Search

Colorectal cancer (CRC) represents a significant health challenge, with incidence increasing by 5.5 % and mortality by 6.1 % from 2018 to 2020. Improving our understanding of the molecular and biological drivers contributing to CRC progression and response to treatment will help generate new diagnostic tools and apply better personalised therapies. Altered metabolism is a well-recognised but highly heterogeneous characteristic of cancers. It is known to be influenced by both oncogenic drivers and the components of the tumour microenvironment (TME) and is likely to contribute to progression and therapy response. One area of metabolic adaptation in cancer that has revived growing attention in recent years is lipid metabolism. Here we aim to investigate the role of the TME on lipid alterations in CRC pre-clinical models and clinical CRC biopsies using a combination of untargeted ambient mass spectrometry imaging (MSI) in a multimodal fashion with a variety of ex vivo techniques. Desorption electrospray ionisation (DESI) MSI, a type of ambient MSI technique, facilitated the investigation of lipidomic changes within biological tissue with little to no sample preparation. With DESI-MSI, hundreds of biomolecular species can be simultaneously detected while retaining their spatial information without the need for labelling reagents, making it an attractive tool for medical settings. Using HCT-116 xenografts that are characterised by an upregulated phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3-K) activity, a common driver in CRC, lipidomic changes were found in association with increased abundance of phosphatidylinositol lipids possibly linked to the PI3-K pathway and multiple fatty acids (FA) and acylcarnitine species, suggesting altered energy metabolism to fatty acid oxidation (FAO) via the β-oxidation pathway. Shifts in oxidative metabolism were also obtained via the analysis of oxygen consumption rate using the Seahorse analyser. This alteration was also apparent in clinical CRC biopsies, were the analysis of primary CRC at different clinical stages as well as primary CRC versus colorectal cancer peritoneal metastasis (CRPM), revealed commonalities in FA species linked to the de novo FA synthesis and dietary FA uptake but differences in FA consumption. Changes in metabolism have also been associated with increased immune cell expression in CRPM compared to the primary CRC biopsies, as determined by the gene expression analysis technique NanoString nCounter. Overall, our data demonstrated that DESI-MSI can be used as a powerful tool to investigate lipidomic alteration in CRC pre-clinical models and clinical biopsies, revealing metabolic changes and key molecular pathways implicated in CRC progression and response to therapy.

Glioblastoma (GBM) is the most common and most aggressive primary brain tumour, with an extremely poor prognosis. Despite extensive research into this aggressive tumour, overall survival has been unaffected in the past 15 years with average survival at 12-15 months with optimal standard of care treatment. Glioblastoma tumour cells exhibit strong proliferative and infiltrative characteristics, which contributes to the infiltration of tumour cells into healthy brain. This feature means safe total resection of all tumour cells is impossible, which contributes to the poor survival and prognosis of GBM. Defining a safe tumour volume for resection and further understanding the underlying biochemistry of glioblastoma tumour cells at the molecular level, are therefore important aspects of research that can help to improve overall patient survival. Investigation of the lipidomic changes associated with tumours is a rapidly growing field, with mass spectrometry imaging techniques widely used for molecular pathology applications. Lipidomic profiling provides downstream analysis of the proteomic and genomic changes that are also used to characterise tumours. Mass spectrometry imaging analysis by desorption electrospray ionisation mass spectrometry (DESI-MS) imaging provides an ideal toll to study investigate lipid profiles and small molecules pre-clinical model of GBM and human GBM biopsies. Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) imaging is highly specified for metal analysis and has also been used to study the metal ion distribution in the pre-clinical model of GBM, particularly that of gadolinium which was used as an MRI contrast agent alongside other endogenous metal species. The work outlined here has utilised DESI-MSI in the analysis of brain tumour samples to examine GBM at a molecular level. Overall survival in GBM has not improved for the last 15 years, so a full understanding of GBM at the molecular level is required in the hope to find new therapeutic strategies. Untargeted DESI-MS acquisition has allowed the generation of hundreds of lipid images, with subsequent multivariate statistical analysis revealing key molecular pathways implicated in GBM. Phosphatidylinositol lipid signalling molecules and overexpression of the PI3-K pathway as well as altered metabolism via beta-oxidation of fatty acids and carnitines have been observed in the pre-clinical models. This altered metabolic activity of GBM utilising fatty acid oxidation for energy production translated to human glioblastoma biopsies, where large number of acyl-carnitines were also detected. The use of histology stains other than H&E have been investigated for compatibility with DESI-MS lipidomic analysis. Cresyl violet, toluidine blue, and methylene blue staining showing compatibility with DESI-MS imaging analysis which could provide an alternative data directed stream for targeted tissue analysis.

