key: cord-339012-4juhmjaj
authors: Hou, Wei; Liu, Fei; van der Poel, Wim H.M.; Hulst, Marcel M.
title: Rapid host response to an infection with Coronavirus. Study of transcriptional responses with Porcine Epidemic Diarrhea Virus
date: 2020-07-28
journal: bioRxiv
DOI: 10.1101/2020.07.28.224576
sha: 
doc_id: 339012
cord_uid: 4juhmjaj

The transcriptional response in Vero cells (ATCC® CCL-81) infected with the coronavirus Porcine Epidemic Diarrhea Virus (PEDV) was measured by RNAseq analysis 4 and 6 hours after infection. Differential expressed genes (DEGs) in PEDV infected cells were compared to DEGs responding in Vero cells infected with Mammalian Orthoreovirus (MRV). Functional analysis of MRV and PEDV DEGs showed that MRV increased the expression level of several cytokines and chemokines (e.g. IL6, CXCL10, IL1A, CXCL8 [alias IL8]) and antiviral genes (e.g. IFI44, IFIT1, MX1, OASL), whereas for PEDV no enhanced expression was observed for these “hallmark” antiviral and immune effector genes. Pathway and Gene Ontology “enrichment analysis” revealed that PEDV infection did not stimulate expression of genes able to activate an acquired immune response, whereas MRV did so within 6h. Instead, PEDV down-regulated the expression of a set of zinc finger proteins with putative antiviral activity and enhanced the expression of the transmembrane serine protease gene TMPRSS13 (alias MSPL) to support its own infection by virus-cell membrane fusion (Shi et al, 2017, Viruses, 9(5):114). PEDV also down-regulated expression of Ectodysplasin A, a cytokine of the TNF-family able to activate the canonical NFKB-pathway responsible for transcription of inflammatory genes like IL1B, TNF, CXCL8 and PTGS2. The only 2 cytokine genes found up-regulated by PEDV were Cardiotrophin-1, an IL6-type cytokine with pleiotropic functions on different tissues and types of cells, and Endothelin 2, a neuroactive peptide with vasoconstrictive properties. Furthermore, by comprehensive datamining in biological and chemical databases and consulting related literature we identified sets of PEDV-response genes with potential to influence i) the metabolism of biogenic amines (e.g. histamine), ii) the formation of cilia and “synaptic clefts” between cells, iii) epithelial mucus production, iv) platelets activation, and v) physiological processes in the body regulated by androgenic hormones (like blood pressure, salt/water balance and energy homeostasis). The information in this study describing a “very early” response of epithelial cells to an infection with a coronavirus may provide pharmacologists, immunological and medical specialists additional insights in the underlying mechanisms of coronavirus associated severe clinical symptoms including those induced by SARS-CoV-2. This may help them to fine-tune therapeutic treatments and apply specific approved drugs to treat COVID-19 patients.

