Himanshu Kaul, our member from University of Toronto, reports the study of a new drug that reduces airway muscle mass in severe asthmatics. The strength of this work is the integration of findings from in vivo clinical trials and in vitro and computational models, demonstrating the key role of in silico clinical trials in the design of new drugs.
It is estimated that around 30 million children and adults under the age of 45 years suffer from asthma in the EU alone. Asthma emerges as a result of complex interactions between a patient’s genes, cells, and environment – a process that is referred to as remodeling. During pathological remodeling, the airways of a patient are invaded by inflammatory cells, add on muscle mass, and become stiffer, which together results in reduced airflow in patients. The disease is a great source of fiscal and emotional burden on society and its prevalence is on the rise despite the availability of treatment options.
Currently, asthma treatments target good asthma control, which includes minimising symptom burden and risk of exacerbation. However, the lack of drugs that can alleviate patients’ breathing difficulties, including a drug that can reduce airway muscle mass in asthmatics, and costs associated with asthma drugs are some of the causes behind asthma’s persistent rise. Tackling asthma requires a mechanistic understanding of how the various elements contributing to the severity and heterogeneity of the disease come together across multiple spatial and temporal scales.
This, however, can only be achieved by integrating computational and experimental investigations. This is because of the wide number and variety of factors responsible for the disease, which are difficult to control experimentally. Computational models, however, offer a more manageable approach where the ‘undesired’ complexity of the system can be reduced by focusing only on parameters that are hypothesied to play a role in the process alone. Secondly, these models are helpful as they enable us to build a dynamic understanding out of the static snapshots we obtain from experimental data. In this manner, they help simulate phenomena that may not be accessible to observation otherwise.
A recent study conducted by the AirPROM consortium, led by Prof Chris Brightling, University of Leicester, integrated computational modeling conducted at the Universities of Sheffield and Leicester by Dr Himanshu Kaul and Prof Rod Smallwood, with a phase II clinical trial using a novel asthma drug, Fevipiprant, to determine the effectiveness of the drug in reducing asthma symptoms. To simulate the clinical trial virtually, the investigators developed a virtual patient using an agent-based model. The application of the agent-based approach was critical, as this approach simulates the time evolution of a system based on interactions between the various model elements. As a result, agent-based models capture the heterogeneity of biological systems and the emergence of higher-level biological phenomena, including function and pathology.
The investigators developed a virtual airway that contained the epithelial, mesenchymal, and inflammatory elements. A set of rules capturing interactions between these various elements was derived from existing literature. In health, these elements work in harmony to ensure effective airflow and appropriate response to external challenges. In asthma, the harmony of these interactions get compromised resulting in an abnormal recruitment of inflammatory cells in the airway, a disrupted epithelium, and increase in muscle mass. When the model was run with ‘normal’ parameters, the model resulted in a normal functioning airway that responded appropriately to challenges (e.g. removing half of the epithelium). However, when these parameters were made to reflect pathology, e.g. an aggressive inflammatory system, the airway remodeling led to the hallmarks of asthma.
The pathological airway was then used to conduct in silico trials by administering virtual drugs. To validate that the model was capturing biological reality, they first administered virtual Mepolizumab. The drug reduced the number of eosinophils, but did not show loss of muscle mass. Both observations were consistent with clinical data. When Fevipiprant was administered, the virtual patient showed a loss in eosinophilia, consistent with clinical data, but the observed reduction in muscle mass was not the same as observed during the clinical trial. This led to the hypothesis that the drug is impacting directly on muscle mass, in addition to its impact via reduction in eosinophilia. Experiments conducted with muscle cells taken from patients suggested that the drug reduced recruitment of cells called myofibroblasts, which add to the airway muscle mass during remodeling. Addition of this secondary feature in the model led to muscle mass reduction in the virtual patient consistent with clinical results.
This study published in Science Translational Medicine reported the first drug to reduce airway muscle mass in severe asthmatics. The drug has been hailed a ‘gamechanger’ in asthma treatment. The drug agnostic virtual patient, on the other hand, represents a milestone in patient-specific computational modeling that will in future consider genomic, tissue, organ-level patient information to predict optimal personalised strategies for patients. The model can play a key role in cost-effective personalised treatment decision-making and design of new drugs, thereby addressing the challenge of persistent rise in asthma.
The full article is available on the Publisher webpage: http://stm.sciencemag.org/content/11/479/eaao6451
