Artificial neural network-based drug design for diabetes mellitus using flavonoids

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Biochemistry

Background:

• Diabetes mellitus is a chronic metabolic disorder characterized by impaired glucose regulation.
• Diabetic complications are often linked to the activity of aldose reductase (AR), an enzyme in the polyol pathway.
• Flavonoids, known for antioxidant and radical scavenging properties, are investigated as potential AR inhibitors (ARIs).

Purpose of the Study:

• To perform quantitative structure-activity relationship (QSAR) analysis on flavones and flavonols.
• To predict novel ARIs with concurrent radical scavenging activity.
• To identify potential therapeutic agents for mitigating diabetic complications.

Main Methods:

• Utilized artificial neural networks for QSAR analysis of flavones and flavonols.
• Investigated the impact of functional groups (H, -OH, -CH3) at eight substitution sites.
• Screened 6561 potential compounds for AR inhibition and radical scavenging activity.

Main Results:

• Predicted 69 potent ARIs and 346 compounds with strong radical scavenging activity.
• Identified 10 overlapping compounds demonstrating both strong AR inhibition (IC50 < 0.3 μM) and radical scavenging activity (IC25 < 1.0 μM).

Conclusions:

• The 10 identified compounds show significant potential as therapeutic agents for diabetic complications.
• These compounds warrant further experimental validation for their efficacy and safety.