Chapter 62Markers of Osteoarthritis
Implications for Early Diagnosis and Monitoring of the Pathological Course and Effects of Therapy
Although joint disease can be diagnosed using routine clinical methods, more accurate and earlier diagnosis may lead to identification of osteoarthritis (OA) before irreversible changes occur within joint tissues. Measuring levels of molecular products of tissue turnover, known as biomarkers, from healthy and diseased cartilage and bone has the potential to achieve early diagnosis and allow a better understanding of the pathophysiology of OA. The potential also exists to monitor disease, especially in response to novel therapeutic agents. Work with biomarkers of articular cartilage and bone in people and horses with OA has yielded promising results. This chapter discusses some of the markers that currently are being evaluated in synovial fluid and serum samples, with a focus on those of potential benefit to the equine industry.
Articular cartilage is a complex tissue with an extensive extracellular matrix. The two main components that define the cartilage matrix are type II collagen and aggrecan (see Chapter 61). A balance of synthesis and degradation orchestrated by the chondrocytes maintains normal populations of these molecules within the cartilage matrix. Osteocytes, osteoblasts, and osteoclasts maintain the structural and functional integrity of bone matrix by regulating synthetic and degradative pathways. Synoviocytes also influence homeostasis in cartilage and bone. OA often is characterized by degradative changes within articular cartilage, bone, and synovium. Direct and indirect factors assault the matrix molecules of these tissues, resulting in degeneration and loss of some macromolecules.
What are direct and indirect molecular markers? Direct molecular markers specifically identify a known molecular process within a given tissue. For example, fibrillar collagens, such as types I and II, are synthesized as immature procollagens that undergo proteolytic changes before conversion to mature collagen fibrils. Peptides at either end of the procollagen molecule are cleaved before the procollagen is incorporated into a mature collagen fibril. Estimations of type II collagen synthesis were obtained from synovial fluid and serum samples by using a specific antibody that recognizes the propeptides cleaved from the carboxyl termini.1 Conversely, an indirect molecular marker reflects more general change that is not clearly definable and may represent contributions from several events and tissues. Indirect markers are cytokines, growth factors, and matrix metalloproteinases (MMPs). OA involves changes in subchondral bone and synovium; therefore assessment of molecular markers from these tissues is relevant.
The carboxyl propeptide of type II collagen is a useful measure of the anabolic process of type II collagen synthesis. Studies have shown that levels of carboxyl propeptide of type II collagen were significantly higher in synovial fluid from people with OA compared with those without OA. Levels peaked early in the radiological progression of the disease and declined in patients with severe radiological changes.2,3 This biomarker also has been shown to change significantly in serum samples from people with OA and rheumatoid arthritis.4,5
Chondroitin sulfate (CS) is a major glycosaminoglycan (GAG) of aggrecan, and measuring specific CS epitopes on newly synthesized proteoglycan (PG) molecules is a useful biomarker for aggrecan synthesis. An epitope called CS-846 that normally is found in fetal tissues, but is almost absent in healthy adult articular cartilage, has been measured in many species. Levels of CS-846 epitope were increased in synovial fluid in people after injury or primary OA compared with levels in synovial fluid from normal joints. Serum levels were elevated in joint disease but to a lesser extent than synovial fluid levels.6,7 Other CS epitopes such as 3B3 and 7D4 were shown to be useful in assessing cartilage injury in animal models and in people with clinical disease.8 Using arthroscopic evaluation, a negative correlation was found between synovial fluid 3B3 concentrations and gross articular damage that was thought to be caused by decreased normal cartilage volume, or inhibition of synthesis, with increasingly severe lesions. Conversely, in people, increased levels of synovial 7D4 epitope were found in diseased knees compared with contralateral normal knees.9
