AI-driven discovery of GPNMB CAR T cells as a multi-cancer therapy

Area of Science:

• Immunology
• Oncology
• Bioinformatics

Background:

• Chimeric antigen receptor (CAR) T cell therapy shows promise for various cancers but faces challenges in identifying effective targets.
• Target discovery is a critical bottleneck limiting the broader application of CAR T cell therapies.

Purpose of the Study:

• To develop an artificial intelligence (AI)-driven pipeline for discovering novel CAR T cell targets.
• To identify and validate new therapeutic targets for multi-cancer CAR T cell treatments.

Main Methods:

• Integrated single-cell RNA sequencing data from human skin cancer and healthy tissues.
• Utilized AI and large language models for target prioritization and nomination.
• Validated target expression and engineered GPNMB-directed CAR T cells.

Main Results:

• Glycoprotein non-metastatic melanoma protein B (GPNMB) was identified as a top candidate target.
• GPNMB expression was confirmed across diverse hematologic and solid tumors.
• Engineered GPNMB-CAR T cells demonstrated significant anti-tumor efficacy in preclinical models.

Conclusions:

• An AI-driven approach provides a scalable pipeline for CAR T cell target discovery.
• GPNMB is a promising pan-cancer target for CAR T cell therapy development.
• GPNMB-directed CAR T cells offer potential as a broad therapeutic strategy for multiple malignancies.