- Email: [email protected]
ort mice is unlikely due to greater vulnerability to mechanical trauma
Osteoarthritis and Cartilage 21 (2013) 756e763
Spontaneous osteoarthritis in Str/ort mice is unlikely due to greater vulnerability to mechanical trauma B. Poulet y *, T.A.T. Westerhof z, R.W. Hamilton x, S.J. Shefelbine z, A.A. Pitsillides y y Lifestyle Research Group, The Royal Veterinary College, Royal College Street, University of London, NW1 0TU, UK z Department of Bioengineering, Imperial College London, South Kensington Campus, London, SW7 2AZ, UK x Department of Materials, Imperial College London, Prince Consort Rd, South Kensington, SW7 2BP, UK
a r t i c l e i n f o
s u m m a r y
Article history: Received 4 September 2012 Accepted 21 February 2013
Objective: Relative contributions of genetic and mechanical factors to osteoarthritis (OA) remain illdefined. We have used a joint loading model found to produce focal articular cartilage (AC) lesions, to address whether genetic susceptibility to OA in Str/ort mice is related to AC vulnerability to mechanical trauma and whether joint loading influences spontaneous OA development. We also develop finite element (FE) models to examine whether AC thickness may explain any differential vulnerability to loadinduced lesions. Methods: Right knees of 8-week-old Str/ort mice were loaded, AC integrity scored and thickness compared to CBA mice. Mechanical forces engendered in this model and the impact of AC thickness were simulated in C57Bl/6 mice using quasi-static FE modelling. Results: Unlike joints in non-OA prone CBA mice, Str/ort knees did not exhibit lateral femur (LF) lesions in response to applied loading; but exhibited thicker AC. FE modeling showed increased contact pressure and shear on the lateral femoral surface in loaded joints, and these diminished in joints containing thicker AC. Histological analysis of natural lesions in the tibia of Str/ort joints revealed that applied loading increased OA severity, proteoglycan loss and collagen type II degradation. Conclusion: Genetic OA susceptibility in Str/ort mice is not apparently related to greater AC vulnerability to trauma, but joint loading modifies severity of natural OA lesions in the medial tibia. FE modelling suggests that thicker AC in Str/ort mice diminishes tissue stresses and protects against load-induced AC lesions in the LF but that this is unrelated to their genetic susceptibility to OA. Ó 2013 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
Keywords: Osteoarthritis Mouse Str/ort Trauma Modeling
Introduction Osteoarthritis (OA) is a complex disease with major genetic and mechanical contributions1,2. Despite attempts to define these contributions, the role of specific elements of the mechanical loading environment remains incompletely defined. It is possible to hypothesise that predisposition to OA is due to greater articular cartilage (AC) susceptibility to mechanical trauma3. Genetic contributions to human OA exhibit Mendelian transmission2,4, and highlight a contribution from environmental, predominantly mechanical factors5e7. These contributions might depend upon the extent to which each gene product interacts with
* Address correspondence and reprint requests to: B. Poulet, Centre for Rheumatology and Connective Tissue Diseases, UCL Medical School, Royal Free Campus, London, NW3 2PF, UK. Tel: 44-207-794-0500x33436. E-mail address: [email protected] (B. Poulet).
joint mechanics, but also upon which specific mechanical factors promote OA. It is known, for instance, that jogging does not modify OA risk5, but that squatting, heavy lifting and repetitive, high impact/intensity sports are strongly associated6,7. Mechanical properties of damaged and OA cartilage differ from normal; AC permeability increases in OA, resulting in decreased stiffness during compression and further damage8. Differences in mechanical properties of OA-prone (knee) and non-prone (ankle) cartilage have also been demonstrated. The latter shows increased equilibrium modulus and dynamic stiffness, and decreased permeability, facilitating greater resistance to mechanical trauma and OA9. It remains unknown, however, whether OA-prone joints are more susceptible to mechanical damage from the outset. The effect of additional joint use on OA-prone AC has been studied previously. Lapvetelainen et al.10,11 used exercise and found that this accelerated OA in ageing C57Bl/6 and transgenic mice harbouring mutated Col2a1. Whilst exercise studies allow some control, they do not allow direct control over the loads applied. We
1063-4584/$ e see front matter Ó 2013 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.joca.2013.02.652
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
have recently characterised a non-surgical model for controlled mechanical loading of murine knees that may help to identify the contribution made by precise mechanical factors to OA12. Applied joint loading in this model has already been found to produce focal AC lesions in the lateral femur (LF) in non-OA prone CBA13 and C57Bl/6 mice, and will be used herein to address whether mice susceptible to spontaneous OA, the Str/ort strain, exhibit greater AC vulnerability to such mechanical load-induced traumatic lesions13,14. Most Str/ort mice develop knee OA lesions spontaneously at 20 weeks of age on the medial tibial plateau (MT)13,15,16. CBA mice, the closest available parental strain17 show, in contrast, very low susceptibility to spontaneous OA13. The MT distribution of natural OA lesions in Str/ort knees clearly differs from the LF location of those produced by applied loads in our model12. This will also be exploited herein to determine, by comparison with known outcomes in aged-matched non-OA prone CBA mice, whether applied joint loading accelerates spontaneous OA development in the MT in young Str/ort mice. Finally, we develop finite element (FE) models of loaded joints to examine the influence of AC thickness on cartilage tissue stresses to define whether AC thickness may contribute to different AC vulnerabilities to load-induced lesions in mice. Materials and method Animals Male Str/ort, CBA and C57Bl/6 (Charles River, UK) mice were kept in polypropylene cages, with light/dark 12-h cycles, at 21 2 C, and fed ad libitum with maintenance diet (Special Diet Services, Witham, UK). All procedures complied with Animals (Scientific Procedures) Act 1986 and local ethics committee. Str/ort mice (from Prof. Roger Mason, London, UK) were maintained by brother/sister pairing. In vivo mechanical loading Right knees of 8-week-old Str/ort mice (n ¼ 8) were loaded as described18. The knee and ankle were placed in custom-made cups in deep flexion. Compression was applied with a servo-hydraulic materials testing machine (Model HC10, Dartec, Stourbridge, UK) through the knee in the upper cup in a trapezoidal wave, with 9 N peak for 0.05 s, rise and fall-time of 0.025 s and 9.9 s baseline holdtime, as previously used in CBA mice12. Baseline 2 N loads maintained tibiae in place between loading episodes. Forty cycles were applied, 3 times/week for 2 weeks to the right knee (left, internal control). Histology Mice were killed 2 days after final loading episode. Dissected right (loaded) and left (contra-lateral) knees were fixed in neutral buffered formalin, decalcified (Immunocal, Quartett, Berlin), waxembedded and 6 mm coronal sections cut. Multiple sections (five/ slide) from 120 mm intervals across the whole joint were stained with Toluidine blue (0.1% in 0.1 M acetate buffer, pH 5.6) and AC lesion severity graded12. Sections from left and right Str/ort joints (n ¼ 8; one section per joint) were stained with Safranin O or used for immunohistochemical detection of collagen-II degradation product, C1, 2C. Sections were dewaxed, endogenous peroxidase activity quenched (3% H2O2 in PBS; 15 min, 37 C), pepsin digested (3 mg/ml in 0.2 M HCl, 45 min, 37 C), blocked in 10% goat serum (1 h), incubated with antiC1, 2C antibody (overnight, 4 C; 1:800; Ibex, Canada) or rabbit IgG
757
(control). After washing, sections were incubated with goat-antirabbit biotinylated antibody (1 h; 1:200; Biotin, Dako, Denmark) and then AvidineBiotin-peroxidase reagent (as per manufacturer’s instructions; Standard Vectastain AvidineBiotin Complex (ABC) kit, Vector Laboratories Inc., CA) followed by colour development with DAB. Grading of AC lesions AC lesion severity was scored in Str/ort mice by the methods of Chambers et al.19, consistent with an internationallyrecognized system20. Briefly, grade 0: normal; grade 1: rough surface or superficial zone lesions; grade 2: lesion down to the intermediate zone; grade 3: lesions down to tidemark or loss of AC; grades 4 and 5: AC loss across between 20% and 50% or 50e 80% of condylar surface; grade 6: loss with subchondral bone exposure. Multiple slides (w10), each containing five 6 mm sections sampled at 120 mm intervals spanning each entire joint were graded. Grading in each joint compartment (lateral/medial, tibia/ femur) allowed for a maximum (most severe) grade to be assigned in each section, and used to produce an overall ‘average’ maximum grade in each group of mice, for the entire joint and for each compartment12. In addition, a mean score was produced for each joint and for each compartment and these similarly used to produce an overall ‘average’ mean grade in each group of mice. Mean grades provide a measure of the ‘extent’ of AC lesion (representing relative volume) in each joint/compartment. Each joint was scored twice, independently by a single blinded observer, and an average used. Assessment of osteophyte maturity and non-AC joint tissues Osteophyte maturity was scored in Str/ort mice as described21. Briefly, multiple Toluidine blue-stained sections (as AC grading), spanning each joint were graded. Grade 0: no osteophyte; grade 1: predominantly cartilaginous; grade 2: mixed cartilage and bone; grade 3: predominantly bone with marrow spaces. Maximum scores on internal and external condylar margins were recorded. Prevalence of obvious morphological deviation from normal histology in synovial, meniscal and ligament tissues was also noted in sections in which AC lesions were graded (w10 slides per joint12). Measurement of AC thickness In order to assess the basis for differences in trauma susceptibility, AC thickness in non-loaded 8-week-old Str/ort and CBA mice (control strain) was measured in Toluidine Blue-stained sections. Distance between AC surface and subchondral bone was measured perpendicularly to the surface in all compartments using an eyepiece graticule (n ¼ 6 animals/group). Four measurements per section in each compartment were made in multiple sections across the entire joint (w10 sections/joint, at 120 mm intervals) and average thickness calculated (lateral/medial, tibia/femur; 40 total measurements/compartment in each joint). Statistical analysis Statistical analysis of AC lesion grades compared loaded (right) and contra-lateral non-loaded (left) joints by paired Wilcoxon’s signed-rank test. Loaded and contra-lateral joints were compared for osteophyte maturity and non-AC tissue changes by Fisher exact test. AC thickness in CBA and Str/ort mouse joints was compared using unpaired t-test. P < 0.05 was considered statistically significant.