Spin traps and scavengers have been used to detect free radicals for decades. They react with free radical molecules forming adducts that either prolong the free radical signal (spin traps) or void it (spin scavengers). This reactivity has been exploited in electron spin resonance (ESR), a technique that despite its low sensitivity it is still considered the gold standard to detect free radicals. In this thesis the aim was to test if spin traps and spin scavengers could be adapted to other analytical modalities with more sensitivity and/or quantification capabilities. The infrastructures available allowed us to consider positron emission tomography (PET) and mass spectrometry (LC-MS and imaging). Firstly the molecules were evaluated in in vitro competition experiments using LC-MS. A simple method to evaluate oxidative damage products was developed, and commercially available spin traps and scavengers were tested. The experimental results agreed with the literature pointing towards DMPO (5,5-dimethyl-1-pyrroline- N oxide) like molecules as suitable tracers. In the second results chapters three F-18 PET tracers with spin trap or scavenger reactivity were produced and tested in cell uptake experiments. The experiments were negative for uptake. The third part of the thesis was planned for investigation of the mechanism of action of the successful PET tracer but also to develop a technique that would allow the identification of radical damaged biomolecules with mass spectrometry imaging. Only a very low intensity peak could be assigned to a possible radical oxidation. The main conclusion of these experiments is that the mass spectrometry imaging (MSI) used might not have enough sensitivity and mass resolution to study these low intensity peaks. The hypothesis that spin traps and spin scavengers could be exploited with other techniques was explored in this thesis in different ways without promising results.

Matrix-assisted laser desorption/ionisation (MALDI) mass spectrometry is an analytical technique used for identifying molecules on the basis of their mass-to-charge ratio, facilitating the analyses of intact large biomolecules through soft ionisation. The technique suits a wide range of biomedical applications, with potential for biomarker discovery. However, quantitative MALDI analysis is very difficult because of the complex variations introduced during sample preparation, the ionisation process and data acquisition. An analysis method was therefore developed based on linear Poisson independent component analysis (LP-ICA) that appropriately addresses signal and noise statistical modelling. It was validated on real MALDI mass spectra that have been pre-processed using in-house algorithms. LP-ICA works by extracting independent components within the mass spectral data set, describing underlying variations in the mass spectra. In order to validate the LP-ICA approach, three data sets were acquired using different binary mixtures of complex biological lipid samples, chosen to mimic the complexity of different types of biological tissues that might be imaged by MALDI-MS. These include cow and goat milk, lamb brain and liver, and lamb brain white and grey matter, at varied relative concentrations to provide known ground truth data sets for the analysis. The resulting quantitative analysis achieved twice the accuracy of the conventional approach using a single mass-to-charge peak associated with a particular biological sample composition. Moreover, it made use of information from the entire mass spectrum, without bias. The application of LP-ICA analysis was then extended to MALDI-MS imaging data, where mass spectra are acquired at an array of locations across a thin tissue section. Extraction of mass spectral components from a post-ischemic stroke rat brain tissue cross-section image was successful, where the component images can distinguish sub-types of brain tissue. The brain contains a number of different types of lipid-rich tissue phenotypes which can be differentiated by biomolecules found to be specific to distinct anatomical regions. LP-ICA is also shown to have potential for the automatic identification and characterisation of healthy and diseased tissue regions.

