A TLR-pathway kinase-inhibitor screen identifies PPM18 as a spontaneous stress-granule inducer with SG-independent...

Stress granules (SG) are cytoplasmic membraneless compartments that regulate cellular stress responses and have been implicated in antiviral defense against Influenza A Virus (IAV). SG stabilization has been reported to be a viable strategy to develop antiviral drug. To expand the repertoire of SG-modulating compounds, we performed a targeted pharmacological screen focusing on inhibitors of toll-like receptor (TLR) signaling pathway, based on our previous work showing that TLR signaling can destabilize SGs. Using the synthetic dsRNA analog polyinosinic:polycytidylic acid (poly(I:C)) to induce SGs in A549 human lung cancer cell line, we screened a panel of TLR signaling pathway inhibitors for their effect on SG stability. We developed a robust image analysis pipeline for high-throughput analysis of SG alterations. The screen identified PPM18 as an inducer of SGs even in the absence of poly(I:C) transfection. Using knockout cell lines our data shows that PPM18 induced SG assembly was dependent on Protein Kinase R-like Endoplasmic Reticulum Kinase, partially on heme regulated inhibitor, and that SG formation is substantially reduced in G3BP1^-/- cells. PPM18 reduced the abundance of IAV RNA and viral ribonucleoproteins (vRNPs) in a dose-dependent manner, with a lowest effective concentration of 12.5 µM. Notably, this antiviral effect was independent of SG formation, as PPM18 robustly reduced IAV RNA and vRNP abundances in G3BP1^-/- cells. PPM18 also induced caspase-independent regulated cell death at 50 µM, as the pan-caspase inhibitor zVAD-fmk did not protect cells. Although the narrow therapeutic window of PPM18 limits its direct translational potential, our study identifies a tool compound for interrogating the relationship between SGs and antiviral activity. More broadly, our image-based screen provides a feasible approach for identifying SG-modulating compounds, although the finding that PPM18's antiviral activity is SG-independent indicates that SG modulation alone may not be sufficient to predict antiviral efficacy.