Chapter 64Models of Equine Joint Disease
Animals have been used extensively as models of joint disease to study clinical conditions in people. However, veterinary researchers have the luxury of using experimental animals from species that are clinically relevant. Unlike in human research, one has no need to assume similarity in findings between species. Equine models of joint disease have been used for several decades to test the effects of drugs and various treatments on joints and to evaluate the pathogenesis of certain diseases. Joint disease can be assessed in horses with clinical disease; however, large numbers are needed for each treatment group to see significant statistical differences in the face of great variation among individual horses. Owner compliance, differing treatment protocols among clinicians, and variability among horses in response to disease and treatment, as well as conformation, limb use, and size, all contribute to this variation. Furthermore, clinical studies take a long time to perform, and the effects of treatment take a long time to be seen. Consequently, in vitro and in vivo models have been developed to give researchers better-controlled studies that can be done in a relatively short time.
The model to be used should be designed to answer a question by using a testable hypothesis. Variability should be reduced as much as possible so that the question can be answered with little outside influence. However, as more and more variables are eliminated, the model becomes less representative of the clinical situation. For instance, the efficacy of oral joint supplements in reducing joint disease cannot rely solely on results from in vitro studies. The drugs must be tested in vivo to determine if and how they work. However, to determine the effects of a drug or chemical on articular cartilage matrix metabolism, a quick, relatively inexpensive in vitro test can be conducted. Therefore the type of model to be used depends on the question to be answered.
Two types of models are used to study equine joint disease. In vitro models can be used to study various treatments, using cells, cell lines, or tissues harvested from joints to test usually one specific pathological pathway or treatment scheme. In vivo systems can be used to test drugs and to determine the pathophysiological response to an insult. Unlike in vitro studies, in vivo studies involve the entire joint, allowing researchers to assess the whole organ to determine truly the clinical efficacy of a drug.
In this chapter the complexity of joint disease and the use of joint models are discussed. The difficulties in modeling joint disease, the rationale for selecting specific models, specific examples of models used for the study of the joint disease, and the current status and future use of equine models of joint disease are considered.
The joint can be considered an organ because it is composed of several different types of tissues that biomechanically and biochemically interact with each other. Joint disease can result from several factors. First, the disease has a direct biochemical effect on all tissues. For instance, with synovitis, inflammatory mediators can be released into the joint space and put the articular cartilage into a catabolic state.1-4 Second, the pain produced by joint disease can result in a change in the character of the gait.5,6 Consequently this change in joint loading alters biomechanical inputs on all tissues, resulting in a biochemical change in the response by tissues.7,8 Third, disease of one tissue can result in a change in the mechanical input on another tissue. For instance, articular cartilage degeneration, which can result in increased stress to subchondral bone, can induce a sclerotic response.9 As another example, subchondral bone sclerosis, which commonly occurs in racehorses, can lead to increased stress on overlying articular cartilage.10 Therefore in the live horse all tissues are affected by one another.
Because several factors can influence joint disease, researchers have attempted to control these influences in experimental studies. However, not all of the factors can be controlled. For instance, several mechanical factors play a role in joint disease; specifically, mechanical input can vary with horse size, exercise intensity, conformation, neurological control, and lameness. Consequently, differences in the stress to joints can result in changes in biochemical pathways in those tissues. For instance, ponies are one third to one half the size of horses and have little naturally occurring joint disease. Therefore they may not be the best equine models of joint disease for joint healing studies and exercise studies because of lower imposed stresses. However, they are still good models if imposed stresses do not play a role in the specific disease being studied, such as induced synovitis models. Tissue material properties also have an influence on joint disease. For instance, weaker tissues undergo greater biomechanical changes than those that are stronger. Factors that control material properties of tissues include genetics, loading history, and age of the animal. The tissue remodeling status and the ability of horses to remodel articular cartilage, subchondral bone, and soft tissues can also influence joint disease, which is also affected by age, loading history, and genetics. The inflammatory response can also change the joint environment. Specifically, differences in the immune system at the time of disease can greatly influence the inflammatory response to the joint and influence the concentration of cytokines released into the joint.
