Mass spectrometry-based analytical platforms in untargeted metabolomics

In the post-genomic era, omics sciences have transformed systems biology, opening new and previously unexplored paths in clinical, pharmaceutical, and food research. These fields rely on multidisciplinary expertise and advanced analytical technologies to enable the comprehensive characterization of entire classes of biomolecules, such as DNA, proteins, or metabolites. The onset of high-resolution mass spectrometry (HRMS), capable of measuring highly accurate mass-to-charge ratios, has been a key driver in this shift. HRMS has enabled a move beyond traditional targeted approaches, focused on a predefined set of compounds, toward untargeted metabolomics, which requires no prior knowledge of sample composition. Whether applied to the characterization of complex food matrices or the discovery of disease-related biomarkers, untargeted metabolomics has become a powerful tool for achieving broad molecular identification. The rapid advancement of lipidomics over the past decade has been largely driven by the development of modern HRMS technologies, which have improved analytical performance in terms of resolving power, sensitivity, and acquisition speed. This technological progress has supported the development of innovative analytical strategies aimed at gaining detailed structural insights into the lipidome, ultimately helping to elucidate the functional and pathological roles of lipids. In the present thesis, the development of novel tandem mass spectrometry methods (MS/MS) addressing two main objectives will be presented. Firstly, several approaches enabling the confident identification of lipid regioisomers through the localization of carbon–carbon double bonds using chemical derivatization techniques will be described. Secondly, effort has been put into identifying disease-associated biomarkers by detecting concentration changes in biofluids using both innovative labeling methods and customized data processing workflows tailored to specific compound classes. The doctoral project was articulated in three main research lines: i) lipid characterization at the C=C location level, ii) lipid characterization in prostate cancer research, and iii) impurity profiling of synthetic oligonucleotide-based therapeutics. After the development of the method, the analytical workflows were eventually employed for one or more applications in the clinical, food, and plant research fields. The photochemical reaction-based analytical platform applied to hempseeds and seaweeds appeared promising for the untargeted characterization of the whole polar lipidome despite dealing with a complex matrix. It has also broadened the knowledge on their polar lipidome, including information on minor lipid classes and fatty acid (FAs) regioisomers that cannot be investigated using conventional lipidomics protocols based on MS/MS technologies. When the same derivatization reaction was optimized and applied to more complex lipids, such as cholesteryl esters (CEs), it was still proven to be effective for the detailed characterization of complex matrices, such as human plasma and animal liver tissue. As regards prostate cancer (PCa) research, integrating a biorthogonal click reaction into an untargeted HRMS-based lipidomics method to annotate isomers at a detailed structural level was found to play a crucial role in distinguishing PCa from benign prostate conditions. An advanced analytical method was developed within this framework to enable comprehensive characterization of acylcarnitines (ACs). As such, combining HRMS, Kendrick mass defect-based data filtering, and retention time prediction in reversed-phase liquid chromatography allowed for in-depth profiling of ACs in PCa tissue. Later, a partial least square–discriminant analysis (PLS-DA) model validated by repeated double cross-validation was built on the dataset of annotated ACs, with classification rates higher than 93% for both groups, and univariate statistical analysis led to the hypothesis that PCa onset may be linked to oxidative modifications primarily affecting short- and medium-chain ACs, while such alterations appear absent in long-chain species. Several previous studies based on targeted MS focused solely on selected high abundance ACs, leaving unnoticed the role of minor compounds. As part of my doctoral secondment at the University of Amsterdam, I undertook a research project focused on developing and optimizing a compact flow-splitting single-heartcut 2DLC method for impurity profiling of oligonucleotide-based therapeutics. Although at first glance this project seemed unrelated to my primary research projects on untargeted metabolomics workflows, it was later integrated into my doctoral work due to shared methodological and conceptual foundations. The research projects presented in this thesis demonstrated the importance of the use of mass spectrometry and separation techniques for the detailed characterization of complex biomolecules, illustrating the flexibility and cross-disciplinary relevance of these analytical tools in nutritional science as well as in cancer research.