Exploring disease biomarkers and mechanisms using metabolomics

Abstract

A metabolome is the complete set of small molecules or metabolites present in a biological system. It is the result of internal, genetically determined processes, as well as external factors. The metabolome can consequently be seen as an interface between the genome and the environment. In contrast to gene transcription and translation which do not necessarily result in active gene products, metabolite levels are a direct consequence of the active phenotype and can therefore provide knowledge about cellular mechanisms in health and disease. The aim of untargeted metabolomics is large-scale detection and quantification of the complete metabolome in a given sample. On a physicochemical level, the metabolome constitutes a complex mixture of compounds, making complete metabolome coverage an analytical challenge. Liquid chromatography-mass spectrometry (LC-MS), which was used throughout this work, is a widely used analytical platform for metabolomics due to its high sensitivity and large metabolome coverage.

In paper I of the present thesis, we identified novel potential metabolite markers for pancreatic cancer, comparing serum and plasma samples from patients with pancreatic ductal adenocarcinoma (PDAC) and chronic pancreatitis (CP). The comparison with chronic pancreatitis is not only clinically relevant, since distinguishing CP and PDAC is a challenge with current diagnostic tools; CP also constitutes an inflammatory control condition of the pancreas and therefore aids in the exclusion of non-specific, general disease markers. The comparison with relevant control groups is an important aspect of biomarker discovery study design in general.

In paper II we showed that the inclusion of non-related disease controls, apart from organ-specific inflammatory controls, can contribute further to the identification of disease-specific biomarkers. We compared the serum metabolic profiles of three non-related diseases with healthy controls and based on overlap analysis of the results we concluded that despite very different etiology and clinical presentation, these three diseases have highly similar effects on the levels of metabolites in serum.

In paper III we moved from blood-based, systemic metabolic profiles to the intracellular level. Combining LC-MS metabolomics with RNA sequencing, stable isotope tracing and viability assays, we characterized metabolic reprogramming associated with drug-resistance in cancer. When the experimental aim is to discover novel and unexpected dysregulations of the metabolic profile without prior knowledge of the compounds involved, as in papers I-III, large metabolome coverage is a key aspect. In paper IV we therefore evaluated the impact of the reconstitution solvent on metabolome coverage.

Taken together, papers I-IV show the large potential of untargeted LC-MS metabolomics as a tool for discovery, to be used as a starting point to trace a chain of molecular events to its origin.