Towards a ML-powered Multiscale Computational Platform Based on QSP and PBPK Modeling to Support the Development of mRNA-based Therapies

mRNA-based therapeutics have emerged as a transformative class of medicines, yet their translation beyond infectious disease vaccines remains challenged by the absence of an integrated pharmacological framework accounting for the tri-component nature of these therapies - the lipid nanoparticle, the mRNA, and the expressed protein. Here, we present a modular, multiscale computational platform integrating two complementary mechanistic models covering the full pharmacological cascade of mRNA-based immunotherapies. The first is a Quantitative Systems Pharmacology (QSP) model describing the immunological response to mRNA vaccines, from antigen expression in antigen-presenting cells through B cell activation and circulating antibody production. The second is a Physiologically Based Pharmacokinetic (PBPK) model tracking whole-body disposition of mRNA-encoded therapeutic antibodies, incorporating a molecular layer resolving LNP uptake, endosomal mRNA escape, and intracellular translation. Both models are informed by a machine learning pipeline that maps IVT-mRNA nucleotide sequences directly onto kinetic parameters, enabling product-specific model simulations. We propose this platform as a step toward the quantitative pharmacological framework that mRNA therapeutics currently lack, and as a practical tool for model-informed design and development of this therapeutic class.