Osteoarthritis

This chapter is dedicated to Dr. Nancy Burton-Wurster, Cornell University, whose intellectual drive and hard work continues to inspire young minds to pursue investigations into the basic science of osteoarthritis in dogs.

Osteoarthritis (OA) is a degenerative process that leads to changes in all components of a synovial joint: articular cartilage, subchondral bone, synovial membrane, synovial fluid, and periarticular soft tissues. Osteoarthritis affects up to 20% of dogs over 1 year of age [1]. Although idiopathic (primary) osteoarthritis is rare in dogs and cats, secondary osteoarthritis is common and usually due to underlying factors such as prior trauma, joint incongruency, joint instability, inflammation, or developmental conditions. Pathologic changes typically consist of variable synovitis, joint capsule thickening, cartilage destruction, subchondral sclerosis, and the production of new bone (osteophytes and enthesiophytes). Because several reviews of OA have been written in veterinary textbooks, this chapter focuses also on information on the subject especially related to biomarkers.

Anatomy and Normal Joint Physiology

Diathrodial joints are an articulation between adjacent bones of the skeleton that are characterized by a surface of articular cartilage, a joint capsule, and synovial fluid secreted into the joint cavity that lubricates the contact surfaces of the joint. Synovial joints can be classified according to the number of articular surfaces they contain (simple or compound joints), the shape or form of the articulating surfaces (plane, ball and socket, ellipsoidal, hinge, condylar, trochoid, or saddle joints) or by the function of the joint [2].

The fibrous joint capsule of a diarthrodial joint is lined by a synovial membrane. The outer fibrous layer consists primarily of collagen but also contains the vascular and nervous supply to the joint. It attaches peripherally to the articular cartilage and blends with the periosteum of each bone [2]. The synovial membrane is vascular and consists primarily of loose areolar tissue with an inner layer of cells that is one to two synoviocytes thick (Fig. 116-1). The two types of synoviocytes are type A, which are primarily phagocytic, and type B, which are primarily secretory. The synovial membrane covers all structures within the joint except the articular cartilage and the contact surfaces of any fibrocartilaginous plates (menisci) and blends with the periosteum as it reflects onto bone [2].

Figure 116-1. A. Photomicrograph of a cross-section of normal synovial membrane. The synovial cavity is between the two pieces of synovial membrane on each side of the photograph. B. Synovial hyperplasia characterized by synovial lining proliferation. C. Chronic synovial hypertrophy characterized by fibroplasia. (A,B,C: Hematoxylin and eosin, X160.) D. Photomicrograph of normal articular cartilage showing tissue sparsely populated with chondrocytes that elaborate a profuse extracellular matrix. Chondrocytes in the deep zone tend to align perpendicular to the articular surface (on top). More intense uptake of stain (which binds the sulfated glycosaminoglycan) is seen pericellularly in the deeper layers. E. Photomicrograph of fibrillated articular cartilage showing decreased staining of the extracellular matrix in a dog in the early stages of hip dysplasia. (D,E: Safranin O/fast green stain, X180.) E. Cloning of chondrocytes in advanced osteoarthritis. F. Photomicrograph of articular cartilage from a dog with advanced osteoarthritis secondary to hip dysplasia. Surface irregularity, fissures, and chondrocyte cloning can be seen (Safranin O/fast green stain, X200).

The primary functions of articular cartilage are to promote motion with minimal friction and to transmit load to the underlying subchondral bone. Hyaline cartilage is composed primarily of matrix and contains relatively few cells (Fig. 116-1). Chondrocytes make up less than 5% of the tissue volume [1]. Each chondrocyte, in association with its pericellular capsule and pericellular matrix, forms a unit described as a "chondron" [3]. Cartilage is composed of 70% to 80% water by weight, and is avascular, aneural, and alymphatic (Fig. 116-2). The primary large molecules comprising articular cartilage matrix are collagen and proteoglycan, and it is these two molecules that give cartilage its unique biochemical and functional properties. Collagen provides tensile strength, whereas proteoglycan provides compressive strength. Both molecules are produced and secreted locally by chondrocytes.

Figure 116-2. A. Pie diagram showing the biochemical composition of the articular cartilage. B. Pie diagram showing the fixed charge density of the proteoglycans, water, and the swelling pressure constrained by the collagen fibrils. C. Diagram of articular cartilage illustrating the selective orientation of the collagen fibrils in normal articular cartilage and the transition through the calcified cartilage to the subchondral bone.

The classic zonal description of articular cartilage is based on chondrocyte organization, orientation of collagen fibrils, and proteoglycan distribution. Zone one, the tangential zone, is the most superficial region and is characterized by relatively few flattened cells, low proteoglycan content, and collagen fibrils that are oriented tangentially to the articular surface (Fig. 116-2). This distribution of collagen and proteoglycan provides the superficial layers of cartilage matrix the greatest ability to withstand tensile forces and resist the swelling pressure exerted by proteoglycan located in the deeper zones [1]. Zones two and three comprise most of the matrix volume. In this region, the density of chondrocytes and proteoglycan content increase with increasing depth in the tissue (Fig. 116-1). The cells become more ovoid and in zone two, the transitional zone, are randomly distributed. The direction of the collagen fibrils changes gradually from a tangential orientation to an oblique orientation, and ultimately the fibrils become radially aligned to the articular surface (Fig. 116-2). In zone three, the radial zone, the cells line up in vertical columns within the matrix. It is the increased proteoglycan content in these deeper regions that provides most of the ability of cartilage to resist compressive load. A distinct separation, known as the tidemark, exists between zones three and four, at the upper limit of zone four, the calcified zone. It contains radially oriented collagen fibrils, but little proteoglycan. The calcified cartilage is separated from the underlying subchondral bone by a cement line. The osteochondral junction is maintained by the morphology of the interdigitating boundary, the undulating nature of which allows shear stresses to be converted into potentially less damaging compressive forces on the subchondral bone [4].