758
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
FE modeling We hypothesized that AC lesions were caused by local contact stresses, which are influenced by AC thickness. FE modeling was used to determine tissue stresses. Right hind legs of three 8-weekold C57Bl/6 mice were stored in PBS-soaked gauze at 90 C until scanned and analyzed. Imaging and loading Thawed samples (n ¼ 3) were positioned in the loading cups identically to the in vivo procedure. The loading rig, adapted to be X-ray translucent, was placed in the microCT (VjTomejX; General Electric). 2 N pre-load was considered the “non-loaded” state as used in the in vivo setting. The knee joint was scanned (85 kV, 111 mA, 19 mm isotropic voxels), forces then increased to 5, 9 and 12 N and re-scanned after each increment. Non-loaded and 12 N scans were used to generate the FE model; the whole range was used to determine rigid body tibia and femur motion. FE model 3D geometrical solid representations were created from microCT scans (Materialise Mimics v14.0 Leuven, Belgium). Femur and tibia were segmented in the non-loaded scan and triangular surfaces meshes made using 3-Matic v5.1 (Materialise, Leuven, Belgium; Fig. 1). The mesh was refined in the joint contact area to create smaller elements and improve subsequent FE calculations. FE software Abaqus v6.10 (Providence, Rhode Island, USA) was used to convert the triangular surface mesh to a tetrahedral volume mesh. Because AC was not seen in scans, the cartilage layers were created by lofting tibial plateau and femoral condyle surfaces to create wedge elements. We made three different models with 20, 40 and 80 mm AC thicknesses. The base of each wedge AC element matched the tetrahedral elements of the subchondral bone, sharing nodes at the interface. The position of the femur relative to the tibia was adjusted for the three different cartilage layers to obtain the same initial gap distance between AC surfaces. The parts were considered as isotropic linear elastic materials with material properties elastic modulus E ¼ 18 GPa and Poisson ratio n ¼ 0.3 for bone22, and E ¼ 6 MPa and n ¼ 0.49 for cartilage23. Though cartilage
is truly biphasic, previous studies have shown that at short time scales it can be accurately represented by linear elastic models24. Surface to surface contact was defined between the tibia and femur AC, with a 0.01 friction coefficient25,26. Femoral displacement relative to the tibia during loading was determined from 3D images of non-loaded and 12 N scans. Surfaces of loaded and non-loaded tibiae were registered in Mimics, generating a rotation matrix and translation vector. This transformation was applied to the loaded scan to align tibiae in the two scans. The change in femoral position with loading was evident when the tibia from both scans were aligned and the femur position compared producing another rotation matrix and translation vector that defined the motion of the femur from unloaded to loaded state (Fig. 1). This femoral motion was used to apply the displacements on the nodes on the proximal end of the femur. Nodes at the distal end of the tibia were constrained in all directions. Because we directly modeled the motion instead of applied force and since the menisci serve little role in stress distribution in this extreme flexed position, menisci and ligaments, which have an influence in defining the motion, were not modeled (Fig. 1). The quasi-static contact model ran for the three different cartilage thicknesses until it reached the specified displacement. Results Applied loading does not induce lesions in the LF of Str/ort knee joints We have shown previously that 2 weeks of applied loading induced localized AC lesions in the LF of 8-week-old CBA mice12. To examine susceptibility of genetically OA-prone AC to load-induced lesions, we loaded 8-week-old Str/ort mice, before overt spontaneous OA was apparent, and examined LF cartilage for mechanical damage. We found that 8-week-old Str/ort mice did not develop significant LF lesions following 2 weeks of applied loading (Fig. 2; n ¼ 8) and that, similar to other strains12, load-induced lesions were lacking in the lateral tibia and medial femur compartments (data not shown).
Fig. 1. FE geometry and mesh and displacement of the femur relative to the tibia during loading. (A) FE with the assigned boundary conditions, bone is in green and cartilage in blue. Arrows show the direction of the loads applied. (B) Segmented data from the microCT scans showing the changes in positions of the femur relative to the tibia before (green) and during loading (red). This was used to determine the direction of the loading for the FE modeling described in A. (C and D) Contact pressure on the femoral (C) and tibial (D) condyles. (E and F) Contact shear on the femoral (E) and tibial (F) condyles. (Femurs are viewed from the posterior side and tibia from the top.)
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
759
Fig. 2. Loading does not induce AC lesions in the LF of 8-week-old Str/ort mice. (A) Mean and maximum AC lesion severity scores are not increased between control (open) and loaded (hashed) Str/ort mouse knee joints (n ¼ 8). (B) AC was thicker in all compartments of the knee joint in 8-week-old Str/ort mice (white) compared to CBA mice (grey). Data shown as mean with 95% CI, statistical significance is marked P < 0.001 between CBA and Str/ort mice.
Greater AC thickness in Str/ort mice may protect from load-induced lesion formation To identify baseline differences in AC structure which may underpin susceptibility to load-induced trauma, we measured LF cartilage thickness. This was significantly thicker in 8-week-old Str/ ort (105.5 mm (95% Confidence Interval (CI): 92.74, 118.2)) compared to CBA mice [76.6 mm (95% CI: 71.15, 82.15); P < 0.0001; n ¼ 6; Fig. 2(B)]. Indeed, AC in Str/ort mice was thicker in all compartments [Fig. 2(B)]. To examine the possible role of AC thickness in protecting against mechanically-induced lesions, a FE model based on loading of C57Bl/6 mice hindlimbs was created with three thicknesses (20, 40 and 80 mm). Femoral movement during loading was determined using microCT images from before and at peak loading, and revealed a rotational movement as well as compression [Fig. 1(B)]. This movement was simulated in the FE model and was found to yield increased contact pressure and shear at the lateral side of femoral and tibial condyles [Fig. 1(CeF)]; the location of contact in the femur resembles that of AC lesions in CBA mice12. No loadinduced AC lesions were found in the lateral tibia in Str/ort mice. Although magnitude of contact stresses were similar in tibia and femur, the FE model showed that their total area of contact was larger in the tibia compared to the femur during loading [Fig. 3(A and B)]. This FE model was also used to examine the effects of increased AC thicknesses on contact stresses in the femur with applied loading. This showed that increased AC thickness resulted in decreased contact pressure and shear, and increased contact area (0.072 mm2 at 20 mm, 0.121 mm2 at 40 mm and 0.267 mm2 at 80 mm thickness) in the LF [Fig. 3(CeH)]. Applied loading accelerates OA progression in the medial tibia of Str/ ort mice To establish the effect of applied joint loading on the development of natural OA, which develops primarily in the MT of Str/ort mice13,27, we analysed the MT of loaded 8-week-old Str/ort mice. AC lesions in the MT were more severe after 2 weeks of applied loading compared to their contra-lateral joints, with higher mean severity in loaded joints (from 0.12 (95% CI: 0.06, 0.31) to 0.77 (95% CI: 0.23, 1.31), P ¼ 0.031; n ¼ 8; Fig. 4). FE modelling showed that promotion of natural OA in response to loading coincided with smaller mechanical stresses in medial AC compartments than the lateral side. To determine whether this increase in AC lesion severity could be due to accelerated degeneration of osteoarthritic AC, we looked
for evidence of AC degradation28,29. Consistent with an acceleration of active OA matrix degradation, the MT of loaded 8-week-old Str/ ort mice showed greater loss of Safranin O staining, a sign of proteoglycan loss, and marked immunolabelling for a collagen type II degradation product, compared with contra-lateral joints (Fig. 4). Applied loading modifies non-AC joint tissues in CBA but Str/ort mice are relatively spared Applied loading did not induce any significant changes in osteophyte formation nor meniscal ossification in Str/ort mice, and only a trend for increased incidence of synovial thickening was observed (Tables I and II; four of eight mice). Loaded joints of Str/ort mice did, however, show increased prevalence of cruciate ligament cell hypertrophy and greater matrix staining intensity (Table II). Discussion Our model allows interplay between mechanical factors, including trauma, and genetic susceptibility to OA to be examined. By showing lower susceptibility to load-induced LF lesions in Str/ ort mice our data suggest that OA-prone joints are not necessarily more vulnerable to AC trauma. We also find using novel FE modeling of mouse knee joints during loading that increased AC thickness may account for this relative protection against loadinduced AC lesions. In addition, we find that joint loading accelerates natural OA lesions in the MT of Str/ort mice, with increased proteoglycan (PG) and collagen type II degradation. Previous studies using our non-invasive loading model have shown that application of mechanical loads creates reproducible AC lesions in the LF of non-OA prone CBA mice12. We now show, somewhat unexpectedly, that Str/ort mouse joints, which develop spontaneous early OA, are resistant to such trauma when an identical regime is used. Although the exact stimulus to the cartilage itself is not quantified, this suggests that this spontaneous OA is not necessarily due to a greater vulnerability of AC to mechanical challenge. Structural AC properties likely contribute to this mechanical resistance, as might greater AC thickness in Str/ort knees (vs CBA). To our knowledge, only Wilson et al.30 have examined whether AC thickness contributes to trauma vulnerability, concluding that thinner AC is more susceptible to tensile strains. Other modifications in AC composition including greater aggrecan levels31 and compressive stiffness may also contribute to this apparent protection27,32. Nevertheless, the characteristics dominating this protection remain undefined and may also include differences in
760
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
Fig. 3. Contact area during loading. (A) Contact pressures on the femoral lateral condyle increase, but remain in the same position during loading. Contour lines indicate relative magnitude of pressure for the contact area, which increases during loading. Black lines in the far right are sums of the contact area for the entire loading cycle and indicate the location of minimum (outer), mean, and maximum (inner) contact areas. This provides an indication of how the contact area moves during loading. (B) In contrast, the contact pressure on the lateral tibia moves during loading, resulting in increased contact area during loading. (CeH) Contour plots from the three different femoral cartilage thicknesses (20, 40 and 80 mm) on the posterior condyles of the femur, indicating highest contact pressure and shear in the thinnest cartilage. (Note: total displacement is similar in each of these three models, but different from the loading shown in [Fig. 1(CeF)] to accommodate the different thickness.)