The lack of knowledge for treating hospitalized SARS-CoV-2 infected patients is one of the pressing problems of the current COVID-19 pandemic. The SARS-CoV-2 virus shows a close genetic similarity to the in April 2003 identified SARS virus (SARS-CoV-1) and to other SARS-related coronaviruses isolated from humans and bats. SARS-CoV-2 induces clinical respiratory symptoms familiar to the 2003 virus, mostly in persons with underlying diseases like COPD, heart failure, diabetes and obesity (1: Wu et al. 2020) . Despite the 2003 SARS-CoV-1 virus has been extensively studied in the last two decades, there are no vaccines available yet, neither there are effective prophylactic and therapeutic treatment regimens with drugs that work equally well for each individual patient with SARS-induced respiratory problems. Such treatments might prevent development of severe disease patterns like "acute respiratory distress syndrome" (ARDS) and other, often fatal complications, and may decrease the case-fatality rate of SARS-CoV-2 infections. In our lab we study the alpha-coronavirus PEDV. PEDV was first detected in pig herds in 1977 in Europe (2: Pensaert and de Bouck 1978) . However, this virus reemerged in the spring of 2013 in North America causing a massive outbreak among pig herds, resulting in the death of about 30% of the suckling piglets due to severe diarrhea and dehydration ( Although several studies concluded that these clinical symptoms were caused by MRV itself, in concordance with the co-existence of MRV3 in PEDV infected piglets, also other MRV serotypes were isolated from hospitalized patients with airway problems diagnosed positive for SARS-CoV-1 (11: Cheng et al. 2009 , 12: Duan et al. 2003 , 13: Zuo et al. 2003 . Recently, a cross-family recombinant coronavirus was isolated in China from bat faeces in which an RNA sequence originating from the S1 segment of MRV was inserted in the coronavirus genome between the N and Ns7a genes, indicating that both viruses were replicating simultaneously in a single cell in bats (14: Huang et al. 2016) . A prevalence study showed that this cross-family recombinant coronavirus circulated in an isolated bat colony in a cave in China (15: Obameso et al. 2017 ). This cooccurrence of MRV with coronaviruses raised the questions whether a synergistic effect between both viruses exists and if such coexistence plays a role in viral pathogenesis. Therefore we studied the host response in cultured cells early (4 and 6 hours) after PEDV and MRV infection using RNAseq. Our original goal was to identify early factors and processes induced by PEDV or MRV that could stimulate or influence the replication and pathogenesis of the other virus. The host, tissue and cell tropism of PEDV differs from SARS-CoV-1 and -2. However, the genomic organization, replication strategy and function of a part of the viral nonstructural proteins share common features among all coronaviruses (16: Brian and Baric 2005) . This applies particularly for interactions in infected cells of nonstructural coronavirus proteins with specific host proteins. Host proteins that are recruited or silenced to support virus replication, assembly and release. In our experiment we used Vero cells (Cercopithecus aethiops epithelial kidney cell line; ATCC® CCL-81) because these cells support efficient infection and replication of both MRV and PEDV. Vero cells are susceptible for many coronaviruses, including SARS-CoV-1 and -2 (17: Chu et al. 2020). They originate from epithelial tissue, in part resembling nasal and bronchial epithelium cells, the prime target cells infected by SARS-CoV-2 in the airways of humans. Recent research showed that SARS-CoV-2 is also able to replicate in epithelial cells of human small intestinal organoids (18: Lamers et al. 2020) . A disadvantage of Vero cells is a deletion in the type I interferon (IFN) gene cluster on chromosome 12 (19: Osada et al. 2014 ). Therefore, these cells lack expression of type I IFNs important for activation of antiviral defense mechanisms. However, research has shown that Vero cells by-pass this IFNactivation route and could mount an antiviral response mediated by interferon regulatory factor 3 (20: Chew et al. 2009 ).

Single infections with PEDV or MRV3 alone and simultaneous (double) infections of Vero cells with both viruses were performed using a maximum multiplicity of infection (MOI) to achieve a synchronized infection of all cells. By RNAseq measured expression levels of mRNA transcripts/genes in infected cells were compared to RNAseq profiles measured from similar treated mock-infected cells harvested at the same time point after infection. The detected sets of differential expressed genes (DEGs) for PEDV and MRV were analyzed by gene set enrichment analysis (GSEA) using functional bioinformatic programs to retrieve biological processes (pathways and Gene Ontology terms [GO-term]) and associations with chemical compounds, including drugs. In addition, we searched the literature for functional information of the PEDV-DEGs to find possible associations with SARS-CoV-2 pathogenesis. Because of the COVID-2 pandemic we gave priority to publish the results of this functional bioinformatical analysis and datamining for the single infected Vero cells with PEDV separate from the results of the double infections with MRV3. In this report we focused on the "very early" host response of epithelial cells to an infection with the coronavirus PEDV and pay less attention to the role of specific viral proteins in this host response to PEDV. In part our results were in agreement with results of a previous RNAseq study comparing SARS-CoV-2 and Influenza host responses by RNAseq (21: Blanco-Melo et al. 2020). But we also found associations with biological processes, and pivotal genes/proteins acting in these processes, that had not been recognized before. This information may contribute to the search for novel or alternative preventive or therapeutic drugs and treatment protocols for this devastating COVID-19 disease.

