RésuméMusculoskeletal diseases, and particularly osteoarthritis (OA), are a concern in human and veterinary medicine. In OA, the progressive alteration of the joint components (i.e. cartilage, subchondral bone, synovium) is responsible for reduced joint function, quality-of-life, or performance.
OA induces progressive changes in three main components of cartilage: increase in water content and loss of type-II collagen and proteoglycans. These biochemical changes are associated with destruction of the zonal organization of collagen, with in fine cartilage defects and exposure of the subchondral bone (SB). SB is also subject to changes in composition (synthesis of abnormal type-I collagen and decrease in minerals deposition) and microarchitecture (increase in the number and thickness of the trabeculae, disorganization of the trabeculae), due to increased remodelling. These changes are responsible for inadequate biomechanical function of the bone, which detrimental for the cartilage.
Several causes can lead to OA such as ageing, joint instability or metabolic conditions. Inflammation is nowadays considered as the common feature of OA-leading mechanisms. Moreover, interplay between joint components may occur. The crosstalk between cartilage and SB has been increasingly studied during the last decade. Chondrocytes are able to increase bone remodelling, bone turn over and to induce a decrease in density of SB. Osteoblasts can modify chondrocyte synthesis and stimulate synthesis of matrix metaloproteinases (i.e. MMP-3, MMP-13); a decrease of cartilage-constitutive molecules (i.e. aggrecan and type-II collagen) is observed. Both articular cartilage and SB changes can be present early after OA induction. The modification of bone biology in animal models has been proved to limit or stop OA initiation and further cartilage damages, suggesting an important role of SB in OA initiation.
Importantly, cartilage has limited ability to self-repair adequately, mainly because of the avascularity of cartilage, with poor access for regenerative cells or for migration of chondrocyte to the damaged area. It is therefore important to provide an early detection of structural and compositional changes associated with OA. The spontaneous healing of cartilage defect mainly leads to the filling with fibrocartilage or cartilage-like tissue, with a slight improvement of tissue organization overtime but without recovery of perfect HC organization. Finally, the tissue produced by naturally occurring repair of cartilage is most of time unable to integrate with the adjacent cartilage leading to inhomogeneities of cartilage surface, to micro-motion and potential damages to the repair tissue itself.
Due to the limited ability of appropriate self-repair of joint tissue, to the multiple structural and compositional changes that can occur, and the complexity of the OA process, it is important to better understand the early pathological processes and to better detect early structural and compositional changes that are associated to the disease. In addition, the validation of in vivo early diagnostic techniques is essential to promote the development of treatments and test them in vivo. Gold standards methods (histology or biochemistry) are invasive and most of time detrimental for the targeted tissue, inducing lesion because of the sampling method (i.e., excision of cartilage sample). Furthermore, in longitudinal studies testing the effect of a treatment, we should ideally be able to assess the structure and composition of cartilage at baseline noninvasively, to optimally identify the changes induced by the treatment itself.
In this context, this thesis aimed:
1/ To compare magnetic resonance imaging (MRI) and computed tomography (CT) techniques in their ability to detect subtle cartilage structural changes.
2/ To review the efficacy of MRI- and CT-based imaging techniques to identify compositional changes of cartilage.
3/ To assess the efficacy of chemical or mechanical subchondral bone stimulation in inducing cartilage compositional or structural changes in an in vivo ovine model
4/ To test MRI T2 mapping and compositional CT to identify the early structural or compositional changes induced by subchondral bone stimulation.
Firstly, the assessment of CT arthrography (CTA) and MRI to detect structural cartilage defect implied two models of subtle naturally occurring cartilage defects (the equine metacarpo-/metatarso-phalangeal joint and the ovine stifle). Initially, MRI and CTA were compared in their ability to detect subtle cartilage defects in a series of equine metacarpo-/metatarso-phalangeal joint collected at a slaughterhouse and without evidence of articular disease. CTA sensitivity and specificity were 82% and 96%, respectively, and were significantly higher than those of MRI (41% and 93%, respectively) to detect cartilage defects in the metacarpo-/metatarso-phalangeal joint. The intra and inter-rater agreements were 0.96 and 0.92, respectively, and 0.82 and 0.88, respectively, for CTA and MRI. As CTA was found to be an adequate tool to detect subtle cartilage defects, we implemented this technique in our animal model: the sheep. We proved that this technique is feasible ex vivo, but also in vivo with sheep under anaesthesia; and gives an accurate assessment of the cartilage defects in the ovine knee, with high sensitivity (81.8%) and high specificity (95.2%). This technique was therefore considered as a useful tool to assess cartilage and to select only animals with healthy cartilage at baseline for experimentation.