Collagen composes 50% of the dry weight of mature articular cartilage. There are at least 27 different types of collagen [5]. Collagen fibrils are composed of collagen monomers, which in turn are composed of three polypeptide alpha chains arranged in a triple helix. Genetically different alpha chains lead to different types of collagen monomers. Type II collagen is the primary form of collagen in articular cartilage. Other types of collagen that are known to be important in the normal function of articular cartilage are types VI, IX, X, and XI. Type VI collagen is found in the pericellular region of the chondron and may be involved in binding the cell surface to matrix constituents. Type IX links type II fibrils together and is important in the association of collagen and proteoglycan within the matrix. Type X collagen is found during development in the hypertrophic cartilage and in the calcified zone of adult cartilage [1].

Proteoglycans make up 22% to 28% of the dry weight of adult articular cartilage (Fig. 116-2). A proteoglycan molecule is composed of a core protein with glycosaminoglycan side chains. The glycosaminoglycans are chains of repeating disaccharide units of variable length and are negatively charged owing to carboxyl and sulfate groups [4]. This intrinsic ionic charge leads them to repel one another, resulting in the "bottle brush" conformation of the proteoglycan molecule and contributing to its hydrophilic nature. The most common glycosaminoglycans in articular cartilage are chondroitin-6-sulfate (mature articular cartilage), chondroitin-4-sulfate (immature articular cartilage), keratan sulfate, and dermatan sulfate. Proteoglycans may be classified as aggregating or non-aggregating, depending on the ability of the proteoglycan monomer to attach to a hyaluronan backbone by a glycoprotein link. The major aggregating proteoglycan of articular cartilage is aggrecan, whereas the major non-aggregating proteoglycans are decorin and biglycan. Hyaluronan is a non-sulfated glycosaminoglycan that has no core protein. It is found in both the articular cartilage extracellular matrix and the synovial fluid. Hyaluronan is secreted into the extracellular matrix by chondrocytes, where it forms a chain by the noncovalent interaction of hyaluronan monomers, and can aggregate with aggrecan monomers. Hyaluranan is secreted into the synovial fluid by type B synoviocytes where it functions in boundary lubrication, but does not appear to be important in cartilage-on-cartilage lubrication [1].

The ionic charge of proteoglycan attracts water (Figure 116-2). Proteoglycans can occupy a volume up to 50 times their dry weight volume when hydrated [1]. However, because proteoglycans are trapped in a collagen matrix, their expansion is limited to within 20% of their potential volume, keeping cartilage turgid. It is this relationship between proteoglycan and collagen and the free motion of fluid that allows cartilage to resist deformation and transmit load. When cartilage experiences a compressive load, fluid flows slowly through the collagen meshwork and the pores created by the proteoglycan molecules. The ability of the cartilage to tolerate load is dependent on the interconnections of collagen fibrils (Fig. 116-2). If these connections are broken, propagation of fissures through the matrix occurs, leading to morphologic changes characteristic of OA. To a certain extent, articular cartilage adapts to the predominant stress level it experiences and is stiffer in areas of high stress and has increased proteoglycan content [4]. Excessive stress in areas of softer cartilage may result in matrix damage and OA.

Because adult articular cartilage is avascular, it is dependent on the synovial fluid for delivery of nutrition, removal of waste, and circulation of synovial lining cells and leukocytes. Synovial fluid is an ultrafiltrate of plasma, to which large molecules such as hyaluronan have been added by synoviocyte secretion. The viscosity of synovial fluid depends on the molecular weight and concentration of hyaluronan[4]. Joint motion leads to the mechanical compression of cartilage and encourages diffusion of nutrient molecules into the cartilage matrix. In immature animals, nutrients may diffuse into the deep zones of articular cartilage from metaphyseal blood vessels because the calcified cartilage and subchondral plate have not formed. Joint lubrication and the ability of cartilage surfaces to deform under load produces a low coefficient of friction. Under conditions of load, some water is forced out of the matrix, resulting in a weeping of fluid onto the articular surface (hydrostatic lubrication), until an equilibrium is reached between the osmotic force of proteoglycans and the compressive force applied [1]. During weight bearing, articular cartilage lubrication is hydrostatic (squeeze-film) and most important under heavy load whereas boundary lubrication predominates under light load. Load leads to fluid exudation from the matrix, resulting in a thin film of fluid that separates articular surfaces. Boundary lubrication is adherence of a molecular film of lubricant to surfaces of articular cartilage and synovium, which separates opposing surfaces. The boundary lubricant of synovial membrane is hyaluronan. The boundary lubricant of articular cartilage is a glycoprotein called lubricin and its associated protein.