Measuring the degradation of type II collagen is of potential benefit in monitoring OA. Antibodies have been developed to identify exposed but previously inaccessible cleaved or denatured type II collagen fragments. Significant elevations in levels of degraded type II collagen were demonstrated in synovial fluid and serum samples from horses, dogs, and rabbits with experimentally induced OA.10,11 Significant increases were detected in the serum of people with OA, with a correlation to disease activity.10
Keratan sulfate (KS), one of the GAGs found on proteoglycan molecules of aggrecan, has been evaluated extensively. In people, elevations in serum levels of KS were associated with OA in some, but not all, studies.7,12,13 Lack of correlation of serum14 and synovial fluid8 KS levels with cartilage damage compromises the value of serum KS as a biomarker of joint disease in people. In dogs, a specific KS epitope (5D4) was of limited value in experimentally induced and naturally occurring cruciate ligament injury.15,16 The usefulness of KS in serum and synovial fluid of horses with osteochondral fragmentation is also questionable.17
In an initial equine study of molecular markers, carboxyl propeptide of type II collagen and the GAG epitopes CS-846 and KS were measured in synovial fluid and serum of horses with and without carpal osteochondral fragments.17 Synovial fluid and serum CS-846 epitope concentrations were significantly higher in joints with osteochondral fragments compared with normal joints and showed good correlation with grades of cartilage damage. Serum concentrations of carboxyl propeptide of type II collagen were elevated in horses with osteochondral fragments, and good correlation between carboxyl propeptide of type II collagen concentration and arthroscopic lesion grade was found. A single blood sample assayed for CS-846 and carboxyl propeptide of type II collagen levels resulted in 79% accuracy for prediction of an osteochondral fragment.
CS-846, KS, and carboxyl propeptide of type II collagen concentrations were measured in synovial fluid of horses with normal joints and those with osteochondrosis.18 Significantly higher levels of carboxyl propeptide of type II collagen and lower amounts of CS-846 and KS epitopes were found in affected joints compared with normal joints.
Cartilage oligomeric protein (COMP) is an abundant noncollagenous protein constituent of cartilage. COMP was once thought to be cartilage specific, but it has also been localized in tendons and synovium. Serum and synovial fluid concentrations of COMP are increased in people with OA.19,20 A positive correlation exists between COMP levels and radiological grading of OA, progression of radiological changes,21 and results of nuclear scintigraphy in people.22 Gene expression of COMP in synoviocytes is up-regulated in OA, suggesting that this marker may be useful to indicate synovitis.
Unlike the elevation of COMP levels in people with OA, in horses initial studies demonstrated that serum and synovial fluid levels of COMP were significantly lower in horses with diseased joints.23 It appears that this discrepancy was caused by the antibody used in the initial equine studies recognizing mainly intact rather than intact and breakdown products of COMP. Thus a subsequent study24 in the horse using an antibody (14G4) recognizing both intact and breakdown fragments confirmed that COMP levels increase with OA.
Anabolic and catabolic cascades exist in bone, but specific markers in normal and disease states are not clearly defined. This section deals only with bone markers thought to be important in joint disease.
Osteocalcin (OCa) is a small noncollagenous protein associated with bone assembly and turnover and has been measured in serum and synovial fluid samples from people with OA. Levels of OCa correlated with bone scan findings and markers of cartilage metabolism.22,25 However, because OCa levels are higher in serum than synovial fluid, OCa in synovial fluid may be derived from peripheral blood and may not reflect local joint disease.25
OCa levels were measured in horses, and, as in people, they appear to vary with age and with the administration of corticosteroids,25,26 but the effect of gender remains unclear.26 General anesthesia affects serum OCa levels for 4 days.27
Bone-specific alkaline phosphatase is an isoform of alkaline phosphatase that is expressed at high levels on the cell surface of the bone-forming osteoblasts and plays an important role in bone formation. In a recent equine study, a correlation was found between synovial fluid levels of bone-specific alkaline phosphatase (BAP), KS-5D4 epitope, and total GAG, as well as between all three biomarkers, and the amount of joint damage defined arthroscopically.28 This supports a putative role for altered subchondral bone metabolism in equine OA.
