Computed tomography imaging and functional in silico modelling of cartilage, bone and tissue engineered cartilage constructs

Knee joint has a complex anatomy and it is the largest joint of the human body. The ends of the bones within the knee joint are covered with a smooth, nearly frictionless articular cartilage enabling smooth movement under normal conditions. Articular cartilage and the underlying bone, consisting of a subchondral bone plate and porous, trabecular bone underneath it, form an osteochondral unit. Articular cartilage consists mainly of interstitial fluid, collagen, and proteoglycans. It has a unique, depth-dependent structure: collagens form an arch-like network and towards the deep cartilage, proteoglycan content increases while water content decreases. The most common joint disease is osteoarthritis (OA), which often affects the knee joint and causes pain and disability for individuals and burden for society. In osteoarthritis, the structure and mechanical properties of the osteochondral unit undergo alterations. Cartilage undergoes a loss of proteoglycans and increased porosity, leading to fibrillation of the superficial tissue and ultimately resulting in complete cartilage loss. In contrast, bone experiences an increase in subchondral bone plate thickness and trabecular bone volume fraction. Despite these changes, a comprehensive understanding of their impact on the mechanical behaviour of the osteochondral unit remains elusive. In silico, i.e. computational, modelling provides a valuable avenue for gaining insights into the mechanical behaviour of tissues. Advanced material models accurately describe tissue mechanical responses, shedding light on internal behaviour−a challenge for conventional experimental testing. Diagnosis of OA is mainly based on radiographic findings and symptoms, such as pain and functional limitation of the joint. A major drawback of the present diagnostic practice is that once the pain is prevalent and osteochondral tissue changes are detectable in plain radiographs, cartilage is already irreversibly degenerated. If the disease could be detected earlier, interventions could be made earlier and the tissue degeneration could be slowed down. For this purpose, a triple contrast agent has been introduced for computed tomography (CT) imaging. The triple contrast agent consists of cationic, iodine-based CA4+, nonionic, gadolinium-based gadoteridol, and bismuth nanoparticles. Principally, CA4+ and gadoteridol diffuse into cartilage, reflecting its composition and functional properties, while bismuth nanoparticles remain in the synovial fluid, allowing cartilage segmentation. Despite its success in characterizing cartilage properties in vitro, the efficacy of the method in a clinical setting for entire knee joints has not been previously investigated. Cartilage has poor regenerative properties −damaged cartilage is unable to repair causing degeneration to be irreversible. Unfortunately, therapies for addressing cartilage deterioration have shown suboptimal long-term results. Tissue engineering, combining biomaterials, cells, and bioactive factors to produce substitutes for damaged tissue, has emerged for repairing and replacing cartilage. Recently, hydrogels with a reinforcing fibre mesh have been shown to have superior mechanical properties compared to other tissue engineering approaches. Properties of the most advanced reinforced hydrogels have mechanical properties approaching those of native cartilage. However, the functional properties of such constructs and how the mechanical stimulation during the tissue culture affects the neotissue formation are not fully understood. The aim of this PhD thesis project was to advance the imaging- and modellingbased characterization of osteoarthritic changes and gain insights into osteochondral tissue mechanical characteristics and their replication by means of tissue engineering. In study I, a finite element model of an osteochondral unit was created based on microcomputed tomography (µCT) of a rabbit knee joint. The model structure of the trabecular bone and subchondral bone plate that are located under the cartilage were altered and the mechanical response of the osteochondral unit was simulated under creep indentation loading. In study II, the triple contrast agent (CA4+, gadoteridol, and bismuth nanoparticles) was for the first time used in a clinical CT setting to image human cadaveric knee joints. Afterwards, osteochondral samples were extracted from the knee joints, imaged with µCT and biomechanically tested to investigate the relationship between contrast agent diffusion and cartilage biomechanical properties and evaluate the accuracy of CT-based segmentation. In studies III and IV, the structural and functional properties of tissue engineered cartilage constructs, consisting of GelMA-HAMA hydrogel reinforced with an mPCL fibre scaffold, were characterized. The biomechanical properties of the constructs with different fibre mesh pore sizes were revealed using fibril-reinforced poroelastic (FRPE) modelling. Constructs were seeded with chondrocytes and mechanically stimulated during tissue culture. The constructs were imaged with µCT using CA4+ as a contrast agent, and X-ray attenuation within the constructs was investigated in a depth-dependent manner and compared to the biochemically obtained glycosaminoglycan (GAG) content. The results of the study I indicate that simulated structural alterations in bone during early osteoarthritis have only a minimal effect on the mechanical stresses and strains experienced by the overlying articular cartilage. Notably, only an unrealistically soft subchondral bone plate (equilibrium modulus of 5 MPa instead of 2690 MPa) resulted in increased maximum principal stress and strain in deep cartilage near the cartilage-bone interface. In study II, the use of bismuth nanoparticles in the synovial fluid facilitated straightforward segmentation of articular cartilage. The cartilage thickness obtained with clinical CT corresponded well to that obtained using µCT (Spearman’s ρ = 0.710, p < 0.001). Additionally, the sitedependent variation in CA4+ concentration in segmented cartilage reflected the sitedependent variation in biomechanical properties: both were higher in femoral than in tibial cartilage ( p < 0.05). In study III, FRPE material could accurately capture (R2 = 0.991) the experimentally observed stress-relaxation behaviour of the mPCLGelMA-HAMA constructs. Computational analysis revealed that fibre-reinforcing the scaffold notably increases the instantaneous interstitial fluid load pressurization of the constructs, reaching the pressurization level of native cartilage. In study IV, CA4+ contrast-enhanced CT identified depth-dependent differences in proteoglycan content between mechanically stimulated and non-stimulated constructs after 42 days of tissue culture, particularly in the superficial layer ( p < 0.05). This localized increase in proteoglycan content, revealed through CT attenuation profiles, demonstrated heterogeneity not observable using bulk biochemistry. In summary, the findings obtained in this thesis provide insights into osteochondral tissue characterization and applications for osteoarthritic cartilage diagnosis and treatment. Study I highlights that early osteoarthritic changes in subchondral bone tissue are unlikely to impact the mechanical loading of cartilage. Meanwhile, computational modelling in study III enables the characterization of mechanical properties in tissue engineered cartilage, shedding light on their correspondence to native tissue properties. Taken together, the obtained computational results enhance the understanding of the mechanical behaviour of tissue engineered cartilage constructs and their interaction with the surrounding environment. Studies II and IV demonstrate the practicality of contrast-enhanced CT imaging in characterizing tissue properties, applied both in clinical CT for cadaveric knee joints, and in the development of tissue engineered cartilage. Looking ahead, contrast-enhanced CT holds promise not only for imaging cartilage defects and assessing degeneration but also for ensuring the adhesion of the tissue engineered cartilage to the native environment and monitoring neotissue formation.