Related Experiment Video
Updated: Sep 15, 2025

Scalable Biomanufacturing Workflow to Produce and Isolate Natural Killer Cell-Derived Extracellular Vesicle-Based Cancer Biotherapeutics
Published on: August 16, 2024
The ExoGAN generative AI framework enables extracellular vesicle-based immunotherapy
Kiley Graim1, Zachary Greenberg1, Tina Salehi Torabi1
1University of Florida.
We developed ExoGAN, an AI framework to design novel tumor neoantigens for cancer immunotherapy. ExoGAN-engineered extracellular vesicles effectively activate T cells, improving anti-tumor immunity and enabling precision cancer treatments.
Area of Science:
- Immunology
- Artificial Intelligence
- Biotechnology
Background:
- Neoantigens (tumor antigenic peptides presented by MHC-I) are crucial for CD8+ T cell activation in cancer immunotherapy.
- Current immunotherapies face challenges due to patient-specific tumor heterogeneity and unpredictable T Cell Receptor (TCR) binding properties.
- Developing novel neoantigens with predictable TCR recognition is essential for effective cancer treatment.
Purpose of the Study:
- To introduce ExoGAN, a generative AI framework designed to guide the creation of novel tumor antigenic peptides.
- To engineer extracellular vesicles (EVs) decorated with ExoGAN-designed peptides for enhanced T cell activation and anti-tumor immunity.
- To improve the predictability and efficacy of cancer immunotherapy through AI-driven neoantigen design.
Main Methods:
- Developed ExoGAN, a generative adversarial network (GAN) integrating HLA physiochemistry and sequence data for peptide design.
- Trained ExoGAN on the largest IEDB dataset of HLA peptides to design HLA-A*02:01-targeting peptides.
- Engineered extracellular vesicles (EVs) to present ExoGAN-designed peptides in MHC complex proteins for targeted delivery and T cell activation.
Main Results:
- ExoGAN successfully designed novel human neoantigens with improved TCR recognition and T cell activation potential.
- EVs decorated with ExoGAN-designed peptides demonstrated consistent antigenic presentation to T cell TCRs.
- Computational and experimental validation confirmed programmable EV agents for effective anti-tumor immunity activation.
Conclusions:
- The ExoGAN framework enables the design of neoantigens with enhanced TCR recognition and T cell activation capabilities.
- EV-based delivery of ExoGAN-designed peptides offers a programmable platform for precision cancer immunotherapy.
- This approach advances mechanistic understanding of neoantigen functionality and accelerates the development of novel immunotherapy agents.
More Related Videos
08:02In Vivo Immunogenicity Screening of Tumor-Derived Extracellular Vesicles by Flow Cytometry of Splenic T Cells
Published on: September 23, 2021
08:35Therapeutic Gene Delivery and Transfection in Human Pancreatic Cancer Cells using Epidermal Growth Factor Receptor-targeted Gelatin Nanoparticles
Published on: January 4, 2012
Related Concept Videos
Tumor Immunotherapy
Gene Therapy
Microorganisms in Medicine and Therapeutics