A first QSAR model for galectin-3 glycomimetic inhibitors based on 3D docked structures

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Biochemistry

Background:

• Galectin-3 is a key target in various diseases.
• Developing selective inhibitors is crucial for therapeutic applications.
• Existing inhibitors lack comprehensive structure-activity relationship data.

Purpose of the Study:

• To develop the first Quantitative Structure-Activity Relationship (QSAR) model for Galectin-3 glycomimetic inhibitors.
• To establish correlations between molecular properties and binding affinity (Kd).
• To provide a guideline for designing novel, potent, and selective Galectin-3 inhibitors.

Main Methods:

• Utilized docked structures of inhibitors to the Galectin-3 carbohydrate recognition domain (CRD).
• Employed quantitative numerical methods like Partial Least Squares (PLS) and Artificial Neural Networks (ANN).
• Applied dimensionality reduction techniques including PLS, Principal Component Analysis (PCA), and Genetic Algorithms (GA) for descriptor selection.

Main Results:

• A QSAR model was established using a dataset of 136 compounds with known binding affinities (Kd).
• Descriptor selection yielded final sets of 56 (PLS), 31 (PCA), and 35 (ANN) descriptors.
• The models demonstrated predictive capabilities for Galectin-3 binding affinity.

Conclusions:

• The developed QSAR model serves as a foundational tool for designing new Galectin-3 inhibitors.
• The study highlights the importance of modifications at C-3' and O-3 positions for inhibitor potency.
• This research paves the way for more targeted therapeutic strategies against Galectin-3 related conditions.