No abstract

Circulating in China and 158 other countries and areas, the ongoing COVID-19 outbreak has caused devastating mortality and posed a great threat to public health. However, efforts to identify effectively supportive therapeutic drugs and treatments has been hampered by our limited understanding of host immune response for this fatal disease. To characterize the transcriptional signatures of host inflammatory response to SARS-CoV-2 (HCoV-19) infection, we carried out transcriptome sequencing of the RNAs isolated from the bronchoalveolar lavage fluid (BALF) and peripheral blood mononuclear cells (PBMC) specimens of COVID-19 patients. Our results reveal distinct host inflammatory cytokine profiles to SARS-CoV-2 infection in patients, and highlight the association between COVID-19 pathogenesis and excessive cytokine release such as CCL2/MCP-1, CXCL10/IP-10, CCL3/MIP-1A, and CCL4/MIP1B. Furthermore, SARS-CoV-2 induced activation of apoptosis and P53 signalling pathway in lymphocytes may be the cause of patients' lymphopenia. The transcriptome dataset of COVID-19 patients would be a valuable resource for clinical guidance on anti-inflammatory medication and understanding the molecular mechansims of host response.

The outbreak of coronavirus disease 2019 (COVID-19) has resulted in a global pandemic due to the rapid spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). At the time of this manuscript’s publication, remdesivir is the only COVID-19 treatment approved by the United States Food and Drug Administration. However, its effectiveness is still under question due to the results of the large Solidarity Trial conducted by the World Health Organization. Herein, we report that the parent nucleoside of remdesivir, GS-441524, potently inhibits the replication of SARS-CoV-2 in Vero E6 and other cell lines. Challenge studies in both an AAV-hACE2 mouse model of SARS-CoV-2 and in mice infected with murine hepatitis virus, a closely related coronavirus, showed that GS-441524 was highly efficacious in reducing the viral titers in CoV-infected organs without notable toxicity. Our results support that GS-441524 is a promising and inexpensive drug candidate for treating of COVID-19 and other CoV diseases.

Generating answer with natural language sentence is very important in real-world question answering systems, which needs to obtain a right answer as well as a coherent natural response. In this paper, we propose an end-to-end question answering system called COREQA in sequence-to-sequence learning, which incorporates copying and retrieving mechanisms to generate natural answers within an encoder-decoder framework. Specifically, in COREQA, the semantic units (words, phrases and entities) in a natural answer are dynamically predicted from the vocabulary, copied from the given question and/or retrieved from the corresponding knowledge base jointly. Our empirical study on both synthetic and realworld datasets demonstrates the efficiency of COREQA, which is able to generate correct, coherent and natural answers for knowledge inquired questions.

Genotoxic stressors, such as radiation, induce cellular damage that activates pre-programmed repair pathways, some of which involve microRNAs (miRNA) that alter gene expression. The let-7 family of miRNA regulates multiple cellular processes including cell division and DNA repair pathways. However, the role and mechanism underlying regulation of let-7 genes in response to stress have yet to be elucidated. In this study we demonstrate that let-7a and let-7b expression decreases significantly following exposure to agents that induce stress including ionizing radiation. This decrease in expression is dependent on p53 and ATM in vitro and is not observed in a p53−/− colon cancer cell line (HCT116) or ATM−/− human fibroblasts. Chromatin Immunoprecipitation (ChIP) analysis showed p53 binding to a region upstream of the let-7 gene following radiation exposure. Luciferase transient transfections demonstrated that this p53 binding site is necessary for radiation-induced decreases in let-7 expression. A radiation-induced decrease in let-7a and let-7b expression is also observed in radiation-sensitive tissues in vivo and correlates with altered expression of proteins in p53-regulated pro-apoptotic signaling pathways. In contrast, this decreased expression is not observed in p53 knock-out mice suggesting that p53 directly repress let-7 expression. Exogenous expression of let-7a and let-7b increased radiation-induced cytotoxicity in HCT116 p53+/+ cells but not HCT116 p53−/− cells. These results are the first demonstration of a mechanistic connection between the radiation-induced stress response and the regulation of miRNA and radiation-induced cytotoxicity and suggest that this process may be a molecular target for anticancer agents.

