Multiple machine learning models combined with virtual screening and molecular docking to identify selective human...

Area of Science:

• Biochemistry and Molecular Biology
• Computational Chemistry
• Oncology

Background:

• Aldehyde dehydrogenases (ALDHs) are oxidoreductase enzymes crucial for aldehyde metabolism.
• Dysregulated ALDH expression is linked to various diseases, notably cancers.
• ALDH1A1 overexpression in tumors contributes to anti-cancer drug resistance.

Purpose of the Study:

• To identify novel, selective small molecule inhibitors targeting ALDH1A1.
• To overcome limitations of existing ALDH1A1 inhibitors, such as poor selectivity and pharmacokinetics.
• To develop potential adjuvant therapies for cyclophosphamide and cisplatin-resistant cancers.

Main Methods:

• Integrated machine learning models for various ALDH isoforms.
• Employed in-silico techniques: virtual screening, molecular docking, ADMET profiling, and molecular dynamics (MD) simulation.
• Focused on identifying inhibitors with selectivity for ALDH1A1.

Main Results:

• Identified ten selective ALDH1A1 inhibitors with diverse chemical scaffolds.
• Selected inhibitors demonstrated appropriate Absorption, Distribution, Metabolism, Excretion, and Toxicity (ADMET) properties.
• These compounds are suitable for further development as potential therapeutic agents.

Conclusions:

• The identified selective ALDH1A1 inhibitors represent promising candidates for overcoming drug resistance in cancer.
• In-silico approaches combined with machine learning effectively identified novel drug leads.
• Further preclinical development could lead to new adjuvant therapies for resistant malignancies.