To identify potential therapeutic stop-gaps for SARS-CoV-2, we evaluated a library of 1,670 approved and reference compounds in an unbiased, cellular image-based screen for their ability to suppress the broad impacts of the SARS-CoV-2 virus on phenomic profiles of human renal cortical epithelial cells using deep learning. In our assay, remdesivir is the only antiviral tested with strong efficacy, neither chloroquine nor hydroxychloroquine have any beneficial effect in this human cell model, and a small number of compounds not currently being pursued clinically for SARS-CoV-2 have efficacy. We observed weak but beneficial class effects of -blockers, mTOR/PI3K inhibitors and Vitamin D analogues and a mild amplification of the viral phenotype with -agonists.

Development of accurate disease models and discovery of immune-modulating drugs is challenged by the immune system's highly interconnected and context-dependent nature. Here we apply deep-learning-driven analysis of cellular morphology to develop a scalable 'phenomics' platform and demonstrate its ability to identify dose-dependent, high-dimensional relationships among and between immunomodulators, toxins, pathogens, genetic perturbations, and small and large molecules at scale. High-throughput screening on this platform demonstrates rapid identification and triage of hits for TGF-β- and TNF-α-driven phenotypes. We deploy the platform to develop phenotypic models of active SARS-CoV-2 infection and of COVID-19-associated cytokine storm, surfacing compounds with demonstrated clinical benefit and identifying several new candidates for drug repurposing. The presented library of images, deep learning features, and compound screening data from immune profiling and COVID-19 screens serves as a deep resource for immune biology and cellular-model drug discovery with immediate impact on the COVID-19 pandemic.

Most DNA viruses replicate in the cell nucleus, although the specific sites of virion assembly are as yet poorly defined. Electron microscopy on freeze-substituted, plastic-embedded sections of murine polyomavirus (PyV)-infected 3T3 mouse fibroblasts or mouse embryonic fibroblasts (MEFs) revealed tubular structures in the nucleus adjacent to clusters of assembled virions, with virions apparently “shed” or “budding” from their ends. Promyelocytic leukemia nuclear bodies (PML-NBs) have been suggested as possible sites for viral replication of polyomaviruses (BKV and SV40), herpes simplex virus (HSV), and adenovirus (Ad). Immunohistochemistry and FISH demonstrated co-localization of the viral T-antigen (Tag), PyV DNA, and the host DNA repair protein MRE11, adjacent to the PML-NBs. In PML −/− MEFs the co-localization of MRE11, Tag, and PyV DNA remained unchanged, suggesting that the PML protein itself was not responsible for their association. Furthermore, PyV-infected PML −/− MEFs and PML −/− mice replicated wild-type levels of infectious virus. Therefore, although the PML protein may identify sites of PyV replication, neither the observed “virus factories” nor virus assembly were dependent on PML. The ultrastructure of the tubes suggests a new model for the encapsidation of small DNA viruses.

Nuclear replication of DNA viruses activates DNA damage repair (DDR) pathways, which are thought to detect and inhibit viral replication. However, many DNA viruses also depend on these pathways in order to optimally replicate their genomes. We investigated the relationship between murine polyomavirus (MuPyV) and components of DDR signaling pathways including CHK1, CHK2, H2AX, ATR, and DNAPK. We found that recruitment and retention of DDR proteins at viral replication centers was independent of H2AX, as well as the viral small and middle T-antigens. Additionally, infectious virus production required ATR kinase activity, but was independent of CHK1, CHK2, or DNAPK signaling. ATR inhibition did not reduce the total amount of viral DNA accumulated, but affected the amount of virus produced, indicating a defect in virus assembly. These results suggest that MuPyV may utilize a subset of DDR proteins or non-canonical DDR signaling pathways in order to efficiently replicate and assemble.

Protrusions are plasma membrane extensions that are found in almost every cell in the human body. Cancer cell filopodial and lamellipodial protrusions play key roles in the integral processes of cell motility and signaling underlying tumor invasion and metastasis. HER2 (ErbB-2) is overexpressed in diverse types of tumors and regulates PI3K-pathway-mediated protrusion growth. It is known that HER2 resides at breast cancer cell protrusions, but how protrusion-based HER2 spatiotemporal dynamics shape cancer signaling is unclear. Here, we study how HER2 location and motion regulate protrusion signaling and growth using quantitative spatio-temporal molecular imaging approaches. Our data highlight morphologically-segregated features of filopodial and lamellipodial protrusions, in in vitro 2D breast cancer cells and in vivo intact breast tumor. Functionalsegregation parallels morphological-segregation, as HER2 and its activated downstream pAKT-PI3K signaling remain spatiallylocalized at protrusions, provoking new protrusion growth proximal to sites of HER2 activation. HER2 in SKBR3 breast cancer cell filopodia exhibits fast, linearly-directed motion that is distinct from lamellipodia and non-protrusion subcellular regions (~3-4 times greater diffusion constant, rapid speeds of 2-3 um 2 /s). Surprisingly, filopodial HER2 motion is passive, requiring no active energy sources. Moreover, while HER2 motion in lamellipodia and non-protrusion regions show hindered diffusion typical of membrane proteins, HER2 diffuses freely within filopodia. We conclude that HER2 activation, propagation, and functional protrusion growth is a local process in which filopodia have evolved to exploit Brownian thermal fluctuations within a barrier-free nanostructure to transduce rapid signaling. These results support the importance of developing filopodia and other protrusion-targeted strategies for cancer.