Computational Mechanobiology in Cartilage and Bone Tissue Engineering: From Cell Phenotype to Tissue Structure - Computational Modeling in Tissue Engineering - page 347

1000 times higher than the inlet velocity. Due to the considerable computational
cost only a part of the full scaffold was modelled.
Sandino et al. [ 100 ] included angiogenesis and tissue differentiation into a lCT
based model of a CaP scaffold section by filling the irregular pore space with a
lattice to simulate cell activities. In vitro seeding was simulated by attaching the
initial MSC population to the struts throughout the scaffold and compared to In
vivo colonisation, where the scaffold was initially cell free and the original MSC
population was modelled to reside at the outside faces of the pore network where
endothelial cells had their origin in both cases as well. In vivo colonisation led to
faster migration and proliferation but affected final tissue distribution and vascu-
larisation of the scaffold only marginally. Despite 70% of the pore volume having
a stimulus favourable for ossification, only 40% of that space filled with bone due
to vascularisation being restricted to the periphery of the scaffold. Deeper vascular
penetration was inhibited by the scaffold walls in the model. When the applied
compression was doubled from 0.5 to 1% strain, cartilage was also predicted in
external pores. Apoptosis, induced by a mechanical stimulus twice as high as that
for fibrous tissue formation, increased from 17 to 22% as the applied strain
increased from 0.5 to 1%.
Stops et al. [ 108 ] simulated MSC differentiation in a collagen-GAG scaffold
based on lCT scans using a combination of FE and CFD analysis. Strain
dependent cell proliferation was included and octahaedral shear strain in combi-
nation with fluid velocity used to determine MSC fate. Applied strains of 1% and
above led to predictions of lower cell densities. By comparing different scaffold
strains and fluid inlet velocities the authors were able to determine specific
combinations of these two parameters that favoured certain phenotypes. While
certain conditions favourable for osteoblasts (final cell fraction 84.9%) and
fibroblasts (final cell fraction 73.9%) could be established, none of the combina-
tions proved particularly suitable for robust chondrocyte differentiation (maximum
cell fraction achieved 56.7%).
Besides its influence on the mechanical conditions inside the scaffold, the pore
size influences cell attachment morphologies. In a combined fluid-elastostatic
analysis, Jungreuthmayer et al. [ 54 ] simulated the effect of cell attachment modes
on the experienced stimuli. Three random sub-volumes of the lCT reconstruction
of a collagen-GAG scaffold were numerically seeded with cells that either attached
flatly to one strut or bridged two struts by means of cellular processes. A steady-
state incompressible Newtonian-fluid flow analysis revealed that pressures and
wall shear stresses experienced by the cells were largely independent of attach-
ment mode. A subsequent linear elastic analysis of cells subjected to the pressure
and shear loads derived from the CFD simulation revealed that bridging cells
underwent approximately 500 times higher displacements than flatly attached
cells. Van Mises stresses in bridging cells were about 26 times higher than in flatly
attached cells. These results can potentially explain why cells seeded onto 3D
scaffolds with different morphologies, which promote various modes of cellular
attachment, elicit a dramatically different biological response when subjected to