COVID-19 emerged as a worldwide pandemic in early 2020, and while the rapid development of safe and efficacious vaccines stands as an extraordinary achievement, the identification of effective therapeutics has been less successful. This process has been limited in part by a lack of human-relevant preclinical models compatible with therapeutic screening on the native virus, which requires a high-containment environment. Here, we report SARS-CoV-2 infection and robust viral replication in PREDICT96-ALI, a high-throughput, human primary cell-based organ-on-chip platform. We evaluate unique infection kinetic profiles across lung tissue from three human donors by immunofluorescence, RT-qPCR, and plaque assays over a 6-day infection period. Enabled by the 96 devices/plate throughput of PREDICT96-ALI, we also investigate the efficacy of Remdesivir and MPro61 in a proof-of-concept antiviral study. Both compounds exhibit an antiviral effect against SARS-CoV-2 in the platform. This demonstration of SARS-CoV-2 infection and antiviral dosing in a high-throughput organ-on-chip platform presents a critical capability for disease modeling and therapeutic screening applications in a human physiology-relevant in vitro system.

The average cost to bring a new drug from its initial discovery to a patient's bedside is estimated to surpass $2 billion and requires over a decade of research and development. There is a need for new drug screening technologies that can parse drug candidates with increased likelihood of clinical utility early in development in order to increase the cost‐effectiveness of this pipeline. For example, during the COVID‐19 pandemic, resources were rapidly mobilized to identify effective therapeutic treatments but many lead antiviral compounds failed to demonstrate efficacy when progressed to human trials. To address the lack of predictive preclinical drug screening tools, PREDICT96‐ALI, a high‐throughput (n = 96) microphysiological system (MPS) that recapitulates primary human tracheobronchial tissue,is adapted for the evaluation of differential antiviral efficacy of native SARS‐CoV‐2 variants of concern. Here, PREDICT96‐ALI resolves both the differential viral kinetics between variants and the efficacy of antiviral compounds over a range of drug doses. PREDICT96‐ALI is able to distinguish clinically efficacious antiviral therapies like remdesivir and nirmatrelvir from promising lead compounds that do not show clinical efficacy. Importantly, results from this proof‐of‐concept study track with known clinical outcomes, demonstrate the feasibility of this technology as a prognostic drug discovery tool.

Despite the relatively common observation of therapeutic efficacy in discovery screens with immortalized cell lines, the vast majority of drug candidates do not reach clinical development. Candidates that do move forward often fail to demonstrate efficacy when progressed from animal models to humans. This dilemma highlights the need for new drug screening technologies that can parse drug candidates early in development with regard to predicted relevance for clinical use. PREDICT96-ALI is a high-throughput organ-on-chip platform incorporating human primary airway epithelial cells in a dynamic tissue microenvironment. Here we demonstrate the utility of PREDICT96-ALI as an antiviral screening tool for SARS-CoV-2, combining the high-throughput functionality of a 96-well plate format in a high containment laboratory with the relevant biology of primary human tissue. PREDICT96-ALI resolved differential efficacy in five antiviral compounds over a range of drug doses. Complementary viral genome quantification and immunofluorescence microscopy readouts achieved high repeatability between devices and replicate plates. Importantly, results from testing the three antiviral drugs currently available to patients (nirmatrelvir, molnupiravir, and remdesivir) tracked with clinical outcomes, demonstrating the value of this technology as a prognostic drug discovery tool.

The COVID-19 pandemic necessitated a rapid mobilization of resources toward the development of safe and efficacious vaccines and therapeutics. Finding effective treatments to stem the wave of infected individuals needing hospitalization and reduce the risk of adverse events was paramount. For scientists and healthcare professionals addressing this challenge, the need to rapidly identify medical countermeasures became urgent, and many compounds in clinical use for other indications were repurposed for COVID-19 clinical trials after preliminary preclinical data demonstrated antiviral activity against SARS-CoV-2. Two repurposed compounds, fluvoxamine and amodiaquine, showed efficacy in reducing SARS-CoV-2 viral loads in preclinical experiments, but ultimately failed in clinical trials, highlighting the need for improved predictive preclinical tools that can be rapidly deployed for events such as pandemic emerging infectious diseases. The PREDICT96-ALI platform is a high-throughput, high-fidelity microphysiological system (MPS) that recapitulates primary human tracheobronchial tissue and supports highly robust and reproducible viral titers of SARS-CoV-2 variants Delta and Omicron. When amodiaquine and fluvoxamine were tested in PREDICT96-ALI, neither compound demonstrated an antiviral response, consistent with clinical outcomes and in contrast with prior reports assessing the efficacy of these compounds in other human cell-based in vitro platforms. These results highlight the unique prognostic capability of the PREDICT96-ALI proximal airway MPS to assess the potential antiviral response of lead compounds.