Machine learning for cross-scale microscopy of viruses.

Despite advances in virological sciences and antiviral research, viruses continue to emerge, circulate, and threaten public health. We still lack a comprehensive understanding of how cells and individuals remain susceptible to infectious agents. This deficiency is in part due to the complexity of viruses, including the cell states controlling virus-host interactions. Microscopy samples distinct cellular infection stages in a multi-parametric, time-resolved manner at molecular resolution and is increasingly enhanced by machine learning and deep learning. Here we discuss how state-of-the-art artificial intelligence (AI) augments light and electron microscopy and advances virological research of cells. We describe current procedures for image denoising, object segmentation, tracking, classification, and super-resolution and showcase examples of how AI has improved the acquisition and analyses of microscopy data. The power of AI-enhanced microscopy will continue to help unravel virus infection mechanisms, develop antiviral agents, and improve viral vectors.

Label-free microscopy for virus infections.

Microscopy has been essential to elucidate micro- and nano-scale processes in space and time and has provided insights into cell and organismic functions. It is widely employed in cell biology, microbiology, physiology, clinical sciences and virology. While label-dependent microscopy, such as fluorescence microscopy, provides molecular specificity, it has remained difficult to multiplex in live samples. In contrast, label-free microscopy reports on overall features of the specimen at minimal perturbation. Here, we discuss modalities of label-free imaging at the molecular, cellular and tissue levels, including transmitted light microscopy, quantitative phase imaging, cryogenic electron microscopy or tomography and atomic force microscopy. We highlight how label-free microscopy is used to probe the structural organization and mechanical properties of viruses, including virus particles and infected cells across a wide range of spatial scales. We discuss the working principles of imaging procedures and analyses and showcase how they open new avenues in virology. Finally, we discuss orthogonal approaches that enhance and complement label-free microscopy techniques.

High-content, arrayed compound screens with rhinovirus, influenza A virus and herpes simplex virus infections.

Viruses are genetically and structurally diverse, and outnumber cells by orders of magnitude. They can cause acute and chronic infections, suppress, or exacerbate immunity, or dysregulate survival and growth of cells. To identify chemical agents with pro- or antiviral effects we conducted arrayed high-content image-based multi-cycle infection screens of 1,280 mainly FDA-approved compounds with three human viruses, rhinovirus (RV), influenza A virus (IAV), and herpes simplex virus (HSV) differing in genome organization, composition, presence of an envelope, and tropism. Based on Z'-factors assessing screening quality and Z-scores ranking individual compounds, we identified potent inhibitors and enhancers of infection: the RNA mutagen 5-Azacytidine against RV-A16; the broad-spectrum antimycotic drug Clotrimazole inhibiting IAV-WSN; the chemotherapeutic agent Raltitrexed blocking HSV-1; and Clobetasol enhancing HSV-1. Remarkably, the topical antiseptic compound Aminacrine, which is clinically used against bacterial and fungal agents, inhibited all three viruses. Our data underscore the versatility and potency of image-based, full cycle virus propagation assays in cell-based screenings for antiviral agents.

Methylene blue, Mycophenolic acid, Posaconazole, and Niclosamide inhibit SARS-CoV-2 Omicron variant BA.1 infection of human airway epithelial organoids.

Sublineages of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) Omicron variants continue to amass mutations in the spike (S) glycoprotein, which leads to immune evasion and rapid spread of the virus across the human population. Here we demonstrate the susceptibility of the Omicron variant BA.1 (B.1.1.529.1) to four repurposable drugs, Methylene blue (MB), Mycophenolic acid (MPA), Posaconazole (POS), and Niclosamide (Niclo) in post-exposure treatments of primary human airway cell cultures. MB, MPA, POS, and Niclo are known to block infection of human nasal and bronchial airway epithelial explant cultures (HAEEC) with the Wuhan strain, and four variants of concern (VoC), Alpha (B.1.1.7), Beta (B.1.351), Gamma (B.1.1.28), Delta (B.1.617.2) (Weiss et al., 2021, Murer et al., 2022). Our results here not only reinforce the broad anti-coronavirus effects of MB, MPA, POS and Niclo, but also demonstrate that the Omicron variant BA.1 (B.1.1.529.1) sheds infectious virus from HAEEC over at least 15 d, and maintains both intracellular and extracellular viral genomic RNA without overt toxicity, suggesting viral persistence. The data emphasize the potential of repurposable drugs against COVID-19.

Identification of broad anti-coronavirus chemical agents for repurposing against SARS-CoV-2 and variants of concern.