The use of secondary ion mass spectrometry (SIMS) for high spatial resolution qualitative mass spectrometry imaging (MSI) offers superior lateral resolution over competing technologies. There are however various caveats imposed for bioanalytical applications, extensive fragmentation and poor ion yields for high mass analytes both require redress. The development of cluster ion sources has increased sputter yields and reduced chemical damage, however the local chemical environment will often enhance or supress ionisation probabilities leading to inaccurate quantitative imaging. Laser post-ionisation sputtered neutral mass spectrometry (L-SNMS) offers the potential to overcome the sample matrix effect by essentially moving the ionisation process into the gas phase. Research is presented here that utilises various laser post-ionisation techniques with the main aim being to increase the available useful ion yields and decrease fragmentation. It was concluded that the 265 nm wavelength gave the best ion yields compared to the other wavelengths investigated for L-SNMS, enhanced ion yields were observed for key diagnostic fragments compared with Au+ SIMS for the pharmaceutics under investigation. The internal energies imparted onto neutrals sputtered by Au+, Au3+ and C60+ primary ion beams are investigated and it was confirmed that the extent of photofragmentation increased proportionally to the energy per nucleon of the primary ion beam projectile. Laser power dependence on the extent of fragmentation decreased as a function of energy per nucleon with sublimed gas-phase analytes having the lowest laser power dependence. Resonance ionisation mass spectrometry is investigated and applied to the analysis of key biological agents of interest. During the analysis of cisplatin it was discovered that an alternative two photon RIS scheme was possible for platinum (presumably a 2+1 scheme with Lambda 1 slightly detuned from the Pt resonance) and that this scheme was more efficient than the 3 photon scheme published in the literature. And finally the use of atmospheric pressure laser ionisation (APLI) for liquid chromatography/mass spectrometry (LC/MS) is investigated. In this research the application of femtosecond laser pulses to APLI is reported for the first time and compared with more mature LC/MS ionisation techniques.

It is well established that lipids play an important role in diseases such as non-alcoholic fatty liver disease and cardiovascular diseases. However, in the past decade, it has come to light that lipids may be important in other diseases; particularly in cancer and neurological disorders. Here, lipid metabolism has been investigated using pre-clinical cancer models for melanoma, glioma, non-small-cell lung cancer and colorectal cancer. The role of lipids in the recovery post-stroke has also been studied. Mass spectrometry imaging offers an ideal tool to study lipids in tissue ex-vivo. Lipids ionise well in a number of mass spectrometry modalities, and hundreds of lipids can be imaged in one mass spectrometry imaging experiment. Furthermore, mass spectrometry imaging offers excellent spatial resolution. In this work, both MALDI-MS and DESI-MS have been used for mass spectrometry imaging. Tumour lipid heterogeneity has been a particular focus of this this project. Heterogeneity exists within tumours, as well as between tumours in the same patient; and this causes major problems for therapy. Owing to the untargeted nature, and high spatial resolution of mass spectrometry imaging, it is an excellent technique to study lipid heterogeneity. Adjacent sections (or in some cases the same section used for mass spectrometry imaging), were used for immunofluorescence and H&E staining. By comparing mass spectrometry images with staining techniques, biological reasons for lipid heterogeneity can be established. Here, a particular focus has been on hypoxia (low oxygen tensions), which is a key contributor to tumour heterogeneity, and is associated with aggressive cancers. Additionally, hypoxia is a feature of ischaemic stroke, and lipids in ischaemic stroke have also been investigated. PET is a non-invasive imaging technique which is able to image a radiolabelled molecule (tracer) in the body. Here, PET has been used as a complementary in-vivo technique to mass spectrometry imaging. The tracers [11C] acetate and [18F]-FTHA have been used to image fatty acid synthase and fatty acid uptake in tumours; both of which are hypothesised to be key in cancer progression. REIMS is a newly established mass spectrometry technique. It is ideal for analysing lipids in cells, as sample preparation is minimal. Here, approaches for cell pellet analysis have been tested, and used to detect lipids in cancer cell lines.

