Deep Learning-Driven Anticancer Drug Discovery: Emodepside as a Potential Therapeutic Candidate for Triple-Negative Breast Cancer.
Summary
This study developed a deep learning framework to accelerate drug discovery for triple-negative breast cancer (TNBC), an aggressive subtype with limited therapeutic options. The model predicted anticancer efficacy, toxicity, and structural similarities of compounds. By screening over 6,000 compounds, the study identified emodepside as a promising candidate. In vitro and in vivo validations confirmed its antitumor efficacy, including inhibiting tumor growth in xenograft models. Multi-omics analyses revealed that emodepside acts by inhibiting NAMPT, promoting TNBC cell apoptosis.
Analysis
CLINIQUE: Triple-negative breast cancer (TNBC) represents a major clinical challenge due to the lack of effective targeted therapies. This study proposes a drug repurposing approach to identify new options. The identification of emodepside, with demonstrated in vivo efficacy in xenograft models (5 mg/kg inhibiting tumor growth), suggests therapeutic potential. While promising, these results are from preclinical models. A phase I/II clinical trial would be justified to assess safety and efficacy in humans, with a potential clinical impact in 5+ years. BIOMOL: The discovery of emodepside is significant as it is a structurally unique molecule, diverging from conventional anticancer agents. The study utilized integrated multi-omics approaches, including transcriptomic sequencing (RNA-seq) and mass spectrometry-based proteomics (CETSA-MS), to elucidate its mechanism of action. These analyses identified NAMPT as the primary target of emodepside, suggesting its anti-TNBC activity occurs via NAMPT inhibition and promotion of cell apoptosis. This discovery paves the way for novel predictive biomarkers or combination therapy strategies. BIOINFO: The study developed a deep neural network framework to accelerate drug discovery, focusing on repurposing. This model is capable of predicting anticancer efficacy, toxicity profiles, and structural similarities of compounds. It was applied to screen over 6,000 compounds from the Drug Repurposing Hub, enabling the identification of promising candidates. The approach demonstrates the capability of deep learning models to discover structurally novel anticancer agents, which could transform drug discovery workflows by prioritizing candidates for experimental testing.