Alzheimer's disease (AD) is a neurodegenerative disorder characterized by pathological hallmarks such as amyloid-beta plaques, tau protein tangles, neuroinflammation, and metabolic dysregulation. Drug repurposing, that seeks new therapeutic uses for existing drugs, is particularly valuable for AD, given the high failure rates of traditional drug development. Polypharmacology, which aim at designing drugs that hit multiple targets involved in AD, is also gaining importance. Given AD multifactorial nature, drugs modulating multiple pathways could be more effective. A computational paradigm which can encompass both approaches is network pharmacology, which analyzes how different drugs can modulate interconnected pathways. Together, network pharmacology, drug repurposing, and polypharmacology offer promising avenues for AD treatment. This thesis leverages computational methods based on the analysis of proteins involved in AD, to formulate hypotheses on drug combinations for polypharmacology approaches. Specifically, network pharmacology methods have been utilized to prioritize drugs for repurposing, leveraging molecular data from DisGeNet, UniProtKB, DrugBank, and others. A curated list of AD-associated genes has been compiled, integrating information from multiple public databases and recent research findings. The set was used as input for diverse computational analyses, to perform the prioritization of repurposable drugs and the selection of their combinations. The results are critically discussed, together with preliminary experimental data on toxicity and neuroprotective effects of three drug combinations. Overall, this work explores the capabilities of computational network-based methods for AD drug repurposing, providing insights over strengths and limitations of recently described pipelines. Furthermore, it proposes an original integration of different data analysis software to address the intricate mechanisms of AD through a curated gene set. It offers a pipeline for drug repurposing and combination therapy, leveraging established methodologies and public data, and flexible enough to be adapted to new data, allowing future iterations that could produce even more promising therapeutic options for AD.
Maria Giulia Prado (2025) studied this question.