, carboxy-terminal procollagen propeptide of type I collagen, osteocalcin, urinary total pyridinoline, bone sialoprotein, and MMP-3.^194,195  Some of these biochemical biomarkers could be used to detect the earliest tissue lesion before the degradation is diagnosed as a clinical disease.

The aims of osteoarthritis management are to educate patients about the disease, to alleviate pain, and to improve joint function. Osteoarthritis should be managed on an individual basis and commonly requires a combination of treatment options. The recommended hierarchy of treatments consists of non-pharmacological treatments first, then drugs, and then, if necessary, surgery. The non-pharmacological approach includes education, weight loss (e.g. diet and exercise), physical therapy (e.g. motion exercise, muscle strengthening and muscle stretching), knee braces, and orthotics.^196,197  The pharmacological approach involves analgesics, such as paracetamol (for mild-to-moderate pain), non-steroidal anti-inflammatory drugs (NSAIDs) (sometimes the first choice),¹⁹⁸  opioid analgesics (an alternative in those patients for whom NSAIDs are contraindicated, ineffective, or poorly tolerated), and intra-articular steroids (acute exacerbations of pain and signs of local inflammation),¹⁹⁹  as well as intra-articular hyaluronan.²⁰⁰  Surgery is indicated when symptoms can no longer be managed by other therapies. Surgery may include arthroscopic debridement and lavage,²⁰¹  osteotomy,²⁰²  joint replacement,²⁰³  allografts, autografts, autologous chondrocyte transplantation,²⁰⁴  and tissue-engineered cartilage transplantation. The latter has received significant attention and is currently focused on selecting the seed cells (e.g. MSCs) and recreating the natural physical environment (e.g. 3-dimensional cell-seeded scaffolds).²⁰⁵ 

Osteoarthritis in humans has been extensively studied clinically. Since articular cartilage is avascular and aneural, a diagnosis of primary osteoarthritis is not made until a patient experiences pain in the joint, which signals later-stage disease. The main focus for human osteoarthritis research is to find a way to diagnose the disease early and accurately. The best-known human imaging project during the recent years is the Osteoarthritis Initiative (OAI), which is a multi-center observational study of human osteoarthritis sponsored by the National Institutes of Health in the USA. The OAI data are available publically on the internet. Many correlational studies have been published using various parts of the OAI data.

In addition to studies on human osteoarthritis, a large number of different domestic animals have been used in osteoarthritis studies.²⁰⁶  However, none of the animal models reproduce all the features of human osteoarthritis.²⁰⁷  A general consensus in the community is that there is no universal animal model to recapitulate the pathogenesis and progression of human osteoarthritis.^206,207  Despite the limitations, animal models are an integral and irreplaceable component of human osteoarthritis research. This is because no disease can be induced ethically in the human; the only alternative for the entire biomedical community is to study the cellular and animal models of the human disease, in order to gain better insights that could be used to produce guidelines for human treatment.

Primary osteoarthritis progresses slowly over several decades. It is the most common form of osteoarthritis and increases in prevalence and severity with the age in both human and non-human animals.²⁰⁷  Compared to the idiopathic and slow development of primary osteoarthritis, post-traumatic osteoarthritis in animals and humans is known to be induced as a consequence of injury, and often develops quite rapidly. Post-traumatic osteoarthritis can be reproduced by mechanical insult or by surgery. Since articular cartilage has a limited capacity for repair, the physiological response to the resulting tissue damage rarely restores a normal articular surface. Similarly, any changes in the cartilage structure caused by abnormal loading in a stable joint, or even by degradative enzymes in the synovium, can lead to the development of osteoarthritis. The assessment of animal models of osteoarthritis traditionally depends on the histological analysis of articular cartilage. Recent advances have seen the emergence of many other effective analyses, such as biomarkers and imaging, for monitoring the progress of osteoarthritis.^{81,154–156,206,208,209} 

Although there is a perceived advantage in using naturally occurring models of osteoarthritis (e.g. Dunkin–Hartley guinea pigs) with a slower onset and progression similar to human osteoarthritis,²¹⁰  many investigators have utilized fast-advancing animal models first reported by Magnuson in 1941.²¹¹  With a division of the medial collateral ligament and the cruciate ligaments in the joint, Magnuson showed osteoarthritis-like changes in articular cartilage. In 1952, Paatsama published a thesis describing the degradation of canine articular cartilage associated with a cranial cruciate ligament rupture.²¹²  Many studies from the 1970s onwards that used a similar model, or only the transection of the anterior cruciate ligament or a removal of the meniscus (menisectomy) showed reproducible histopathological changes including fibrillated articular surface, depleted proteoglycan content, and cloning of chondrocytes.^206,207,210 

