Experimental modeling of osteoarthritis in animals

K. R. N. Pritzker (1994) defined an experimental animal model of any disease as "a homogeneous group of animals that exhibit an inherited, naturally acquired, or experimentally induced biological process, subject to scientific investigation, that is similar in one or more respects to the disease in man." Animal models of osteoarthritis are useful for studying the evolution of structural changes in joint tissues, to determine how various risk factors initiate or promote these changes, and to evaluate therapeutic measures. It is important to remember that osteoarthritis is not a disease of just one tissue, the articular cartilage, but of all tissues of the affected joint, including the subchondral bone, synovial membrane, menisci, ligaments, periarticular muscles, and afferent nerves with endings both outside and inside the joint capsule. Studies of pharmacological agents in animal models focus primarily on their effects on articular cartilage. It is impossible to evaluate the main symptom of osteoarthrosis in humans - joint pain - in experimental models. At the same time, when modeling osteoarthrosis in animals, a number of important factors contributing to the development and progression of osteoarthrosis are not taken into account (for example, the vertical position of the human body, weakness of the periarticular muscles, etc.).

Of course, the most illustrative model of the disease is the one that has the greatest similarity to the changes in human osteoarthritis. Animal models of osteoarthritis are of greatest interest in terms of studying the effectiveness of disease-modifying OA drugs (DMOAD). Although a number of drugs in this group prevent the development or slow the progression of experimentally induced or spontaneous osteoarthritis in animals, all of them were ineffective when studying their effects in humans.

Animal models of osteoarthritis

Modeling mechanism	Animal species	Inducing factor/agent	Source
Spontaneous osteoarthritis	Guinea pigs	Age/overweight	Bendele AM et al., 1989
	Mice STR/ORT, STR/INS	Genetic predisposition	Das-Gupta EP et al., 1993 Dunham J. etal., 1989 Dunham J. etal., 1990
	Black mice C57	Genetic predisposition	OkabeT., 1989 StabescyR. etal., 1993 Takahama A.. 1990 van der Kraan PM etal., 1990
	Mice	Collagen II mutation	GarofaloS. et al., 1991
	Mice	Collagen IX mutation	NakataK. et al., 1993
	Dogs	Hip dysplasia	SmaleG. et al., 1995
	Primates	Genetic predisposition	Alexander C.J., 1994 Carlson C.S. etal., 1994 Chateauvert J.M. et al., 1990
Chemically induced osteoarthritis	Chickens	Iodoacetate premium*	Kalbhen D.A., 1987
	Rabbits	Papain premium	Marcelon G. etal., 1976 Coulais Y. etal., 1983 Coulais Y. et al., 1984
	Guinea pigs	Papain premium	Tanaka H. et al., 1992
	Dogs	Chymopapain v/s	Leipold HR et al., 1989
	Mice	Papain premium	Van der Kraan PM et al., 1989
	Mice	Collagenase premium	Van der Kraan PM et al., 1989
	Mice	TFR-R v/s	Van den Berg WB. 1995
	Rabbits	Hypertonic NaCI solution	VasilevV. et al.. 1992
Physically (surgically) induced osteoarthritis	Dogs	Anterior cruciate ligament transection (unilateral)	Marshall JL et al., 1971 Brandt KD, 1994
	Dogs	Anterior cruciate ligament transection (bilateral)	Marshall KW Chan AD, 1996
	Rabbits	Anterior cruciate ligament transection	Christensen S.B., 1983 VignonE. et al., 1991
	Sheep	Meniscectomy	Ghosh P. et al., 1993
	Rabbits	Meniscectomy	FamA.G. etal., 1995 Moskowitz RW, Goldberg VM, 1987
	Guinea pigs	Meniscectomy	Bendele A.M., 1987
	Guinea pigs	Myectomy	ArseverC.L, BoleG.G., 1986 LaytonM.W. etal., 1987 Dedrick DK etal., 1991
	Rabbits	Patella contusion	Oegema T. R. J., et al., 1993 Mazieres B. et al., 1990
	Rabbits	Immobilization	Langenskiold A. et al., 1979 Videman T., 1982
	Dogs	Immobilization	Howell D.S. etal., 1992 Ratcliffe A. et al., 1994 PalmoskiM., Brandt KD, 1981
	Dogs	Denervation followed by transection of the anterior cruciate ligament	VilenskyJA et al., 1994

* intra-articular - intra-articular.