Pathophysiology and Pathology of Osteoarthritis

Osteoarthritis can originate by one of two mechanisms: normal forces acting on an abnormal joint (e.g., in the case of osteochondrosis, hip or elbow dysplasia, or patellar luxation) or abnormal forces acting on a normal joint (e.g., joint trauma resulting in fracture or luxation) (Fig. 116-3). Regardless of the specific inciting cause, alterations in the function of the joint lead to specific molecular and cellular changes that result in the ultimate dysfunction of cartilage and periarticular structures that is observed clinically [1]. Changes in cartilage gene expression have been detected as early as 2 weeks after cranial cruciate ligament transaction in a canine model of OA, prior to detectable changes in glycosaminoglycan or collagen content, or evidence of gross or histologic pathology [6]. Joint insult results in release of pro-inflammatory agents such as interleukin 1α ( IL-1α) and β and tumor necrosis factor-α (TNF-α) by chondrocytes, synoviocytes, and infiltrating inflammatory cells, resulting in synovitis. The synovitis is marked by synoviocyte proliferation more than neutrophilic infiltration. Synovial lining cells and leukocytes release destructive enzymes, free radicals, cytokines, and prostaglandins. Persistent inflammation alters cartilage metabolism and cartilage degradation follows. At least some of the pain associated with OA has been attributed to the synovitis [7].

Figure 116-3. Photographs of joint with developmental orthopedic conditions that all result in secondary osteoarthritis owing to surface irregularity or abnormal mechanical loading. A. The proximal radial head and ulnar articular surfaces in the elbow joint showing a slightly elevated triangular-shaped, fragmented medial coronoid process between the medial coronoid of the ulna and the radial head. The full-thickness articular cartilage defect at bottom right is artifact. B. Photograph of an osteochondrosis lesion on the medial humeral condyle in a dog’s elbow. C. Photograph of the femoral head of a dog with early hip dysplasia. Notice perifoveal articular cartilage and hypertrophy of the round ligament of the femoral head (bottom center). D. Photograph of the stifle of a dog with medial patellar luxation. Note full-thickness articular cartilage loss along the axial medial trochlea and the articulating surface of the patella (top right). E. Photograph of an osteoarthritic humeral head with full-thickness articular loss (ulcer) caudodorsally. F. Photograph of a stifle joint taken 6 months following experimental transaction of the cranial cruciate ligament. Osteophytes are present along the medial trochlear ridge. (Compliments of Dr. Steven Budsberg, University of Georgia).

Synovial fluid becomes less viscous as hyaluronan concentration drops and joint lubrication suffers. Release of cartilage fragments and proteinases worsens the inflammatory response and induces further breakdown of collagen cross links [8]. Changes in synovial membrane and synovial fluid may be reversible. Articular cartilage damage is usually irreversible and self-perpetuating. Osteoarthritis is a self-perpetuating destructive cycle involving all components of the joint. The release of leukocytes, prostaglandins, lysozomal enzymes, hyaluronidase, interleukin-1, leukotrienes, and proteinases propagate joint tissue destruction. Synovial fluid viscosity decreases owing to alteration, breakdown, and dilution of hyaluronan and other proteins. Biomechanical properties of synovial fluid are altered with suboptimal lubrication, and decreased diffusion of nutrients into articular cartilage ensues.

Biochemically, OA is characterized by reduction in aggrecan content, alteration in collagen fibril size and structure, and increased synthesis and degradation of matrix macromolecules in cartilage. Proteoglycan synthesis by chondrocytes increases initially but then proteoglycan levels drop [9]. Finally, interleukin-1 and TNF-α induce proteoglycan depletion in articular cartilage by increasing the rate of proteoglycan degradation, decreasing synthesis by chondrocytes, or both. Matrix catabolism is mediated by the effects of matrix metalloproteinases, aggrecanases, cathepsins (acidic proteinases), interleukin-1, TNF-α, hyaluronidase, and prostaglandins. Decreased proteoglycans are accompanied by increased water content, increased compressibility, and decreased stiffness of the articular cartilage. Loss of matrix support that results from collagen fibril degradation permits the matrix to swell.

In the early stages of OA, cartilage fibrillation occurs. Fibrillation (loss of the superficial layer [zone 1]) fundamentally alters the biomechanical properties of articular cartilage [10]. Fibrillation occurs early in the pathogenesis of OA; full-thickness articular cartilage loss follows. Abnormal stresses cause fissures to propagate to deeper layers (Fig. 116-1). Chondrocytes cluster and increase in cell size (Fig. 116-1). Erosion (uniform surface loss of articular cartilage) follows. Finally, in regions of full-thickness cartilage loss, subchondral bone becomes exposed and eburnated (the polished appearance of sclerotic subchondral bone).

In young dogs with articular cartilage fibrillation owing to hip dysplasia (Fig. 116-3), the subchondral and epiphyseal bone mineral content is increased as measured by quantitative computed tomography [11]. In dogs with cranial cruciate ligament deficiency, a marked and sustained rise in urinary excretion of collagen pyridinium cross links was reported, which are thought to derive at least in part from the degradation of mature collagen in bone. Calcitonin, a potent inhibitor of osteoclastic bone resorption, markedly reduced the urinary levels of pyridinium cross links in OA dogs [12]. In canine experimental OA, subcutaneous injection of calcitonin reduced the severity of cartilage osteoarthritis lesions assessed both grossly and histologically [12]. Interestingly, mean bone mineral density of the placebo group was 80% of the calcitonin-treated group [13].

Periarticular osteophytes can be seen as early as 2 weeks from the onset of osteoarthritis (Fig. 116-4 and Fig. 116-5). Osteophytes are bony exostoses typically developing at sites of joint capsular attachment. Osteophytes probably represent an attempt by the body to compensate for increased tension on joint capsular attachments owing to chronic synovial effusion, concomitant joint capsular distension, persistent joint instability, and a proliferative repair response. Osteophytes eventually become canalized and the marrow cavity communicates with that of the epiphyseal bone. Enthesiophytes are bone growths or dystrophic mineralization that extends into capsular, ligamentous, and regional soft tissue attachments (Fig. 116-4 and Fig. 116-5).

