Arrhythmogenic cardiomyopathy (ACM) is an inherited heart muscle disease characterized by progressive fibrofatty replacement and electrical instability, for which current pharmacological treatments remain palliative rather than curative. This thesis presents a network medicine approach to identify FDA-approved or experimental drugs capable of modifying the progression of the underlying pathology by treating ACM as a "network failure" disease. The study employs a human interactome model and applies the Random Walk with Restart (RWR) algorithm to quantify the functional relevance of drugs with respect to the ACM disease module. To address the topological hub bias inherent in scale-free networks, the methodology integrates a degree-preserving permutation test with a phenotype-based context score. This dual-filter approach effectively prioritizes candidates that are both topologically significant and phenotypically relevant to the cardiac context. The framework demonstrated high predictive performance, achieving an Area Under the Curve (AUC) of 0.861 in the retrieval of drugs with known clinical relevance or use in treating ACM, and the topological characterization revealed that effective target genes are not generic hubs, but rather network "bridges" with high betweenness and eigenvector centrality. Furthermore, transcriptomic validation using the Connectivity Map (CMap) on five independent datasets identified compounds capable of reversing the pathogenic gene expression signatures. The analysis highlighted Isradipine and Fostamatinib as promising new candidates, along with the rediscovery of established therapies such as Amiodarone and Flecainide. This work provides an open-source and adaptable computational pipeline, as well as a prioritized list of drugs that can be further validated in vitro.
La cardiomiopatia aritmogena (CMA) è una malattia ereditaria del muscolo cardiaco caratterizzata da progressiva sostituzione fibroadiposa e instabilità elettrica, per la quale gli attuali trattamenti farmacologici rimangono palliativi piuttosto che curativi. Questa tesi presenta un approccio di medicina di rete per identificare farmaci approvati dalla FDA o sperimentali in grado di modificare la progressione della patologia, trattando la CMA come una malattia da "fallimento di rete". Lo studio utilizza un modello di interattoma umano e applica l'algoritmo del Random Walk with Restart (RWR) per quantificare la rilevanza funzionale dei farmaci per il modulo di malattia della CMA. Per affrontare il bias topologico degli hub insito nelle reti scale-free, la metodologia integra un test di permutazione che preserva il grado con un punteggio di contesto basato sul fenotipo. Questo approccio a doppio filtro dà effettivamente priorità ai candidati che sono sia topologicamente significativi che fenotipicamente rilevanti per il contesto cardiaco. Il metodo ha dimostrato elevate prestazioni predittive, raggiungendo un'area sotto la curva (AUC) di 0.861 nel recupero di farmaci noti nel trattare la CMA, e la caratterizzazione topologica ha rivelato che i geni bersaglio efficaci non sono hub generici, ma piuttosto "ponti" di rete con elevate betweenness e eigenvector centrality. Inoltre, la validazione trascrittomica utilizzando la Connectivity Map (CMap) su cinque set di dati indipendenti ha identificato farmaci in grado di invertire le firme di espressione genica patogena. L'analisi ha dato priorità a Isradipina e Fostamatinib come nuovi candidati promettenti, insieme alla riscoperta di terapie consolidate come Amiodarone e Flecainide. Questo lavoro fornisce una pipeline computazionale open source e adattabile, nonché un elenco prioritario di farmaci che possono essere ulteriormente validati in vitro.
A network medicine approach for drug repurposing in arrhythmogenic cardiomyopathy
Vido, Aurora
2024/2025
Abstract
Arrhythmogenic cardiomyopathy (ACM) is an inherited heart muscle disease characterized by progressive fibrofatty replacement and electrical instability, for which current pharmacological treatments remain palliative rather than curative. This thesis presents a network medicine approach to identify FDA-approved or experimental drugs capable of modifying the progression of the underlying pathology by treating ACM as a "network failure" disease. The study employs a human interactome model and applies the Random Walk with Restart (RWR) algorithm to quantify the functional relevance of drugs with respect to the ACM disease module. To address the topological hub bias inherent in scale-free networks, the methodology integrates a degree-preserving permutation test with a phenotype-based context score. This dual-filter approach effectively prioritizes candidates that are both topologically significant and phenotypically relevant to the cardiac context. The framework demonstrated high predictive performance, achieving an Area Under the Curve (AUC) of 0.861 in the retrieval of drugs with known clinical relevance or use in treating ACM, and the topological characterization revealed that effective target genes are not generic hubs, but rather network "bridges" with high betweenness and eigenvector centrality. Furthermore, transcriptomic validation using the Connectivity Map (CMap) on five independent datasets identified compounds capable of reversing the pathogenic gene expression signatures. The analysis highlighted Isradipine and Fostamatinib as promising new candidates, along with the rediscovery of established therapies such as Amiodarone and Flecainide. This work provides an open-source and adaptable computational pipeline, as well as a prioritized list of drugs that can be further validated in vitro.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_03_Vido_Tesi.pdf
accessibile in internet per tutti
Descrizione: Tesi
Dimensione
15.8 MB
Formato
Adobe PDF
|
15.8 MB | Adobe PDF | Visualizza/Apri |
|
2026_03_Vido_Executive_Summary.pdf
accessibile in internet per tutti
Descrizione: Executive Summary
Dimensione
1.84 MB
Formato
Adobe PDF
|
1.84 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/252123