underlying bone stiffness. Structural differences33,34 or age-related changes in bone architecture35, rather than increased AC vulnerability to trauma, may also explain the genetic susceptibility to spontaneous OA in Str/ort mice. FE analysis was used to model the role of AC thickness. This showed that thicker AC in Str/ort mice might generate decreased contact stresses to protect against load-induced LF trauma. Although AC thickness is uniform in the FE models, it clearly varies across the joint, and for this reason, multiple histological measures were made (across the entire joint) and mean thickness reported. However, the critical parameter in these FE models is the AC thickness at the area of contact during applied loading. Conveniently, the position of this contact area is relatively restricted and reproducible in our loading model, which provides greater confidence in making predictions based on our modeling. Although FE modeling showed stresses of similar magnitude in the lateral surfaces of tibia and femur, no AC lesions developed in the lateral tibia. Our findings indicate that stresses in the tibia were spread over a greater surface area than in the femur during applied loading, suggesting that AC damage in the femur is due to concentration of abnormal stresses over a small area. Eight-week-old Str/ort mice do not normally show any spontaneous, overt OA. We found that load application led to increased MT lesion severity in these young mice, with PG loss and Collagen type II degradation, suggesting an acceleration of OA. Previous studies have shown that OA cartilage has decreased compressive and tensile stiffness and may therefore be particularly sensitive to loading8,36,37. It is also possible that once superficial zone integrity is lost, as in early OA, the underlying AC experiences higher
stresses37. Indeed, our observations are consistent with lesion progression and modification of AC mechanical properties in the tibio-femoral joints of rabbits in response to chronic loading38. Our FE modeling showed that mechanical stresses generated by applied loading in the MT were lower than on the lateral side, but were nonetheless elevated compared to daily loading from ambulation. Thus applied loading may accelerate spontaneous medial OA. This would be consistent with previous findings indicating that increased mechanical stimulation of OA chondrocytes, engendered by changes in matrix composition, produce increased catabolism39e44. Alternatively, this may not be due to the direct effects of loading, but rather due to increased catabolic factors in the synovial fluid arising from tissues such as the synovium45,46, which also showed load-related intimal layer thickening. Our earlier studies revealed load-induced osteophyte formation at the external LF in young CBA mouse joints12. Osteophyte formation was, however, unaffected by loading in young Str/ort joints, possibly due to presence of mild natural osteophytes diluting their apparent induction. Prevalence of synovial intimal hyperplasia in loaded Str/ort mouse joints, although not significantly increased in this study, is also consistent with previous data12,47,48. Loadinduced meniscal chondrogenesis was not detected in Str/ort joints, suggesting that they are relatively resistant. The cruciate ligament, however, showed load-induced changes that were similar to those in CBA mice12. These data, together with the relative protection of LF AC against load-induced lesions, suggest that Str/ort mice are more resistant to joint trauma, but that the susceptibility of the cruciate ligaments does not differ.
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
761
Table I Effect of loading on osteophyte maturity
Medial tibia External Internal Medial femur External Internal Lateral tibia External Internal Lateral femur External Internal
Left non-loaded
Right loaded
0.06 (0.08, 0.21) Nil
0.68 (0.05, 1.42) 0.37 (0.05, 0.80)
0.06 (0.08, 0.21) 0.12 (0.17, 0.42)
0.62 (0.04, 1.20) 0.62 (0.19, 1.05)
Nil Nil
0.25 (0.13, 0.63) 0.12 (0.17, 0.42)
0.06 (0.08, 0.21) Nil
0.25 (0.06, 0.56) 0.12 (0.17, 0.42)
Osteophyte maturity scores (mean with 95% CI) at each of eight defined locations (average of maximum scores at each location; n ¼ 8). No statistical significance was noted between control and loaded knees (Wilcoxon’s signed-rank test).
Fig. 4. Loading increases AC degeneration in OA-prone medial tibia of 8-week-old Str/ ort mice. (A) Mechanical loading increased AC lesion severity mean and maximum scores in the medial tibia. Data shown as mean with 95% CI. (BeE) Safranin O staining was decreased around AC lesions in the MT of right loaded (C) vs left non-loaded (B) Str/ort mice (proteoglycan (PG) loss; B and C), and collagen degradation products, detected by immunohistochemistry (C1, 2C; D and E), were strongly labelled in the extracellular matrix (ECM) surrounding AC lesions in MT of right loaded (E) compared to left non-loaded (D) joints of Str/ort mice. (F) Section incubated with rabbit IgG lacking primary antibody (control for E).
Limitations must be considered. Applied loads increase severity of natural OA lesions in the MT and it is tempting to speculate that such loading will be deleterious at all OA stages. Advanced OA may, however, show distinct interplay with joint mechanics and studies in older mice are therefore required. With regards joint mechanics, it is also pertinent that the model used herein applies compressive load through a fully flexed knee joint and may represent abnormal loading at an abnormal location. Our studies are also, like most in vivo models, limited by an assumption that they are not complicated by changes in habitual gait; this is currently being
examined. Nor do our studies examine the contribution of subchondral bone architecture. In addition, possible inaccuracies may arise from using a single mouse strain (C57Bl/6) in the FE modeling, as differences in joint architecture, articulation and size could produce differences in contact stresses. Although this loading model is relatively controllable, we cannot ignore the fact that other tissues or joint shape may explain the lack of load-induced AC lesions in Str/ort mice. Use of a joint from one strain for FE modeling is unlikely to be a major problem as joint position is likely reproducible during loading; with cup design and holding loads (2 N) constraining limb bones to a specific position. This is supported by the reproducible location of load-induced AC lesions12. Ongoing studies show that AC lesion location is identical in different mouse strains and co-localize with maximal LF stresses described herein. Thus, use of a single joint is unlikely to undermine our conclusion that cartilage thickness plays a critical role in determining joint stresses. Further FE modeling and validation will strengthen our conclusions and may inform adjustments needed to enhance the physiological relevance of our loading model. Due to FE model simplifications some structures were not taken into account. Menisci and ligaments, critical to maintaining joint stability, were not modeled. We have shown in previous studies that flexion in our model places most contact area between femoral and tibial surfaces, with limited meniscal contribution12. The loading conditions (displacement) applied to our FE model were taken directly from an ex vivo specimen with menisci and ligaments intact, these structures have therefore been implicitly accounted for in the relative motion of the AC surfaces. Our FE modeling also represented the femur and tibia as solid deformable bodies without distinct trabecular structure. Our focus on the AC meant that inclusion of
Table II Effect of loading on prevalence of pathological change in non-AC joint tissues
Synovium Intimal layer hyperplasia Meniscus Chondrogenesis (lateral) Chondrogenesis (medial) Cruciate ligament Cell hypertrophy Matrix stain Cell clusters
Left, control
Right, loaded
0/8
4/8
0/8 0/8
2/8 0/8
0/8 0/8 0/8
6/8* 6/8* 4/8
Incidence (number affected/number in group) of pathologic changes determined histologically in joints loaded for 2 weeks (n ¼ 8 mice). Statistical significance between left control and right loaded joints is denoted * for P ¼ 0.007 (Fisher exact test).
762
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
trabecular structures would likely not affect the location of high magnitude stresses or the effects of AC thickness in this model. Furthermore, AC was modeled as isotropic elastic, rather than poroelastic material without simulating the joint capsule and surrounding synovial fluid. This is consistent with Carter and Beaupre24, supporting use of such properties when cyclic loads, as in our model, exceed 0.1 Hz. While previous studies suggest that a biphasic model, including fluid flow through the AC49, should be used, the high loading rate (0.1 s loading) and low AC permeability make it unlikely that significant fluid flow occurs24. It is possible that inclusion of synovial fluid would dampen the loading effects, thereby reducing stress magnitude but unlikely that it would alter the location of highest stresses and the trends observed with AC thickness. In addition, our FE model assumes homogenous cartilage properties with depth and uses previously cited bulk parameters to avoid relying on many unknown variables50. These assumptions may represent an oversimplification, as AC properties vary with depth and we therefore only reported contact/surface stresses. Future work will examine the effect of varying cartilage thickness across the joint and of poroelastic material properties on cartilage stresses. This simplified model does, however, inform us that joint stresses co-localize with lesions and are modified by AC thickness, and therefore indicates how cartilage lesions may be directly related to load-induced stresses. In conclusion, use of our in vivo murine knee joint loading model has allowed us to show that genetic susceptibility to OA is not necessarily linked to a greater AC vulnerability to mechanical damage, but that load application appears to accelerate OA which arises spontaneously in another compartment. We also found by direct measurement and via the generation of FE models that this protection from mechanical trauma may be related to greater AC thickness in Str/ort mouse knee joints. Together, these data indicate that genetic susceptibility to OA, at least in Str/ort mice, is not necessarily linked to a greater AC vulnerability to mechanical trauma.
3.
4.
5.
6.
7.
8.
9.
10.
11. Author contributions Conception and design: Pitsillides, Poulet, Shefelbine. Collection and assembly of data: Poulet, Westerhof, Hamilton. Analysis and interpretation of data: all co-authors. Statistical analysis: Poulet Drafting of article: Poulet, Westerhof Critical revision: Pitsillides, Poulet, Shefelbine, Hamilton Final Approval: Pitsillides, Poulet, Shefelbine Obtaining Funding: Pitsillides Competing interests The authors have no conflict of interest to declare.
12.
13. 14.
15.
Acknowledgments We are grateful to the Biotechnology and Biological Sciences and Sciences Research Councils, UK, and Arthritis Research UK (grant numbers: 16454, 18768). We remain indebted to Prof. Roger Mason (Imperial College London, UK) for providing us with our original Str/ort mice and to Dr Mark Chambers for his advice in their use.
16.
17.
