OC 24.1 - Using artificial intelligence to disrupt the vWF- GPIBα interaction in platelet-type von Willebrand disease

Background: Platelet glycoprotein-1B alpha (GPIBα) is a surface glycoprotein that is functionally defective (GPIBαFD) in platelet-type von Willebrand disease (PT-VWD). Gain-of-function mutations within the β-sheet of the glycoprotein convert GPIBα to an open conformation resulting in the excessive association of GPIBαFD with its receptor von Willebrand clotting factor (vWF), and counterintuitively a bleeding phenotype. The β-sheet mutations (clustered within the Trp230 - Met239 window) therefore represent an attractive site to develop therapeutics. We recently developed the In-Silico Protein Synthesizer (InSiPS) – an artificial intelligence that designs small peptides which interfere with important protein-protein interactions (PPIs).

Aims: We are using InSiPS to design small peptides that disrupt the interaction between GPIBαFD and vWF.

Methods: We used InSiPS to generate peptides specific to GPIBαFD. The resulting peptides were validated using docking simulations and in-vitro binding assays. A total of 23 peptides were generated and assessed for target affinity using AlphaFold2 software. Peptide specificity towards GPIBαG233V, M239V was assessed in-vitro using differential scanning fluorimetry (DSF), aggregation assays, and a pulldown assay.

Results: Derived from our InSiPS AI (Figure 1), one peptide (G14) displayed high affinity towards GPIBαG233V, M239V (Figure 2A). Via DSF, G14 was found to increase the melt peak of GPIBαFD (Figure 2B) in a dose-dependent manner. Furthermore, G14 disrupts the PPI between GPIBαG233V, M239V and vWF via a latex bead aggregation assay (Figures 2C and D) in a dose-dependent manner. These results are supported by a pulldown assay, which indicates that G14 disrupts the interaction between GPIBαG233V, M239V and vWF.

Conclusion(s): This study demonstrates that the AI-generated G14 peptide can bind to and disrupt the interaction between GPIBαG233V, M239V and vWF. Consequently, we introduce G14 as a potential drug candidate for PT-VWD. Future work will assess the efficacy of the G14 candidate in mammalian systems.