(W-071) Mathematical modeling can help to successfully translate preclinical findings in mice models of asthma into first-in-human trials

Théo Galland, MSc – QSP Modeler, Novadiscovery; Caterina Sansone, PharmaD – Clinical Research Head, OM Pharma; Anne Vaslin Chessex, PhD – Head of Preclinical Research, OM Pharma; Christian Pasquali, PhD – Senior Scientific Liaison Director, OM Pharma; Lorenz Lehr, PhD – Head of Clinical and Preclinical Development, OM Pharma; Alexander Kulesza, QSP Modeler – Executive Manager, Novadiscovery; Simon Arsène, PhD – Executive Manager, Novadiscovery

Objectives: Experimental mouse models of asthma are instrumental for investigating involved immunological pathways and candidate drugs, but they are not without limitations. This has led to resources spent on drug candidates that show promise in preclinical settings, but often fail in clinical trials. Detecting such false-positive candidates would improve the efficiency of drug development pipelines. Mathematical models could bridge preclinical results with the clinical context by incorporating species-specific characteristics and can be informed with previous treatment investigations. Here we report the development of a translational model of asthma which captures the dynamics of key immune markers following an allergen challenge, both in mice and humans.

Methods: The model is constructed as a system of ordinary differential equations that describe the kinetics of 12 biomarkers including different cell types (CD103+ and CD11+ DCs, Tregs, ILC2s, Th2, eosinophils), cytokines (IL-5, -10, -13, -33), mucus overproduction and airway hyperresponsiveness. To inform the behavior of these variables, we collected a large dataset from heterogeneous sources composed of time dynamics following allergen challenge in mice and humans, as well as fold-change of these markers in knock-out mice strains. We parameterized the model to reproduce the mouse dataset and then applied allometric scaling for the translation to humans. We also added three cell populations specific to the human model: tissue-resident memory Th2 and Treg cells, and memory-like ILC2 cells. This follows the hypothesis that the tissue-resident memory compartment in humans is highly activated compared to mice due to years of repeated allergen exposure. Finally, we determined the values of the associated parameters to reproduce the human allergen challenge dataset.

Results: Our translational model can replicate the kinetics of 12 key immune markers in response to an allergen challenge both in humans, wild-type mice, and in knock-out mice with CD11+ DCs, IL-33 or Tregs depletion. By conducting simulations with the model, we can experimentally augment, decrease or deplete, alone or in combination, different cells and cytokines in both species, and measure the effect this would have in the late asthmatic reaction to an allergen challenge. Such experiments can help inform which pathways are more susceptible to generate false positive responses when targeted for treatment.

Conclusions: We developed a translational model of the late asthmatic response to an allergen challenge that can be broadly used in the drug development pipeline for new asthma treatments. In early stages, the model could help generate new hypotheses and guide preclinical experiments design. The model could also explore how candidate drugs with strong preclinical support could perform in first-in-human trials, while in later stages of drug development it could give insights for clinical trial design.

Citations: Holmes AM, Solari R, Holgate ST. Animal models of asthma: value, limitations and opportunities for alternative approaches. Drug Discov Today. 2011 Aug;16(15-16):659-70. doi: 10.1016/j.drudis.2011.05.014. Epub 2011 Jun 23. PMID: 21723955.