References 18. 1. Conaghan PG. Update on osteoarthritis part 1: current concepts and the relation to exercise. Br J Sports Med 2002;36(5): 330e3. 2. Felson DT, Couropmitree NN, Chaisson CE, Hannan MT, Zhang Y, McAlindon TE, et al. Evidence for a Mendelian gene in
a segregation analysis of generalized radiographic osteoarthritis: the Framingham Study. Arthritis Rheum 1998;41(6): 1064e71. Thambyah A. A hypothesis matrix for studying biomechanical factors associated with the initiation and progression of posttraumatic osteoarthritis. Med Hypotheses 2005;64(6): 1157e61. Spector TD, Cicuttini F, Baker J, Loughlin J, Hart D. Genetic influences on osteoarthritis in women: a twin study. BMJ 1996;312(7036):940e3. Felson DT, Niu J, Clancy M, Sack B, Aliabadi P, Zhang Y. Effect of recreational physical activities on the development of knee osteoarthritis in older adults of different weights: the Framingham Study. Arthritis Rheum 2007;57(1):6e12. Kujala UM, Kettunen J, Paananen H, Aalto T, Battie MC, Impivaara O, et al. Knee osteoarthritis in former runners, soccer players, weight lifters, and shooters. Arthritis Rheum 1995;38(4):539e46. Amin S, Goggins J, Niu J, Guermazi A, Grigoryan M, Hunter DJ, et al. Occupation-related squatting, kneeling, and heavy lifting and the knee joint: a magnetic resonance imaging-based study in men. J Rheumatol 2008;35(8):1645e9. Knecht S, Vanwanseele B, Stussi E. A review on the mechanical quality of articular cartilage e implications for the diagnosis of osteoarthritis. Clin Biomech (Bristol, Avon) 2006;21(10):999e 1012. Treppo S, Koepp H, Quan EC, Cole AA, Kuettner KE, Grodzinsky AJ. Comparison of biomechanical and biochemical properties of cartilage from human knee and ankle pairs. J Orthop Res 2000;18(5):739e48. Lapvetelainen T, Hyttinen MM, Saamanen AM, Langsjo T, Sahlman J, Felszeghy S, et al. Lifelong voluntary joint loading increases osteoarthritis in mice housing a deletion mutation in type II procollagen gene, and slightly also in non-transgenic mice. Ann Rheum Dis 2002;61(9):810e7. Lapvetelainen T, Nevalainen T, Parkkinen JJ, Arokoski J, Kiraly K, Hyttinen M, et al. Lifelong moderate running training increases the incidence and severity of osteoarthritis in the knee joint of C57BL mice. Anat Rec 1995;242(2):159e65. Poulet B, Hamilton RW, Shefelbine S, Pitsillides AA. Characterizing a novel and adjustable noninvasive murine joint loading model. Arthritis Rheum 2011;63(1):137e47. Walton M. Degenerative joint disease in the mouse knee; histological observations. J Pathol 1977;123(2):109e22. Poulet B, Collinson RS, Pitsillides AA. Non-invasive loading of murine knee joints: a model for the study of OA? Trans Orthop Res Soc 2008;33:O214. Price JS, Chambers MG, Poole AR, Fradin A, Mason RM. Comparison of collagenase-cleaved articular cartilage collagen in mice in the naturally occurring STR/ort model of osteoarthritis and in collagen-induced arthritis. Osteoarthritis Cartilage 2002;10(3):172e9. Poulet B, Ulici V, Stone TC, Pead M, Gburcik V, Constantinou E, et al. Time-series transcriptional profiling yields new perspectives on susceptibility to murine osteoarthritis. Arthritis Rheum 2012;64(10):3256e66. Sokoloff L, Crittenden LB, Yamamoto RS, Jay Jr GE. The genetics of degenerative joint disease in mice. Arthritis Rheum 1962;5: 531e46. De Souza RL, Matsuura M, Eckstein F, Rawlinson SC, Lanyon LE, Pitsillides AA. Non-invasive axial loading of mouse tibiae increases cortical bone formation and modifies trabecular organization: a new model to study cortical and cancellous compartments in a single loaded element. Bone 2005;37(6): 810e8.
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
19. Chambers MG, Kuffner T, Cowan SK, Cheah KS, Mason RM. Expression of collagen and aggrecan genes in normal and osteoarthritic murine knee joints. Osteoarthritis Cartilage 2002;10(1):51e61. 20. Glasson SS, Chambers MG, Van Den Berg WB, Little CB. The OARSI histopathology initiative e recommendations for histological assessments of osteoarthritis in the mouse. Osteoarthritis Cartilage 2010;18(Suppl 3):S17e23. 21. Little CB, Barai A, Burkhardt D, Smith SM, Fosang AJ, Werb Z, et al. Matrix metalloproteinase 13-deficient mice are resistant to osteoarthritic cartilage erosion but not chondrocyte hypertrophy or osteophyte development. Arthritis Rheum 2009;60(12):3723e33. 22. Silva MJ, Brodt MD, Hucker WJ. Finite element analysis of the mouse tibia: estimating endocortical strain during three-point bending in SAMP6 osteoporotic mice. Anat Rec A Discov Mol Cell Evol Biol 2005;283A(2):380e90. 23. Beaupre GS, Stevens SS, Carter DR. Mechanobiology in the development, maintenance, and degeneration of articular cartilage. J Rehabil Res Dev 2000;37(2):145e51. 24. Carter DR, Beaupre GS. Linear elastic and poroelastic models of cartilage can produce comparable stress results: a comment on Tanck et al. (J Biomech 32:153e161, 1999). J Biomech 1999;32(11):1255e7. 25. Linn FC. Lubrication of animal joints. I. The arthrotripsometer. J Bone Joint Surg Am 1967;49(6):1079e98. 26. Linn FC, Radin EL. Lubrication of animal joints. 3. The effect of certain chemical alterations of the cartilage and lubricant. Arthritis Rheum 1968;11(5):674e82. 27. Mason RM, Chambers MG, Flannelly J, Gaffen JD, Dudhia J, Bayliss MT. The STR/ort mouse and its use as a model of osteoarthritis. Osteoarthritis Cartilage 2001;9(2):85e91. 28. Schmitz N, Laverty S, Kraus VB, Aigner T. Basic methods in histopathology of joint tissues. Osteoarthritis Cartilage, 18 Suppl 3:S113e6. 29. Billinghurst RC, Dahlberg L, Ionescu M, Reiner A, Bourne R, Rorabeck C, et al. Enhanced cleavage of type II collagen by collagenases in osteoarthritic articular cartilage. J Clin Invest 1997;99(7):1534e45. 30. Wilson W, van Burken C, van Donkelaar C, Buma P, van Rietbergen B, Huiskes R. Causes of mechanically induced collagen damage in articular cartilage. J Orthop Res 2006;24(2):220e8. 31. Gaffen JD, Gleave SJ, Crossman MV, Bayliss MT, Mason RM. Articular cartilage proteoglycans in osteoarthritic STR/Ort mice. Osteoarthritis Cartilage 1995;3(2):95e104. 32. Lyyra T, Arokoski JP, Oksala N, Vihko A, Hyttinen M, Jurvelin JS, et al. Experimental validation of arthroscopic cartilage stiffness measurement using enzymatically degraded cartilage samples. Phys Med Biol 1999;44(2):525e35. 33. Walton M, Elves MW. Bone thickening in osteoarthrosis. Observations of an osteoarthrosis-prone strain of mouse. Acta Orthop Scand 1979;50(5):501e6. 34. Pasold J, Engelmann R, Keller J, Joost S, Marshall RP, Frerich B, et al. High bone mass in the STR/ort mouse results from increased bone formation and impaired bone resorption and is associated with extramedullary hematopoiesis. J Bone Miner Metab 2013;31(1):71e81. 35. Uchida K, Urabe K, Naruse K, Kozai Y, Onuma K, MikuniTakagaki Y et al. Differential age-related bone architecture
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
763
changes between female and male STR/Ort mice. Exp Anim 61(1):59e66. Guilak F, Ratcliffe A, Lane N, Rosenwasser MP, Mow VC. Mechanical and biochemical changes in the superficial zone of articular cartilage in canine experimental osteoarthritis. J Orthop Res 1994;12(4):474e84. Arokoski JP, Hyttinen MM, Helminen HJ, Jurvelin JS. Biomechanical and structural characteristics of canine femoral and tibial cartilage. J Biomed Mater Res 1999;48(2):99e107. Roemhildt ML, Beynnon BD, Gardner-Morse M, Badger G, Grant C. Changes induced by chronic in vivo load alteration in the tibiofemoral joint of mature rabbits. J Orthop Res 30(9): 1413e22. Alexopoulos LG, Williams GM, Upton ML, Setton LA, Guilak F. Osteoarthritic changes in the biphasic mechanical properties of the chondrocyte pericellular matrix in articular cartilage. J Biomech 2005;38(3):509e17. Alexopoulos LG, Youn I, Bonaldo P, Guilak F. Developmental and osteoarthritic changes in Col6a1-knockout mice: biomechanics of type VI collagen in the cartilage pericellular matrix. Arthritis Rheum 2009;60(3):771e9. Lin PM, Chen CT, Torzilli PA. Increased stromelysin-1 (MMP-3), proteoglycan degradation (3B3- and 7D4) and collagen damage in cyclically load-injured articular cartilage. Osteoarthritis Cartilage 2004;12(6):485e96. Kurz B, Lemke AK, Fay J, Pufe T, Grodzinsky AJ, Schunke M. Pathomechanisms of cartilage destruction by mechanical injury. Ann Anat 2005;187(5e6):473e85. Bevill SL, Briant PL, Levenston ME, Andriacchi TP. Central and peripheral region tibial plateau chondrocytes respond differently to in vitro dynamic compression. Osteoarthritis Cartilage 2009;17(8):980e7. Nicodemus GD, Bryant SJ. Mechanical loading regimes affect the anabolic and catabolic activities by chondrocytes encapsulated in PEG hydrogels. Osteoarthritis Cartilage 2010;18(1): 126e37. Muroi Y, Kakudo K, Nakata K. Effects of compressive loading on human synovium-derived cells. J Dent Res 2007;86(8): 786e91. Raif el M. Effect of cyclic tensile load on the regulation of the expression of matrix metalloproteases (MMPs -1, -3) and structural components in synovial cells. J Cell Mol Med 2008;12(6A):2439e48. Cake MA, Smith MM, Young AA, Smith SM, Ghosh P, Read RA. Synovial pathology in an ovine model of osteoarthritis: effect of intraarticular hyaluronan (Hyalgan). Clin Exp Rheumatol 2008;26(4):561e7. Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A, et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52(11):3492e501. Wilson W, Van Donkelaar C, Van Rietbergen R, Huiskes R. The role of computational models in the search for the mechanical behavior and damage mechanisms of articular cartilage. Med Eng Phys 2005;27(10):810e26. Cao L, Youn I, Guilak F, Setton LA. Compressive properties of mouse articular cartilage determined in a novel microindentation test method and biphasic finite element model. J Biomech Eng 2006;128(5):766e71.