In addition to cartilaginous changes, Radin and Rose¹⁹¹  showed alterations in the sub-chondral bone and the calcified zone of cartilage after application of repetitive sub-impact loads to the patellofemoral joint. Responses to traumatic injury were also studied using a direct impact applied to the patella using a "drop tower" type of apparatus in order to produce high-energy damage similar to that seen in human injuries. Intra-articular injections of many proteolytic enzymes such as trypsin, papain, and collagenase have also been used to trigger model osteoarthritis in mice, rats, and rabbits. However, the mechanisms responsible for cartilage degradation in these models, particularly those in which papain or trypsin was injected, may deviate significantly from those that normally occur in the human disease. Chemically induced osteoarthritis models use intra-articular injection of sodium iodoacetate to study acute cartilage toxicity and degradation and joint pain; however, this has many limitations as a model of osteoarthritis.²¹³  For example, since sodium iodoacetate is a metabolic poison, this model exhibits extensive chondrocyte cell death, unlike naturally occurring osteoarthritis in humans. However, it does provide an in vivo model of rapid cartilage degradation mirroring some of the events observed in vitro in organ culture screening studies.²¹⁴ 

In contrast to in vivo animal studies concentrating on the outcomes of post-traumatic joints and progression of osteoarthritis, many in vitro models have been used to study the effect of cell death and specific degradative mechanisms in well-defined loading and culture environments.^{208,215–220}  Many enzymes have been used to induce cartilage degradation in vitro. For example, to investigate the potential use of imaging as a diagnostic tool for osteoarthritis, purified collagenase, trypsin, or chondroitinase ABC^217,218  have been used to digest bovine patellar cartilage in order to generate spatial and temporal changes in the cartilage matrix. Similar changes in structure and collagen network were observed in proteoglycan-depleted tissue and correlated directly with the loss of compressive strength.²¹⁵ 

As an alternative to enzymatic cleavage, many studies have used in vitro explant injury models to study the molecular mechanisms of chondrocyte death and matrix degradation in injured cartilage. The relatively low compressive moduli and the compression-induced stiffening in the superficial zone are closely related to cell death following a blunt impact or repeated mechanical insults. This was well demonstrated in vitro in a study of chondrocyte necrosis, where chondrocyte death occurred only in the superficial zone when two cartilage disks were positioned articular-surface-to-articular-surface and subjected to 1.0 MPa cyclic compression for 24 h, as shown in Figure 1.12.²¹⁶  The vulnerability of cells in the superficial zone is a feature often seen in early osteoarthritic cartilage and is significant in the initiation of post-traumatic osteoarthritis.

Some aspects of cartilage repair can also be examined in vitro by osteochondral models, such as defects of different depths created using a dermal biopsy punch and a scalpel.²⁰⁹  Lin et al.²⁰⁸  observed progressive changes in cell viability, collagen cleavage, and proteoglycan loss by cyclically loading cartilage explants with 1 and 5 MPa stresses for 24 h. Thibault et al.²²¹  subjected cartilage explants to high but physiological cyclic load levels and characterized the resulting damage using a sequence of unconfined-compression stress-relaxation tests. These studies indicated that acute injury in articular cartilage can induce an upregulation of reactive oxygen species and pro-inflammatory cytokines. These are important areas of research, since the prevention of cell death and the inhibition of matrix-degrading enzymes in the injured joint are significant for the prevention of posttraumatic osteoarthritis. Together, these in vitro explant models provide effective systems to study biomechanical and mechanobiological factors involved in initiating cartilage injury; biochemical factors associated with cell death and matrix degradation; and gene regulation critical for the advance of post-traumatic osteoarthritis.

Y.X. thanks the four R01 grants from the National Institutes of Health (NIH NIAMS AR045172, AR052353, AR069047), which support his cartilage research at Oakland University. K.I.M. acknowledges cartilage research funding from the Australian Research Council (Discovery Project grant DP0880346) and Australian Institute of Nuclear Science and Engineering (grants ALNGRA14051 and ALNGRA15025). Z.C. thanks Dr Yong Lu and the travel grant from Shanghai JiaoTong University School of Medicine (Shanghai 200025, China), which enables her travel to Oakland University (Michigan, USA) for cartilage research.