Secondly, a review of the literature highlighted multiple imaging techniques focusing on water, glycosaminoglycan or collagen of cartilage. CT-based imaging technique (contrast-enhanced computed tomography (CECT), or quantitative CT requires the injection of a contrast medium, in order to assess the change in cartilage X-Ray attenuation between pre- and post-contrast images. This change in attenuation is associated to the ability of the contrast medium to penetrate cartilage depending primarily on its glycosaminoglycan content. MRI-based techniques such as glycosaminoglycan chemical-exchange saturation transfer and sodium imaging have been proved to correlate to glycosaminoglycan content, but require specific device and/or high field magnets, that are seldom available. Long acquisition time are sometimes require for other imaging techniques (T1 rho imaging or diffusion weighted imaging), limiting their use in practice. The review of the literature revealed that nearly all compositional imaging technique require validation for veterinary research or clinical practice, and that delayed gadolinium-enhanced MRI of cartilage and T2 mapping appear to be the most applicable methods for compositional imaging of animal cartilage.
Thirdly, two methods of stimulation of the SB was performed to induce further cartilage changes and to describe bone and cartilage changes by histology, biochemistry, micro-computed tomography, T2 mapping and CECT. The sheep was selected as an animal model for in vivo experiments because of its availability, its easy handling, its similar anatomy to human, and its appropriate size to either harvest samples for multiple analysis, or use clinical medical imaging devices (MRI, CT) and to assess locomotion. The SB of the medial femoral condyle was reached by drilling a track from the medial aspect of this condyle to avoid injury to the other joint’s components (ligaments, cartilage and synovium). The stimulation technique was performed either mechanically (impact with a metal rod through the drill track) or chemically (ethanol injection with a needle through the drill track) depending on group allocation (N=9/group). The sheep were clinically observed and euthanized after 2, 4, 6 and 8 weeks (N=2/time-point and 1 control sheep euthanized at day 0). Several salient findings were observed, such as: the increase in apparent bone density and in trabecular thickness at micro-computed tomography, in both stimulation groups (ethanol and impact). Two opposite histological trends were observed between both groups. Ethanol-stimulated limbs showed a progressive alteration of cartilage at histology over time (i.e. increase in OARSI scores for cartilage) that could be explained by the change in subchondral bone stiffness induced by ethanol dehydration. Impact-stimulated limbs showed the inverse trend, with an improvement of cartilage histological characteristics, as suggested by the decrease in OARSI scores for cartilage over time. This inverse trend could be associate to the healing effect previously described for drilling, osteostixis or micro-picking.
The last aim of this thesis was to test advanced imaging techniques (CECT and T2 mapping) to image the alterations of cartilage following subchondral bone stimulation. We proved that these techniques were feasible in living sheep, but we were unable to identify significant differences between groups, or to correlate imaging findings with the histological changes. This could be due to multiple factors such as: the limited changes after stimulation, the limitations in imaging modalities to detect these limited changes, the restriction in limb positioning in living animals, the requirement for reduced acquisition time (to limit anaesthesia time and associated risks) and the subsequent compromise between short acquisition time and image quality (resolution, signal-to-noise ratio, volume averaging). Most of these techniques have been previously validated on osteochondral plugs, overpassing the encountered limitations, or on joint to compare fibrous cartilage to hyaline cartilage-like tissue, suggesting a larger range of changes between conditions than what was observed after subchondral bone stimulation in the ovine model.
This PhD thesis highlighted several perspectives for research and clinical applications. We demonstrated that the detection of cartilage defects was more accurate with CTA than MRI, ex vivo in the equine metacarpo-/metatarso-phalangeal joint and in vivo in the ovine stifle. However, the requirement of general anaesthesia (and associated risks), size of the body region of interest and the thickness of the structure of interest (and subsequent concerns about image resolution) could sometimes limit the use of powerful MRI or CT scans in practice.
In research, noninvasive baseline assessment of the joint should be part of the inclusion criteria.
Compositional imaging techniques (T2 mapping and CECT) were unable to demonstrate a correlation with the subtle changes induced by chemical or mechanical stimulation. In the future, nevertheless, compositional imaging should be continuously improved since early noninvasive identification of joint changes is important for research and development of therapies.
The subchondral bone insult model developed in this PhD thesis could be intensified. The main advantage of this model was the selective stimulation of the subchondral bone (without damages to the synovium, cartilage or ligaments), therefore limiting confusing factors such as synovitis. The optimal doses should be found for ethanol-stimulation techniques to induce significant changes. Nevertheless, the histological differences that we identified between both types of stimulation were interesting since they may be highlighted two different mechanisms, one leading to pathology and the other to repair. This deserves to be investigated further to better understand how disease develops but also how subchondral bone stimulation (like drilling, micro-picking, or other) can be useful.
|la date de réponse||23 août 2018|
|Superviseur||Jean-Michel Vandeweerd (Promoteur), Peter Clegg (Copromoteur), Simon Tew (Jury), Charles Nicaise (Jury), Mandy Peffers (Jury), Sarah Taylor (Jury), René Van Weeren (Jury) & Simon Tew (Jury)|
- animal model
- magnetic resonance imaging
- computed tomography
- subchondral bone
Attachement à un institut de recherche reconnus à l'UNAMUR