Type I collagen C-telopeptides (CTX) may be useful markers of bone resorption. CTX levels in people with rheumatoid arthritis were positively correlated with indices of disease activity and joint destruction.29,30 The marker was influenced by the administration of corticosteroids. CTX is present in equine serum, although its usefulness as a marker of pathological processes is unknown.31
Human bone sialoprotein is found only in adult bone, and levels are seven times higher at the interface of cartilage and bone compared with other locations in bone.32 Serum levels are elevated significantly in people with clinically apparent OA and those with bone scans consistent with OA.32,33 Equine bone sialoprotein has yet to be characterized, but development of an assay is currently underway. The hope is that this will be useful in identifying subchondral bone damage in horses with OA.
Limited data are available regarding the use of biomarkers to diagnose and monitor equine joint disease. Factors influencing levels of biomarkers include liver and kidney clearance, circadian rhythms, intestinal peristalsis, exercise level, age, breed, diet, sex, drug administration, surgery, and general anesthesia. Methods of sample collection and storage also may be influential.
Although biomarkers may have a role in diagnosis and monitoring equine OA, a combination of markers likely will be required, especially because so many factors influence activity. Proof-of-principle work has been completed showing that biomarkers significantly change in the face of experimentally induced OA, and this change is significantly greater than with exercise alone.34,35 Specifically, synovial fluid concentration for eight of eight biomarkers was significantly increased in OA-affected joints of horses undergoing exercise compared with sham joints of similarly exercised horses. Likewise, serum from OA-affected horses had a significant increase in six of eight biomarkers compared with serum from similarly exercised horses. Using biomarker levels from either synovial fluid or serum, horses could be correctly categorized into the appropriate group (OA-affected or sham) 100% of the time within 14 days of OA induction using discriminant analysis, suggesting great promise for the use of biomarkers.
To date, in a clinical setting, several cross-sectional studies have looked at numerous biomarkers in both synovial fluid and serum for the potential to be useful in noninvasive prediction of disease severity. Specifically, COMP was significantly altered in serum in horses with clinical OA compared with control horses.36 Differences in synovial fluid COMP levels were identified when making similar comparisons in horses with OA.24,37,38 Another cross-sectional study39 using the contralateral limb as a control demonstrated that concentrations of BAP, KS, KS/GAG ratio, and hyaluronan were all significantly different in early OA compared with contralateral joints. It was also observed that BAP and KS concentrations and the KS/GAG ratio had good correlations with articular cartilage pathology. These cross-sectional studies provide good examples for the potential use of biomarkers in equine medicine; prospective longitudinal studies are the next step for providing further proof of principle.
One such study has been completed through collaboration of the Equine Research Center at Colorado State University and equine veterinarians practicing at southern California Thoroughbred racetracks. Two- and 3-year-old racehorses were entered in the study (N = 238).40 Horses had monthly musculoskeletal examinations, and blood was collected and stored for later biomarker examination. Horses were followed for a maximum of 10 months and were considered to have sustained an injury if they were out of training for more than 30 days. Horses with solitary musculoskeletal injuries and completion of more than 2 months in the study were analyzed for biomarker levels, along with a randomly selected control population of uninjured horses. The following were considered musculoskeletal injuries: intraarticular fragmentation, tendon or ligamentous injury, stress fractures, and dorsal metacarpal disease. Fifty-nine horses sustained a single musculoskeletal injury; 71 acted as uninjured controls. The greatest change in biomarker levels was 4 to 6 months before injury. Using sophisticated statistical modeling, it was possible, based on biomarker levels in this group of horses, to accurately predict horses that would sustain an injury 73.9% of the time. Given these promising results, another study has been instituted in Western performance horses, which have a less tightly controlled training pattern compared with Thoroughbred racehorses. Because exercise is known to affect biomarker levels, this will be an important test of the “real world” application of biomarkers in musculoskeletal disease.