Figure 116-4. A. Cranial-caudal radiograph showing subchondral lucency and osteophyte formation on the medial elbow joint. B. Lateral radiograph of the same elbow, showing osteophyte formation on the proximal anconeal process and cranial proximal radial head. (Compliments of Dr. Peter Scrivani, Cornell University).

Figure 116-5. A. Photograph of the cranialcaudal view of the stifle showing osteophyte formation on the lateral tibial plates and soft-tissue proliferation on the medial stifle. These findings are consistent with cranial cruciate ligament disruption. B. Lateral radiograph of an osteoarthritic stifle joint of the same dog as in Figure 5A showing increased intra-articular mass and obliteration of fat pad, fabellar remodeling, and osteophytes in the proximal and distal patella and along the nonarticulating surface of the femoral trochlea. (Compliments of Dr. Peter Scrivani, Cornell University).

Synovial Fluid Analysis

Laboratory evaluation of synovial fluid may be useful in confirming a diagnosis of OA and ruling out other inflammatory or infectious joint diseases. Synovial fluid is evaluated grossly for color, clarity, and viscosity, and microscopically for cell count, cell type, and the presence of infectious agents [4]. The amount of synovial fluid varies according to the joint sampled and the size of the patient. The average volume of synovial fluid in the normal stifle joint of adult dogs, for example, ranges from 0.2 to 2.0 mL [2]. Synovial fluid volume in OA is generally increased, but may vary according to the chronicity of the condition and the degree of synovitis. Normal synovial fluid is clear, colorless to pale yellow, and contains less than 1000 nucleated cells/μl, whereas synovial fluid from an OA joint may be a clear to hazy pale yellow with nucleated cell counts ranging between 3000 and 5000 cells/μl. The predominant nucleated cell observed in both normal and OA synovial fluid should be the monocyte, and neutrophils should account for less than 10% of the nucleated cells observed (Fig. 116-6) [14]. Nucleated cell counts in synovial fluid greater than 5000 cells/μl, or cell counts with a higher proportion of neutrophils, indicate arthritis of an infectious or inflammatory etiology that may include an acute exacerbation of a chronic condition. In dogs, the normal synovial fluid cell counts may vary among joints, with the highest cell counts found in synovial fluid from the shoulder and stifle joints [15]. Synovial fluid cell counts from normal and OA stifle and shoulder joints in cats fall within the ranges described above [16].

Figure 116-6. Cells found in smears of synovial fluid. A. Polymorphonuclear cells should be less than 10% of the total number. B. Lymphocytes C. Synovial lining cell (probably macrophagic). D. Signet cells indicating joint injury or degeneration. E. Mitotic figure. F. Signet cell indicating joint injury. (Wright Giemsa stain, X160) (Compliments of Dr. Kathleen Freeman, Scotland).

Biomarkers

Because significant limitations exist in the ability of routine diagnostic methods to detect OA in its early stages or to evaluate subtle changes in progression of the disease in an individual, interest in the identification of endogenous substances as biomarkers for the OA process has arisen. Non- or minimally invasive evaluation of one or more of these substances may eventually allow better identification of individuals in the early stages of the disease, quantification of the severity of the disease and its progression, and objective evaluation of the efficacy of treatment. Biomarkers are typically products of articular cartilage synthesis or degradation. Biomarkers may be anabolic (a product of a synthetic process) or catabolic (a product of degradation) indicators. Alterations in the biomechanics or biochemistry of the joint lead to an imbalance of matrix degradation and synthesis, and the resulting substances that are released may be measured in cartilage, synovial fluid, and in other body fluids such as blood or urine. Concentrations of biomarkers in individual samples of synovial fluid are affected by the presence of joint effusion, clearance of the marker from the joint, exercise level, and circadian rhythms [4]. Although serum concentrations of markers are less affected by dilution, the value of systemic biomarker concentrations in the diagnosis and monitoring of an OA joint is controversial, as the concentrations in urine or blood will more likely reflect cartilage turnover in all joints and not solely a joint or joints of interest [17]. Additionally, although differences exist in the average values of these markers between normal and OA populations, the variation between individuals is large and the extensive overlap limits their use as a solitary diagnostic tool in any one individual patient. Better discrimination is obtained when markers are evaluated longitudinally or multiple markers are used in combination [18,19].

Three broad categories of molecular markers have been identified based on their origin and function during the OA process: agents related to mechanisms of cartilage degradation (matrix metalloproteinase-1, matrix metalloproteinase-3, tissue inhibitors of metalloproteinase [TIMPs], interleukin-1, and interleukin-6), the degradation products of cartilage (keratan sulfate, chondroitin sulfate, aggrecan fragments, cartilage oligomeric matrix protein, cartilage matrix glycoprotein), and components of articular cartilage released as part of an anabolic response to OA (chondroitin sulfate epitopes 3B3(-) and 7D4, link protein, collagen X, fibronectin) (Table 116-1). Fibronectin is an extracellular matrix glycoprotein through which cells interact with their surrounding matrix. Fibronectin isoforms ED-A and ED-B have previously been described as biomarkers in the rheumatic diseases [20,21]. Total fibronectin concentrations have been demonstrated to be elevated in the cartilage of dogs with OA secondary to hip dysplasia [22] and in the synovial fluid of dogs with OA secondary to cruciate ligament insufficiency [23]. In the latter study, total fibronectin concentration in synovial fluid was inversely correlated with the duration of clinical signs, suggesting that this marker is an indicator of early repair phase post-injury [23]. Elevation of total fibronectin in the synovial fluid of human patients with OA has also been reported.24 However, because fibronectins are found in nearly all body tissues, this lack of specificity makes the total fibronectin concentration in serum a poor solitary biomarker for osteoarthritis. (V+C)- fibronectin is a splice variant specific to cartilage so that its presence in bodily fluids implies a cartilage origin.25 It constitutes 50% to 80% of the total fibronectin present in articular cartilage [26]. A 10-fold elevation of this isoform in OA cartilage has been reported; however, a study of spontaneous OA in the stifles of clinical canine patients failed to demonstrate a significant elevation in synovial fluid (V+C)- concentration as compared with synovial fluid from a control population [23,27]. However, like cartilage oligomeric matrix protein (COMP) evaluated in the same study, (V+C)- fibronectin was elevated in the contralateral stifles, and its elevation could be an indicator of early disease [23].

