Author ORCID Identifier

0000-0001-7113-5111

Document Type

Theses, Ph.D

Disciplines

Biophysics

Publication Details

A thesis submitted for the degree of Doctor of Philosophy, School of Physics and Clinical & Optometric Sciences, Technological University Dublin, January 2026.

 doi:10.21427/z3ax-7p87

Abstract

Cellular metabolism and its intricate biochemical mechanisms which allow the conversion of nutrients into energy to sustain life are well understood, leading to numerous breakthroughs in science and technology, human health and wellbeing. From an industrial perspective the pathway kinetics are significant for monitoring cellular bioprocesses, screening desired metabolite phenotypes in drug discovery processes, etc.

An initial review of glycolysis pathway kinetics highlighted the lack of sophisticated approaches to monitor the metabolic pathway kinetics as a function of time in a non-destructive, non-invasive, label-free manner. Fluxomics, a metabolomics based approach providing high-throughput insights, can be considered a gold standard approach, although, being destructive in nature, it is limited to providing a snapshot of metabolic insights. Kinetic glycolysis assays, on the other hand, can provide kinetic insights into the cellular metabolism in a non-destructive and non-invasive manner, although they are also limited, in this case to the pathway end-point kinetics. Vibrational spectroscopy can potentially overcome the limitations of both the approaches, as it can provide high-content insights in a non-destructive, non-invasive and label-free manner and thus its potential to monitor the cellular metabolic kinetics as a function of time was explored in this study.

Firstly, the cellular glycolysis pathway kinetics under the Control, Stimulation (using oligomycin drug) and Inhibition (using 2-deoxyglucose drug) conditions were elucidated using the kinetic glycolysis assay to obtain a baseline for the vibrational spectroscopic experiments. The assay reproduced the expected outcome and a simplistic numerical, rate equation based model was developed to simulate the pathway end-point kinetics. The model predicted the kinetics of the cascading steps leading to the end-point kinetics, enabling simulation of the metabolic kinetics beyond the assay’s sensitivity, comparison of the pathway modulations in numeric terms, and provided a better understanding of the cellular metabolic regulation under the modulated conditions compared to the simple assay end-point kinetics.

Following the assay, vibrational spectroscopy was used to elucidate the extracellular metabolic kinetics using the same experimental conditions. Since, Raman spectroscopy, sensitive to polarisable molecules, is relatively less sensitive to the water molecules with a dipole movement, it was selected for experimentation over the Fourier transform infrared spectroscopy. Partial least squares regression indicated Raman spectroscopy was sensitive to monitor metabolites in biological range (below 20 mM) and the multivariate curve resolution- alternating least squares (MCR-ALS) method was deemed suitable for datamining spectral fingerprints. The MCR-ALS analysis of the extracellular medium indicated three components were evolving as a function of time in all three metabolic conditions (Control, Stimulation and Inhibition) and evolved in a similar manner with time as observed from the kinetic assay. The resolved components demonstrated the complexity of the interaction between cells and their extracellular medium which was missed in the kinetic assay.

Next, 30 cellular spectra for each time point were acquired from a similar experiment in biological and technical replicates for a 3-hour timeframe. A principal components analysis showcased the complexity of the data, as no resolution could be seen among the different metabolic conditions, whereas a small degree of differentiation could be seen among the different timepoints of individual metabolic conditions. Since, MCR-ALS could not accurately resolve the components from the complex dataset, the inbuilt constraints in the toolbox were tested using simulated experimental data. The simulated data indicated that the complex dataset with higher cellular weighting overlaying the metabolic modulations, does not resolve the data quantitively, although the resolved spectral fingerprints are accurate qualitatively. It was deduced that MCR-ALS could not estimate the components evolving in the dataset in the MCR part of the algorithm and thus struggled to datamine the spectral fingerprints. This was overcome by manually providing the initial estimates in the MCR, better enabling qualitative datamining. The time evolutions of the resolved components indicated that, since vibrational spectroscopy is label-free in nature, it also captured the cellular kinetic features not normally affiliated with the glucose metabolism. The cellular spectroscopic analysis augmented the insights from the extracellular experiments.

DOI

https://doi.org/10.21427/z3ax-7p87

Funder

Science Foundation Ireland

Creative Commons License

Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.

Download
Contact the Author


Included in

Biophysics Commons

Share

COinS