Physically and chemically induced models of osteoarthrosis are currently very popular, but they reflect processes observed in secondary osteoarthrosis in humans rather than idiopathic osteoarthrosis. An alternative to them are models of spontaneous osteoarthrosis in bipedal primates and quadrupeds.

Some authors are quite skeptical about modeling osteoarthrosis in animals in general. Thus, according to MEJ Billingham (1998), the use of models for the discovery of osteoarthrosis-modifying drugs is "...an expensive gamble."

[ 1 ], [ 2 ], [ 3 ], [ 4 ], [ 5 ], [ 6 ], [ 7 ], [ 8 ]

Models of spontaneous osteoarthritis

Almost all inbred mouse strains develop osteoarthrosis of varying severity and localization. The highest incidence of osteoarthrosis and the most severe course of the disease are observed in mice of the STR/ORT and STR/INS strains. Among STR/ORT mice, the disease is more common, and it is more severe in males than in females. Primary damage to the articular cartilage develops in the medial part of the tibial plate. It was assumed that the appearance of changes in the cartilage is preceded by displacement of the patella, however, R. G. Evans et al. (1994), C. Collins et al. (1994) found that in all mice of this strain, cartilage damage develops by 11 months, but not all of them showed displacement of the patella. The same authors found that changes in articular cartilage in STR/ORT mice are often preceded by chondrocyte-osteoblastic metaplasia of tendon and ligament cells around the affected knee joints, indicating that these changes are primary in the pathogenesis of osteoarthritis in this model. It is possible that the initial calcification of ligaments and tendons alters the mechanical stress on intra-articular structures and that subsequent changes in articular cartilage reflect an attempt to maintain normal joint loading. Unlike guinea pig and macaque models, in which cartilage degeneration is preceded by changes in the subchondral bone, subchondral sclerosis appears later in STR/ORT and STR/INS mice.

The advantage of this osteoarthrosis model is the small size of the animals, which requires minimal consumption of the tested pharmacological agent. However, the size is also a disadvantage, since biochemical and pathohistological analysis of cartilage in mice is difficult.

The studies of A.M. Bendele, J.E. Hulman (1988), A.M. Bendel et al. (1989), and S.C.R. Meacock et al. (1990) devoted to the study of the natural course of spontaneous osteoarthrosis in guinea pigs have stimulated interest in this model of the disease. Beginning at the age of 13 months, all male Dunkin Hurtley guinea pigs develop degeneration of the articular cartilage. Similar changes in females appear somewhat later and are milder. At the age of 1 year, complete loss of articular cartilage is observed in the region of the medial condyle of the femur and the tibial plate. An increase in the body weight of Dunkin Hurtley guinea pigs aggravates the course of the disease, and a decrease in body weight to 900 g or less improves the course of osteoarthrosis. At the age of 8 weeks, changes in the subchondral bone are already detected in this model, i.e. the latter precede cartilage damage. Changes in the cruciate ligaments of the knee joints can accelerate bone remodeling.

Spontaneous osteoarthrosis develops in rhesus and cynomolgus macaques. A very important advantage of primates over other animals used to create an experimental model of osteoarthrosis is their bipedalism. The disease develops in middle-aged/older individuals. Early histological findings include thickening of the subchondral bone followed by fraying of the articular cartilage in the region of the medial plate of the tibia. Later, the lateral plate is also involved in the process. Notably, articular cartilage degeneration begins to develop only after the subchondral bone thickness reaches 400 μm. The prevalence and severity of osteoarthrosis in macaques increases with age, but these indicators are not affected by sex and body weight. To date, primate models of osteoarthrosis have not been used to study the effectiveness of DMOADs.