Endemic human coronaviruses (hCoVs) 229E and OC43 cause respiratory disease with recurrent infections, while severe acute respiratory syndrome (SARS)-CoV-2 spreads across the world with impact on health and societies. Here, we report an image-based multicycle infection procedure with α-coronavirus hCoV-229E-eGFP in an arrayed chemical library screen of 5440 clinical and preclinical compounds. Toxicity counter selection and challenge with the ß-coronaviruses OC43 and SARS-CoV-2 in tissue culture and human airway epithelial explant cultures (HAEEC) identified four FDA-approved compounds with oral availability. Methylene blue (MB, used for the treatment of methemoglobinemia), Mycophenolic acid (MPA, used in organ transplantation) and the anti-fungal agent Posaconazole (POS) had the broadest anti-CoV spectrum. They inhibited the shedding of SARS-CoV-2 and variants-of-concern (alpha, beta, gamma, delta) from HAEEC in either pre- or post exposure regimens at clinically relevant concentrations. Co-treatment of cultured cells with MB and the FDA-approved SARS-CoV-2 RNA-polymerase inhibitor Remdesivir reduced the effective anti-viral concentrations of MB by 2-fold, and Remdesivir by 4 to 10-fold, indicated by BLISS independence synergy modelling. Neither MB, nor MPA, nor POS affected the cell delivery of SARS-CoV-2 or OC43 (+)sense RNA, but blocked subsequent viral RNA accumulation in cells. Unlike Remdesivir, MB, MPA or POS did not reduce the release of viral RNA in post exposure regimen, thus indicating infection inhibition at a post-replicating step as well. In summary, the data emphasize the power of unbiased, full cycle compound screens to identify and repurpose broadly acting drugs against coronaviruses.

Microscopy deep learning predicts virus infections and reveals mechanics of lytic-infected cells.

Imaging across scales reveals disease mechanisms in organisms, tissues, and cells. Yet, particular infection phenotypes, such as virus-induced cell lysis, have remained difficult to study. Here, we developed imaging modalities and deep learning procedures to identify herpesvirus and adenovirus (AdV) infected cells without virus-specific stainings. Fluorescence microscopy of vital DNA-dyes and live-cell imaging revealed learnable virus-specific nuclear patterns transferable to related viruses of the same family. Deep learning predicted two major AdV infection outcomes, non-lytic (nonspreading) and lytic (spreading) infections, up to about 20 hr prior to cell lysis. Using these predictive algorithms, lytic and non-lytic nuclei had the same levels of green fluorescent protein (GFP)-tagged virion proteins but lytic nuclei enriched the virion proteins faster, and collapsed more extensively upon laser-rupture than non-lytic nuclei, revealing impaired mechanical properties of lytic nuclei. Our algorithms may be used to infer infection phenotypes of emerging viruses, enhance single cell biology, and facilitate differential diagnosis of non-lytic and lytic infections.

A high-content image-based drug screen of clinical compounds against cell transmission of adenovirus.

Human adenoviruses (HAdVs) are fatal to immuno-suppressed individuals, but no effective anti-HAdV therapy is available. Here, we present a novel image-based high-throughput screening (HTS) platform, which scores the full viral replication cycle from virus entry to dissemination of progeny and second-round infections. We analysed 1,280 small molecular weight compounds of the Prestwick Chemical Library (PCL) for interference with HAdV-C2 infection in a quadruplicate, blinded format, and performed robust image analyses and hit filtering. We present the entire set of the screening data including all images, image analyses and data processing pipelines. The data are made available at the Image Data Resource (IDR, idr0081). Our screen identified Nelfinavir mesylate as an inhibitor of HAdV-C2 multi-round plaque formation, but not single round infection. Nelfinavir has been FDA-approved for anti-retroviral therapy in humans. Our results underscore the power of image-based full cycle infection assays in identifying viral inhibitors with clinical potential.

The FDA-Approved Drug Nelfinavir Inhibits Lytic Cell-Free but Not Cell-Associated Nonlytic Transmission of Human Adenovirus.