A time-dependent infection experiment was performed with cultured Vero cells. Details are described in supplementary file 1 (material and methods) and visually displayed in this file. Briefly, overnight cultured Vero cells grown in 2 cm 2 wells were mock-infected, infected with MRV3 strain WBVR (7: Hulst et al. 2017) or PEDV strain CV777 (2: Pensaert and de Bouck 1978, 22: Rasmussen et al. 2018] ) with a multiplicity of infection of ≥1 for 30 min at 4°C. For PEDV and corresponding mockinfected cells, 10 µg/ml of trypsin in serum-free medium was used to facilitate infection of Vero cells during the whole experiment. All virus and mock-infected timepoints were performed in quadruplicate. After incubation for 30 min at 4°C, virus was discarded and cells were washed twice and supplied with fresh culture medium. Cells were incubated for 0, 2, 4, 6, and 16h at 37°C and 5% CO2. After incubation for the indicated times, cells were placed on ice before total RNA was isolated from three of the quadruplicate wells. The replication of both viruses in Vero cells was monitored using virus-specific RT-qPCR tests ( Fig.1 : methods and primers used for PCR are provided in supplementary file 1). In addition, cells in one of the quadruplicate wells incubated for 16h were fixated and stained with antibodies directed against the S2 spike protein of PEDV and the S1 attachment protein (α1) of MRV3. A decrease in CT-values for PEDV was not observed before 6 h post inoculation (6 h.p.i), indicating that replication in PEDV infected cells started later than was observed for MRV (at 4 h.p.i). Staining of the cells after 16h indicated that nearly all Vero cells were infected with MRV3 and more than 50% with PEDV. Also more than 50% of the cells in 16hwells appeared as fused cells (syncytia), confirming that more than 50% of the cells were infected with PEDV. Quality control of the total RNA isolated from infected cells using an Agilent Bioanalyzer showed that RNAs isolated from PEDV infected wells at 16 h.p.i. were partially degraded (RIN values below 9), making them unsuitable for RNAseq analysis. Therefore, only 0, 4 and 6h timepoints were analyzed using RNAseq. stained with a monoclonal antibody directed against the S2 spike protein. MRV and mock infected cells were stained with a polyclonal rabbit serum raised against a peptide sequence of the S1-attachment protein of MRV serotype 3. Nuclei were stained blue with the Hoechst, 4',6-diamidino-2-phenylindole dye.

Equal amounts of total RNA isolated from triplicate wells were pooled and subjected to RNAseq analysis by GenomeScan B.V.(Leiden, The Netherlands) using Next Generation Sequencing (NGS) (see supplementary file 2a for details). Mapping of NGS reads to the Cercopithecus aethiops reference genome and preparation of datafiles with calculated Fold Change (FC) of expression levels of mapped mRNAs, were performed for each comparison at 0, 4, and 6h by GenomeScan (see supplementary file 2b). From these datafiles we extracted lists of DEGs with a FC>2 and p-value of <0.05. After accessing the NCBI, Panther or KEGG databases for human orthologs, not annotated Cercopithecus aethiops DEGs were annotated with an HUGO official gene symbols (http://www.genenames.org). In supplementary file 3 sheet "PEDV-MRV DEGs FC>2", lists of all annotated PEDV and MRV DEGs are presented with their FC. In a separate sheet "PEDV-DEGs functional info" all 266 individual DEGs regulated by PEDV at 4 and 6 h.p.i. are presented with their FC, information about their function and the types of human cells in which expression of the gene is relatively high compared to other human cells (retrieved from the "Primary Cell Atlas" dataset of BioGPS: http://biogps.org/). Note that all tables in these Excel sheets of supplementary file 3 are sortable using the headers. In all results paragraphs beneath information about the biological function of DEGs was retrieved by consulting the "GeneCards" (Weizmann Institute of Science: https://www.genecards.org/) and NCBI Gene reports (Entrez Gene: https://www.ncbi.nlm.nih.gov/gene/), and literature linked to these reports (for references about these biological functions of genes/proteins we refer to publications cited in these reports: "GeneCards" weblinks are provided in supplementary file 3).

