Artificial Intelligence-Based Computational Screening and Functional Assays Identify Candidate Small Molecule...

Area of Science:

• Biochemistry
• Oncology
• Drug Discovery

Background:

• PTPmu (receptor protein tyrosine phosphatase mu) is involved in cell adhesion and signaling.
• Proteolytic downregulation of PTPmu in glioblastoma leads to fragments that may promote cancer progression.
• Targeting these PTPmu fragments offers a potential therapeutic strategy for glioblastoma.

Purpose of the Study:

• To identify novel compounds that interact with PTPmu fragments.
• To evaluate the efficacy of identified compounds in inhibiting PTPmu-mediated cell adhesion and glioblastoma growth.

Main Methods:

• Utilized AtomNet®, a deep learning platform, to screen millions of compounds for predicted interaction with PTPmu.
• Conducted cell-based assays to assess PTPmu-dependent cell aggregation and glioma sphere formation.
• Validated compound interactions with PTPmu extracellular fragments.

Main Results:

• Identified 76 candidate compounds predicted to interact with PTPmu.
• Four compounds inhibited PTPmu-mediated Sf9 cell aggregation; six inhibited glioma sphere formation.
• Two compounds showed efficacy in both assays, with one notably inhibiting PTPmu aggregation and glioma growth down to 25 micromolar.
• The lead compound demonstrated direct interaction with PTPmu extracellular fragments.

Conclusions:

• The identified lead compound is a promising starting point for developing PTPmu-targeting cancer therapeutics.
• This approach offers a novel strategy for treating glioblastoma and other cancers involving PTPmu dysregulation.