Introduction

In preparing this abstract I reflected on invitations to speak at 2 previous ICRS meetings – Barcelona 2010 and Montreal 2012. At both my topic was on use of animal models particularly genetically modified (GM) mice to advance knowledge on OA pathophysiology and develop new therapies. Sadly today as then, there remains no cure for OA, no registered therapies that prevent or even slow OA disease progression or effectively manage the associate pain. So, in 2022 it seemed pertinent to reflect on the use of animal models of OA in research and previous ICRS presentations, and ask in the ensuing decade: have we made any progress; has or should anything change in the use of OA-animal models in pathophysiology and/or therapeutic-development research; indeed, is animal model use in OA research appropriate and valuable?

Content

In a 2010 review of the available literature, 79 GM mouse strains were identified in which OA had specifically been studied, ~30% evaluated in induced post-traumatic (pt)OA and ~70% in spontaneous disease: 50 GMs had a significant detrimental effect and 8 significantly decreased disease (the latter all in ptOA models).¹ Subsequent reviews only 2-3 years later,^2,3 revealed 109 separate GM strains in which OA was studied, again ~75% of reported GMs had significant effects, ~60% of studies in induced ptOA models. In 2020 the most recent review (and associated searchable online database: http://skeletalvis.ncl.ac.uk/OATargets/)⁴ revealed studies on >450 genes/gene-products in OA, with >300 of these shown to have a significant protective or detrimental effect. Moving from a widely held belief of OA as a mechanical “wear-and-tear” disease, to the first in vivo evidence in 2005 for structural-disease modification despite ongoing joint biomechanical derangement by targeting a biological pathway (Adamts5),^5,6 and now >300 identified OA-effector genes/gene-products, is extraordinary knowledge-gain and progress. Why has this not translated into a booming and successful OA therapeutic-development pipeline?

One suggested issue is that the models most commonly used in animals do not align with the target patient population – the “mismatched OA-phenotype debate”.^2,7 Our reviews a decade ago showed that where a GM was evaluated in different models, ~70% had the same significant effect across models.³ However some genes only played a significant role in one model and not another, and one (Il6) even showed opposing (beneficial and worsening) roles in different models. This was some of the first evidence that different OA phenotypes (e.g. ptOA versus age-associate OA; OA with more or less synovitis) may have distinct molecular pathophysiology. In 2020 the most common OA models continued to be spontaneous/age-associated and ptOA (predominantly surgically-induced), and in 83% of cases where both models were studied for a given gene/gene-product, the same effect was observed.⁴ This data provides reassurance that ptOA models routinely used in drug discovery are likely predictive of effects in other phenotypes. However, that 20-30% of genes/gene-products seen as viable targets for therapeutic-development were only effective in particular OA-phenotypes, suggests a critical step in a drug-development pipeline should be testing high-priority candidates in multiple models/phenotypes. Furthermore, differences in outcome between/in key known patient risk-factors/sub-groups should be also explored, including direct comparison of the effects of age, sex and common co-morbidities encountered in the human OA patient population (e.g. obesity, diabetes, hypertension, CVD). Most OA-animal studies have and continue to use young otherwise healthy males while clinical trials have a preponderance of aged, overweight female patients.

There has been extensive commentary/discussion that disease-modelling in animals, and mice (which has become predominant in the last 15 years) in particular, may be problematic/inappropriate for translating to human efficacy due to genetic, anatomical and biomechanical differences. Again, the existing review evidence does not support this contention in OA research. Analysis of 15 available genome-wide profiles identifying differentially expressed genes in human OA, found statistically significant overlap between the protein coding OA-genes identified in mice regardless of OA model, with >70% differentially expressed in at least one human OA dataset. Where effects of specific GMs in mice were further evaluated using exogenous/pharmacologic approaches either in mice or other species (most commonly rat or rabbit), the same outcome was observed in 70-80% of studies.^2,4 Collectively this indicates that the majority of molecular mechanisms identified in mice appear to be similarly active in human OA, and that GM models are strong predictors of therapeutic/interventional and cross-species efficacy.

