From tissue to network Computational modeling of drug-refractory epilepsies

This thesis investigates the pathobiology of drug-resistant epilepsies, including mTORopathies and temporal lobe epilepsy with hippocampal sclerosis (TLE-HS), using machine learning and data-driven frameworks. By modeling transcriptional profiles from patient-derived tissues, Chapter 2 identifies a global molecular signature of epilepsies. It highlights hallmarks such as impaired gene regulatory networks involved in energy metabolism, brain extracellular matrix, neuronal support and myelination, neuronal function, and neuroinflammation, offering new drug discovery targets. Chapter 3 focuses on mTORopathies, specifically tuberous sclerosis complex (TSC), revealing disruptions in mitochondrial function and calcium regulation. These findings were validated through in vitro and electron microscopy studies. Chapters 4 and 5 examine neuroinflammation and neurotransmission in mTORopathies and low-grade epilepsy-associated tumors (LEATs), exploring the role of IL-1β in GABAergic dysfunction and the modulatory effects of anti-inflammatory mediators like IL-10. Chapter 6 addresses neuropsychiatric comorbidities in TSC, a critical yet often underestimated aspect of these disorders. It identifies miRNAs and isomiRs as circulating biomarkers that predict and differentiate these conditions, providing valuable insights for improving patient care and quality of life. Finally, Chapter 7 discusses these findings, emphasizing the translation of computational models into experimental validation. It outlines future directions for personalized therapeutic strategies and the optimization of clinical trials for drug-resistant epilepsies. This thesis highlights the potential of computational and molecular approaches in unveiling disease mechanisms and fostering innovative treatments.