Spontaneous osteoarthritis in Str/ort mice is unlikely due to greater vulnerability to mechanical trauma B. Poulet y *, T.A.T. Westerhof z, R.W. Hamilton x, S.J. Shefelbine z, A.A. Pitsillides y y Lifestyle Research Group, The Royal Veterinary College, Royal College Street, University of London, NW1 0TU, UK z Department of Bioengineering, Imperial College London, South Kensington Campus, London, SW7 2AZ, UK x Department of Materials, Imperial College London, Prince Consort Rd, South Kensington, SW7 2BP, UK
a r t i c l e i n f o
s u m m a r y
Article history: Received 4 September 2012 Accepted 21 February 2013
Objective: Relative contributions of genetic and mechanical factors to osteoarthritis (OA) remain illdefined. We have used a joint loading model found to produce focal articular cartilage (AC) lesions, to address whether genetic susceptibility to OA in Str/ort mice is related to AC vulnerability to mechanical trauma and whether joint loading influences spontaneous OA development. We also develop finite element (FE) models to examine whether AC thickness may explain any differential vulnerability to loadinduced lesions. Methods: Right knees of 8-week-old Str/ort mice were loaded, AC integrity scored and thickness compared to CBA mice. Mechanical forces engendered in this model and the impact of AC thickness were simulated in C57Bl/6 mice using quasi-static FE modelling. Results: Unlike joints in non-OA prone CBA mice, Str/ort knees did not exhibit lateral femur (LF) lesions in response to applied loading; but exhibited thicker AC. FE modeling showed increased contact pressure and shear on the lateral femoral surface in loaded joints, and these diminished in joints containing thicker AC. Histological analysis of natural lesions in the tibia of Str/ort joints revealed that applied loading increased OA severity, proteoglycan loss and collagen type II degradation. Conclusion: Genetic OA susceptibility in Str/ort mice is not apparently related to greater AC vulnerability to trauma, but joint loading modifies severity of natural OA lesions in the medial tibia. FE modelling suggests that thicker AC in Str/ort mice diminishes tissue stresses and protects against load-induced AC lesions in the LF but that this is unrelated to their genetic susceptibility to OA. Ó 2013 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
Keywords: Osteoarthritis Mouse Str/ort Trauma Modeling
Introduction Osteoarthritis (OA) is a complex disease with major genetic and mechanical contributions1,2. Despite attempts to define these contributions, the role of specific elements of the mechanical loading environment remains incompletely defined. It is possible to hypothesise that predisposition to OA is due to greater articular cartilage (AC) susceptibility to mechanical trauma3. Genetic contributions to human OA exhibit Mendelian transmission2,4, and highlight a contribution from environmental, predominantly mechanical factors5e7. These contributions might depend upon the extent to which each gene product interacts with
* Address correspondence and reprint requests to: B. Poulet, Centre for Rheumatology and Connective Tissue Diseases, UCL Medical School, Royal Free Campus, London, NW3 2PF, UK. Tel: 44-207-794-0500x33436. E-mail address: [email protected] (B. Poulet).
joint mechanics, but also upon which specific mechanical factors promote OA. It is known, for instance, that jogging does not modify OA risk5, but that squatting, heavy lifting and repetitive, high impact/intensity sports are strongly associated6,7. Mechanical properties of damaged and OA cartilage differ from normal; AC permeability increases in OA, resulting in decreased stiffness during compression and further damage8. Differences in mechanical properties of OA-prone (knee) and non-prone (ankle) cartilage have also been demonstrated. The latter shows increased equilibrium modulus and dynamic stiffness, and decreased permeability, facilitating greater resistance to mechanical trauma and OA9. It remains unknown, however, whether OA-prone joints are more susceptible to mechanical damage from the outset. The effect of additional joint use on OA-prone AC has been studied previously. Lapvetelainen et al.10,11 used exercise and found that this accelerated OA in ageing C57Bl/6 and transgenic mice harbouring mutated Col2a1. Whilst exercise studies allow some control, they do not allow direct control over the loads applied. We
1063-4584/$ e see front matter Ó 2013 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.joca.2013.02.652
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
have recently characterised a non-surgical model for controlled mechanical loading of murine knees that may help to identify the contribution made by precise mechanical factors to OA12. Applied joint loading in this model has already been found to produce focal AC lesions in the lateral femur (LF) in non-OA prone CBA13 and C57Bl/6 mice, and will be used herein to address whether mice susceptible to spontaneous OA, the Str/ort strain, exhibit greater AC vulnerability to such mechanical load-induced traumatic lesions13,14. Most Str/ort mice develop knee OA lesions spontaneously at 20 weeks of age on the medial tibial plateau (MT)13,15,16. CBA mice, the closest available parental strain17 show, in contrast, very low susceptibility to spontaneous OA13. The MT distribution of natural OA lesions in Str/ort knees clearly differs from the LF location of those produced by applied loads in our model12. This will also be exploited herein to determine, by comparison with known outcomes in aged-matched non-OA prone CBA mice, whether applied joint loading accelerates spontaneous OA development in the MT in young Str/ort mice. Finally, we develop finite element (FE) models of loaded joints to examine the influence of AC thickness on cartilage tissue stresses to define whether AC thickness may contribute to different AC vulnerabilities to load-induced lesions in mice. Materials and method Animals Male Str/ort, CBA and C57Bl/6 (Charles River, UK) mice were kept in polypropylene cages, with light/dark 12-h cycles, at 21 2 C, and fed ad libitum with maintenance diet (Special Diet Services, Witham, UK). All procedures complied with Animals (Scientific Procedures) Act 1986 and local ethics committee. Str/ort mice (from Prof. Roger Mason, London, UK) were maintained by brother/sister pairing. In vivo mechanical loading Right knees of 8-week-old Str/ort mice (n ¼ 8) were loaded as described18. The knee and ankle were placed in custom-made cups in deep flexion. Compression was applied with a servo-hydraulic materials testing machine (Model HC10, Dartec, Stourbridge, UK) through the knee in the upper cup in a trapezoidal wave, with 9 N peak for 0.05 s, rise and fall-time of 0.025 s and 9.9 s baseline holdtime, as previously used in CBA mice12. Baseline 2 N loads maintained tibiae in place between loading episodes. Forty cycles were applied, 3 times/week for 2 weeks to the right knee (left, internal control). Histology Mice were killed 2 days after final loading episode. Dissected right (loaded) and left (contra-lateral) knees were fixed in neutral buffered formalin, decalcified (Immunocal, Quartett, Berlin), waxembedded and 6 mm coronal sections cut. Multiple sections (five/ slide) from 120 mm intervals across the whole joint were stained with Toluidine blue (0.1% in 0.1 M acetate buffer, pH 5.6) and AC lesion severity graded12. Sections from left and right Str/ort joints (n ¼ 8; one section per joint) were stained with Safranin O or used for immunohistochemical detection of collagen-II degradation product, C1, 2C. Sections were dewaxed, endogenous peroxidase activity quenched (3% H2O2 in PBS; 15 min, 37 C), pepsin digested (3 mg/ml in 0.2 M HCl, 45 min, 37 C), blocked in 10% goat serum (1 h), incubated with antiC1, 2C antibody (overnight, 4 C; 1:800; Ibex, Canada) or rabbit IgG
757
(control). After washing, sections were incubated with goat-antirabbit biotinylated antibody (1 h; 1:200; Biotin, Dako, Denmark) and then AvidineBiotin-peroxidase reagent (as per manufacturer’s instructions; Standard Vectastain AvidineBiotin Complex (ABC) kit, Vector Laboratories Inc., CA) followed by colour development with DAB. Grading of AC lesions AC lesion severity was scored in Str/ort mice by the methods of Chambers et al.19, consistent with an internationallyrecognized system20. Briefly, grade 0: normal; grade 1: rough surface or superficial zone lesions; grade 2: lesion down to the intermediate zone; grade 3: lesions down to tidemark or loss of AC; grades 4 and 5: AC loss across between 20% and 50% or 50e 80% of condylar surface; grade 6: loss with subchondral bone exposure. Multiple slides (w10), each containing five 6 mm sections sampled at 120 mm intervals spanning each entire joint were graded. Grading in each joint compartment (lateral/medial, tibia/ femur) allowed for a maximum (most severe) grade to be assigned in each section, and used to produce an overall ‘average’ maximum grade in each group of mice, for the entire joint and for each compartment12. In addition, a mean score was produced for each joint and for each compartment and these similarly used to produce an overall ‘average’ mean grade in each group of mice. Mean grades provide a measure of the ‘extent’ of AC lesion (representing relative volume) in each joint/compartment. Each joint was scored twice, independently by a single blinded observer, and an average used. Assessment of osteophyte maturity and non-AC joint tissues Osteophyte maturity was scored in Str/ort mice as described21. Briefly, multiple Toluidine blue-stained sections (as AC grading), spanning each joint were graded. Grade 0: no osteophyte; grade 1: predominantly cartilaginous; grade 2: mixed cartilage and bone; grade 3: predominantly bone with marrow spaces. Maximum scores on internal and external condylar margins were recorded. Prevalence of obvious morphological deviation from normal histology in synovial, meniscal and ligament tissues was also noted in sections in which AC lesions were graded (w10 slides per joint12). Measurement of AC thickness In order to assess the basis for differences in trauma susceptibility, AC thickness in non-loaded 8-week-old Str/ort and CBA mice (control strain) was measured in Toluidine Blue-stained sections. Distance between AC surface and subchondral bone was measured perpendicularly to the surface in all compartments using an eyepiece graticule (n ¼ 6 animals/group). Four measurements per section in each compartment were made in multiple sections across the entire joint (w10 sections/joint, at 120 mm intervals) and average thickness calculated (lateral/medial, tibia/femur; 40 total measurements/compartment in each joint). Statistical analysis Statistical analysis of AC lesion grades compared loaded (right) and contra-lateral non-loaded (left) joints by paired Wilcoxon’s signed-rank test. Loaded and contra-lateral joints were compared for osteophyte maturity and non-AC tissue changes by Fisher exact test. AC thickness in CBA and Str/ort mouse joints was compared using unpaired t-test. P < 0.05 was considered statistically significant.