SARS-CoV-2 causes the pandemic of COVID-19 and no effective drugs for this disease are available thus far. Due to the high infectivity and pathogenicity of this virus, all studies on the live virus are strictly confined in the biosafety level 3 (BSL3) laboratory but this would hinder the basic research and antiviral drug development of SARS-CoV-2 because the BSL3 facility is not commonly available and the work in the containment is costly and laborious. In this study, we constructed a reverse genetics system of SARS-CoV-2 by assembling the viral cDNA in a bacterial artificial chromosome (BAC) vector with deletion of the spike (S) gene. Transfection of the cDNA into cells results in the production of an RNA replicon that keeps the capability of genome or subgenome replication but is deficient in virion assembly and infection due to the absence of S protein. Therefore, such a replicon system is not infectious and can be used in ordinary biological laboratories. We confirmed the efficient replication of the replicon by demonstrating the expression of the subgenomic RNAs which have similar profiles to the wild-type virus. By mutational analysis of nsp12 and nsp14, we showed that the RNA polymerase, exonuclease, and cap N7 methyltransferase play essential roles in genome replication and sgRNA production. We also created a SARS-CoV-2 replicon carrying a luciferase reporter gene and this system was validated by the inhibition assays with known anti-SARS-CoV-2 inhibitors. Thus, such a one-plasmid system is biosafe and convenient to use, which will benefit both fundamental research and development of antiviral drugs.

Engineered tumor-targeted anthrax lethal toxin proteins have been shown to strongly suppress growth of solid tumors in mice. These toxins work through the native toxin receptors tumor endothelium marker-8 and capillary morphogenesis protein-2 (CMG2), which, in other contexts, have been described as markers of tumor endothelium. We found that neither receptor is required for tumor growth. We further demonstrate that tumor cells, which are resistant to the toxin when grown in vitro, become highly sensitive when implanted in mice. Using a range of tissue-specific loss-of-function and gain-of-function genetic models, we determined that this in vivo toxin sensitivity requires CMG2 expression on host-derived tumor endothelial cells. Notably, engineered toxins were shown to suppress the proliferation of isolated tumor endothelial cells. Finally, we demonstrate that administering an immunosuppressive regimen allows animals to receive multiple toxin dosages and thereby produces a strong and durable antitumor effect. The ability to give repeated doses of toxins, coupled with the specific targeting of tumor endothelial cells, suggests that our strategy should be efficacious for a wide range of solid tumors.anthrax toxin | tumor targeting | angiogenesis | CMG2 | TEM8 R ecognition that aberrant activation of the RAS and PI3K pathways is often the mechanism of human tumorigenesis has inspired development of many small molecule inhibitors of these pathways and has led to improved treatments in certain cancers (1). However, these therapies are effective only in patients having defects in the specific targets of these drugs, and the therapies are rarely curative due to the development of resistance through acquisition of additional oncogenic mutations (2). Therefore, strategies are needed that are effective against a broad spectrum of cancers and that act through features that are not subject to development of resistance. This unmet need has fostered continued interest in strategies that target host-derived tumor vasculature.Anthrax toxin, a major virulence factor of Bacillus anthracis, consists of three individually nontoxic proteins: the cellular binding component, protective antigen (PA), and two enzymatic moieties, lethal factor (LF) and edema factor (EF) (3). PA binds to two host cell-surface integrin-like proteins: tumor endothelium marker-8 (TEM8) [also termed anthrax toxin receptor 1 (ANTXR1)] and capillary morphogenesis protein-2 (CMG2 or ANTXR2) (4, 5). Receptor-bound PA is processed by the ubiquitous cell-associated furin protease to a 63-kDa fragment (PA63), which then forms LFand EF-binding competent PA oligomers. Three or four molecules of LF or EF bind to the PA oligomers, and the complexes are then endocytosed (6-8). The acidic pH within the endosomes causes the PA oligomer to form a pore in the endosomal membrane, allowing translocation of LF or EF into the cytosol of cells to exert their cytotoxic actions (9). Thus, LF plus PA forms lethal toxin and EF plus PA forms edema toxin, with both toxins playing essenti...

By reason of being able to obtain natural language responses, natural answers are more favored in real-world Question Answering (QA) systems. Generative models learn to automatically generate natural answers from large-scale question answer pairs (QA-pairs). However, they are suffering from the uncontrollable and uneven quality of QA-pairs crawled from the Internet. To address this problem, we propose a curriculum learning based framework for natural answer generation (CL-NAG), which is able to take full advantage of the valuable learning data from a noisy and uneven-quality corpora. Specifically, we employ two practical measures to automatically measure the quality (complexity) of QA-pairs. Based on the measurements, CL-NAG firstly utilizes simple and low-quality QApairs to learn a basic model, and then gradually learns to produce better answers with richer contents and more complete syntaxes based on more complex and higher-quality QA-pairs. In this way, all valuable information in the noisy and unevenquality corpora could be fully exploited. Experiments demonstrate that CL-NAG outperforms the state-of-the-art, which increases 6.8% and 8.7% in the accuracy for simple and complex questions, respectively.