Surya Gyawali (MSc Physics, 68090700701)

MSc in Physics, KMUTT, Bangkok, Thailand

Advisor: Dr. Thana

Project umbrella: Applications of Molecular Modeling and AI-Based Techniques to Screen Antibodies Specific for Diseases

1. Short Overview

This thesis is part of a larger research project that combines molecular modeling, artificial intelligence (AI), and experimental validation to design next-generation therapeutics for G protein–coupled receptors (GPCRs).

My work focuses on the GLP-1 receptor (GLP-1R), a key target in treatments for type 2 diabetes and obesity. By using molecular dynamics (MD) simulations, Principal Component Analysis (PCA), and AI-based structure prediction tools (such as AlphaFold-Multimer, MaSIF, and DeepInteract), I aim to:

• understand how GLP-1R moves in different signaling states (Gs vs β-arrestin), and
• contribute to the rational design and in silico screening of peptide-fused nanobodies that preferentially stabilize Gs-biased, therapeutically favorable conformations.

WhatsApp Video 2025-11-21 at 13.05.50_ffeb0abd.mp4

2. Background – Why GLP-1R and Nanobodies?

• GLP-1R in metabolism:

  GLP-1R is a class B1 GPCR that regulates insulin secretion, appetite, and blood glucose. GLP-1R agonists such as semaglutide have transformed type 2 diabetes and obesity treatment by increasing cAMP production, promoting insulin release, and reducing appetite.

• The problem of side effects:

  Many GLP-1R agonists activate:

  • Gs-protein signaling → desired: cAMP production, better glucose control
  • β-arrestin signaling → often linked to side effects (nausea, receptor desensitization, internalization)

  Therapies that favor Gs over β-arrestin (Gs-biased agonism) can potentially prolong beneficial signaling and reduce adverse effects.

• Biased signaling:

  Different ligands can bind the same receptor but push it into different conformational ensembles. These subtle structural differences lead to biased signaling: one ligand might strongly favor Gs coupling, another might more easily recruit β-arrestin. Understanding which motions correspond to which pathway is essential for rational drug design.

• Peptide-fused nanobodies as next-generation therapeutics:

  A promising strategy is to create peptide-fused nanobodies:

  • a GLP-1-like agonist peptide
  • covalently linked to a nanobody that binds the GLP-1R extracellular domain (ECD) or other specific regions.

  This dual engagement can “unlock” the receptor, stabilize active conformations, prolong half-life, and potentially bias signaling toward Gs. Existing examples (e.g., Everestmab, Glutazumab) show that such fusion constructs can give long-lasting glucose control and enhanced selectivity, but their full structural and dynamic mechanisms are not yet public.

This motivates a computational + AI-guided pipeline that connects receptor dynamics, nanobody design, and experimental testing.

3. My Role Within the Larger Grant

The grant project has three major components: