Early changes in osteochondral tissues in a rabbit model of post‐traumatic osteoarthritis

Concurrent osteoarthritic (OA) manifestations in bone and cartilage are poorly known. To shed light on this issue, this study aims to investigate changes in subchondral bone and articular cartilage at two time points after anterior cruciate ligament transection (ACLT) in a rabbit model. 2 (N = 16) and 8 (N = 10) weeks after ACLT, the subchondral bone structure, cartilage thickness, Osteoarthritis Research Society International (OARSI) score, fixed charged density (FCD), and collagen orientation angle were analyzed. OA related changes were evaluated by comparing the ACLT to the contralateral (C‐L) and control knees. Already 2 weeks after ACLT, higher trabecular number in the medial femoral condyle and femoral groove, greater OARSI score in the femoral condyles, and thinner trabeculae in the lateral tibial plateau and femoral groove were observed in ACLT compared to C‐L knees. Only minor changes of cartilage collagen orientation in the femoral condyles and femoral groove and smaller FCD in the femoral condyles, medial tibial plateau, femoral groove and patella were observed. 8 weeks post‐ACLT, the surgical knees had thinner subchondral plate and trabeculae, and smaller trabecular bone volume fraction in most of the knee locations. OARSI score was greater in the femoral condyle and lateral tibial plateau cartilage. FCD loss was progressive only in the femoral condyle, femoral groove, and patellar cartilage, and minor changes of cartilage collagen orientation angle were present in the femoral condyles, femoral groove, and lateral tibial plateau. We conclude that ACLT induces progressive subchondral bone loss, during which proteoglycan loss occurs followed by their partly recovery, as indicated by FCD results.


| INTRODUCTION
Post-traumatic osteoarthritis (PTOA) is a clinical phenotype of OA 1 caused by joint trauma. Substantial ligament or capsule injuries, or articular fractures, increase the risk of PTOA. 2 Anterior cruciate ligament transection (ACLT) is a commonly used method for inducing PTOA in animal models. ACLT leads to progressive cartilage degeneration, 3,4 as characterized by an increase in chondrocyte apoptosis, proteoglycan (PG) depletion, 5 roughening of the articular surfaces initially, and cartilage delamination and excavation later in the disease process. 6 These structural changes lead to decreased intrinsic moduli and increased hydration and hydraulic permeability of articular cartilage. 7 Cartilage fibrillation has been reported in the medial femoral condyle 2 weeks after ACLT, followed by osteophyte formation, cartilage erosion, and full thickness ulceration 8 weeks post-ACLT in the femoral condyles of rabbit knees. 8

weeks after a
unilateral ACLT, the number of apoptotic chondrocytes in the superficial zone of the cartilage increased, and cell apoptosis appeared in the middle zone. 9 In addition, ACLT can also induce changes in the composition and collagen organization of articular cartilage in rabbit knees. 9 weeks after ACLT, the glycosaminoglycan content decreased by 11%, and the water content increased by 7% in articular cartilage from the medial femoral condyle. 10 Similarly, the collagen network was changed and the PG content was reduced in the superficial zone of articular cartilage 11 9 weeks post-ACLT, and significant alterations in the collagen network were extended deeper in the lateral than the medial femoral condyle. 12 Subchondral bone is a mechanosensitive tissue providing mechanical support for articular cartilage. 13 Structural changes in the subchondral bone have been thought to occur prior to any observable degeneration in articular cartilage. 14 It is still unclear how the cartilage and subchondral bone change concurrently in the onset and progression of early OA. In primary (idiopathic) OA, a positive association between the Osteoarthritis Research Society International (OARSI) grade and the subchondral bone plate thickness was discovered, and aberrant bone formation was observed in the proximal tibias of total knee arthroplasty patients. 15,16 Studies on the human tibial plateaus from OA patients and the medial condyles from guinea pigs with spontaneous OA showed a reduction in trabecular rods and an increase in trabecular thickness. 17 During OA progression, the bone turnover, thinning of the trabecular structure, osteophytes, bone marrow lesions, and sclerosis of the subchondral bone increased. 18,19 There have been also various studies on animal models of PTOA. These have reported bone loss in the femoral condyles and tibial plateaus 4 and 8 weeks after ACLT surgery, 4 accompanied by a reduced bone mineral density. 20 Based on the previous observations, research on timing of structural changes of bone and cartilage after ACLT in rabbits is lacking. 21 In PTOA rabbit models, cartilage fibrillation, minor collagen orientation changes and significant fixed charged density (FCD) loss were observed 2 weeks after ACLT, till full thickness cartilage ulceration 8 weeks post-ACLT. 8,22 The ACLT-induced morphological degeneration of articular cartilage and subchondral bone was retained the same from 8 to 16 weeks. 21 Thus, 2 weeks post-ACLT seemed an appropriate earliest time point to study the onset and very early changes of PTOA, and 8 weeks post-ACLT should provide good information about the progress of these earliest OA changes. Furthermore, most previous studies focused on changes in either articular cartilage or subchondral bone, or were limited to a few locations and one time point, and therefore, they could not provide a comprehensive site-specific understanding of the initiation and progression of PTOA.
The aim of this study was to characterize the concurrent changes in subchondral morphology, as well as the collagen orientation and PG content of articular cartilage in PTOA rabbit knee joints 2 (early OA) and 8 (advanced OA) weeks after ACLT. Since thickness of cartilage and loading mode change in respect of anatomical region of knee, we hypothesized that ACLT induced changes are site-specific in the experiments. All the animals were housed in single cages and unilateral ACLT surgery was conducted on a random knee of 14 random rabbits. 23 The main procedure of the sample processing is shown in Figure 1. Animals were sacrificed 2 or 8 weeks after the ACLT surgery.
The detailed procedure of anesthesia, operative approach, postoperative analgesia, and euthanasia have been described previously. 23 Shortly, all rabbits in the surgery group were injected subcutaneously with a tranquilizer and painkiller 30 min before anesthesia. Rabbits then were subjected to deep surgical anesthesia using 5% isoflurane/oxygen initially followed by a maintenance flow of 1%-2% isoflurane/oxygen for the surgery. To access the ACL, a 2.5-3 cm incision was made at approximately 2 mm posterior to the patella and the patellar ligament through the joint capsule. The ACL was then visualized and a hooked probe was placed around the ACL, and the ACL was transected with a single cut along the probe using a scalpel. Following transection, the joint capsule was sutured with a 4-0 interrupted silk suture and the wound was closed with a simple continuous suture using 3-0 Vicryl followed by a mattress suture on the surface. The rabbits recovered postoperatively on a heating pad covered with a blanket until they woke up and went back to their cages. Rabbits were injected subcutaneously with painkillers twice daily for 2 days. At the appropriate time points, rabbits were sacrificed by intracardiac injection of pentobarbital sodium for the collection of knee samples. Knees were harvested by cutting the femur and tibia approxi-

