Prediction of high-performing spleen-targeted lipid nanoparticles using a deep learning model for robust anticancer...

Area of Science:

• Biotechnology
• Nanomedicine
• Immunotherapy

Background:

• Messenger RNA (mRNA) therapeutics offer broad medical potential.
• Lipid nanoparticles (LNPs) are advanced mRNA delivery vehicles but face challenges like off-target effects and liver accumulation.
• Optimizing ionizable lipids for LNPs is crucial but experimentally intensive.

Purpose of the Study:

• To develop a deep learning model for accelerated virtual prediction of high-performing ionizable cationic lipids.
• To synthesize and validate novel bis-hydroxyethylamine derived lipids (BDLs) for enhanced mRNA delivery.
• To evaluate the efficacy and safety of lead candidate YK-407 for targeted mRNA delivery and therapeutic applications.

Main Methods:

• A deep learning model was created to predict ionizable lipid performance.
• 24 bis-hydroxyethylamine derived lipids (BDLs) were synthesized based on model predictions.
• In vitro and in vivo studies assessed LNP transfection efficiency, organ targeting (spleen, liver), and anti-tumor efficacy in a mouse model.

Main Results:

• The AI model successfully identified promising ionizable lipid candidates.
• YK-407 demonstrated superior in vitro transfection and in vivo spleen-specific mRNA delivery, targeting antigen-presenting cells (APCs).
• YK-407 LNPs significantly inhibited tumor growth and induced CD8+ T cell responses in a mouse model, outperforming existing standards with minimal toxicity.

Conclusions:

• Artificial intelligence can significantly accelerate LNP development for mRNA therapeutics.
• YK-407 represents a promising novel LNP formulation with enhanced targeting, efficacy, and safety profiles.
• YK-407 holds potential for advancing mRNA-based treatments, particularly in cancer immunotherapy.