Clinically-relevant computational model of airway remodelling to design and optimise interventions to treat asthma

Introduction: Quantifying the mechanisms underpinning asthma requires computational models that offer tightly controlled parameters and can be easily manipulated. Such models, following validation, can offer insights into drug pharmacodynamics and guide drug design & optimisation. We report a model that can predict impact of therapies targeting asthma. Methods: An agent-based model of a representative human airway featuring epithelial, mesenchymal and inflammatory agents was developed. Parametric analyses were conducted by altering model variables independently and collectively. For clinical validation, the following served as the hallmarks of asthma: 10 eosinophils/mm2 sub-mucosa & >10% muscle mass/wall area. Results: The most parsimonious set of boundary conditions (25 sets studied) to capture all three hallmarks was selected as the virtual patient and intervened by (i) inducing eosinophil apoptosis & (ii) reducing eosinophil recruitment. A range of ‘doses’ were tested. The dose inducing apoptosis in 15% eosinophil population reduced patient eosinophil activity by 54.1%, concurring with the clinical impact of Mepolizumab (55% loss). The dose reducing eosinophil recruitment by 40% lowered patient eosinophil activity by 81.4%, consistent with the clinical impact of Fevipiprant (79.6% loss). When predicting loss in muscle mass, the pro-apoptosis model showed 35% reduction for the highest dose. For the same dose the anti-recruitment model reported 70% reduction. Conclusions: Model suggests the anti-recruitment therapy to be more efficacious, demonstrating its potential as a robust predictive tool to guide drug development in respiratory medicine.