| Articular cartilage assessment
Before µCT imaging, the thickness of articular cartilage was measured by using optical coherence tomography (Ilumien PCI Optimization System; Jt. Jude Medical) as previously described. 22 After the µCT imaging, samples were processed for histology with ethylenediaminetetraacetic acid decalcification, followed by dehydration and paraffin embedding. To avoid artefacts in the structural analysis caused by nonhomogenous staining, unstained sections were used for the assessment of collagen orientation. 29 The depth-wise collagen orientation angle (0°-parallel to cartilage surface, 90°-perpendicular to cartilage surface) was evaluated from unstained sections with a quantitative polarized light microscope 30 (Supplementary Material). The distribution profiles of the collagen orientation angle and FCD were laterally averaged for each section, and then the averaged profiles were averaged again over these three sections. The finally acquired averaged profiles were analyzed from a 150-µm-wide region (from the cartilage surface to the cartilage-bone interface) to form the depth-wise collagen orientation angle and FCD profiles.

| Statistical analysis
For the depth-wise statistical comparisons of cartilage structure, the FCD and collagen orientation profiles were linearly interpolated to 100 points. A linear mixed regression model without confidence interval adjustment was used for testing differences in the bone morphology, cartilage thickness, OARSI score, as well as the depth-wise collagen orientation angle and FCD between the ACLT, C-L and CNTRL groups 2 or 8 weeks after surgery. In addition to the analysis of depth-wise collagen orientation angle and FCD, the averaged collagen orientation angle and averaged FCD in the superficial layer of articular cartilage (within 0%-20% of cartilage depth) were also analyzed. In the model, animal identification and knee side (left/right) were defined as random effects, and the experimental groups (i.e., the ACLT, C-L, and CNTRL groups) as fixed effects. All the results are presented in the format of mean ± standard deviation and the significance level for all tests was α = 0.05.