758
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
FE modeling We hypothesized that AC lesions were caused by local contact stresses, which are influenced by AC thickness. FE modeling was used to determine tissue stresses. Right hind legs of three 8-weekold C57Bl/6 mice were stored in PBS-soaked gauze at 90 C until scanned and analyzed. Imaging and loading Thawed samples (n ¼ 3) were positioned in the loading cups identically to the in vivo procedure. The loading rig, adapted to be X-ray translucent, was placed in the microCT (VjTomejX; General Electric). 2 N pre-load was considered the “non-loaded” state as used in the in vivo setting. The knee joint was scanned (85 kV, 111 mA, 19 mm isotropic voxels), forces then increased to 5, 9 and 12 N and re-scanned after each increment. Non-loaded and 12 N scans were used to generate the FE model; the whole range was used to determine rigid body tibia and femur motion. FE model 3D geometrical solid representations were created from microCT scans (Materialise Mimics v14.0 Leuven, Belgium). Femur and tibia were segmented in the non-loaded scan and triangular surfaces meshes made using 3-Matic v5.1 (Materialise, Leuven, Belgium; Fig. 1). The mesh was refined in the joint contact area to create smaller elements and improve subsequent FE calculations. FE software Abaqus v6.10 (Providence, Rhode Island, USA) was used to convert the triangular surface mesh to a tetrahedral volume mesh. Because AC was not seen in scans, the cartilage layers were created by lofting tibial plateau and femoral condyle surfaces to create wedge elements. We made three different models with 20, 40 and 80 mm AC thicknesses. The base of each wedge AC element matched the tetrahedral elements of the subchondral bone, sharing nodes at the interface. The position of the femur relative to the tibia was adjusted for the three different cartilage layers to obtain the same initial gap distance between AC surfaces. The parts were considered as isotropic linear elastic materials with material properties elastic modulus E ¼ 18 GPa and Poisson ratio n ¼ 0.3 for bone22, and E ¼ 6 MPa and n ¼ 0.49 for cartilage23. Though cartilage
is truly biphasic, previous studies have shown that at short time scales it can be accurately represented by linear elastic models24. Surface to surface contact was defined between the tibia and femur AC, with a 0.01 friction coefficient25,26. Femoral displacement relative to the tibia during loading was determined from 3D images of non-loaded and 12 N scans. Surfaces of loaded and non-loaded tibiae were registered in Mimics, generating a rotation matrix and translation vector. This transformation was applied to the loaded scan to align tibiae in the two scans. The change in femoral position with loading was evident when the tibia from both scans were aligned and the femur position compared producing another rotation matrix and translation vector that defined the motion of the femur from unloaded to loaded state (Fig. 1). This femoral motion was used to apply the displacements on the nodes on the proximal end of the femur. Nodes at the distal end of the tibia were constrained in all directions. Because we directly modeled the motion instead of applied force and since the menisci serve little role in stress distribution in this extreme flexed position, menisci and ligaments, which have an influence in defining the motion, were not modeled (Fig. 1). The quasi-static contact model ran for the three different cartilage thicknesses until it reached the specified displacement. Results Applied loading does not induce lesions in the LF of Str/ort knee joints We have shown previously that 2 weeks of applied loading induced localized AC lesions in the LF of 8-week-old CBA mice12. To examine susceptibility of genetically OA-prone AC to load-induced lesions, we loaded 8-week-old Str/ort mice, before overt spontaneous OA was apparent, and examined LF cartilage for mechanical damage. We found that 8-week-old Str/ort mice did not develop significant LF lesions following 2 weeks of applied loading (Fig. 2; n ¼ 8) and that, similar to other strains12, load-induced lesions were lacking in the lateral tibia and medial femur compartments (data not shown).
Fig. 1. FE geometry and mesh and displacement of the femur relative to the tibia during loading. (A) FE with the assigned boundary conditions, bone is in green and cartilage in blue. Arrows show the direction of the loads applied. (B) Segmented data from the microCT scans showing the changes in positions of the femur relative to the tibia before (green) and during loading (red). This was used to determine the direction of the loading for the FE modeling described in A. (C and D) Contact pressure on the femoral (C) and tibial (D) condyles. (E and F) Contact shear on the femoral (E) and tibial (F) condyles. (Femurs are viewed from the posterior side and tibia from the top.)
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
759
Fig. 2. Loading does not induce AC lesions in the LF of 8-week-old Str/ort mice. (A) Mean and maximum AC lesion severity scores are not increased between control (open) and loaded (hashed) Str/ort mouse knee joints (n ¼ 8). (B) AC was thicker in all compartments of the knee joint in 8-week-old Str/ort mice (white) compared to CBA mice (grey). Data shown as mean with 95% CI, statistical significance is marked P < 0.001 between CBA and Str/ort mice.
Greater AC thickness in Str/ort mice may protect from load-induced lesion formation To identify baseline differences in AC structure which may underpin susceptibility to load-induced trauma, we measured LF cartilage thickness. This was significantly thicker in 8-week-old Str/ ort (105.5 mm (95% Confidence Interval (CI): 92.74, 118.2)) compared to CBA mice [76.6 mm (95% CI: 71.15, 82.15); P < 0.0001; n ¼ 6; Fig. 2(B)]. Indeed, AC in Str/ort mice was thicker in all compartments [Fig. 2(B)]. To examine the possible role of AC thickness in protecting against mechanically-induced lesions, a FE model based on loading of C57Bl/6 mice hindlimbs was created with three thicknesses (20, 40 and 80 mm). Femoral movement during loading was determined using microCT images from before and at peak loading, and revealed a rotational movement as well as compression [Fig. 1(B)]. This movement was simulated in the FE model and was found to yield increased contact pressure and shear at the lateral side of femoral and tibial condyles [Fig. 1(CeF)]; the location of contact in the femur resembles that of AC lesions in CBA mice12. No loadinduced AC lesions were found in the lateral tibia in Str/ort mice. Although magnitude of contact stresses were similar in tibia and femur, the FE model showed that their total area of contact was larger in the tibia compared to the femur during loading [Fig. 3(A and B)]. This FE model was also used to examine the effects of increased AC thicknesses on contact stresses in the femur with applied loading. This showed that increased AC thickness resulted in decreased contact pressure and shear, and increased contact area (0.072 mm2 at 20 mm, 0.121 mm2 at 40 mm and 0.267 mm2 at 80 mm thickness) in the LF [Fig. 3(CeH)]. Applied loading accelerates OA progression in the medial tibia of Str/ ort mice To establish the effect of applied joint loading on the development of natural OA, which develops primarily in the MT of Str/ort mice13,27, we analysed the MT of loaded 8-week-old Str/ort mice. AC lesions in the MT were more severe after 2 weeks of applied loading compared to their contra-lateral joints, with higher mean severity in loaded joints (from 0.12 (95% CI: 0.06, 0.31) to 0.77 (95% CI: 0.23, 1.31), P ¼ 0.031; n ¼ 8; Fig. 4). FE modelling showed that promotion of natural OA in response to loading coincided with smaller mechanical stresses in medial AC compartments than the lateral side. To determine whether this increase in AC lesion severity could be due to accelerated degeneration of osteoarthritic AC, we looked
for evidence of AC degradation28,29. Consistent with an acceleration of active OA matrix degradation, the MT of loaded 8-week-old Str/ ort mice showed greater loss of Safranin O staining, a sign of proteoglycan loss, and marked immunolabelling for a collagen type II degradation product, compared with contra-lateral joints (Fig. 4). Applied loading modifies non-AC joint tissues in CBA but Str/ort mice are relatively spared Applied loading did not induce any significant changes in osteophyte formation nor meniscal ossification in Str/ort mice, and only a trend for increased incidence of synovial thickening was observed (Tables I and II; four of eight mice). Loaded joints of Str/ort mice did, however, show increased prevalence of cruciate ligament cell hypertrophy and greater matrix staining intensity (Table II). Discussion Our model allows interplay between mechanical factors, including trauma, and genetic susceptibility to OA to be examined. By showing lower susceptibility to load-induced LF lesions in Str/ ort mice our data suggest that OA-prone joints are not necessarily more vulnerable to AC trauma. We also find using novel FE modeling of mouse knee joints during loading that increased AC thickness may account for this relative protection against loadinduced AC lesions. In addition, we find that joint loading accelerates natural OA lesions in the MT of Str/ort mice, with increased proteoglycan (PG) and collagen type II degradation. Previous studies using our non-invasive loading model have shown that application of mechanical loads creates reproducible AC lesions in the LF of non-OA prone CBA mice12. We now show, somewhat unexpectedly, that Str/ort mouse joints, which develop spontaneous early OA, are resistant to such trauma when an identical regime is used. Although the exact stimulus to the cartilage itself is not quantified, this suggests that this spontaneous OA is not necessarily due to a greater vulnerability of AC to mechanical challenge. Structural AC properties likely contribute to this mechanical resistance, as might greater AC thickness in Str/ort knees (vs CBA). To our knowledge, only Wilson et al.30 have examined whether AC thickness contributes to trauma vulnerability, concluding that thinner AC is more susceptible to tensile strains. Other modifications in AC composition including greater aggrecan levels31 and compressive stiffness may also contribute to this apparent protection27,32. Nevertheless, the characteristics dominating this protection remain undefined and may also include differences in
760
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
Fig. 3. Contact area during loading. (A) Contact pressures on the femoral lateral condyle increase, but remain in the same position during loading. Contour lines indicate relative magnitude of pressure for the contact area, which increases during loading. Black lines in the far right are sums of the contact area for the entire loading cycle and indicate the location of minimum (outer), mean, and maximum (inner) contact areas. This provides an indication of how the contact area moves during loading. (B) In contrast, the contact pressure on the lateral tibia moves during loading, resulting in increased contact area during loading. (CeH) Contour plots from the three different femoral cartilage thicknesses (20, 40 and 80 mm) on the posterior condyles of the femur, indicating highest contact pressure and shear in the thinnest cartilage. (Note: total displacement is similar in each of these three models, but different from the loading shown in [Fig. 1(CeF)] to accommodate the different thickness.)
