Exploring and grouping COVID-19 pharmaceutical interventions to determine mechanisms of action and endotypes of response

Research question Do three novel therapeutic COVID-19 therapies work through different biological mechanisms and can we predict responders and non-responders to each therapy. Background SARS-CoV-2 has caused over 4 million deaths globally. Despite the success of vaccine development, vaccination is not ubiquitously available. The inequity of vaccination dissemination across the world, and the evolution of new strains capable of escaping vaccine protection has meant that novel drugs are still necessary to combat severe disease. This project utilises the ACCORD phase II clinical trial data that assesses the rapid testing and efficacy of novel drugs for COVID-19 treatment against the Standard of Care (SoC). The 3 novel drugs are: Bemcentinib, MEDI3506 'Äì Tozorikumab and Zilucoplan, of which their biological mechanisms of action in COVID-19 are still unknown. Aims and objective We aim to identify biological mechanisms underpinning the efficacy of three novel therapies for COVID-19. We will identify which features of the disease predict clinical response and explore the biological differences leading to different treatment outcomes. We will further build prognostic models to predict patient response to each novel drug and assess the most salient features driving drug response. The ability to predict and understand response to these novel drugs will improve patient recovery time and ease strain on healthcare resources. Methods Multiple data types have been generated from patients receiving one of the three COVID-19 therapies compared to patients receiving contemporaneous SoC. Patients'Äô clinical responses to the therapies have already been analysed, enabling the comparison of clinical response and quantification of clinical efficacy. All three treatments demonstrated a signal of clinical benefit however these differed in the size and nature of effect. This project aims to improve our understanding of these clinical effects through the analysis of RNA-Sequencing, inflammatory/immune proteomics, viral load, pharmacokinetics and patient phenotype. Bioinformatics analysis will first be applied to transcriptomic data to identify differentially expressed genes and the biological mechanisms that are impacted by the trial treatments. Gene expression results will be correlated across other data types. In addition machine learning will be applied to ascertain the most important predictors of response to each treatment to if patients are likely to respond well to a particular treatment and to create new stratification tools to target treatments. Timelines for delivery This project will last 13 months. To deliver the project in this timeframe activities will be coordinated in parallel. Anticipated impact and dissemination It is anticipated that this work will lead to multiple high impact publications as well as conference presentations. More practically, this work will lead the way in determining how novel COVID-19 drugs biologically impact patients and could lead to a reduced burden on healthcare systems through the more efficient prescription of COVID-19 therapies. These results will directly inform the design of Phase 3 trials of the individual therapies including patient inclusions and endpoint selections. They will also inform on the potential application of the treatments for other respiratory infections beyond COVID-19.