Cartilage oligomeric matrix protein (COMP) is a pentameric matrix protein in the thrombospondin family. It is found primarily in cartilage, but has also been identified in ligament, synovium, tendon, and meniscus [28]. Its function is not known, but it may have a role in chondrogenesis and the interaction of the chondrocyte with its surrounding matrix [23]. Increases in serum and synovial fluid COMP have been reported in human and canine patients with OA [29-31]. Another canine study found that COMP levels in the synovial fluid of stifles with OA secondary to cranial cruciate ligament insufficiency were not significantly elevated compared with synovial fluid from a control population, but that COMP was elevated in the contralateral (unaffected) stifles of these patients. Given the increased risk of cranial cruciate ligament injury in the contralateral stifle when one stifle is already affected, this elevation may indicate preclinical disease and COMP may be a marker of early cartilage injury [23]. However, within the joint, COMP is also secreted by the synovium and increased concentration of COMP may reflect synovitis. Additionally, owing to its presence in other tissues, serum elevation of COMP is not considered specific for cartilage degradation and individual variation is high; therefore, the utility of COMP as a solitary biomarker is limited.

Chondroitin sulfate is a glycosaminoglycan that is a major constituent of aggrecan. In early OA, newly synthesized aggrecan molecules contain chondroitin sulfate side chains with increased chain length and altered structure that are identifiable utilizing monoclonal antibodies [32]. Two epitopes (3B3(-) and 7D4) have been most thoroughly investigated, but significant difference exists in the timing of elevation in synovial fluid concentration of these epitopes between species. Multiple canine studies have demonstrated significant elevations in the concentration of synovial fluid chondroitin sulfate after experimental induction of OA. Significant increases in the 3B3(-) epitope of chondroitin sulfate have also been demonstrated in the synovial fluid from OA joints in humans [17]. Expression of 7D4 appears to be more prevalent in OA cartilage and synovial fluid from humans, monkeys, and guinea pigs with OA joints and was found to be more elevated in patients with acute injury and early osteoarthritis [32]. In another study, synovial fluid concentrations of 3B3(-), but not 7D4, were elevated in human patients with chronic knee OA [33]. Although the 3B3(-) form of chondroitin sulfate is absent in healthy canine articular cartilage, it has been identified in canine cartilage during early OA [34,35]. A longitudinal study in a canine model found that an increase in 7D4 concentration preceded the rise in 3B3(-) concentration [36]. A correlation between 3B3(-) and 7D4 epitope concentrations in synovial fluid from a canine model of naturally occurring OA contradicted a study in humans that found that the synovial fluid concentrations of these two epitopes varied independently of each other [32,37].

Keratan sulfate is a glycosaminoglycan that is a major constituent of aggrecan. Numerous studies with different models of induced and naturally occurring OA have produced a wide variation in results, with synovial fluid concentration of keratan sulfate rising significantly post-injury in some models, a lack of significant change, and even a significant decrease in keratan sulfate concentrations in others [17].

Matrix metalloproteinases (MMPs) are a family of zinc-dependent endopeptidases. They and their inhibitors, tissue inhibitors of matrix metalloproteinases (TIMPs), play an important role in normal degradation of the extracellular matrix. Once activated, MMPs become susceptible to inhibition by TIMPs, which are secreted by chondrocytes and synoviocytes. Because MMPs appear to play a substantial role in OA, potential treatment strategies include regulation of their activity by control of gene expression and secretion, deactivation of pro-MMPs, and inhibition of active MMPs [38]. Most MMPs are present in articular tissues in an inactive form [39] but during OA, MMPs are present in higher concentrations than the TIMPs and the latent forms of MMPs become activated [40]. Aggrecanases are members of the ADAMT (a disintegrin and metalloproteinase with thrombospondin motif) family. Several aggrecanases have been described, 6 of which have been associated with aggrecan core protein cleavage. Current thinking is that as many as 11 MMPs are responsible for normal articular cartilage matrix turnover, and the aggrecanases contribute to pathologic degradation [41].

Stromelysin (MMP-3) degrades collagen II, IX, and X, proteoglycan, and fibronectin. This MMP is released from chondrocytes during OA in response to cytokine stimulation. Stromelysin concentrations have been demonstrated in a number of in vitro and in vivo studies to be elevated in the tissues and synovial fluid of OA joints in dogs, horses, and humans [39,42-45] so that stromelysin concentration in articular cartilage and/or synovial fluid may be a biomarker for OA.17 Collagenase (MMP-1) also degrades collagen II, IX, and X. Gelatinases (MMP-2 and MMP-9) were detected in the synovial fluid of healthy joints, while a canine study demonstrated elevated MMP-2 and MMP-9 activity in the synovial fluid of OA joints that paralleled the increase in MMP-9 activity seen in horses with OA [40,46,47]. Although molecular markers of osteoarthritis do not currently have direct clinical application, numerous substances are under investigation as potential biomarkers, and significant advances in understanding the molecular processes of OA have been made. Further identification and understanding of the OA process may ultimately allow for earlier diagnosis and treatment opportunities.