Along with the now rejected “wear-and-tear” pathophysiology paradigm, was a focus on OA as a primary disease of cartilage. A great deal of animal model research and associated histopathologic scoring systems reflected this cartilage-centric view.^1-3,8-15 It is now well accepted that OA is a disease of the entire joint with pathology in and cross-talk between different joint tissues, albeit with individual/patient, temporal, and disease-phenotype variability in the extent of different-tissue pathologies. Importantly, outcome evaluation in animal models is beginning to reflect the OA-joint-organ concept with more studies reporting on sub-chondral bone remodeling, osteophyte formation and synovitis in addition to cartilage pathology.^2-4,8 This joint-wide approach has provided new understanding of the significant association between pathologies in different joint tissues, and importantly how this differs with time/disease-progression and OA phenotype.¹⁶ While tissue-pathology associations in animal models reflect those in human OA (e.g. cartilage loss and subchondral sclerosis), interventions (genetic or exogenous) possible in animal studies have shed new light on the nature of these associations.^1,2,4 While significant amelioration of pathology in one tissue (e.g. cartilage) by a particular GM/exogenous-treatment may be accompanied by coordinate reduction in another (e.g. sub-chondral bone sclerosis, osteophytes), this is NOT always the case, even with other GMs/treatments targeting the same primary joint tissue and evaluated in the same OA model in the same species, age and sex. This has important implications for OA animal model research in discovery/pathophysiology (pathology in all tissues should be analysed and reported) and therapeutic-/drug-development (targeting one tissue will not necessarily improve whole-joint disease; optimal molecular pathways and targets should have beneficial effects in multiple joint tissues).

A final and important change in OA-animal model research has been the emergence of pain as both a key research topic and outcome measure.^3,8,17-20 Despite pain/disability being the major issue for OA-patients (human and veterinary), OA-animal-model research has tended to primarily focus on structural disease and its modification. Numerous validated quantitative pain outcome measures have been developed for use in small and large animals. Their more routine use is providing important insights into the association between OA structural pathology and pain, how this mimics associations seen in patients, but importantly how this changes with disease stage and initiating-cause/phenotype.²¹ Animal models are enabling dissection of cellular and molecular mechanisms of OA-pain, and again have provided novel insight into how this changes with disease stage/chronicity and may differ even in joint with apparently similar late-stage disease pathology and pain but incited by different factors – again “phenotype matters”.^19-21 As for structural pathology and its molecular drivers, ongoing work is needed to confirm the predictive validity of the pain measures in animals for clinically-relevant patient outcomes, that similar pathophysiological pathways are activated across species, ages and sexes, and how these may change/differ with disease phenotype.

So in answer to the questions raised: we have made extraordinary progress in understanding the pathophysiology of structural and symptomatic OA, and animal models have and will continue to play a key role in this. The organism-wide regulation of OA structural and symptomatic disease and the psychosocial modifiers of the pain/disability experience cannot be modelled in vitro or in silico. We have numerous well-validated animal models, and while none individually address the spectrum of disease in OA-patients, all can provide important information. The issue is how we as researchers better select, use and interpret the data from OA-animal models, to better align this with target patient populations/phenotypes to improve translation. This issue and a suggested checklist to help in selecting animal models and outcome measures has been addressed in recent review.²² I am very optimistic that adopting a more nuanced approach to selection and use of OA-animal will improve the quality of research evidence and its translational value and enable us in another 10 years to reflect on their key role in the development of successful OA therapies in clinical trial and even clinical practice.

References

References.

1. Little, C.B. & Fosang, A.J. Is cartilage matrix breakdown an appropriate therapeutic target in osteoarthritis--insights from studies of aggrecan and collagen proteolysis? Curr Drug Targets 11, 561-575 (2010).