Models of physically (surgically) induced osteoarthritis

Osteoarthritis models based on surgically induced knee laxity that alters mechanical stress on the knee joint are most commonly used in dogs and rabbits. The most widely used model is the one with cruciate ligament transection in dogs. Surgical models of osteoarthritis in rabbits involve transection of the cruciate ligaments with or without excision of the medial and collateral ligaments, total or partial meniscectomy, and surgical tear of the menisci. Surgical models of osteoarthritis in guinea pigs have been described that involve transection of the cruciate and collateral ligaments and partial meniscectomy. Partial meniscectomy in guinea pigs results in osteophyte formation within 2 weeks and excessive degeneration of articular cartilage within 6 weeks.

Until recently, the canine model of osteoarthritis following anterior cruciate ligament transection was viewed with skepticism due to the absence of cartilage ulceration and marked disease progression observed in human osteoarthritis. J. L. Marshall and S. - E. Olsson (1971) found that changes in the tissues of dog knee joints 2 years after surgery were virtually identical to those recorded immediately after surgery. The authors suggested that mechanical factors (e.g., fibrosis of the joint capsule and osteophyte formation) stabilize the postoperatively loosened knee joint and prevent further progression of articular cartilage destruction. It was also suggested that this model be considered a model of cartilage damage and repair, rather than a model of osteoarthritis. However, the results of studies conducted by KD Brandt et al. (1991), who studied the dynamics of changes in the tissues of knee joints destabilized by the intersection of the anterior cruciate ligaments for a longer period, refuted the assumptions of previous authors.

S.A. McDevitt et al. (1973, 1977) found that already during the first days after the cruciate ligament transection the synthesis of proteoglycans by chondrocytes of articular cartilage increases. During 64 weeks after surgical induction of knee joint instability the thickness of articular cartilage was higher than normal, although biochemical, metabolic and histological changes in it corresponded to those in osteoarthrosis. This thickening of cartilage was associated with increased synthesis of proteoglycans and their high concentration in articular cartilage. Using magnetic resonance imaging (MRI), M.E. Adams and K.D. Brandt (1991) showed that after the transection of cruciate ligaments the hypertrophy of cartilage is maintained for 36 months, then progressive loss of cartilage occurs, so that after 45 months most of the articular surfaces are devoid of cartilage. Morphological examination of the cartilage 54 months after surgery confirmed the MRI findings. Thus, M.E. Adams and K.D. Brandt (1991) demonstrated that surgically induced instability of the stifle joints in dogs can be considered a model of OA.

The phenomenon of hypertrophic reparation of articular cartilage is well illustrated by the above-described model of osteoarthrosis in dogs. However, it is known that this phenomenon is not unique to it. Hypertrophy of articular cartilage, which was of a reparative nature, was first described in patients with osteoarthrosis by EGL Bywaters (1937), and later by LC Johnson. It is also found in other models of osteoarthrosis - in rabbits after partial meniscectomy (Vignon E. et al., 1983), in rhesus macaques, cartilage hypertrophy develops spontaneously.

Modern descriptions of pathogenesis focus mainly on the progressive "loss" of cartilage, but authors often overlook its thickening and increased synthesis of proteoglycans, which corresponds to the homeostatic phase of stabilized osteoarthrosis. During this phase, cartilage reparation compensates for its loss and can maintain the joint in a functional state for a long time. But reparative tissue often cannot cope with the mechanical load imposed on it in the same way as healthy articular cartilage, which leads to the inability of chondrocytes to maintain the normal composition of the matrix and a decrease in proteoglycan synthesis. The final stage of osteoarthrosis develops.

