A Quantitative System Pharmacology model for the development and optimization of mRNA vaccines.
Lorenzo Dasti (1), Giada Fiandaca (1,2), Natascia Zangani (1), Lorena Leonardelli (3), Elio Campanile (3,4), Elisa Pettina’ (1), Stefano Giampiccolo (3), Luca Marchetti (1,3).
1. Department of Cellular, Computational and Integrative Biology (CIBIO), University of Trento, Italy; 2. Current affiliation: Macbes team, Inria Center at Université Côte d'Azur, France; 3. Fondazione the Microsoft Research – University of Trento Centre for Computational and Systems Biology, Rovereto, Italy; 4. Department of Mathematics, University of Trento, Italy.
Introduction: The necessity of reducing time and costs in the development phase of a vaccine has been an open challenge for decades, but the COVID-19 pandemic has raised this problem to the top of the general interest. This was the first time mRNA vaccines – a technology that has been around for years – could prove their incredible potential. Mathematical modeling is turning out to be one of the best tools for supporting drug development, and it can be used to boost the search in this promising field further, as confirmed by the growing interest in modeling and simulation applications by the European Medicines Agency (EMA) and the U.S. Food and Drug Administration (FDA).
Objectives: We present a Quantitative System Pharmacology (QSP) model describing the events following the administration of an mRNA vaccine, from product injection to the appearance of antibodies in the blood. It consists of two linked model layers: a molecular one that accounts for product-specific parameters (like antigen molecular weight and translation rate of mRNA) together with an extracellular level that comprehends different compartments (injection site, draining lymph node, blood) and describes the main actors involved in the immune response at the whole-body scale. The objective is to create a computational tool capable of reproducing the most important events occurring after an mRNA vaccine administration and – by simulating different scenarios – can guide the optimization of several crucial aspects like identification of optimal dosing and appropriate scheduling windows.
Methods: The model has been implemented in MATLAB and lies on a previous one introduced by the group [1]. Each of the two model layers consists of a system of ordinary differential equations simulated via MATLAB ode15s; the model parameters have been estimated through a nonlinear least-squares method. The model development process has been informed by literature evidence through a text-mining procedure driving the model design towards a knowledge-oriented approach [2]. For calibration and validation, we used data from preclinical and clinical works, like counts of antigen-presenting and immune cells in different tissues [3] and time series of serum antibodies according to product doses and administration schedules [4,5].
Results: The model exhibited good capabilities in reproducing the IgG dynamics in the blood of different COVID-19 mRNA vaccines. In the case of the PfizerBioNTech BNT-162b2 vaccine, we estimated the model parameters by fitting two doses (1 μg and 30 μg) [6], which were successfully validated by testing the model capability of predicting the generated antibodies with two other doses (10 μg and 20 μg) of the same vaccine. We then used literature data on the Moderna mRNA-1273 vaccine to test the level of adaptability of our platform in supporting the development of a novel vaccine. Considering the parameters estimates previously computed for the Pfizer vaccine, data from one dose (100 μg) of the Moderna vaccine was enough to refine the parameter estimates and adapt the computational model, successfully validated on other two doses of the same product (25 μg [7] and 50 μg [8]), finding a satisfactory agreement with the literature data. This fact indicates that our computational tool can reproduce saturation phenomena that may happen in vaccine dose escalation studies, a key feature to inform dose-finding applications.
Conclusions: The presented model can quickly reproduce different in silico scenarios. It demonstrates a good representation of the initial events following the administration of an mRNA vaccine, together with a predictive power concerning the saturating behavior and the identification of optimal doses. Its multiscale and accurate design makes the model adequate for testing in silico a significant number of tailored customizations – e.g., changes in the mRNA sequence, in the LNP formulation, and in the administration protocol – providing a first step toward a multiscale QSP model to support optimal dose finding for a wide range of mRNA vaccine products.
References:
[1] Selvaggio G., Leonardelli L. et al. “A quantitative systems pharmacology approach to support mRNA vaccine development and optimization”. CPT Pharmacometrics Syst Pharmacol. 2021;10:1448–1451.
[2] Leonardelli L. et al. “Literature mining and mechanistic graphical modelling to improve mRNA vaccine platforms”. Front Immunol. 2021;12:738388.
[3] Liang F. et al. “Efficient targeting and activation of antigen-presenting cells in vivo after modified mRNA vaccine administration in rhesus macaques”. Mol Ther. 2017;25:2635-2647.
[4] Goel R. R. et al. “Distinct antibody and memory B cell responses in SARS-CoV-2 naïve and recovered individuals after mRNA vaccination”. Sci. Immunol. 6, eabi6950 (2021).
[5] Keshavarz B. et al. (2022) “Trajectory of IgG to SARS-CoV-2 After Vaccination With BNT162b2 or mRNA-1273 in an Employee Cohort and Comparison With Natural Infection. Front. Immunol. 13:850987.
[6] Sahin U., Muik A., Vogler I. et al. “BNT162b2 vaccine induces neutralizing antibodies and poly-specific T cells in humans”. Nature 595, 572–577 (2021).
[7] Jochum S. et al. (2022) “Clinical Utility of Elecsys Anti-SARS-CoV-2 S Assay in COVID-19 Vaccination: An Exploratory Analysis of the mRNA-1273 Phase 1 Trial”. Front. Immunol. 12:798117.
[8] Kirste I., Hortsch S., Grunert V.P. et al. “Quantifying the Vaccine-Induced Humoral Immune Response to Spike-Receptor Binding Domain as a Surrogate for Neutralization Testing Following mRNA-1273 (Spikevax) Vaccination Against COVID-19”. Infect Dis Ther. 12, 177–191 (2023).