2. Little, C.B. & Hunter, D.J. Post-traumatic osteoarthritis: from mouse models to clinical trials. Nat Rev Rheumatol 9, 485-497 (2013).

3. Little, C.B. & Zaki, S. What constitutes an "animal model of osteoarthritis"--the need for consensus? Osteoarthritis Cartilage 20, 261-267 (2012).

4. Soul, J., Barter, M.J., Little, C.B. & Young, D.A. OATargets: a knowledge base of genes associated with osteoarthritis joint damage in animals. Ann Rheum Dis 80, 376-383 (2020).

5. Glasson, S.S., et al. Deletion of active ADAMTS5 prevents cartilage degradation in a murine model of osteoarthritis. Nature 434, 644-648 (2005).

6. Stanton, H., et al. ADAMTS5 is the major aggrecanase in mouse cartilage in vivo and in vitro. Nature 434, 648-652 (2005).

7. Hunter, D.J. & Little, C.B. The great debate: Should Osteoarthritis Research Focus on "Mice" or "Men"? Osteoarthritis Cartilage 24, 4-8 (2016).

8. Blaker, C.L., Clarke, E.C. & Little, C.B. Using mouse models to investigate the pathophysiology, treatment, and prevention of post-traumatic osteoarthritis. J Orthop Res 35, 424-439 (2017).

9. Glasson, S.S., Chambers, M.G., Van Den Berg, W.B. & Little, C.B. The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the mouse. Osteoarthritis Cartilage 18 Suppl 3, S17-23 (2010).

10. Little, C.B., et al. The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in sheep and goats. Osteoarthritis Cartilage 18 Suppl 3, S80-92 (2010).

11. Cook, J.L., et al. The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the dog. Osteoarthritis Cartilage 18 Suppl 3, S66-79 (2010).

12. Gerwin, N., Bendele, A.M., Glasson, S. & Carlson, C.S. The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the rat. Osteoarthritis Cartilage 18 Suppl 3, S24-34 (2010).

13. Kraus, V.B., Huebner, J.L., DeGroot, J. & Bendele, A. The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the guinea pig. Osteoarthritis Cartilage 18 Suppl 3, S35-52 (2010).

14. Laverty, S., Girard, C.A., Williams, J.M., Hunziker, E.B. & Pritzker, K.P. The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the rabbit. Osteoarthritis Cartilage 18 Suppl 3, S53-65 (2010).

15. McIlwraith, C.W., et al. The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the horse. Osteoarthritis Cartilage 18 Suppl 3, S93-105 (2010).

16. Zaki, S., Smith, M.M., Smith, S.M. & Little, C.B. Differential patterns of pathology in and interaction between joint tissues in long-term osteoarthritis with different initiating causes: phenotype matters. Osteoarthritis Cartilage 28, 953-965 (2020).

17. Malfait, A.M. & Little, C.B. On the predictive utility of animal models of osteoarthritis. Arthritis Res Ther 17, 225 (2015).

18. Malfait, A.M., Little, C.B. & McDougall, J.J. A commentary on modelling osteoarthritis pain in small animals. Osteoarthritis Cartilage 21, 1316-1326 (2013).

19. Miller, R.E. & Malfait, A.M. Osteoarthritis pain: What are we learning from animal models? Best Pract Res Clin Rheumatol 31, 676-687 (2017).

20. Malfait, A.M., Miller, R.E. & Miller, R.J. Basic Mechanisms of Pain in Osteoarthritis: Experimental Observations and New Perspectives. Rheum Dis Clin North Am 47, 165-180 (2021).

21. Zaki, S., Smith, M.M. & Little, C.B. Pathology-pain relationships in different osteoarthritis animal model phenotypes: it matters what you measure, when you measure, and how you got there. Osteoarthritis Cartilage 29, 1448-1461 (2021).

22. Zaki, S., Blaker, C.L. & Little, C.B. OA foundations - experimental models of osteoarthritis. Osteoarthritis Cartilage (2021).