Correlating single cell form and function under the influence of nanotopography

Cutiongco, Marie Francene (2019) Correlating single cell form and function under the influence of nanotopography. PhD thesis, University of Glasgow.

Abstract

Topographical cues patterned on surfaces activate specific cell morphologies, functions and behaviours. This leads to patterned surfaces emerging as in vitro platforms to study cell behaviour and in vivo devices for tissue regeneration. Yet the current practice in the field merely screens pre-made topography libraries to obtain biological hits, providing little to no biological intuition in design and severely limiting diversity of topographies. This thesis presents an alternative methodology to topography design that exploits morphological and functional responses of cells.
First, we established the relationship between nanopit topography structure, cell form and function. An image-based profile of focal adhesions, actin and chromatin were measured from single cells to create a morphological profile of cells termed the ‘morphome’. Using the morphome as predictors, 6 musculoskeletal cell types (accuracy of 98-99%) were classified using logistic regression. Meanwhile, nanopit diameter and disorder from a perfect square grid were accurately predicted (mean absolute error of 2-10%) from the morphome using linear regression. The morphome clearly reflected functional response to topography by robustly predicting (mean absolute error of 10-21%) expression levels of 14 musculoskeletal genes across 24 combinations of cell type and topography using linear regression.
Next, we utilized the topography structure and cell function information encoded in the morphome to demonstrate functionally-guided topography design. A Bayesian optimisation algorithm attained maximum osteogenic gene expression and the corresponding optimum morphome, from which pit topography characteristics were predicted. Pits with hexagonal geometry, 255 nm diameter and 183 nm disorder or square geometry, 398 nm diameter and 399 nm disorder were therefore predicted to maximize osteogenic gene expression. Compared to a random approach (average of 43), topography design through optimizing cell function also reduced the number of experiments (average of 27) needed to achieve the optimum topography.
Finally, we studied the mechanism through which the morphome linked topography structure and cell function. Mechanosensing and functional cell regimes were altered by a library of disordered nanopits. Nanopits with disorder of 20, 50 and 120 nm induced strong osteogenic response that correlated with formation of actin stress fibers, and assembly of large adhesions with longer lifetimes and localization of tension-related components. Cells on nanotopographies with disorder of 0 and 80 nm manifested lamellipodial protrusions containing small adhesions with high turnover, precluding the development of tension required for osteogenesis. Our results implicate intracellular tension generation as a key biological process reflected in the morphome.
The work presented here furthers a rational approach to topography design using the structural and biological information encoded in the morphome. The morphome enables broader and more efficient design of biologically valuable topographies compared to the prevailing trial and error method.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Additional Information: The first chapter in the thesis has been deposited as a pre-print on bioRxiv https://doi.org/10.1101/495879
• Keywords: topography, machine learning, Bayesian optimisation, cell-material interaction.
• Subjects: Q Science > Q Science (General)
• Colleges/Schools: College of Science and Engineering > School of Engineering > Biomedical Engineering
• Funder's Name: European Research Council (ERC)
• Supervisor's Name: Gadegaard, Prof. Nikolaj
• Date of Award: 2019
• Embargo Date: 25 October 2022
• Depositing User: Marie Francene Cutiongco
• Unique ID: glathesis:2019-75126
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 28 Oct 2019 11:11
• Last Modified: 05 Mar 2020 22:28
• Thesis DOI: 10.5525/gla.thesis.75126
• URI: https://theses.gla.ac.uk/id/eprint/75126

Related URLs

• Article DOI
• Article DOI
• Enlighten Publications Record