Computational models of neurovascular coupling

Type of content
Theses / Dissertations
Publisher's DOI/URI
Thesis discipline
Mechanical Engineering
Degree name
Doctor of Philosophy
Publisher
University of Canterbury
Journal Title
Journal ISSN
Volume Title
Language
English
Date
2015
Authors
Dormanns, Katharina
Abstract

The human brain critically relies on a continuous blood supply to ensure its function. Impaired cerebral perfusion is associated with diseases, such as Alzheimer’s Disease, cortical spreading depression and stroke. Cerebral cells work together to ensure a finely regulated blood supply. However, many of the underlying mechanisms of this regulation, termed neurovascular coupling (NVC), are still not fully understood. The development of mathematical models of cell signalling is a promising approach that complements experimental studies by investigating the contribution of key components and pathways and guiding further experiments. This dissertation reports the development of a lumped parameter compartment model of a neurovascular unit (NVU), comprising neurons, astrocytes, smooth muscle cells and endothelial cells. The model couples neuronal activity to vasodilation/contraction models via the an astrocytic mediated perivascular K+ and the a smooth muscle Ca2+ pathway. It successfully relates a neuronal input signal to the corresponding vessel reaction. It was found that the voltage-operated Ca2+ channels are, due to the hyperpolarisation from the K+ efflux of the SMC, almost entirely closed and do not seem to play a significant role during neuronal activity. The current model shows that in contrast to a number of experiments, outlining the importance of astrocytic Ca2+ in NVC, this Ca2+ pathway is not the only one mediating NVC. Model inputs from the endothelial side of the blood vessel in form of luminal agonists provide a flux of IP3 into the EC cytosol. EC/SMC coupling through gap junctions has a substantial effect on the NVC dynamics. The IP3 influx induces Ca2+ release from the SMC stores resulting in Ca2+induced Ca2+ release oscillations which were shown to inhibit NVC. The incorporation of nitric oxide (NO) signalling pathways in the NVU model shows the vasodilating effects of NO in the resting state. Furthermore, dilation during neuronal activation is enhanced with NO dynamics included. Results show that K+ release is responsible for the fast onset of vascular response, whereas NO-modulated mechanisms maintain dilation. The govern- ing source of NO that diffuses into the SMC, which provides the contractile force, depends on neuronal activation. In the resting state the EC provides the major contribution of NO release for vasorelaxation, whereas during neuronal stimulation NO produced by the NE dominates. In a scaled model, embedded in a parallel computing environment, multiple NVUs were connected to a space filling binary tree, facilitating the simulation of a perfusing vascular tree. The model couples the NVUs to the vasculature via stretch-mediated Ca2+ channels on both the EC and SMC . The coupling between the vasculature and the set of NVUs was found to be relatively weak for the case with agonist induced where only the Ca2+ in cells inside the activated area becomes oscillatory. However, the radii of vessels both inside and outside of activated areas oscillate (albeit small for those outside). In addition, simulations reveal a different oscillation profile when comparing coupled and decoupled states with the time required to refill the cytosol with decreasing Ca2+ and increasing frequency with coupling. The solution algorithm is shown to have excellent weak and strong scaling. Results have been generated for tissue slices containing up to 4096 NVUs, which corresponds to an area of 2.5 cm x 2.5 cm of cerebral tissue.

Description
Citation
Keywords
Ngā upoko tukutuku/Māori subject headings
ANZSRC fields of research
Rights
All Right Reserved