Genetics

Many developmental diseases in dogs are a result of a complex interaction between genetic susceptibility, nutrition, and other environmental factors. Polygenic modes of inheritance have been proposed for hip dysplasia, osteochondritis dissecans, and elbow dysplasia [4]. Osteoarthritis is a secondary disease process that occurs as a result of these joint abnormalities, but these developmental diseases result in variable degrees of OA in a given individual. Heritability of elbow dysplasia has been reported at 10% to 45%, and osteochondrosis of the shoulder joint has been estimated at 55% to 70% in various breeds of dogs. One study showed dogs with hip OA to be predisposed to elbow and stifle osteoarthritis [48]. For further information on the genetics of developmental and acquired orthopedic traits in dogs, see other reviews [4,49].

Diagnostic Imaging

Radiography

In one study, radiographic evidence of DJD was evident in 90% of older cats evaluated as part of a diagnostic workup for other disease. Other studies of cats of varied ages have reported overall incidences of OA of 20% to 30% [50]. The elbow is the most commonly affected joint in cats [16,50]. In radiographic evaluation of human knee OA, within-observer reproducibility is greater than between observers, and osteophyte score, joint space narrowing, and bony contour are acceptably repeatable, but subchondral sclerosis and subchondral cyst scores suffer from poor reliability [51]. To meaningfully assess joint-space narrowing, a weight-bearing radiograph is required [52]. In assessment of progression of OA in canine stifles, effusion, osteophytosis, and intra-articular mineralization were found to be more reliable than assessment of subchondral sclerosis [53].

Computed Tomography

When conventional radiography fails to demonstrate changes characteristic of OA, or fails to reveal the underlying etiology when OA is recognized, computerized tomography can be employed to help identify the underlying cause of lameness. Perhaps the most common use of this modality is to identify fragmented medial coronoid process of the elbow. Computed tomography is more sensitive than conventional radiography to early bone changes and three-dimensional reconstructions can be useful in evaluating the extent of bone lesions.

Nuclear Scintigraphy

Nuclear Scintigraphy is most useful when the cause of lameness is unknown. A bone-seeking radionuclide compound, such as technecium-99m-labeled methylene diphosphonate, will be taken up in areas of increased bone and periarticular metabolic activity. Therefore, it reflects both the vascular supply to the region and the rate of mineralization as it is adsorbed to the mineralization front. Any process that disturbs the normal balance of bone production and resorption can produce an abnormality on a bone scan. A bone scan is a highly sensitive but relatively nonspecific diagnostic tool. When administered to an animal with a non-localized lameness, areas of increased activity aid in identifying the affected joint(s) and allow for more specific evaluation using conventional radiographic techniques, computerized tomography, or magnetic resonance imaging in order to further identify the underlying etiology.

Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) is infrequently used as an imaging tool for clinical cases of joint disease in veterinary medicine. However, this modality may have its greatest potential use for the evaluation of subtle lesions and assessing cartilage thickness or damage as part of a clinical research protocol. MRI can detect subtle changes in cartilage structure and subchondral bone long before changes in the underlying bone or periarticular structures are detected by conventional radiographic methods. Well hydrated tissues are best imaged with MRI. Changes including cartilage thinning, surface irregularities, and increased signal intensity (called a bone bruise) in subchondral bone have been noted at 4 weeks post-injury in a canine experimental model of OA [6].

Treatment

Osteoarthritis leads to decreased patient flexibility, increased joint stiffness, and pain; the resultant disuse atrophy of periarticular supporting muscle results in decreased overall strength. As a result, treatment goals include pain relief, reduction of any inflammatory component, improvement of joint function, maintenance of muscle mass, and if possible, preventing or minimizing progression of disease. If an underlying cause of the OA can be identified, treatment is aimed both at correction of the initiating cause and management of the subsequent degenerative cartilage changes.

Surgical management of OA is primarily aimed either at manipulation of joint congruency and stability (e.g., stabilization of a cruciate-deficient stifle or removal of a fragmented coronoid process) early in the disease process in an attempt to minimize progression of OA, or by salvage procedures in end-stage disease (total hip replacement, pancarpal arthrodesis) to reduce pain and improve overall limb function. Medical management of OA may be broadly conceptualized as having four main components: weight management, pharmacologic treatment with analgesics and anti-inflammatories, exercise modification/physical therapy, and slow-acting disease-modifying osteoarthritis agents (SDMOAs). Slow-acting, disease-modifying agents are also described as chondroprotective agents and are purported to have three primary effects: enhancement of chondrocyte and synoviocyte metabolism, inhibition of degradative enzymes within the synovial fluid and matrix, and inhibition of thrombus formation in the small blood vessels supplying the joint [54].

Obesity is a significant risk factor for the development and progression of OA in people, and may be associated with increased stresses on weight-bearing joints or metabolic alterations [55]. Weight loss in dogs with OA has been demonstrated to have a profound positive effect on the clinical management of the disease [55]. In addition, it has been demonstrated that long-term limitation of caloric intake results in decreased radiographic prevalence and severity of OA in dogs [48,56]. Management recommendations to clients should include long-term efforts to keep patients as lean as possible to reduce force transmission through OA joints. More recently, nutritional supplements and commercial diets with specific ratios of omega-3 and omega-6 fatty acids have been introduced as components of management of the OA patient. By shifting the prostaglandin synthesis pathway from arachidonic acid metabolites to dihydroxy-eicosapentaenoic acid metabolites, omega-3 fatty acids are thought to decrease inflammation and decrease the expression and activity of proteoglycan degrading enzymes [57]. The ideal dietary ratio of omega-6:omega-3 fatty acids for dogs is controversial but the current recommendation is between 10:1 and 5:1 [54].

