Gene expression analysis of subchondral bone, cartilage, and synovium in naturally occurring equine palmar/plantar osteochondral disease

Osteoarthritis (OA) is a disease of the entire joint but the relationship between pathological events in various joint tissues is poorly understood. We examined concurrent changes in bone, cartilage, and synovium in a naturally occurring equine model of joint degeneration. Joints (n = 64) were grossly assessed for palmar/plantar osteochondral disease (POD) in racehorses that required euthanasia for unrelated reasons and assigned a grade of 0 (n = 34), 1 (n = 17), 2 or 3 (n = 13) using a recognized grading scheme. Synovium, cartilage, and subchondral bone were collected for histological and gene expression analysis. Relations between POD grade, cartilage histological score, and gene expression levels were examined using one‐way analysis of variance or Kruskal–Wallis test and Spearman's correlation coefficient with corrections for multiple comparisons. Cartilage histological score increased in joints with POD grade 1 (p = 0.002) and 2 or 3 (p < 0.001) compared to 0. At grade 1, expression of COL1A1, COL2A1, and MMP1 increased and BGN decreased in subchondral bone while expression of BGN and ACAN decreased in cartilage. These changes further progressed at grades 2 and 3. POD grades 2 and 3 were associated with decreased expression of osteoclast inhibitor OPG and increased markers of cartilage degeneration (MMP13, COL1A1). Expression of the vascular endothelial growth factor decreased with POD grade and negatively correlated with cartilage histological score. Synovium showed no histological or transcriptomic changes related to pathology grade. Cartilage degeneration in POD is likely to be secondary to remodeling of the subchondral bone. Limited activation of proinflammatory and catabolic genes and moderate synovial pathology suggests distinct molecular phenotype of POD compared with OA.


| INTRODUCTION
Osteoarthritis (OA) is a chronic disease that can affect all joint tissues. Major pathological changes observed in OA are degradation of the articular cartilage, thickening or sclerosis of the subchondral bone, synovial inflammation, and hypertrophy of the joint capsule. 1,2 The structural changes in articular cartilage during OA result from disruption of tissue homeostasis at the molecular level. OA is characterized by the shift from extracellular matrix (ECM) synthesis to catabolism, promoted by proinflammatory mediators that are released due to mechanical insult or individual risk factors; such as obesity and ageing. 3,4 Tissue degradation is followed by inadequate repair response, involving chondrocyte hypertrophy 5 and changes in bone and cartilage ECM that affect biomechanical properties of the joint tissues. 6,7 Pathological changes in one joint tissue can affect other tissues due to close biomechanical and biochemical association and so it can be difficult to determine where the primary pathology is initiated. 8,9 There is growing evidence that subchondral bone alterations precede degeneration of the overlying articular cartilage in some of the OA phenotypes 2,10-12 and this sequence of pathological events is thought to result from both biomechanical and biochemical impact of subchondral bone perturbation on articular cartilage. 11,13 Therefore, better understanding of the crosstalk between bone and cartilage within healthy and diseased joints may help develop strategies to mitigate pathological changes or arrest OA progression.
Animal models that mimic symptoms and pathological features of human OA are essential for studying molecular events underlying joint tissue degeneration and developing interventions that could mitigate disease progression. [14][15][16] Various experimental animal models of OA have been shown to be effective in inducing cartilage and subchondral bone lesions characteristic of the late stage joint degeneration in humans. However, induction of OA through surgical procedure or enforced loading can induce rapid extensive changes (affecting subchondral bone and cartilage, simultaneously), which make it difficult to identify the early stages of disease development. 15,17,18 Horses can provide a naturally occurring model of OA that resembles human disease on the macroscopic, histological, and molecular level through increased expression of proinflammatory cytokines, and catabolic enzymes in the synovium, synovial fluid, and articular cartilage. 14,19,20 Joint disease in racehorses, associated with regular, intensive exercise, provides a model for the study of repetitive stress induced microstructural changes in articular cartilage and subchondral bone. 21 24 This can be attributed to the variations in loading of metacarpo-and metatarsophalangeal joints, due to morphology of the distal condyles of Mc/MtIII. 24 In early stages of POD, subchondral bone abnormalities occur whilst the articular cartilage remains grossly intact. 21,24,25 Therefore, POD might provide a plausible model for studying association between subchondral bone injury and development of other OA changes in other tissues such as articular cartilage and synovium.
The main aim of this study was to examine microstructural and molecular changes in different joint tissues associated with different stages of POD in racehorses. Furthermore, we aimed to compare identified changes to those previously reported in OA. We hypothesize that (i) subchondral bone is the predominantly affected tissue in early POD and that development of cartilage and synovial pathologic lesions follow the alterations in bone metabolic equilibrium as POD progresses and (ii) POD could serve as a model for early OA induced by repetitive mechanical stress, due to overlapping structural and molecular phenotype in POD and early OA. The potential impact of this study includes validation of POD as a model for human OA research and improving knowledge of relations between bone, cartilage, and synovium pathologic lesions concurrent with development of OA. Further details on sample and data collection and the racing population at Hong Kong jockey club were previously described by Pinchbeck et al. 24 Mc/MtPh joints were examined by gross observation and assigned a POD score using the grading system described by Barr et al. 26 Briefly, grade 0 represents no gross pathological lesions, grade 1-discolouration (bruising) of subchondral bone with none or minimal disruption of articular cartilage, grade 2-mild to moderate disruption of articular cartilage, and grade 3-disruption/collapse of articular surface ( Figure 1A).   Quantitative real-time PCR (qRT-PCR) was performed on an ABI 7300 system using MESA Blue SYBR Green reagent (Eurogentec) to examine gene expression in equine joint tissues (bone, cartilage, and synovium) as described before. 20 Gene expression levels were calculated using the -Δ 2 Ct method with glyceraldehyde 3-phosphate dehydrogenase as the reference gene, using the primers described in Table S1.

| Statistical analysis
All statistical analyses were performed using GraphPad Prism 6.0 (GraphPad Software). Due to the low sample number of POD

| POD is associated with degradation of cartilage but not synovitis
With increasing POD grade there was an increase in Mankin score (p < 0.05). This suggests that articular cartilage damage was present even at grade 1 POD ( Figure 1D) and became more pronounced at higher grades (p < 0.001). No difference was observed in histological synovitis score across POD grades ( Figure 1D).    Gene expression profiles associated with joint degenerative disease has been relatively well characterized in cartilage through human and animal model studies. 20,32,34,37,40,41 However, knowledge of the transcriptomic changes affecting the subchondral bone is still limited. In human OA research, subchondral bone sample collection is usually limited to the end-stage disease (joint replacement). 44,45 Mouse model studies can offer insight into early stages of joint degenration but they are less suitable for gene expression analysis in specific joint tissues due to their small size. 38 It has been previously reported that microstructural damage can be present in equine Mc/MtPh joints despite normal macroscopic appearance. 21,41 In this study, we aimed to account for that by using two outcome measures to assess joint degeneration, gross joint scoring and histological scoring of articular cartilage, and gene ex-