Advancing antiviral strategies against emerging RNA viruses by phenotypic drug discovery

Pathogenic RNA viruses can emerge from unexpected sources at unexpected times and cause severe disease in humans, as exemplified by the ongoing coronavirus disease 19 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and Ebola virus (EBOV), Crimean-Congo hemorrhagic fever virus (CCHFV) and Zika virus (ZIKV) outbreaks from the past decade. Despite the increasing impact of emerging viruses to health and economy worldwide, our preparedness to stand against these diseases is hampered by the lack of approved and effective antiviral therapies. Thus, the development of novel antivirals is of urgent need. To date, antiviral drug discovery has primarily focused on targeting specific viral proteins, but these treatments often suffer from viral resistance and are limited to only one or few viruses. Instead, phenotypic drug discovery enables the identification of drug candidates that are active in the disease-relevant model and not restricted to previously characterized biological processes. As RNA viruses are highly dependent on the host cell pathways due to their relatively small genome, targeting virus vulnerabilities within the host cell has been a promising antiviral strategy for broad spectrum antivirals, but is relatively unexplored so far. In fact, phenotypic approaches can additionally identify host-directed antivirals due to the unbiased nature. The focus of this doctoral thesis was to identify novel antiviral compounds with broad spectrum activity and investigate the compound mechanism of action and target pathways from the host cell and virus perspective. To achieve these goals, multiple cutting-edge phenotype-based methodologies were implemented that additionally advanced the antiviral drug discovery landscape. In Paper I, we developed an image-based phenotypic antiviral assay and screened our in-house chemical library targeting cellular oxidative stress and nucleotide metabolism pathways in Hazara virus (HAZV)-infected cells. Screening hit compounds TH3289 and TH6744 activity was validated by their therapeutic window and both compounds were also active beyond HAZV, especially TH3289 that was tested and displayed activity against EBOV, CCHFV, SARS-CoV-2 and a common cold coronavirus 229E (CoV-229E). We also excluded the intended target 8-oxoguanine DNA glycosylase (OGG1) protein to be responsible for TH6744 antiviral activity and characterized host cell chaperone and co-chaperone network as target pathways of TH6744 by implementing thermal proteome profiling methodology. In Paper II, we transferred our image-based phenotypic assay to ZIKV-infected brain cells in order to screen structural analogs of TH3289 and TH6744 against a pathogenic RNA virus. TH3289 and TH6744 again appeared among the screening hits and presented a promising therapeutic window in various cellular models, further confirming their broad activity. Moreover, TH6744 reduced ZIKV infection and progeny release in cerebral organoid model and impressively rescued ZIKV-induced cytotoxicity in organoids. Additionally, treatment with TH6744 rapidly diminished ZIKV progeny release during late replication cycle stages, elucidating the antiviral mechanism of action. In Paper III, we established an untargeted morphological profiling method to provide in-depth host cell responses during antiviral screening. We combined the Cell Painting protocol with antibody-based virus detection in a single assay followed by automated image analysis pipeline providing segmentation and classification of infected cells and extraction of cell morphological features. We demonstrated how our assay reliably distinguished CoV-229E infected human lung fibroblasts from non-infected controls based on cellular morphological features. Furthermore, our method can be applied in phenotypic drug screening as validated by nine host- and virus-targeting antivirals. Effective antivirals Remdesivir and E-64d treatment reversed the infection-specific signatures in host cells. Thereby, the developed method can be implemented for antiviral phenotypic drug discovery by morphological profiling of drug candidates