A Combined Chemoinformatics- and Machine Learning-Based Approach Identifies Chlormidazole as a Drug Repurposing...

Area of Science:

• Computational chemistry
• Oncology
• Drug discovery

Background:

• Treatment options for advanced, therapy-resistant, and metastatic prostate cancer (PCa) are limited.
• There is a need for novel therapeutic strategies and drug repurposing in PCa management.

Purpose of the Study:

• To develop and validate an integrated chemoinformatics and machine learning (ML) workflow for drug repurposing in prostate cancer.
• To identify potential drug candidates for advanced PCa through virtual screening.

Main Methods:

• Developed an integrated chemoinformatics and ML workflow incorporating ligand-based similarity filtering.
• Validated the workflow on independent external datasets to assess performance and reduce false positives.
• Screened the DrugBank database to identify repurposing candidates with antiproliferative activity against PCa cells.

Main Results:

• The integrated workflow demonstrated improved virtual screening performance and reduced false positives.
• Five drug repurposing candidates with confirmed antiproliferative activity in 2D and 3D PCa models were identified.
• The antifungal agent chlormidazole showed significant tumor-selective, cytostatic activity, inducing G1 arrest via p57 upregulation and reduced Rb phosphorylation.
• Chlormidazole enhanced docetaxel's antiproliferative effect in PCa models, enabling comparable efficacy at lower drug concentrations.

Conclusions:

• Chlormidazole is a promising repurposing candidate for prostate cancer treatment.
• The integration of chemoinformatics and ML is a valuable strategy for drug repurposing and virtual screening in oncology.
• This approach can accelerate the identification of novel therapies for challenging diseases like advanced PCa.