Multi-Scale Force Transmission to and Within the Nucleus.
Abstract
The mechanical state of cells, controlled primarily by cytoskeletal (CSK) networks
(actin, microtubules and intermediate filaments) is a critical component of maintaining
healthy function. Forces transmitted through the cytoskeleton influence
the organisation and state of nuclear material, leading to changes in gene expression.
This thesis aims to increase our understanding of the role of the CSK
networks, specifically the intermediate filament keratin, and their interplay in integrating
mechanical forces. We primarily use immunofluorescence imaging of the
CSK networks and the nucleus, supported by Atomic Force Microscopy. We work
in human epidermal keratinocytes (HEKs), as they are rich in keratin, whose role
in cytoskeletal force transmission is under-studied.
Since drugs to disrupt keratin are scarce, we first established that Withaferin-A, a
compound previously used to disrupt vimentin intermediate filaments, can disrupt
keratin at non cyto-toxic doses; impacting cell mechanics and migration.
Following from this, Withaferin-A was used alongside established cyto-modulatory
drugs to disrupt CSK networks, quantifying a range of properties describing their
organisation. These data were fitted to nuclear parameters that described opposing
functions on the nuclear state of HEKs for keratin and tubulin, with keratin
protecting the nucleus from mechanical force.
Finally, machine and deep learning techniques were used to expand the mathematical
modelling of data. By training networks to predict nuclear location from
only CSK images, a causative relationship between CSK organisation and nuclear
location can be derived. In addition, we develop new models to rapidly analyse
Atomic Force Microscopy curves and generate synthetic cell images.
These results demonstrate the important role of keratin in protecting the nucleus
from mechanical force and that deep learning techniques can be used in the study
of cell mechanics to gain new insights.
Authors
Keeling., Michael.Collections
- Theses [4192]