Gianluca Selvaggio (1), Giada Fiandaca (2), Lorena Leonardelli (1), Stefano Giampiccolo (1), Luca Marchetti (1,2)
(1) Fondazione The Microsoft Research - University of Trento Centre for Computational and Systems Biology (COSBI), Rovereto, Italy. (2) Department of Cellular, Computational and Integrative Biology (CIBIO), University of Trento, Trento, Italy.
Introduction: Over the past two years mRNA vaccine platforms have proven to be a formidable tool to quickly develop and deploy enough doses to carry the world out of a pandemic event. Furthermore, the possibility to quickly update the vaccine target upon the discovery of a new variant of the pathogen, further consolidates their pivotal role. However, despite the mRNA vaccine astonishing reduced developing time from benchwork to approval, dose selection was mainly guided by previous knowledge and the necessity to limit adverse effects. While the humoral immune response occurring after the antigen presentation is expected to be similar to traditional vaccines, the interaction between the human innate immune cells and mRNA molecules, or delivery vectors, appears to be fragmented amid the literature, making it difficult to formulate new hypotheses to be tested in clinical settings and ultimately improve this technology platform. Additionally, the difficulties of designing an adequate clinical trial with doses and scheduling optimization greatly affects the already short time window of a pandemic.
Mathematical models and computational tools can be instrumental in guiding vaccines development from the mRNA design to the scheduling and dosing selection eventually simulating boosting scenarios on virtual populations with patients’ cohorts of adequate size to properly inform the clinical trial design.
Objectives:
- Develop a QSP model able of simulating the induced immunogenicity at systemic level, depending on different vaccine doses, delivery systems and animal species.
- Infer dose scheduling and boosting scenarios through in silico simulation.
Methods: The QSP model is implemented in MATLAB as a set of ODEs and leverages on the work of Chen et al. 2014 [1], which was extended to represent the early events after injection according to the information retrieved through a Natural Language Processing (NLP) pipeline for automated knowledge extraction, developed to gather biological evidence about mRNA vaccines. Vaccine uptake and antigen expression were modelled based upon experimental data from Liang et al. 2017 [2].
Results: A QSP model previously developed [4] was further extended to describe in greater details the cellular dynamics happening at the injection sites after the vaccination. The model represents the whole immunogenic process providing, as endpoint to the simulation, the antibody titers. From the analysis of the QSP model, the main tuneable properties of the vaccine were identified, and a sensitivity analysis carried out to predict how prioritize drug design strategies and inform dose regimens during drug development.
Conclusions: In silico tools are becoming increasingly relevant and crucial not only to cut losses and support candidate lead selection, but also informing dosing and scheduling. In a fast-paced environments, like vaccine development during a pandemic, the herein presented model can be used starting from different vaccine specifics (e.g., targets, mRNA properties and delivery system) to virtually explore the drug optimization landscape and quantitatively address core issues in the pipeline of development and optimization of mRNA-based products [5].
References:
[1] Chen et al. CPT Pharmacometrics Syst. Pharmacol. (2014) 3 (133).
[2] Liang et al. Mol. Ther. (2017) 25:2635-2647.
[3] Leonardelli et al. Front Immunol. (2021) 12:738388.
[4] Selvaggio et al. CPT Pharmacometrics Syst Pharmacol. (2021) 10(12):1448-1451.
[5] The Wellcome Leap R3 project, https://wellcomeleap.org/r3/
Reference: PAGE 30 (2022) Abstr 10115 [www.page-meeting.org/?abstract=10115]
Poster: Drug/Disease Modelling - Infection