Variables such as age, size, conformation, and neurological status can be controlled in experimental in vivo studies. However, loading history and the presence of subclinical disease are virtually impossible to control. With the advent of more sophisticated imaging equipment, such as computed tomography (CT) and magnetic resonance imaging (MRI), more information on loading history and subclinical disease can be obtained. For instance, subchondral bone density is indicative of loading history.11 Therefore researchers at the Equine Orthopaedic Research Center (EORC) at Colorado State University often perform prestudy CT examinations to determine subchondral bone density as an indicator of loading history. Another benefit of CT and MRI is that some subclinical diseases may be more easily detectable. Researchers at the EORC also initiate a controlled exercise program on a high-speed treadmill before starting the study. The hope is that this can normalize the loading histories of experimental horses. Equally important as loading history and subclinical disease is the biomechanical and biochemical status of the articular cartilage, which at this time is difficult to assess noninvasively at the beginning of a project. With the advent of MRI and pressure probes, articular cartilage and bone matrix structure can be assessed more readily.
Several types of models have been used to study joint disease. In vitro models have been used and are increasing in use for several reasons. Researchers in university settings are constantly driven to develop in vitro studies to reduce the use of live animals and to address the humane issues that surround live animal research. Currently, in vivo models are essential for testing new drugs. However, animal care and use committees at universities critically evaluate in vivo research projects to be sure that they are necessary. These committees are under constant pressure to ensure that research animal use falls within appropriate ethical guidelines. In vitro systems for studying joint disease can be performed on cells from the tissues of joints. Culture systems can be created for cells such as synoviocytes and chondrocytes in two-dimensional and three-dimensional matrices. Isolated cells also can be evaluated, within the individual matrices, or within artificial matrices. Isolated cells are used to evaluate cellular response to therapeutic agents. Specific outcomes are analyzed to determine cell metabolism and proliferation. Some cell cultures are allowed or stimulated to produce the matrix, and specific cell matrix interaction and matrix metabolism can be studied. Cells are often cultured within artificial matrices to evaluate cell-matrix interactions and cell proliferation within the matrix.
Tissues from living systems are also used for in vitro study of joint diseases. The advantages of such a system over isolated cells are that cells can be maintained within the natural matrix and the experiments are relatively easy to perform. However, tissues must equilibrate in the culture medium to a steady state. This metabolic state may not truly reflect the in vivo state, because no axial load exists and the tissue edges are unconfined. The change in stress patterns and bathing in artificial media then influence articular cartilage matrix metabolism, often leading to increased articular cartilage matrix degradation. Tissues that are harvested from cadavers and specimens are placed within the medium, which is changed every 24 to 48 hours. These tissues then can be manipulated for study, and the medium and tissue can be evaluated. Media collected during the study can be analyzed for release or degradation of matrix components and inflammatory mediators. Tissues can be assessed for changes in cell proliferation, matrix synthesis, and characteristics of matrix degradation. Various molecular techniques can also be performed on the tissues.
Two types of tissue cultures are available. Single tissues can be studied to determine biochemical, molecular, mechanical, and histological changes that occur with certain influences.12,13 However, coculture systems also can be evaluated to determine interactions between tissues. Investigators in the EORC have studied synovium–articular cartilage and articular cartilage–subchondral bone coculture systems. Exposure of articular cartilage matrix to subchondral bone caused a significant reduction in articular cartilage matrix metabolism.14 Furthermore, similar findings also were seen in the synovium–articular cartilage coculture systems, in that exposure of articular cartilage matrix to subintimal tissues and vessels led to a significant reduction in articular cartilage matrix metabolism.15 From these studies one can conclude that exposure of various depths of synovium and subchondral bone to articular cartilage can lead to release of mediators that can change articular cartilage metabolism. Unlike the equilibration period for in vitro experiments, this change in metabolism is severe and long lasting. However, when articular cartilage is cocultured with synoviocytes, there is a protective effect on articular cartilage degradation. Therefore one can conclude that cocultured systems using pure cells can be used to measure simultaneous effects from both tissues.16
Examples of in vitro studies that have been performed on tissues and cells include study of enrofloxacin on articular cartilage explants,17 glucosamine and chondroitin on explants,18,19 glucosamine alone on cells,20,21 the effect of single-impact injury on articular cartilage explants,22 the effects of nonsteroidal antiinflammatory drugs and herbal preparations on tissues,23 and evaluation of corticosteroids and growth factors on articular cartilage.24 These studies demonstrate the robustness of in vitro systems to study various medications and pathological events.