Sets of PEDV and MRV DEGs were analyzed using the GSEA program GeneAnalytics (LifeMap Sciences, Inc.) and pathways (for MRV and PEDV), GO-terms (not for MRV), and associations with compounds/drugs (not for MRV) with a high or medium score (p-value <0.05) were retrieved and listed in 3 separate sheets in supplementary file 3 (sheets "MRV-PEDV pathways", "PEDV G0-terms" and "PEDV Compounds"). Similar and related pathways retrieved for both PEDV and MRV, and remarkable PEDV pathways, GO-terms and compound associations are summarized in Table 1 . For PEDV all DEGs within these pathways are provided with their regulation, up (green) or down (red). For MRV only DEGs in common with PEDV-DEGs were listed in Table 1 (see sheet "MRV-PEDV pathways" in supplementary file 3 for all MRV-DEGs acting in these pathways). Subsets of PEDV-DEGs were selected matching the terms "chemokines-cytokines", "antiviral" , and terms related to the pathogenesis of COVID-2 (explained below) using the genotyping program VarElect (LifeMap Sciences, Inc.) and displayed in supplementary file 3 in separate sheets: "chemokines-cytokines", "(anti)-viral", etc. Based on these selections we prepared a set of PEDV KEY-DEGs consisting of genes regulated with a FC of >10 (up) or <-10 (down) or playing an important role in biological processes induced by PEDV and related to COVID-19 pathology. In beneath results sections we tried to give as much as possible meaningful information about the function of KEY-DEGs for which we found an association with SARS-CoV-2 infections. We emphasize that further dedicated experimental and in-silico research is necessary to confirm the involvement of the proteins encoded by these genes for pathogenesis of this viral disease. Table 1 . Enriched pathways, GO-terms and compound associations of PEDV-DEGs.

*PEDV enriched pathways (A), GO-terms (B), and associations with compounds and drugs (C) with a high and medium score and with at least 2 matching genes were retrieved from GeneAnalytics. Common pathways for MRV were included in Table 1A . A full list of pathways with DEGs, retrieved for MRV at 4 and 6h, is provided in supplementary file 3 (sheet MRV-PEDV pathways). A possible function or process related to specific DEGs, pathways, GO-term, or compounds/drugs is provided in blue text between brackets. $ Official Gene-symbols (HUGO abbreviations) are listed for DEGs. Down-regulated DEGs were colored red and up-regulated DEGs were colored green. In section A the number of DEGs regulated by MRV in a pathway and the common DEGs are provided between brackets. DEGs regulated by both PEDV and MRV are underlined.

Compared to MRV, only a few genes involved in "cytokine/chemokine signaling" were regulated at 4 and 6h by PEDV. In Fig. 2A the regulation of cytokines/chemokines in PEDV and MRV infected Vero cells are displayed. This indicated that MRV increased the transcription of a broad set of cytokines/chemokines, including interferonmediated cytokines like CXCL10 and CXCL8 (alias IL8), already at 4 h.p.i., whereas PEDV did not, even not when replication of PEDV RNA was detected by RT-qPCR at 6 h.p.i.. For MRV, this cytokine/chemokine response at 4 h.p.i. was followed by high up-regulation of "hallmark" antiviral genes at 6 h.p.i. (see Fig. 2B : e.g. interferoninduced genes [IFI] and OASL) and chemokines that attract T cells, monocytes, granulocytes, including basophils (e.g. CXCL8, CXCL11 and CCL2). PEDV infection up-regulated only a few genes coding for proteins with cytokine activity (CTF1 and EDN2), and also did not elevated gene expression of these "hallmark" antiviral genes. In contrast, PEDV down-regulated expression of 6 genes (out of 9 in total) coding for Zinc Finger Proteins (out of 9 in total), all 6 with an antiviral activity towards Herpes simplex virus 1 ( Fig 2C) Binding of thrombin to F2RL2 reduces inflammation, activates platelets and increases vasodilation and permeability of the vascular wall (see also below in the section "platelets activation"). CSF1R is a receptor for the cytokine colony stimulating factor 1, a cytokine that regulates differentiation and function of macrophages, and in the CNS, the density and distribution of microglia cells. The BLNK gene codes for a cytosolic protein that passes on B-cell receptor signals in the signaling cascade that activates B-cell development and function. Gene expression of genes coding for essential components of this B-cell signaling, like "spleen associated tyrosine kinase" (SYK) and "LYN proto-oncogene"(LYN) were not regulated by PEDV, nor by MRV.

