Drug repurposing and combination therapy for glioblastoma treatment

Abstract

Introduction

Glioblastoma (GBM) remains one of the most aggressive brain cancers, with standard treatment unchanged since 2005. Challenges like the blood-brain barrier, tumor invasiveness, and resistance to standard therapies have hindered the discovery of new treatment options. However, personalized drug repurposing (PDR) and combination drug therapy (CDT) may offer promising solutions. PDR repurposes existing medicines, leveraging known efficacy for faster, cost-effective treatment discovery for GBM in a patient-tailored manner. However, treatment resistance to monotherapy in GBM is prevalent due to its genetic and phenotypic heterogeneity. To combat this, CDT targets multiple pathways simultaneously, improving treatment efficacy and patient outcomes.

Methods

We developed SPARC (Synergistic Prediction for Anticancer Repurposing and Combination), a tool integrating PDR and CDT to identify personalized drug combinations. SPARC leverages patient-derived GBM tissues and drug screening data. It incorporates two models: (i) Mechanisms of Action Landscape (MOAL), prioritizing drug pairings based on mechanisms of action, and (ii) Augmented Cancer Drug Atlas (ACDA), a machine learning model predicting drug synergy using the ALMANAC dataset of 100+ FDA-approved drugs.

Results

Using leave-one-out cross-validation, SPARC correctly predicted synergy in 60% of synergistic drug pairs and 65% of non-synergistic drug pairs. These findings suggest that SPARC can serve as a valuable tool for prioritizing drug combinations for experimental validation, ultimately contributing to more efficient drug discovery and personalized treatment strategies.

Conclusion

The SPARC model is a promising tool to identify personalized drug combinations for glioblastoma. Currently, in silico predicted synergistic drug combinations are being validated on patient-derived GBM cell cultures. SPARC will be extended to include advanced deep learning and predictive models to enhance the functionality.