In addition to study of specific tissues and cells, the isolated joints can be used to study various factors. For instance, Hardy and colleagues developed an isolated perfused limb to study the physiological response of inflammation on joints.25,26 In addition, Bragdon and colleagues used the same system to evaluate drug delivery.27 This system is shown to work well to study the physiological response of the joint to various factors. This is mostly focused on the vascular and cell permeability response. In addition, cadaver limbs can also be used to study various physical responses of joints and limbs. Easton and I used an isolated forelimb to study loading patterns and the correlation of subchondral bone density in the fetlock joint of horses.28 Briefly, the isolated forelimb was placed in a loading system and the preparation was loaded to variable fetlock joint angles, including that corresponding to galloping. Dye staining was then used to characterize the loading pattern in the third metacarpal bone, the proximal sesamoid bones, and the proximal phalanx. It has long been established that fractures can be studied with this type of system, as illustrated by distal radial fracture repair.29
In vitro experiments are relatively straightforward to perform and allow large numbers of repeats. Consequently the numbers are strong for statistical analysis. However, only one small portion of the disease process usually can be evaluated. Furthermore, the more tissues are removed from the natural joint environment, the greater the change in matrix metabolism and response to treatment.
In vivo studies have been used for decades to study equine joint disease. The advantages of in vivo systems are that the cells and matrices are kept within the native environment and can be studied without the influence of harvest and culturing, and normal interaction among tissues can be maintained. The disadvantages are that live animals are humanely destroyed, and the costs can be prohibitive. In vivo studies are used for evaluating medications and articular cartilage healing techniques and for determining pathological responses to disease. An understanding of the clinical disease and an appreciation for strict experimental design are needed to produce useful in vivo models. In vivo studies in the horse have gained popularity because Frisbie and colleagues showed that equine articular cartilage is more similar to human cartilage than is cartilage from other species.30
Various inciting mechanisms can be used to induce experimental joint disease. For instance, synovitis models are used to evaluate the sequence of events that occurs and the influence of various drugs and medications. Change in gait, measurement of inflammatory mediators, assessment of articular cartilage matrix metabolism, and various clinical tests have been studied. Examples of these studies include injection of lipopolysaccharide,1,31 interleukin-1,26 sodium monoiodoacetate,32,33 filipin,2,34 polyvinyl alcohol foam particles,35 carrageenan,36 Freund’s adjuvant,37blood,38 and amphotericin6 into joints. Of these models, those that use the natural inflammatory mediators, such as lipopolysaccharide and interleukin-1, seem to be the best for promoting the natural cycle of events in inflammation. The advantages of such studies are that they are quick, relatively well investigated, and replicate a common clinical problem.
Instability models also have been studied in horses. The purposes of these studies have been to determine the sequence of events that occurs within tissues because of instability and to induce osteoarthritis (OA).15,25,39 Once OA has been induced, various treatments can be examined. An example of an instability model includes cutting the collateral and collateral sesamoidean ligaments in the metacarpophalangeal joints of horses.39 Unlike in the dog, complete surgical transection of the cranial cruciate ligament in horses has not resulted in progressive OA.40 Mild osteophyte formation occurred in these joints, but no progressive articular surface changes or lameness were appreciated. The results are unlike those in horses with naturally occurring cranial cruciate ligament damage, in which some horses are lame and substantial articular cartilage damage may result. Unlike synovitis models, instability models are relatively long term and represent an example of chronic, progressive disease.41,42
Forced exercise can also lead to changes in joint environment. Examples include evaluation of osteochondral tissues in response to various levels of exercise. Significant changes have been found in articular cartilage matrix biochemical and biomechanical properties in exercised horses compared with horses not exercised.43 Strenuous exercise led to significant increases in calcified cartilage thickness, significant decreases in articular cartilage mechanical properties, and significant increases in fibronectin at the sites of degradation.43-45 Treadmill exercise led to significant increase in clinical disease in the metacarpophalangeal joints of horses. This was detected grossly and using various imaging techniques.46 Currently several studies are ongoing to evaluate the effects of exercise to determine the level of exercise most appropriate for protecting the tissues of the musculoskeletal system. At the EORC horses with experimentally induced OA are being compared with young normal horses that are exercised, in the hope of differentiating clinical and diagnostic test results in horses undergoing exercise-induced adaptation from those in horses with joint disease. There have been recent studies on the effects of exercise on foals, weanlings, and yearlings, to investigate whether early exercise will lead to stronger musculoskeletal tissues, with generally negative results.47-53
The osteochondral fragment model is a blend of in vivo models and involves inducing a clinically relevant disease (Editors’ note: for racehorses, but not necessarily for other equine athletes) through creation of an osteochondral fragment and imposed exercise. Researchers at the EORC have used this model extensively, which induces progressive OA yet is benign enough to induce grade 1 to grade 3 lameness.54 In addition, osteochondral fragments have been taken from the tarsocrural joint and implanted into the carpus and metacarpophalangeal joint to induce inflammation.55
Using the osteochondral fragment model various medications have been studied, including betamethasone,56 intravenous hyaluronan,57 triamcinolone acetonide,58 and methylprednisolone acetate,59 as have gene therapy for joint disease,60 change of biomarkers with exercise,61 and the effects of exercise on imaging outcomes.62 The response to avocado and soybean unsaponifiables,63 intraarticular autologous conditioned serum,64 and the interleukin-1 receptor antagonist gene65 have also been evaluated. Mechanical nociceptive thresholds have been determined.66
Models of disuse also have been shown to induce osteochondral damage. Disuse is a clinically relevant problem that induces articular cartilage and subchondral bone atrophy. Examples include a significant reduction in articular cartilage matrix metabolism in horses with a lower limb cast.67 Furthermore, application of a cast for 7 weeks, followed by treadmill exercise, has been shown to lead to a significant increase in lameness in the limb with the cast and a significant decrease in bone formation.68-70 Disuse models are important, because casts are used clinically and can lead to substantial problems after removal. Often the race is between healing and the degradative changes caused by disuse.