Nonsteroidal Anti-inflammatory Drugs (NSAIDs)

Although the inflammation associated with OA is variable, NSAIDS are used to reduce pain and synovitis as a result of OA. Nonsteroidal anti-inflammatory drugs decrease prostaglandin synthesis by inhibition of cyclooxygenase (Fig. 116-7). The two primary isoforms described as cyclooxygenase, COX-1 and COX-2, are simplistically described as the constitutive form and the inducible form, respectively. NSAIDs inhibit sensitization and stimulation of peripheral nociceptors and spinal cyclooxygenase activity. NSAIDs reportedly inhibit nitric oxide-induced apoptosis independent of cyclooxygenase activity. These signaling pathways involve nuclear-factor kappaB and caspase activation. COX-2 and PGE2 production are affected by the non-COX mechanism [58].

Figure 116-7. Products and enzymes of arachidonic acid metabolism.

Nonsteroidal anti-inflammatory drugs can be divided into COX nonspecific, COX-2 preferential, and COX-2 specific. The COX nonspecific drugs include aspirin, phenylbutazone, ibuprofen, and naproxen. Aspirin (acetylsalicylic acid), a first generation NSAID, is a nonspecific COX inhibitor. It has adverse effects on the gastrointestinal tract, inhibits platelet aggregation, can cause renal failure in susceptible patients, and is used at an approximate dose of 1 adult aspirin (375 mg) per 60 lb body weight up to 3 times a day with a little food.

COX-2 preferential drugs are generally well tolerated by dogs, with fewer side effects reported. Their analgesic action is related to inhibition of spinal nociceptive transmission and attenuation of peripheral inflammation. They include carprofen (Rimadyl, Pfizer), which is approved for use in dogs at 2.2 mg/kg twice daily or 4 mg/kg once daily; etodolac (Etogesic, Fort Dodge), which is approved for use in dogs at 10 to15 mg/kg once daily; and meloxicam (Metacam, Merial/Boehringer Ingelheim) used at 0.2 mg/kg as an oral liquid suspension. Carprofen is generally considered to be either COX 2 preferential or nonselective. In cell culture, it has low COX-2 inhibition. It may have an alternative pathway for its analgesic activity. Peak plasma concentration is reached in 1 to 3 hours and plasma elimination half-life is 7 to 9 hours. Carprofen has good analgesic, antipyretic, and anti-inflammatory properties. Idiosyncratic hepatic toxicity has been reported, especially in Labrador retrievers. Minimal gastrointestinal toxicity has been observed in experimental studies. Carprofen probably has no negative impact on articular cartilage metabolism. In the cranial cruciate transaction model of OA in dogs, carprofen was given 4 weeks after induction of disease and continued for 8 weeks at 2.2 and 4.4 mg/kg bid. Carprofen reduced morphologic changes in articular cartilage and subchondral bone normally associated with OA [59]. Carprofen improves limb function in dogs based on both subjective and objective gait evaluation.

Etodolac reaches peak plasma concentration in 1 hour, with an elimination half-life of 10 to 14 hours, and is generally considered COX-2 preferential but that is debated. It has been shown to improve hind limb function in dogs with hip OA [60]. Etodolac is reported to spare collagen synthesis by chondrocytes but data conflict on its effect on proteoglycan synthesis. It exhibits minimal gastrointestinal toxicity. Meloxicam exhibits peak plasma concentration at 8 hours and has an elimination half-life of 23 hours. It performs comparably to carprofen in clinical trials for lameness in dogs with minimal side effects reported [61]. Derracoxib (Derramax, Novartis) is a COX-2 selective drug which reduced postoperative pain and lameness in dogs given at 1 to 2 mg/kg. Celecoxib (Celebrex) had a positive effect on hyaluronan and proteoglycan synthesis in human OA cartilage explants. Minimal side effects have been so far reported.

In an acute stifle synovitis model in dogs, butorphanol (0.2 mg/kg, IV), etodolac (17 mg/kg, PO), carprofen (4 mg/kg, PO), and meloxicam (0.2 mg/kg, PO) were compared. A Latin cross-over design and 3-week washout was followed. Treatments were given 3 hours after monosodium urate injections into the stifle joint. Ground reaction force (GRF), orthopedic exams, and C-reactive protein were measured [62]. The greatest improvement in vertical GRF was in the carprofen group. Etodolac had the fastest onset of action. Compared with butorphanol, only carprofen and etodolac had decreased stifle pain scores. There were fewer non-responders in the carprofen and meloxicam groups.

Leukotrienes can also be inhibited by combined cyclooxygenase/lipoxygenase inhibitors (Fig. 116-7). Leukotriene B4, for example, induces chemotaxis, aggregation, degranulation, and increased cytokine production by leukocytes, hyperalgesia, and bone resorption. Other leukotrienes affect smooth muscle function, mucus secretion, vascular permeability, and airway inflammation. Cysteinyl leukotrienes may enhance gastric mucosal injury by causing local vasoconstriction. Prostaglandins and leukotrienes have complementary effects but lipoxins can inhibit the inflammatory effects of leukotrienes. Tepoxalin (Zubrin, Schering Plough) inhibits both COX activity and lipoxygenase 5, the activity of which results in leukotriene synthesis. It has potent anti-inflammatory activity with excellent gastric tolerance. The tables are rapidly disintegrating. A similar drug used in human medicine called licofelone given for 8 weeks beginning the day after surgery in a model of OA induced by cranial cruciate ligament transaction prevented abnormal subchondral bone cell metabolism, reduced PGE2 production in synovial fluid, inhibited collagenase 1 production in articular cartilage, interleukin 1-β, and leukotriene B4 in synovial membrane, and reduced chondrocyte cell death (decreased caspase 3 activity) probably owing to lower nitric oxide and PGE2.