underlying bone stiffness. Structural differences33,34 or age-related changes in bone architecture35, rather than increased AC vulnerability to trauma, may also explain the genetic susceptibility to spontaneous OA in Str/ort mice. FE analysis was used to model the role of AC thickness. This showed that thicker AC in Str/ort mice might generate decreased contact stresses to protect against load-induced LF trauma. Although AC thickness is uniform in the FE models, it clearly varies across the joint, and for this reason, multiple histological measures were made (across the entire joint) and mean thickness reported. However, the critical parameter in these FE models is the AC thickness at the area of contact during applied loading. Conveniently, the position of this contact area is relatively restricted and reproducible in our loading model, which provides greater confidence in making predictions based on our modeling. Although FE modeling showed stresses of similar magnitude in the lateral surfaces of tibia and femur, no AC lesions developed in the lateral tibia. Our findings indicate that stresses in the tibia were spread over a greater surface area than in the femur during applied loading, suggesting that AC damage in the femur is due to concentration of abnormal stresses over a small area. Eight-week-old Str/ort mice do not normally show any spontaneous, overt OA. We found that load application led to increased MT lesion severity in these young mice, with PG loss and Collagen type II degradation, suggesting an acceleration of OA. Previous studies have shown that OA cartilage has decreased compressive and tensile stiffness and may therefore be particularly sensitive to loading8,36,37. It is also possible that once superficial zone integrity is lost, as in early OA, the underlying AC experiences higher
stresses37. Indeed, our observations are consistent with lesion progression and modification of AC mechanical properties in the tibio-femoral joints of rabbits in response to chronic loading38. Our FE modeling showed that mechanical stresses generated by applied loading in the MT were lower than on the lateral side, but were nonetheless elevated compared to daily loading from ambulation. Thus applied loading may accelerate spontaneous medial OA. This would be consistent with previous findings indicating that increased mechanical stimulation of OA chondrocytes, engendered by changes in matrix composition, produce increased catabolism39e44. Alternatively, this may not be due to the direct effects of loading, but rather due to increased catabolic factors in the synovial fluid arising from tissues such as the synovium45,46, which also showed load-related intimal layer thickening. Our earlier studies revealed load-induced osteophyte formation at the external LF in young CBA mouse joints12. Osteophyte formation was, however, unaffected by loading in young Str/ort joints, possibly due to presence of mild natural osteophytes diluting their apparent induction. Prevalence of synovial intimal hyperplasia in loaded Str/ort mouse joints, although not significantly increased in this study, is also consistent with previous data12,47,48. Loadinduced meniscal chondrogenesis was not detected in Str/ort joints, suggesting that they are relatively resistant. The cruciate ligament, however, showed load-induced changes that were similar to those in CBA mice12. These data, together with the relative protection of LF AC against load-induced lesions, suggest that Str/ort mice are more resistant to joint trauma, but that the susceptibility of the cruciate ligaments does not differ.
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
761
Table I Effect of loading on osteophyte maturity
Medial tibia External Internal Medial femur External Internal Lateral tibia External Internal Lateral femur External Internal
Left non-loaded
Right loaded
0.06 (0.08, 0.21) Nil
0.68 (0.05, 1.42) 0.37 (0.05, 0.80)
0.06 (0.08, 0.21) 0.12 (0.17, 0.42)
0.62 (0.04, 1.20) 0.62 (0.19, 1.05)
Nil Nil
0.25 (0.13, 0.63) 0.12 (0.17, 0.42)
0.06 (0.08, 0.21) Nil
0.25 (0.06, 0.56) 0.12 (0.17, 0.42)
Osteophyte maturity scores (mean with 95% CI) at each of eight defined locations (average of maximum scores at each location; n ¼ 8). No statistical significance was noted between control and loaded knees (Wilcoxon’s signed-rank test).
Fig. 4. Loading increases AC degeneration in OA-prone medial tibia of 8-week-old Str/ ort mice. (A) Mechanical loading increased AC lesion severity mean and maximum scores in the medial tibia. Data shown as mean with 95% CI. (BeE) Safranin O staining was decreased around AC lesions in the MT of right loaded (C) vs left non-loaded (B) Str/ort mice (proteoglycan (PG) loss; B and C), and collagen degradation products, detected by immunohistochemistry (C1, 2C; D and E), were strongly labelled in the extracellular matrix (ECM) surrounding AC lesions in MT of right loaded (E) compared to left non-loaded (D) joints of Str/ort mice. (F) Section incubated with rabbit IgG lacking primary antibody (control for E).
Limitations must be considered. Applied loads increase severity of natural OA lesions in the MT and it is tempting to speculate that such loading will be deleterious at all OA stages. Advanced OA may, however, show distinct interplay with joint mechanics and studies in older mice are therefore required. With regards joint mechanics, it is also pertinent that the model used herein applies compressive load through a fully flexed knee joint and may represent abnormal loading at an abnormal location. Our studies are also, like most in vivo models, limited by an assumption that they are not complicated by changes in habitual gait; this is currently being
examined. Nor do our studies examine the contribution of subchondral bone architecture. In addition, possible inaccuracies may arise from using a single mouse strain (C57Bl/6) in the FE modeling, as differences in joint architecture, articulation and size could produce differences in contact stresses. Although this loading model is relatively controllable, we cannot ignore the fact that other tissues or joint shape may explain the lack of load-induced AC lesions in Str/ort mice. Use of a joint from one strain for FE modeling is unlikely to be a major problem as joint position is likely reproducible during loading; with cup design and holding loads (2 N) constraining limb bones to a specific position. This is supported by the reproducible location of load-induced AC lesions12. Ongoing studies show that AC lesion location is identical in different mouse strains and co-localize with maximal LF stresses described herein. Thus, use of a single joint is unlikely to undermine our conclusion that cartilage thickness plays a critical role in determining joint stresses. Further FE modeling and validation will strengthen our conclusions and may inform adjustments needed to enhance the physiological relevance of our loading model. Due to FE model simplifications some structures were not taken into account. Menisci and ligaments, critical to maintaining joint stability, were not modeled. We have shown in previous studies that flexion in our model places most contact area between femoral and tibial surfaces, with limited meniscal contribution12. The loading conditions (displacement) applied to our FE model were taken directly from an ex vivo specimen with menisci and ligaments intact, these structures have therefore been implicitly accounted for in the relative motion of the AC surfaces. Our FE modeling also represented the femur and tibia as solid deformable bodies without distinct trabecular structure. Our focus on the AC meant that inclusion of
Table II Effect of loading on prevalence of pathological change in non-AC joint tissues
Synovium Intimal layer hyperplasia Meniscus Chondrogenesis (lateral) Chondrogenesis (medial) Cruciate ligament Cell hypertrophy Matrix stain Cell clusters
Left, control
Right, loaded
0/8
4/8
0/8 0/8
2/8 0/8
0/8 0/8 0/8
6/8* 6/8* 4/8
Incidence (number affected/number in group) of pathologic changes determined histologically in joints loaded for 2 weeks (n ¼ 8 mice). Statistical significance between left control and right loaded joints is denoted * for P ¼ 0.007 (Fisher exact test).
762
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
trabecular structures would likely not affect the location of high magnitude stresses or the effects of AC thickness in this model. Furthermore, AC was modeled as isotropic elastic, rather than poroelastic material without simulating the joint capsule and surrounding synovial fluid. This is consistent with Carter and Beaupre24, supporting use of such properties when cyclic loads, as in our model, exceed 0.1 Hz. While previous studies suggest that a biphasic model, including fluid flow through the AC49, should be used, the high loading rate (0.1 s loading) and low AC permeability make it unlikely that significant fluid flow occurs24. It is possible that inclusion of synovial fluid would dampen the loading effects, thereby reducing stress magnitude but unlikely that it would alter the location of highest stresses and the trends observed with AC thickness. In addition, our FE model assumes homogenous cartilage properties with depth and uses previously cited bulk parameters to avoid relying on many unknown variables50. These assumptions may represent an oversimplification, as AC properties vary with depth and we therefore only reported contact/surface stresses. Future work will examine the effect of varying cartilage thickness across the joint and of poroelastic material properties on cartilage stresses. This simplified model does, however, inform us that joint stresses co-localize with lesions and are modified by AC thickness, and therefore indicates how cartilage lesions may be directly related to load-induced stresses. In conclusion, use of our in vivo murine knee joint loading model has allowed us to show that genetic susceptibility to OA is not necessarily linked to a greater AC vulnerability to mechanical damage, but that load application appears to accelerate OA which arises spontaneously in another compartment. We also found by direct measurement and via the generation of FE models that this protection from mechanical trauma may be related to greater AC thickness in Str/ort mouse knee joints. Together, these data indicate that genetic susceptibility to OA, at least in Str/ort mice, is not necessarily linked to a greater AC vulnerability to mechanical trauma.
3.
4.
5.
6.
7.
8.
9.
10.
11. Author contributions Conception and design: Pitsillides, Poulet, Shefelbine. Collection and assembly of data: Poulet, Westerhof, Hamilton. Analysis and interpretation of data: all co-authors. Statistical analysis: Poulet Drafting of article: Poulet, Westerhof Critical revision: Pitsillides, Poulet, Shefelbine, Hamilton Final Approval: Pitsillides, Poulet, Shefelbine Obtaining Funding: Pitsillides Competing interests The authors have no conflict of interest to declare.
12.
13. 14.
15.
Acknowledgments We are grateful to the Biotechnology and Biological Sciences and Sciences Research Councils, UK, and Arthritis Research UK (grant numbers: 16454, 18768). We remain indebted to Prof. Roger Mason (Imperial College London, UK) for providing us with our original Str/ort mice and to Dr Mark Chambers for his advice in their use.
16.
17.