Models of articular cartilage healing have been studied extensively because of the need to test various modes of treatment in vivo. These models also have been stimulated by the cutting edge research being done by several equine research laboratories around the world and the fact that the horse is becoming more accepted as a model of articular cartilage healing for people. If an implant or technique can stimulate and maintain articular cartilage healing in a horse, then the thought is that the implant or technique should work in people. Several equine models are used to study joint healing. Besides the type of treatment that is tested, the models vary in the depth and size of articular cartilage defect formation. Partial-thickness and full-thickness articular cartilage, articular and calcified cartilage, and osteochondral defects have been evaluated in vivo and treated with various techniques.
The size and location of osteochondral lesions appear to have an effect on healing.71 Small lesions in weight-bearing areas healed better than large lesions and lesions in non–weight-bearing areas. The physical characteristics of the lesions also affect healing. For instance, subchondral cystic lesions developed in horses with linear articular cartilage lesions but not in horses with elliptical lesions.72 The subchondral defects did not fill in with bone. Trauma to the subchondral bone led to cyst formation.73 The presence or absence of calcified cartilage also plays a role in articular cartilage healing. Calcified articular cartilage has reduced defect filling compared with defects without the presence of calcified cartilage.74 However, the conclusion from that study and others was that calcified cartilage may provide support for improved healing. Consequently the influence of calcified cartilage on healing is currently being investigated.75
Various articular cartilage resurfacing treatments have been tested in vivo, including periosteal grafts,76,77 sternal cartilage grafts,78,79 mosaicplasty,80 subchondral micropicking,74 and cell-based grafting.77 Sternal cartilage grafts have produced short-term benefits; however, significant degradation occurred after 4 months, with subchondral cyst formation.78,79 Mosaicplasty has shown promise; however, harvest sites are needed to obtain tissue for implantation.80,81 Subchondral bone micropicking may improve healing of osteochondral defects and is simple to perform.74 Cell-based grafting may be beneficial, but it requires special equipment and advanced training to perform.82,83 Other techniques that have been assessed in the horse include radiofrequency and mechanical debridement84,85; various growth factors in gene therapy86,87; and stem cell therapy.88 The horse may be an ideal model for evaluation of stem cell therapy for human joints.89
Models of infectious arthritis also have been studied (see Chapter 65). For instance, studies have aimed at identifying the ability of certain medications to potentiate infection when injected along with a subinfective dose of bacteria. This work proved the clinical impression that polysulfated glycosaminoglycans increased the chances of infectious arthritis unless given with antibiotics.90,91 Comparisons of various treatment methods used to treat infectious arthritis have also been performed.92,93
Computerized models currently are being developed to study joint disease not only in people, but also in horses and other animals. Specifically, these models are derived from the geometry of the joints, the forces imparted by limb loading and muscle force, and the material properties of tendons, ligaments, and articular cartilage. Consequently, joint disease can be imitated on a computer and the resulting change in joint forces determined. These models also can be used to determine what changes in loading may be expected in horses with clinical disease. Surgical procedures can be inserted into the program, and the resulting forces evaluated.
In vivo work will not be replaced in the near future because it is the best means of evaluating tissue response to disease and treatment. However, newer in vitro systems and computer models are becoming more precise and better accepted by the research community and clinicians. In musculoskeletal modeling to characterize the forces that surround the joint94-99 the goal is to create a finite element model of the joint of interest and input various loading parameters that are seen during exercise and may lead to injury.