In Silico Discovery and Repurposing Of Dual Monoamine Oxidase Inhibitors for Glioblastoma Treatment Through Machine Learning and ADMET Profiling

Creative Commons License

Arslanbaş A. M., Güler H. İ.

MolBiyoKon’26 10 th International Congress of the Molecular Biology Association of Turkey, İstanbul, Türkiye, 17 - 19 Eylül 2026, ss.107, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.107
  • Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
  • Karadeniz Teknik Üniversitesi Adresli: Evet

Özet

In Silico Discovery And Repurposing Of Dual Monoamine Oxidase Inhibitors For Glioblastoma Treatment Through Machine Learning And ADMET Profiling


Ayse Mina ARSLANBAS1, 2, Halil Ibrahim GULER1

1 Karadeniz Technical University, Faculty of Science, Department of Molecular Biology and Genetics, 61080, Trabzon, Turkiye
2 Karadeniz Technical University, Graduate Institute of Natural and Applied Sciences, Department of Molecular Biology and Genetics, 61080, Trabzon, Turkiye

 

Background/Aim

Glioblastoma is an aggressive central nervous system (CNS) malignancy necessitating innovative treatments. Monoamine oxidase A (MAO-A) and B (MAO-B) enzymes are highly promising therapeutic targets. This study aims to identify effective, CNS-penetrable dual MAO inhibitors by executing two distinct computational strategies: drug repurposing and novel lead discovery via machine learning (ML).

Materials and Methods 

Two parallel in silico workflows were systematically executed. Data preprocessing and Extended Connectivity Fingerprint 4 (ECFP4) extractions utilized KNIME and RDKit. The ML pipeline employed Random Forest and CatBoost algorithms, augmented by SHAP (SHapley Additive exPlanations) analysis for model explainability. The drug repurposing approach screened 5000 FDA-approved drugs from the ChEMBL database, while the novel lead discovery screened extensive chemical libraries. Promising hits underwent comprehensive Absorption, Distribution, Metabolism, Excretion, and Toxicity (ADMET) profiling via ADMETlab 3.0, SwissADME, and ProTox-II. Tanimoto similarity against Safinamide was calculated to ensure the structural novelty of the candidates.

Results 

The dual-strategy pipeline yielded three elite candidates for future validation. The drug repurposing workflow identified the FDA-approved drug Chlorzoxazone, possessing a highly safe Class 4 toxicity profile (LD50: 1230 mg/kg) suitable for neuro-oncology. The novel lead workflow identified two distinct compounds: CHEMBL2106281 and CHEMBL2105544. Tanimoto scores (ranging from 0.09 to 0.14) confirmed these novel leads do not structurally resemble existing MAO inhibitors. ADMET profiling revealed that all three candidates possess high blood-brain barrier (BBB) penetration probabilities (>0.900) alongside exceptionally low major organ toxicity profiles.

Conclusion 

Executing parallel computational workflows integrating KNIME, CatBoost, and SHAP analysis successfully identified three equally valuable, CNS-penetrable dual MAO-A/B inhibitor candidates: one repurposed drug (Chlorzoxazone) and two structurally novel chemical leads. These compounds establish a robust foundation for Glioblastoma therapy and are prioritized for upcoming structural validation via molecular docking and molecular dynamics (MD) simulations.

 

 

Keywords:

Glioblastoma, Drug Repurposing, Novel Lead Discovery, Machine Learning, ADMET