| The assessment of subchondral structure deterioration
Representative 3D visualization of the subchondral plate and trabecular bone in each group are shown in Figure 3. 2 weeks after the ACLT F I G U R E 2 Visualization of cut histology images (Safranin-O staining) of osteochondral structure from the lateral and medial femoral condyles and tibial plateaus, femoral groove and patella of rabbit knees in the CNTRL, C-L, and ACLT groups 2 and 8 weeks after the surgery. The articular cartilage degenerates in the femoral condyles of the 2-week ACLT group against the C-L group and degenerates in the femoral condyles and lateral tibial plateau of the 8-week ACLT group against the C-L group. The degeneration of articular cartilage in the medial tibial plateau of the CNTRL and C-L group is much more than that of the ACLT group. The size of shown images: 2 mm × height. The images before cutting were used for the OARSI score evaluation. ACLT, anterior cruciate ligament transection; C-L, contralateral; CNTRL, age-matched control; OARSI, Osteoarthritis Research Society International [Color figure can be viewed at wileyonlinelibrary.com] surgery, a significantly higher trabecular number in the medial femoral condyle (2.14%, p = 0.003) and femoral groove (9.09%, p = 0.019), and lower trabecular thickness in the lateral tibial plateau (−6.67%, p = 0.007) and the femoral groove (−10.91%, p = 0.015) were discovered in the ACLT group, compared to the C-L group (Table 1)  Note: *Significant difference to the CNTRL group, *p < 0.05, **p < 0.01, ***p < 0.001; mean ± SD. # Significant difference between the ACLT and C-L group, # p < 0.05, ## p < 0.01, ### p < 0.001; mean ± SD.
compared to the C-L group ( Table 3). The OARSI score was higher in the lateral (p = 0.025) and medial (p = 0.042) femoral condyle in the ACLT group compared to the C-L group. The OARSI score was also higher in the medial femoral condyle (p = 0.006) when compared to the CNTRL group 2 weeks after ACLT. As an indication of primary OA in the control group knees at the 2-week time point, we observed a higher OARSI score in the medial tibial plateau of the CNTRL group  6.00 ± 4.69 6.13 ± 5.57 Note: *Significant difference to the CNTRL group, *p < 0.05, **p < 0.01, ***p < 0.001; mean ± SD. # Significant difference between the ACLT and C-L group, # p < 0.05, ## p < 0.01, ### p < 0.001; mean ± SD.
HUANG ET AL.

| 2563
in trabecular bone microstructure, cartilage composition, and histopathology as early as 2 weeks after the ACLT surgery, and these changes were tissue-, and location-specific. During the study period, we observed novel dynamics of PTOA progression with bone degeneration becoming more severe from 2 to 8 weeks, while site-specific recovery of PG content was observed in the patella and femoral groove during the same observation period.
In our previous study the bone loss was observed in the medial femoral condyle of the ACLT group compared to the C-L group, without differences in the lateral femoral condyles 4 weeks after  The response of the subchondral plate thickness to the ACLT surgery was insensitivity, which could be attributed to contradictory changes in the thickness of subchondral bone plate and calcified cartilage. Calcified cartilage thickness has been found to increase during the initiation and development of OA in post-ACLT animal models 15,35 and in our previous study. 36 It is possible that the ob- This study has some limitations. The ACLT animal model we used in our study does not mimic perfectly human ACL injuries and subsequent tissue alterations, because clinically ACL rupture is often accompanied by bone bruises and meniscus tears, which might contribute to apoptosis, bone changes, and the progression of joint degeneration. The difference in tissue turnover and loading patterns between our animal model and humans produces challenges for the translation of our results to clinical applications. However, the fast tissue turnover in rabbits makes it easier to study OA within a short time period. The use of animal models also allows investigating OA in a controlled environment that is not possible with humans.
The resolution of the µCT imaging we used was moderate and was selected to allow comparisons to our previous studies. 12 To detail further changes in the subchondral bone plate, a highresolution µCT is needed to separate the subchondral bone plate porosity and calcified cartilage. 38 The OARSI score in the medial tibial plateau was greater in the CNTRL group compared to the C-L and ACLT groups at the 2-week time point. We can only speculate that this may be ascribed to the presence of spontaneous OA in the CNTRL group at the beginning of the experiments. We cannot exclude the possibility that spontaneous OA may have also occurred in the other groups. It has been reported that approximately 12.5% skeletally mature rabbits have spontaneous OA in hips and knees as early as 1 year of age. 39 The current study is limited to structural changes in the osteochondral unit at two time points after surgery. More time points are required to study time-dependent sequences of joint changes.
HUANG ET AL.

| 2565
The synovial component should also be analyzed for a comprehensive understanding of inflammatory responses. Possible roles of upregulated inflammatory responses of cartilage and increased proportion of proinflammatory fatty acids in the same animal model of OA have been discussed in some of our earlier works. 23,40 Finally, some studies suggest that invasive ACLT may cause exacerbated bone and PG loss compared to less invasive mechanical ACL rupture. 32 In our study, the lack of a sham control group makes it challenging to assess possible surgery-induced changes (pain, inflammatory response, and unloading) to the osteochondral unit. Thus, in future studies it could be relevant to add a sham-operation group. 41

| CONCLUSION
We conclude that our rabbit model of PTOA shows very early and progressive subchondral bone deterioration. PG loss is progressive in the femoral condyles, but in the medial tibial plateau, femoral groove and patella, there is a very early loss of PG followed by partly recovery. The tibial plateaus seem to be sensitive to changes in the bone morphology, while the femoral condyles are more sensitive to changes in cartilage composition and structure. These findings provide vital clues to the location-specific diagnosis and clinical study on human OA at the early stage.