Rational Immunogenicity Modulation of Virus-Like Particles by VLPIM

Virus-like particles (VLPs) are versatile platforms for drug delivery and vaccine development, but precisely customizing their immune profiles for specific clinical applications remains challenging. Here, we present a computational framework that integrates complementary artificial-intelligence-driven tools for protein design, structure prediction, and antigenicity assessment to enable rational immunogenicity modulation of VLPs while preserving their structural integrity. Leveraging the central role of CD4⁺ T cells in adaptive immunity, this framework first combines class II major histocompatibility complex (MHC-II)-restricted epitope prediction with structure-guided sequence optimization to generate a focused mutation library. Subsequently, its virtual screening workflow combines immunogenicity predictions and structural modeling to pinpoint variants likely to attenuate or enhance CD4⁺ T cell activation while preserving VLP integrity. This framework enables a broader and more efficient exploration of the mutation space to identify VLP structural variants with tailored immunogenicity profiles; hence, we term it the VLPIM (Virus-Like Particle Immunogenicity Modulation) platform. Using hepatitis B core VLPs as proof of concept, VLPIM generated a mutation library of 120 variants, computationally filtered down to 4 candidates, and ultimately identified 2 variants that retained native virus-like morphology, as confirmed by transmission electron microscopy. In vitro immunogenicity assays confirmed that epitope substitutions could either suppress or enhance CD4⁺ T cell activation. This generalizable pipeline permits precise immunogenicity modulation of VLPs and other multimers, accelerating their development as targeted drug carriers and next-generation vaccine platforms. VLPIM is available as both a locally executable tool on GitHub (https://github.com/RuijinHospitalVNAR/VLPIM) and an online web server (https://ruijinhospitalvnar.github.io/Toolboxes/VLPIM_Web_services/).