Positron emission tomography (PET) is a quantitative and non-invasive investigative tool, which permits the identification of pivotal biomarkers and their role in disease onset, transformation and progression. Quantitative detection of these disease biomarkers has qualified PET as a pioneering approach to rapid diagnosis. The tool is a powerful approach to patient diagnosis and investigation of inter- and intra- patient disease heterogeneity thereby supporting patient stratification and forging a truly personalised medicine approach. The sensitivity of PET is complemented by use of highly selective radiopharmaceuticals; this has fuelled the popularity of radiolabelled affinity peptides and proteins (APPs). Inspired by the high selectivity and affinity of biological protein-protein interactions, APP-based radiotracers are an increasingly popular class of radiotracer. Monoclonal antibodies (mAbs) are an important category of APP-based radiotracer; their prevalence in the field is attributed to their high target selectivity and affinity characteristics, which, will likely, uphold their popularity in the field. Yet, advancements in protein engineering has transformed the landscape of APP-based radiotracers, indicated by a preference for small radiolabelled APPs which exhibit rapid pharmacokinetics whilst retaining high target specificity and affinity. Their fast pharmacokinetics is matched by the half-life of 18F which has been defined as an ideal PET radionuclide. 18F APP radiolabelling hasn't yet met its full potential due to its inherent challenges; many approaches have been adopted, ranging from non-site specific radiolabelling with amine reactive prosthetic groups to site-specific methods including oxime bond formation. The radiochemistry of 18F APP radiolabelling is diverse and evolving; yet ideal goals are apparent including site-specificity, fast reaction kinetics, mild labelling conditions and applicability to automation. Automation is challenging for multi-step 18F APP radiolabelling methods, but is a key developmental step. Automation helps with conformance to good manufacturing practice (GMP) by enhancing process robustness, consistency and reliability. It also permits the radiosynthesis of clinically relevant radiotracer doses. The development of an automated generic platform, or a method that adopts a general approach, for APP radiolabelling with 18F would assist in [18F]APPs meeting their potential in PET and, importantly, aid their translation from bench-to-bedside.

Cancer is one of the most devastating human diseases that cause a vast number of mortalities worldwide each year. Cancer research is one of the largest fields in the life sciences and despite many astounding breakthroughs and contributions over the past few decades, there is still a considerable amount to unveil on the function of cancer that would improve diagnostics, prognostics and therapy. Since cancer is known to involve a wide range of processes, applying methods to study it from a systems perspective could reveal new properties of cancer. Systems biology is becoming an increasingly popular tool in the life sciences. The approach has been applied to many biological and biomedical analyses drawing upon recent advancements in technology that make high throughput analyses of samples and computational modelling possible. In this thesis, the effect of hypoxia inducible factor-1 (HIF-1) on cancer metabolism, the entity considered most closely related to phenotype has been investigated. This transcription factor is known to regulate a multitude of genes and proteins to promote survival in a low oxygen environment that is prevalent in solid tumours. However its effect on the metabolome is less well characterised. By revealing the effect of HIF-1 on the metabolome as a system it is hoped that phenotypic signatures, key metabolic pathways indicative of cancer function and potential targets for future cancer therapy, can be revealed.The system has been studied using two cell models: mouse hepatocellular carcinoma and human colon carcinoma, whereby metabolism has been profiled using a range of analytical platforms. In each model, wild type cells have been compared to cells deficient in HIF-1 to reveal its effect on cellular metabolism. Gas chromatography mass spectrometry (GC MS) and ultra high performance liquid chromatography - mass spectrometry (UHPLC MS) have been employed for metabolic profiling of cells exposed to a range of oxygen conditions. Additionally, time-of-flight secondary ion mass spectrometry (ToF SIMS) has been employed for imaging mass spectrometric analysis of multicellular tumour spheroids cultured from wild type cells and cells with dysfunctional HIF-1 to represent small initiating tumours. Using these techniques in metabolic profiling it has been possible to reveal metabolites associated with the effect of oxygen and HIF-1 on cancer metabolism along with key pathways and hubs that could be targeted in future therapy. Using imaging mass spectrometry it has been possible to localise metabolites in situ revealing how tumour structure relates to function. Finally, a novel approach to consider how metabolites are correlated with one another in the response to oxygen level or presence or absence of functional HIF-1 has been undertaken to better understand the systems properties of cancer metabolism. Metabolites found to be differently correlated with respect to oxygen and/or HIF-1 have been mapped onto a human metabolic network to determine their network-based origins. This allowed the simulation of sub-networks of metabolism most affected by oxygen and HIF-1, highlighting the key mechanisms in HIF 1 mediated cancer cell survival.