Adenoviruses (AdVs) are prevalent and give rise to chronic and recurrent disease. Human AdV (HAdV) species B and C, such as HAdV-C2, -C5, and -B14, cause respiratory disease and constitute a health threat for immunocompromised individuals. HAdV-Cs are well known for lysing cells owing to the E3 CR1-ß-encoded adenovirus death protein (ADP). We previously reported a high-throughput image-based screening framework and identified an inhibitor of HAdV-C2 multiround infection, nelfinavir mesylate. Nelfinavir is the active ingredient of Viracept, an FDA-approved inhibitor of human immunodeficiency virus (HIV) aspartyl protease that is used to treat AIDS. It is not effective against single-round HAdV infections. Here, we show that nelfinavir inhibits lytic cell-free transmission of HAdV, indicated by the suppression of comet-shaped infection foci in cell culture. Comet-shaped foci occur upon convection-based transmission of cell-free viral particles from an infected cell to neighboring uninfected cells. HAdV lacking ADP was insensitive to nelfinavir but gave rise to comet-shaped foci, indicating that ADP enhances but is not required for cell lysis. This was supported by the notion that HAdV-B14 and -B14p1 lacking ADP were highly sensitive to nelfinavir, although HAdV-A31, -B3, -B7, -B11, -B16, -B21, -D8, -D30, and -D37 were less sensitive. Conspicuously, nelfinavir uncovered slow-growing round HAdV-C2 foci, independent of neutralizing antibodies in the medium, indicative of nonlytic cell-to-cell transmission. Our study demonstrates the repurposing potential of nelfinavir with postexposure efficacy against different HAdVs and describes an alternative nonlytic cell-to-cell transmission mode of HAdV.

Label-Free Digital Holo-tomographic Microscopy Reveals Virus-Induced Cytopathic Effects in Live Cells.

Cytopathic effects (CPEs) are a hallmark of infections. CPEs are difficult to observe due to phototoxicity from classical light microscopy. We report distinct patterns of virus infections in live cells using digital holo-tomographic microscopy (DHTM). DHTM is label-free and records the phase shift of low-energy light passing through the specimen on a transparent surface with minimal perturbation. DHTM measures the refractive index (RI) and computes the refractive index gradient (RIG), unveiling optical heterogeneity in cells. We find that vaccinia virus (VACV), herpes simplex virus (HSV), and rhinovirus (RV) infections progressively and distinctly increased RIG. VACV infection, but not HSV and RV infections, induced oscillations of cell volume, while all three viruses altered cytoplasmic membrane dynamics and induced apoptotic features akin to those caused by the chemical compound staurosporine. In sum, we introduce DHTM for quantitative label-free microscopy in infection research and uncover virus type-specific changes and CPE in living cells with minimal interference.IMPORTANCE This study introduces label-free digital holo-tomographic microscopy (DHTM) and refractive index gradient (RIG) measurements of live, virus-infected cells. We use DHTM to describe virus type-specific cytopathic effects, including cyclic volume changes of vaccinia virus infections, and cytoplasmic condensations in herpesvirus and rhinovirus infections, distinct from apoptotic cells. This work shows for the first time that DHTM is suitable to observe virus-infected cells and distinguishes virus type-specific signatures under noninvasive conditions. It provides a basis for future studies, where correlative fluorescence microscopy of cell and virus structures annotate distinct RIG values derived from DHTM.

Concepts in Light Microscopy of Viruses.

Viruses threaten humans, livestock, and plants, and are difficult to combat. Imaging of viruses by light microscopy is key to uncover the nature of known and emerging viruses in the quest for finding new ways to treat viral disease and deepening the understanding of virus–host interactions. Here, we provide an overview of recent technology for imaging cells and viruses by light microscopy, in particular fluorescence microscopy in static and live-cell modes. The review lays out guidelines for how novel fluorescent chemical probes and proteins can be used in light microscopy to illuminate cells, and how they can be used to study virus infections. We discuss advantages and opportunities of confocal and multi-photon microscopy, selective plane illumination microscopy, and super-resolution microscopy. We emphasize the prevalent concepts in image processing and data analyses, and provide an outlook into label-free digital holographic microscopy for virus research.

Plaque2.0-A High-Throughput Analysis Framework to Score Virus-Cell Transmission and Clonal Cell Expansion.

Classical plaque assay measures the propagation of infectious agents across a monolayer of cells. It is dependent on cell lysis, and limited by user-specific settings and low throughput. Here, we developed Plaque2.0, a broadly applicable, fluorescence microscopy-based high-throughput method to mine patho-biological clonal cell features. Plaque2.0 is an open source framework to extract information from chemically fixed cells by immuno-histochemistry or RNA in situ hybridization, or from live cells expressing GFP transgene. Multi-parametric measurements include infection density, intensity, area, shape or location information at single plaque or population levels. Plaque2.0 distinguishes lytic and non-lytic spread of a variety of DNA and RNA viruses, including vaccinia virus, adenovirus and rhinovirus, and can be used to visualize simultaneous plaque formation from co-infecting viruses. Plaque2.0 also analyzes clonal growth of cancer cells, which is relevant for cell migration and metastatic invasion studies. Plaque2.0 is suitable to quantitatively analyze virus infections, vector properties, or cancer cell phenotypes.

