Definition of axonal injury tolerances across scales
A computational multiscale approach
Time: Fri 2020-02-14 10.00
Location: T2, Hälsovägen 11C, Huddinge (English)
Doctoral student: Annaclaudia Montanino , Neuronik
Opponent: Associate Professor Kacy Cullen, University of Pennsylvania
Supervisor: Professor Svein Kleiven, Neuronik, Flygteknik, Farkost- och flygteknik, Bioteknologi; Alessandra Villa, ; Xiaogai Li, Neuronik
Traumatic brain injury (TBI) is today regarded as a global health challenge. Revealing how external mechanical loads translate into tissue and cellular damage is necessary, not only for the development of better preventive measures, but also for the definition of treatments that could spare the patients from suffering TBI's devastating consequences. Significant advancements have been made in the past decades in the understanding of the biomechanical basis of TBI. Finite element (FE) head models, among others, have proved valuable in clarifying the relation between head kinematic and brain deformations patterns. Nevertheless, a comprehensive picture of TBI pathophysiology across the multiple length scales involved is still lacking.
In this thesis, the multiscale nature of TBI was explicitly considered with the aim of, first, ruling out a mechanically plausible axonal injury mechanism and, second, of defining axonal injury tolerances at different scales. To do so, in Study I, a composite FE model of the axon was developed. The vulnerability of its components was tested in a typical injury scenario. The large and nonhomogeneous deformations observed in the axonal membrane motivated Study II, where the FE axonal model was used in cascade with a molecular model of the axonal plasma membrane (or lipid bilayer). It is at this level --the molecular one-- that mechanoporation can be observed and thresholds can be established in dependence of axonal strain and strain rate.
In Study III, potential mechanistic differences in thresholds derived with single-cell or tissue injury models were investigated. The axon FE model was here expanded in a tissue-like model, where the axon is not only surrounded by matrix, but also by other axons using PBCs. The previously derived molecular-level thresholds were used as a benchmark and tissue-injury models were found to have higher tolerances than single-cell models. In Study IV an experimental approach was adopted to characterize the mechanical behavior of glial tissue (derived from the squid giant axon) at large strains and dynamic rate.
Finally, in Study V, a framework for the multiscale analysis of concussive impacts was proposed. Kinematic data from a real concussion case served as boundary conditions to a subject-specific head FE model. Tissue strains were then used as input to histology-informed tissue-like models of the corpus callosum's subregions. Resulting membrane strains were eventually compared against mechanoporation thresholds to infer about the injury outcome.
In summary, this thesis increases our understanding of the possible mechanical cues behind axonal injury. By using a computational approach bridging the organ-to-molecule length scales, this work proposes a new way of non-invasively predicting axonal damage. Although further experimental evidence is required, such an approach lays the foundation for increasingly complex and potentially revealing simulations of axonal injury.