PII-122 - A COMPUTATIONAL PLATFORM TO SUPPORT THE DEVELOPMENT OF MRNA-BASED VACCINES AND THERAPEUTICS.

L. Dasti¹, E. Campanile^2,3, E. Pettinà¹, G. Fiandaca^1,4, N. Zangani¹, S. Giampiccolo², L. Leonardelli², L. Marchetti^5,1; ¹University of Trento, Department of Cellular, Computational and Integrative Biology (CIBIO), Trento, Italy, ²Fondazione The Microsoft Research - University of Trento Centre for Computational and Systems Biology (COSBI), Rovereto (TN), Italy, ³University of Trento, Department of Mathematics, Trento, Italy, ⁴Macbes team, Inria Center at Université Côte d'Azur, Sophia Antipolis, France, ⁵Fondazione The Microsoft Research - University of Trento Centre for Computational and Systems Biology (COSBI), Rovereto (TN), Italy.

Background: In recent years, mRNA vaccines and therapeutics have rapidly responded to disease outbreaks and pandemics. We introduce a computational platform based on mathematical modeling to support and accelerate the development of these products.

Methods: Our platform employs two mathematical modeling approaches based on systems of Ordinary Differential Equations (ODEs). The model for mRNA vaccines extends a Quantitative Systems Pharmacology (QSP) model [PMID: 34672423] and was trained and validated using clinical data from studies of PfizerBioNTech BNT-162b2 and Moderna mRNA-1273 vaccines [PMID: 34044428, 35386716, 34514454, 35746484, 36104370]. The model for mRNA therapeutics extends a Physiologically Based Pharmacokinetic (PBPK) model [PMID: 31079322] and has been trained and validated on public data of two mRNA-encoded antibodies: a Bi-specific T-cell engager (BiTE) and Pembrolizumab [PMID: 36403209, 35069904].

Results: The platform provides two modeling solutions to support the development of mRNA vaccines and therapeutics that can be tailored for specific mRNA-based products and parameterized for multiple animal models and specific human phenotypes (Figure 1). The QSP model for mRNA vaccines describes the main events post-vaccination, from antigen-presenting cell recruitment to antibody generation. The PBPK model for mRNA therapeutics describes mRNA trafficking, translation, and protein trafficking in tissues.

Conclusion: The platform enables rapid testing of numerous in-silico scenarios, significantly reducing the need for wet lab experiments, thereby shortening product development timelines and lowering costs. Its applications include identifying molecular processes that require optimization, defining experimental priorities, and computing optimal conditions such as species-specific dosing and scheduling, or translating treatment protocols across animal species and to humans. By training and validating our models with individual time-series data, the platform also serves as a powerful tool in precision medicine, allowing the identification of tailored treatment protocols that suit the unique characteristics of specific patient populations, such as elderly individuals or immunocompromised subjects.