Chemical induction of unfolded protein response enhances cancer cell killing through lytic virus infection.

UNLABELLED: Cancer cells are susceptible to oncolytic viruses, albeit variably. Human adenoviruses (HAdVs) are widely used oncolytic agents that have been engineered to produce progeny within the tumor and elicit bystander effects. We searched for host factors enhancing bystander effects and conducted a targeted RNA interference screen against guanine nucleotide exchange factors (GEFs) of small GTPases. We show that the unfolded protein response (UPR), which is readily inducible in aggressive tumor cells, enhances melanoma or epithelial cancer cell killing upon HAdV infection. UPR was triggered by knockdown of Golgi-specific brefeldin A-resistant guanine nucleotide exchange factor 1 (GBF-1) or the GBF-1 inhibitor golgicide A (GCA) and stimulated HAdV infection. GBF-1 is a GEF for ADP ribosylation factors (Arfs) regulating endoplasmic reticulum (ER)-to-Golgi apparatus and intra-Golgi apparatus membrane transport. Cells treated with GCA enhanced HAdV-induced cytopathic effects in epithelial and melanoma cancer cells but not normal cells, if the drug was applied several hours prior to HAdV inoculation. This was shown by real-time label-free impedance measurements using the xCELLigence system. GCA-treated cells contained fewer incoming HAdVs than control cells, but GCA treatment boosted HAdV titers and spreading in cancer cells. GCA enhanced viral gene expression or transgene expression from the cytomegalovirus promoter of B- or C-species HAdVs but did not enhance viral early region 1A (E1A) expression in uninfected cell lines or cells transfected with plasmid reporter DNA. The UPR-enhanced cell killing required the nuclease activity of the UPR sensor inositol-requiring enzyme 1 (IRE-1) and X box binding protein 1 (XBP-1), which alleviate ER stress. The collective results show that chemical UPR induction and viruses boost tumor cell killing by enhancing oncolytic viral efficacy. IMPORTANCE: Cancer is difficult to combat. A wide range of oncolytic viruses show promise for killing cancer cells, yet the efficacy of oncolytic killing is low. We searched for host factors enhancing adenovirus cancer cell killing and found that the knockdown of Golgi-specific brefeldin A-resistant guanine nucleotide exchange factor 1 (GBF-1) or chemical inhibition of GBF-1 enhanced adenovirus infection by triggering the IRE-1/XBP-1 branch of the unfolded protein response (UPR). IRE-1/XBP-1 promote cell survival and enhanced the levels of the adenoviral immediate early gene product E1A, virus spreading, and killing of cancer cells. Aggressive tumor cells depend on a readily inducible UPR and, hence, present prime targets for a combined strategy involving adenoviruses and small chemicals inducing UPR.

Tracking viral genomes in host cells at single-molecule resolution.

Viral DNA trafficking in cells has large impacts on physiology and disease development. Current methods lack the resolution and accuracy to visualize and quantify viral DNA trafficking at single-molecule resolution. We developed a noninvasive protocol for accurate quantification of viral DNA-genome (vDNA) trafficking in single cells. Ethynyl-modified nucleosides were used to metabolically label newly synthesized adenovirus, herpes virus, and vaccinia virus vDNA, without affecting infectivity. Superresolution microscopy and copper(I)-catalyzed azide-alkyne cycloaddition (click) reactions allowed visualization of infection at single vDNA resolution within mammalian cells. Analysis of adenovirus infection revealed that a large pool of capsid-free vDNA accumulated in the cytosol upon virus uncoating, indicating that nuclear import of incoming vDNA is a bottleneck. The method described here is applicable for the entire replication cycle of DNA viruses and offers opportunities to localize cellular and viral effector machineries on newly replicated viral DNA, or innate immune sensors on cytoplasmic viral DNA.

Isotherm of ligand adsorption on DNA at multiplicative noise.

Fluctuations of the number of ligands adsorbed on macromolecules are considered in the case when the number of ligands in the solution fluctuates under the action of fluctuations of the external medium (external noise). For the case of small filling, the multiplicative type of stochastic differential equation is obtained, describing the time variation of the number of ligands adsorbed on macromolecules. The isotherm of adsorption of ligands on DNA is obtained. It is shown that at small ligand concentrations, for some relations between adsorption parameters and the intensity of the external noise, no macromolecule adsorption of ligands takes place.