UCLA

Background

Many respiratory viruses attack the airway epithelium and cause a wide spectrum of diseases for which we have limited therapies. To date, a few primary human stem cell-based models of the proximal airway have been reported for drug discovery but scaling them up to a higher throughput platform remains a significant challenge. As a result, most of the drug screening assays for respiratory viruses are performed on commercial cell line-based 2D cultures that provide limited translational ability.

Methods

We optimized a primary human stem cell-based mucociliary airway epithelium model of SARS-CoV-2 infection, in 96-well air-liquid-interface (ALI) format, which is amenable to moderate throughput drug screening. We tested the model against SARS-CoV-2 parental strain (Wuhan) and variants Beta, Delta, and Omicron. We applied this model to screen 2100 compounds from targeted drug libraries using a high throughput-high content image-based quantification method.

Results

The model recapitulated the heterogeneity of infection among patients with SARS-CoV-2 parental strain and variants. While there were heterogeneous responses across variants for host factor targeting compounds, the two direct-acting antivirals we tested, Remdesivir and Paxlovid, showed consistent efficacy in reducing infection across all variants and donors. Using the model, we characterized a new antiviral drug effective against both the parental strain and the Omicron variant.

Conclusion

This study demonstrates that the 96-well ALI model of primary human mucociliary epithelium can recapitulate the heterogeneity of infection among different donors and SARS-CoV-2 variants and can be used for moderate throughput screening. Compounds that target host factors showed variability among patients in response to SARS-CoV-2, while direct-acting antivirals were effective against SARS-CoV-2 despite the heterogeneity of patients tested.