Massively parallel simulations of coupled arterial cells : Ca2+ dynamics and atherosclerosis.

Abstract

Ischaemic heart disease (IHD) is the most common cardiovascular disease, and is a major cause of mortality globally. The underlying process of IHD involves the development of atherosclerotic plaques on the arterial wall. These plaques can subsequently rupture and release thrombogenic species into circulation, or can occlude the vessel downstream following detachment. Such complications result in ischaemia—a restriction of blood supply to tissue that results in a shortage of vital cellular nutrients, such as oxygen and glucose. Recent publications hypothesise that cellular ionised calcium (Ca2+) concentrations play an important role in atherogenesis. There has been a significant amount of research on cardiovascular disease within multiple sub areas, including: in vivo and in vitro experimental work, computational fluid dynamics (CFD) simulations, and computational modelling of pathological behaviour. However, a combination of these fields will provide a greater understanding of the conditions that promote plaque development. The research presented in this thesis consists primarily of massively parallel simulations of arterial bifurcations which used CFD to generate the input agonist maps for each arterial mesh. Micro-scale dynamics of coupled endothelial cells (ECs) and smooth muscle cells (SMCs) were modelled in bifurcation surfaces containing over one million cells. In particular, the effect bifurcation angulation may have on atherosclerosis development was investigated. A number of improvements were introduced to the original coupled cells model to perform these simulations. A surface-mesh generation pipeline capable of creating geometrically varying 3-D surfaces, including EC and SMC layers, was implemented. These surfaces were used in CFD simulations to generate agonist input maps, and to define the EC and SMC layers on which the dynamics in our simulations are mapped. A detailed inositol triphosphate (IP3) pathway and gap-junction currents were introduced to the coupled cells model. These additions were to ensure our simulations present physiologically accurate results when compared to related experimental research and computational modelling. Finally, the parallel implementation that enabled our simulations to be conducted at the macro scale was improved by the introduction of Open-Multi-Processing (OpenMP). The massively parallel simulations displayed propagating Ca2+ waves in SMCs and steady-state concentrations of Ca2+ in ECs. Particularly complex SMC Ca2+ behaviour was observed in the lateral regions where the main stem meets the branches. Waves propagated in a slower, sporadic manner, and over significantly shorter distances. Further, we observed lower time-averaged Ca2+ concentrations in arterial geometries with wider bifurcation angles compared to those with narrower bifurcation angles. The regions of low EC and SMC Ca2+ concentrations correspond to the sites re- search utilising CFD agrees are those most likely to experience plaque development due to flow detachment. Furthermore, we noted the low Ca2+ concentrations in these areas are more prominent in arterial geometries with wider bifurcation angles. These results suggest bifurcation angulation may have a significant effect on the susceptibility of arterial regions to atherosclerosis development.