Trabecular bone remodeling in the aging mouse: A micro-multiphysics agent-based in silico model using single-cell mechanomics.

Bone remodeling is regulated by the interaction between different cells and tissues across many spatial and temporal scales. Notably, in silico models are regarded as powerful tools to further understand the signaling pathways that regulate this intricate spatial cellular interplay. To this end, we have established a 3D multiscale micro-multiphysics agent-based (micro-MPA) in silico model of trabecular bone remodeling using longitudinal in vivo data from the sixth caudal vertebra (CV6) of PolgA(D257A/D257A) mice, a mouse model of premature aging. Our in silico model includes a variety of cells as single agents and receptor-ligand kinetics, mechanomics, diffusion and decay of cytokines which regulate the cells' behavior. We highlighted its capabilities by simulating trabecular bone remodeling in the CV6 of five mice over 4 weeks and we evaluated the static and dynamic morphometry of the trabecular bone microarchitecture. Based on the progression of the average trabecular bone volume fraction (BV/TV), we identified a configuration of the model parameters to simulate homeostatic trabecular bone remodeling, here named basal. Crucially, we also produced anabolic, anti-anabolic, catabolic and anti-catabolic responses with an increase or decrease by one standard deviation in the levels of osteoprotegerin (OPG), receptor activator of nuclear factor kB ligand (RANKL), and sclerostin (Scl) produced by the osteocytes. Our results showed that changes in the levels of OPG and RANKL were positively and negatively correlated with the BV/TV values after 4 weeks in comparison to basal levels, respectively. Conversely, changes in Scl levels produced small fluctuations in BV/TV in comparison to the basal state. From these results, Scl was deemed to be the main driver of equilibrium while RANKL and OPG were shown to be involved in changes in bone volume fraction with potential relevance for age-related bone features. Ultimately, this micro-MPA model provides valuable insights into how cells respond to their local mechanical environment and can help to identify critical pathways affected by degenerative conditions and ageing.

Motion grading of high-resolution quantitative computed tomography supported by deep convolutional neural networks.

Image quality degradation due to subject motion confounds the precision and reproducibility of measurements of bone density, morphology and mechanical properties from high-resolution peripheral quantitative computed tomography (HR-pQCT). Time-consuming operator-based scoring of motion artefacts remains the gold standard to determine the degree of acceptable motion. However, due to the subjectiveness of manual grading, HR-pQCT scans of poor quality, which cannot be used for analysis, may be accepted upon initial review, leaving patients with incomplete or inaccurate imaging results. Convolutional Neural Networks (CNNs) enable fast image analysis with relatively few pre-processing requirements in an operator-independent and fully automated way for image classification tasks. This study aimed to develop a CNN that can predict motion scores from HR-pQCT images, while also being aware of uncertain predictions that require manual confirmation. The CNN calculated motion scores within seconds and achieved a high F1-score (86.8 ± 2.8 %), with good precision (87.5 ± 2.7 %), recall (86.7 ± 2.9 %) and a substantial agreement with the ground truth measured by Cohen's kappa (κ = 68.6 ± 6.2 %); motion scores of the test dataset were predicted by the algorithm with comparable accuracy, precision, sensitivity and agreement as by the operators (p > 0.05). This post-processing approach may be used to assess the effect of motion scores on microstructural analysis and can be immediately implemented into clinical protocols, significantly reducing the time for quality assessment and control of HR-pQCT scans.

An in silico micro-multiphysics agent-based approach for simulating bone regeneration in a mouse femur defect model.

Bone defects represent a challenging clinical problem as they can lead to non-union. In silico models are well suited to study bone regeneration under varying conditions by linking both cellular and systems scales. This paper presents an in silico micro-multiphysics agent-based (micro-MPA) model for bone regeneration following an osteotomy. The model includes vasculature, bone, and immune cells, as well as their interaction with the local environment. The model was calibrated by time-lapsed micro-computed tomography data of femoral osteotomies in C57Bl/6J mice, and the differences between predicted bone volume fractions and the longitudinal in vivo measurements were quantitatively evaluated using root mean square error (RMSE). The model performed well in simulating bone regeneration across the osteotomy gap, with no difference (5.5% RMSE, p = 0.68) between the in silico and in vivo groups for the 5-week healing period - from the inflammatory phase to the remodelling phase - in the volume spanning the osteotomy gap. Overall, the proposed micro-MPA model was able to simulate the influence of the local mechanical environment on bone regeneration, and both this environment and cytokine concentrations were found to be key factors in promoting bone regeneration. Further, the validated model matched clinical observations that larger gap sizes correlate with worse healing outcomes and ultimately simulated non-union. This model could help design and guide future experimental studies in bone repair, by identifying which are the most critical in vivo experiments to perform.