Whether NSAIDs can affect the progression of OA is an open question. Prostaglandin inhibition by NSAIDS may have a negative effect on chondrocytes and the cartilage matrix [63]. The detrimental effects of NSAIDs on chondrocytes are partly mediated by inhibition of glycosyltransferase activity, uncoupling of mitochondrial oxidative phosphorylation, activation of cAMP-dependent kinase A, and disruption of protein interactions at the cell surface. The majority of positive effects are due to suppression of inflammation characterized by inhibition of COX-2, which is highly expressed in OA tissues, the inhibitory effects on IL-1 production or IL-1 receptor expression, decreased PGE2 production, reduced inducible nitric oxide synthase (iNOS) synthesis by interleukin-1, and decreased nitric oxide. Nitric oxide enhances MMP activity, decreases proteoglycan synthesis, and inhibits interleukin 1β receptor antagonists. Inhibition of matrix metalloproteinase activity will encourage maintenance of the extracellular matrix and inhibit chondrocyte apoptosis, a critical feature of OA progression.

Slow-Acting Disease-Modifying Osteoarthritis Agents

Disease-modifying agents of OA comprise a large, diverse group of compounds, many of which are poorly characterized as to their efficacy and/or mechanism of action, and their production and administration are poorly regulated, with varying degrees of quality control. Disease-modifying OA agents are also commonly referred to as "nutraceuticals"; however, the only thing shared in common is that they are neither foods nor drugs recognized by the FDA, and as such undergo no pre-market approval process.64 The only FDA-approved SDMOA in dogs is injectable Adequan (Luitpold Pharmaceuticals), which is a polysulfated glycosaminoglycan (PSGAG). Because it is a heparin analogue, injectible PSGAG has the potential to affect coagulation. In cats, injectable PSGAG has been demonstrated to produce a prolongation in the activated partial thromboplastin time and should be avoided in animals with bleeding disorders or on concurrent NSAIDs that exhibit antithromboxane effects [65]. The proposed mechanisms of action include inhibition of serine proteases, PGE2, elastase, stromelysin, MMPs, and hyaluronidases [66]. In one study, dogs treated with an injectable PSGAG from a young age and studied until skeletal maturity had better hip conformation and fewer joint abnormalities than did controls [67]. Pentosan polysulfate is a polysaccharide sulfate ester and is thought to modify disease progression by its antithrombotic and fibrinolytic effects in addition to improving subchondral and synovial membrane blood flow [4,68]. However, a clinical study evaluating the effects of pentosan polysulfate failed to demonstrate a significant clinical improvement over control in the postoperative progression of OA after surgical stabilization of canine stifles with cranial cruciate ligament insufficiency [68]. Glucosamine salt supplements are commonly found in combination products with chondroitin sulfate and manganese ascorbate. Glucosamine is an amino sugar that is a precursor to matrix glycosaminoglycans [54]. Chondroitin sulfate is a glycosaminoglycan found naturally within the extracellular matrix of articular cartilage. It has been proposed that glucosamine supplementation benefits OA articular cartilage by promoting formation and repair of cartilage, whereas chondroitin sulfate is thought to promote water retention and elasticity in cartilage as well as to inhibit degradative enzymes. Clinical and experimental studies support the use of glucosamine-chondroitin-manganese combinations or as individual components [54]. Oral glucosamine-chondroitin sulfate formulations may be obtained over the counter. However, clients should be warned about significant variations among measured concentrations and label claims in both glucosamine and chondroitin sulfate in products on the market [69,70]. In addition, differences in the bioavailability of these substances may be significant between formulations and between species. Oral bioavailability of glucosamine hydrochloride in humans is 84%, but the bioavailability of the sulfate salt is only 47% [64]. The oral bioavailability of glucosamine in dogs was reported to be 10% [71]. The oral bioavailability of chondroitin sulfate varies inversely with molecular weight, and molecules of chondroitin sulfate with a molecular weight of 17000 have the most favorable permeability coefficient [64]. In addition to concerns about variations in formulation and bioavailability, accurate pharmacokinetic and pharmacodynamic information for most products is lacking, as are studies on safety and efficacy. Efficacy in veterinary patients is particularly difficult to assess; however, the placebo effect in animals is likely equal or greater than the 30% to 40% placebo effect that is reported in human studies evaluating pain. Recently, it was shown that glucosamine and chondroitin sulfate regulate expression of matrix-degrading enzymes and the inhibitors at the transcriptional level [72]. Products containing glucosamine and chondroitin sulfate have been shown to cause significant decreases in red blood cell and platelet indices in dogs and cats, but these differences stay within clinically normal ranges and are unlikely to be clinically relevant [73,74].

The use of other nutritional supplements such as elk velvet antler, green-lipped mussels, type-II collagen, and milk protein concentrate have been preliminarily described in veterinary patients for the treatment of OA; however, further evaluation is needed before these substances can be recommended [75-78]. Hyaluronan, methylsulfonylmethane, dimethylsulfoxide, and doxycycline have all been used to treat osteoarthritis, but supporting evidence of their success is lacking.