The metabolic signature of autophagy modulators : an untargeted metabolomics approach for early mechanistic evaluation

Autophagy is an evolutionary conserved cellular degradation process that maintains homeostasis by recycling cytoplasmic components such as damaged organelles and protein aggregates. Its tight regulation is essential for cellular health, and its dysfunction has been linked to a wide range of diseases, including age-related neurodegenerative diseases, inherited neuropathies, cardiomyopathies, and cancer. As pharmacological modulation of autophagy gains interest as a therapeutic strategy, there is a growing demand for robust tools to systematically assess how compounds influence this complex pathway. This work investigated untargeted metabolomics as a powerful approach to characterize autophagy modulation during early-stage mechanistic evaluation. To that end, a robust and reproducible in vitro untargeted metabolomics and lipidomics workflow was developed, specifically optimized for autophagy research. This included tailored cell culture conditions, a biphasic extraction method suitable for both polar and apolar metabolites, LC-(DTIM)-HRMS analysis across multiple platforms, and a QA/QC framework. Special attention was given to normalization strategies, reproducibility, and the creation of a multidimensional in-house spectral library to improve metabolite annotation confidence. These methodological advances were guided by metabolomics best practices and designed to support both hypothesis generation and mechanistic interpretation. The optimized workflow was applied to characterize the metabolic and lipidomic fingerprints of two widely used autophagy inducers, Torin1 (mTORC1/2 inhibitor) and Tat-Beclin1 (Beclin1 complex activator). Despite both increasing autophagic flux, they produced distinct metabolic profiles, revealing how upstream mechanisms of induction influence downstream biochemical pathways. Shared signatures were also identified which may represent conserved metabolic responses to autophagy. The case study with compound 5j, a novel biarylacetamide derivate, further illustrated how metabolomics can be integrated with proteomics and autophagy validation assays to gain mechanistic insight into novel autophagy modulators during early mechanistic evaluation. In conclusion, this thesis provides both a conceptual and practical foundation for integrating untargeted metabolomics into autophagy research. The presented workflow and reference fingerprints lay the groundwork for future studies aimed at screening novel modulators, understanding compound mechanisms, and identifying translational metabolic biomarkers. Moving forward, combining untargeted metabolomics with complementary omics technologies, validation assays, and more pathophysiological relevant models will further enhance the clinical potential of this approach.