The study of Charcot arthropathy has led to the development of a method for neurogenic acceleration of surgically induced osteoarthrosis modeling. Charcot arthropathy is characterized by severe joint destruction, joint "mice", joint effusion, ligament instability, and the formation of new bone and cartilage tissue within the joint. The general concept of the pathogenesis of Charcot (neurogenic) arthropathy is the interruption of sensory signals from the limb proprioceptors and nociceptors to the central nervous system (CNS). To accelerate the progression of osteoarthrosis induced by transection of the anterior cruciate ligaments in dogs, ganglionectomy or excision of the nerve innervating the joint is performed before surgery, which leads to the appearance of cartilage erosions already in the first week after surgery. Interestingly, the new DMOAD diacerein was effective when used in a slowly progressive (neurologically intact) model of osteoarthritis, but was ineffective in neurogenically accelerated experimental osteoarthritis.

In conclusion, it should be noted that it is impossible to fully assess the identity of the experimental model of osteoarthrosis and osteoarthrosis in humans, since the etiology and precise mechanisms of pathogenesis of the disease have not yet been clarified. As stated earlier, the main purpose of using experimental models of osteoarthrosis in animals is to use them to assess the effectiveness of new drugs, mainly of the "disease-modifying" group. The likelihood of how much the results of treatment in an animal will coincide with the results of using an experimental pharmacological agent in humans is also impossible to determine. NS Doherty et al. (1998) emphasized the significant differences between the species of animals used to model osteoarthrosis in terms of different development of pathology, various mediators, receptors, enzymes, which leads to an objective extrapolation of the therapeutic activity of new drugs used in animals to humans. An example is the high effectiveness of NSAIDs in modeling inflammatory arthritis in rodents. This has led to a re-evaluation of the efficacy of NSAIDs in humans, in whom prostaglandins do not play the fundamental role in disease pathogenesis that they do in rodents, and the clinical efficacy of NSAIDs is limited to symptom treatment rather than disease modification.

At the same time, underestimation of new pharmacological agents when studying their effectiveness in animal models may lead to the loss of potentially effective therapeutic agents in humans. For example, gold salts, penicillamine, chloroquine and sulfasalazine, which have some effect in the treatment of rheumatoid arthritis, are absolutely ineffective in animals used for screening antirheumatic drugs.

The difference in response between an animal model of osteoarthrosis and a patient with osteoarthrosis to DMOAD treatment largely depends on collagenase, an enzyme that is believed to be actively involved in the pathogenesis of osteoarthrosis. Inhibitors of interstitial collagenase (collagenase-1 or matrix metalloproteinase (MMP)-1) are often found in rodents with model OA, but a homologue of human collagenase-1 has not been found in rodents and may not exist. Thus, specific inhibitors of human collagenase-1 will not show therapeutic efficacy in rodents with experimental osteoarthrosis. Most MMP inhibitors created to date are non-selective and therefore inhibit collagenase-3 (MMP-13), which is involved in the pathogenesis of experimental osteoarthrosis in rodents. Moreover, as studies by NRA Beeley et al. (1994), JMP Freije et al. (1994) have shown, human collagenase-3 is expressed in the articular cartilage of patients with osteoarthritis and may play a role in the pathogenesis of the disease.

It can be assumed that these mediators, receptors or enzymes play a similar role in the pathogenesis of modeled osteoarthrosis in a particular animal and in humans. An example is the chemotactic capacity of leukotriene B4, which is considered the same in humans, mice and rabbits, but the activity of antagonists of this biologically active substance differs by 1000 times between animal species. In order to avoid such inaccuracies in experiments, it is necessary to create methods that allow studying pharmacodynamics in vivo. For example, it is possible to study the effect of any substances on the activity of exogenous enzymes or mediators in humans. This technique was used by V Ganu et al. (1994) to assess the activity of MMP inhibitors by determining the ability of drugs to inhibit the release of proteoglycans from articular cartilage after injection of human stromelesin into the knee joint of a rabbit.

Although the results obtained in the experimental model of osteoarthrosis may lead to an incorrect assessment of potential DMOADs, animal models of osteoarthrosis play an important role in basic research. A final decision on the effectiveness of pharmacological agents in the treatment of human diseases can only be made after conducting phase III clinical trials in humans.