Introduction: Hypoxia is found in solid cancerous tumours. The presence of hypoxia within tumours inhibits anti-cancer treatment strategies such as chemotherapy from being completely effective and it is suspected that multiple mechanisms contribute to the resistance. Methods: In this project a systems biology approach was applied to determine how the toxicity of doxorubicin is affected by hypoxia at the metabolome level. A multitude of analytical techniques were applied to analyse the intracellular metabolism of a monolayer of cancer cells (MDA-MB-231). Metabolic profiling was used to determine metabolite markers related to hypoxia-induced chemoresistance. For this gas chromatography mass spectrometry (GC-MS) and ultra high performance liquid chromatography mass spectrometry (UHPLC-MS) were used. Furthermore, network-based correlation analysis was developed as a novel tool to bridge the gap between metabolomics dataset and systems biology modelling. This methodology was applied to elucidate novel metabolic pathways as potential therapeutic targets to overcome hypoxia-induced chemoresistance. This algorithm determines significant correlation differences between different physiological states, and through applying graph-theory on large genome scale models; it is possible to construct a metabolic network of the pathways connecting the pair-wise correlation. Finally, imaging mass spectrometry using time-of-flight secondary ion mass spectrometry (ToF-SIMS) was developed as a tool for in situ metabolite analysis to investigate the metabolic response to chemotherapy in multi-tumour spheroids (MTSs). Results: Metabolic fingerprinting analysis characterised a snapshot of cells exposed to various environmental perturbations. Metabolite markers associated with hypoxia-induced chemoresistance were related to metabolic pathways including gluconeogenesis, DNA synthesis and fatty acid synthesis. Furthermore, network-based correlation analysis revealed specific metabolites in the fatty acid synthesis pathways were contributing to drug resistance, which included malonyl-CoA, 3-oxoeicosanoyl-CoA, stearoyl-CoA and octadecanoic acid. To facilitate the detection of metabolites in ToF SIMS datasets, a series of metabolites standard spectra were acquired. Hypoxic metabolite markers detected in ToF-SIMS data of cell lysates included glycine, lactic acid and succinic acid, which were also shown to be metabolite markers in GC-MS metabolic data. Furthermore, MTS sections were imaged using ToF-SIMS to profile the chemical response to chemotherapy treatment within the oxygen gradient. Loadings from image PCA were explored to determine the metabolic response in the highly oxygenated outer region and hypoxic inner region of the MTS. Conclusion: A multitude of analytical techniques were able to contribute to elucidating the metabolic mechanisms associated with hypoxia-induced chemoresistance. Metabolic profiling combined with a systems biology approach was further able to identify potential underlying metabolic regulation of resistance. Finally ToF-SIMS was developed as a tool for metabolite analysis in complex biological systems in situ.