References 18. 1. Conaghan PG. Update on osteoarthritis part 1: current concepts and the relation to exercise. Br J Sports Med 2002;36(5): 330e3. 2. Felson DT, Couropmitree NN, Chaisson CE, Hannan MT, Zhang Y, McAlindon TE, et al. Evidence for a Mendelian gene in
a segregation analysis of generalized radiographic osteoarthritis: the Framingham Study. Arthritis Rheum 1998;41(6): 1064e71. Thambyah A. A hypothesis matrix for studying biomechanical factors associated with the initiation and progression of posttraumatic osteoarthritis. Med Hypotheses 2005;64(6): 1157e61. Spector TD, Cicuttini F, Baker J, Loughlin J, Hart D. Genetic influences on osteoarthritis in women: a twin study. BMJ 1996;312(7036):940e3. Felson DT, Niu J, Clancy M, Sack B, Aliabadi P, Zhang Y. Effect of recreational physical activities on the development of knee osteoarthritis in older adults of different weights: the Framingham Study. Arthritis Rheum 2007;57(1):6e12. Kujala UM, Kettunen J, Paananen H, Aalto T, Battie MC, Impivaara O, et al. Knee osteoarthritis in former runners, soccer players, weight lifters, and shooters. Arthritis Rheum 1995;38(4):539e46. Amin S, Goggins J, Niu J, Guermazi A, Grigoryan M, Hunter DJ, et al. Occupation-related squatting, kneeling, and heavy lifting and the knee joint: a magnetic resonance imaging-based study in men. J Rheumatol 2008;35(8):1645e9. Knecht S, Vanwanseele B, Stussi E. A review on the mechanical quality of articular cartilage e implications for the diagnosis of osteoarthritis. Clin Biomech (Bristol, Avon) 2006;21(10):999e 1012. Treppo S, Koepp H, Quan EC, Cole AA, Kuettner KE, Grodzinsky AJ. Comparison of biomechanical and biochemical properties of cartilage from human knee and ankle pairs. J Orthop Res 2000;18(5):739e48. Lapvetelainen T, Hyttinen MM, Saamanen AM, Langsjo T, Sahlman J, Felszeghy S, et al. Lifelong voluntary joint loading increases osteoarthritis in mice housing a deletion mutation in type II procollagen gene, and slightly also in non-transgenic mice. Ann Rheum Dis 2002;61(9):810e7. Lapvetelainen T, Nevalainen T, Parkkinen JJ, Arokoski J, Kiraly K, Hyttinen M, et al. Lifelong moderate running training increases the incidence and severity of osteoarthritis in the knee joint of C57BL mice. Anat Rec 1995;242(2):159e65. Poulet B, Hamilton RW, Shefelbine S, Pitsillides AA. Characterizing a novel and adjustable noninvasive murine joint loading model. Arthritis Rheum 2011;63(1):137e47. Walton M. Degenerative joint disease in the mouse knee; histological observations. J Pathol 1977;123(2):109e22. Poulet B, Collinson RS, Pitsillides AA. Non-invasive loading of murine knee joints: a model for the study of OA? Trans Orthop Res Soc 2008;33:O214. Price JS, Chambers MG, Poole AR, Fradin A, Mason RM. Comparison of collagenase-cleaved articular cartilage collagen in mice in the naturally occurring STR/ort model of osteoarthritis and in collagen-induced arthritis. Osteoarthritis Cartilage 2002;10(3):172e9. Poulet B, Ulici V, Stone TC, Pead M, Gburcik V, Constantinou E, et al. Time-series transcriptional profiling yields new perspectives on susceptibility to murine osteoarthritis. Arthritis Rheum 2012;64(10):3256e66. Sokoloff L, Crittenden LB, Yamamoto RS, Jay Jr GE. The genetics of degenerative joint disease in mice. Arthritis Rheum 1962;5: 531e46. De Souza RL, Matsuura M, Eckstein F, Rawlinson SC, Lanyon LE, Pitsillides AA. Non-invasive axial loading of mouse tibiae increases cortical bone formation and modifies trabecular organization: a new model to study cortical and cancellous compartments in a single loaded element. Bone 2005;37(6): 810e8.
B. Poulet et al. / Osteoarthritis and Cartilage 21 (2013) 756e763
19. Chambers MG, Kuffner T, Cowan SK, Cheah KS, Mason RM. Expression of collagen and aggrecan genes in normal and osteoarthritic murine knee joints. Osteoarthritis Cartilage 2002;10(1):51e61. 20. Glasson SS, Chambers MG, Van Den Berg WB, Little CB. The OARSI histopathology initiative e recommendations for histological assessments of osteoarthritis in the mouse. Osteoarthritis Cartilage 2010;18(Suppl 3):S17e23. 21. Little CB, Barai A, Burkhardt D, Smith SM, Fosang AJ, Werb Z, et al. Matrix metalloproteinase 13-deficient mice are resistant to osteoarthritic cartilage erosion but not chondrocyte hypertrophy or osteophyte development. Arthritis Rheum 2009;60(12):3723e33. 22. Silva MJ, Brodt MD, Hucker WJ. Finite element analysis of the mouse tibia: estimating endocortical strain during three-point bending in SAMP6 osteoporotic mice. Anat Rec A Discov Mol Cell Evol Biol 2005;283A(2):380e90. 23. Beaupre GS, Stevens SS, Carter DR. Mechanobiology in the development, maintenance, and degeneration of articular cartilage. J Rehabil Res Dev 2000;37(2):145e51. 24. Carter DR, Beaupre GS. Linear elastic and poroelastic models of cartilage can produce comparable stress results: a comment on Tanck et al. (J Biomech 32:153e161, 1999). J Biomech 1999;32(11):1255e7. 25. Linn FC. Lubrication of animal joints. I. The arthrotripsometer. J Bone Joint Surg Am 1967;49(6):1079e98. 26. Linn FC, Radin EL. Lubrication of animal joints. 3. The effect of certain chemical alterations of the cartilage and lubricant. Arthritis Rheum 1968;11(5):674e82. 27. Mason RM, Chambers MG, Flannelly J, Gaffen JD, Dudhia J, Bayliss MT. The STR/ort mouse and its use as a model of osteoarthritis. Osteoarthritis Cartilage 2001;9(2):85e91. 28. Schmitz N, Laverty S, Kraus VB, Aigner T. Basic methods in histopathology of joint tissues. Osteoarthritis Cartilage, 18 Suppl 3:S113e6. 29. Billinghurst RC, Dahlberg L, Ionescu M, Reiner A, Bourne R, Rorabeck C, et al. Enhanced cleavage of type II collagen by collagenases in osteoarthritic articular cartilage. J Clin Invest 1997;99(7):1534e45. 30. Wilson W, van Burken C, van Donkelaar C, Buma P, van Rietbergen B, Huiskes R. Causes of mechanically induced collagen damage in articular cartilage. J Orthop Res 2006;24(2):220e8. 31. Gaffen JD, Gleave SJ, Crossman MV, Bayliss MT, Mason RM. Articular cartilage proteoglycans in osteoarthritic STR/Ort mice. Osteoarthritis Cartilage 1995;3(2):95e104. 32. Lyyra T, Arokoski JP, Oksala N, Vihko A, Hyttinen M, Jurvelin JS, et al. Experimental validation of arthroscopic cartilage stiffness measurement using enzymatically degraded cartilage samples. Phys Med Biol 1999;44(2):525e35. 33. Walton M, Elves MW. Bone thickening in osteoarthrosis. Observations of an osteoarthrosis-prone strain of mouse. Acta Orthop Scand 1979;50(5):501e6. 34. Pasold J, Engelmann R, Keller J, Joost S, Marshall RP, Frerich B, et al. High bone mass in the STR/ort mouse results from increased bone formation and impaired bone resorption and is associated with extramedullary hematopoiesis. J Bone Miner Metab 2013;31(1):71e81. 35. Uchida K, Urabe K, Naruse K, Kozai Y, Onuma K, MikuniTakagaki Y et al. Differential age-related bone architecture
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
763
changes between female and male STR/Ort mice. Exp Anim 61(1):59e66. Guilak F, Ratcliffe A, Lane N, Rosenwasser MP, Mow VC. Mechanical and biochemical changes in the superficial zone of articular cartilage in canine experimental osteoarthritis. J Orthop Res 1994;12(4):474e84. Arokoski JP, Hyttinen MM, Helminen HJ, Jurvelin JS. Biomechanical and structural characteristics of canine femoral and tibial cartilage. J Biomed Mater Res 1999;48(2):99e107. Roemhildt ML, Beynnon BD, Gardner-Morse M, Badger G, Grant C. Changes induced by chronic in vivo load alteration in the tibiofemoral joint of mature rabbits. J Orthop Res 30(9): 1413e22. Alexopoulos LG, Williams GM, Upton ML, Setton LA, Guilak F. Osteoarthritic changes in the biphasic mechanical properties of the chondrocyte pericellular matrix in articular cartilage. J Biomech 2005;38(3):509e17. Alexopoulos LG, Youn I, Bonaldo P, Guilak F. Developmental and osteoarthritic changes in Col6a1-knockout mice: biomechanics of type VI collagen in the cartilage pericellular matrix. Arthritis Rheum 2009;60(3):771e9. Lin PM, Chen CT, Torzilli PA. Increased stromelysin-1 (MMP-3), proteoglycan degradation (3B3- and 7D4) and collagen damage in cyclically load-injured articular cartilage. Osteoarthritis Cartilage 2004;12(6):485e96. Kurz B, Lemke AK, Fay J, Pufe T, Grodzinsky AJ, Schunke M. Pathomechanisms of cartilage destruction by mechanical injury. Ann Anat 2005;187(5e6):473e85. Bevill SL, Briant PL, Levenston ME, Andriacchi TP. Central and peripheral region tibial plateau chondrocytes respond differently to in vitro dynamic compression. Osteoarthritis Cartilage 2009;17(8):980e7. Nicodemus GD, Bryant SJ. Mechanical loading regimes affect the anabolic and catabolic activities by chondrocytes encapsulated in PEG hydrogels. Osteoarthritis Cartilage 2010;18(1): 126e37. Muroi Y, Kakudo K, Nakata K. Effects of compressive loading on human synovium-derived cells. J Dent Res 2007;86(8): 786e91. Raif el M. Effect of cyclic tensile load on the regulation of the expression of matrix metalloproteases (MMPs -1, -3) and structural components in synovial cells. J Cell Mol Med 2008;12(6A):2439e48. Cake MA, Smith MM, Young AA, Smith SM, Ghosh P, Read RA. Synovial pathology in an ovine model of osteoarthritis: effect of intraarticular hyaluronan (Hyalgan). Clin Exp Rheumatol 2008;26(4):561e7. Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A, et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52(11):3492e501. Wilson W, Van Donkelaar C, Van Rietbergen R, Huiskes R. The role of computational models in the search for the mechanical behavior and damage mechanisms of articular cartilage. Med Eng Phys 2005;27(10):810e26. Cao L, Youn I, Guilak F, Setton LA. Compressive properties of mouse articular cartilage determined in a novel microindentation test method and biphasic finite element model. J Biomech Eng 2006;128(5):766e71.