Genes involved in amino acid, protein translation and metabolism of immuno-active compounds.

PEDV DEGs coding for enzymes involved in the metabolism of the non-essential amino acids histidine, phenylalanine, tryptophan and proline were found enriched in the PEDV dataset (see supplementary file 3, sheet PEDV-compounds). Remarkable were the DEGs coding for amine oxidases involved in the catabolism of the biogenic amines histamine, tryptamine and phenylethylamine, their derivates and related substrates/products of these enzymes (Fig. 3, AOC1 , MAOA, IL4I1). None of these amino oxidase genes were regulated by MRV. Using the genotyping program VarElect, PEDV-DEGs with an association with these biogenic amines were retrieved (supplementary file 3, sheet "Biogenic amines"). Three enzymes clustered in the "Histidine metabolism" pathway (https://www.kegg.jp/keggbin/show_pathway?hsa00340+4128) with histamine and reaction products generated from this biogenic amine (Fig. 3) . Also most association of PEDV-DEGs were found by VarElect for histamine. The gene coding for the amine oxidase "Interleukin 4 Induced 1" (IL4I1) was strongly down-regulated (30-fold) 4h after infection with PEDV. Besides catalysis of L-Phenylaniline into Phenylpyruvate (Fig.3) , IL4I1 also fulfills an important role in signaling in "synaptic clefts" formed between antigen presenting cells (APC) and T cells (so-called "immune cleft": see also below) ( Expression of the Prostaglandin-Endoperoxide Synthase 2 (PTGS2) gene was upregulated by MRV at 6h p.i. PTGS2 synthesizes prostaglandin endoperoxide H2 (PGH2), an compound with a short half-life and the precursor of many biological active prostaglandins: e.g. Thromboxane-A2 (mediates activation of platelets), PGI2 and PGE2. In contrast, PEDV increased the expression of the gene coding for Prostaglandin E Synthase (PTGES) which converts PGH2 into PGE2. PGE2 is a direct vasodilator, but does not inhibit platelet aggregation. PGE2 also suppresses T cell receptor signaling. PEDV decreased expression of the Gamma-Glutamyltransferase 1 gene (GGT1) after 4h (2-fold), but increased expression of this gene 2 hours later to a 4-fold level compared to mock infected cells. GGT synthesizes Leukotriene D4 (LTD4) from LTC4. IgE-activated mast cells may secrete LTD4 and LTC4, together with histamine and platelets activating factor (PAF). This vesicle mediated secretion by mast cells (degranulation) results in stimulation of mucus production, and similar to histamine, increases the permeability and smooth muscle contraction of the vascular wall. In persons suffering from asthma this degranulation leads to an immediate allergic response (bronchospasm, airflow obstruction and forming of edema).

Genes involved in "Cilia and Synaptic cleft" formation.

GSEA detected "Axon Guidance" as the GO-term with the highest score for PEDV (see Table 1 ). In addition, PEDV-DEGs were enriched coding for proteins involved in calcium ion-dependent exocytosis from vesicles into the "synaptic clefts" between two cells (e.g. between axons and dendrites), and DEGs coding for proteins involved in formation of cilia. Cilia protruding from cells are found in many forms. They can have a static (structural) function, e.g. in forming of clefts between two cells (see Fig. 4 ), or a motile function. Motile cilia on the surface of ciliated cells lining up the epithelial layers in the nose, trachea and bronchia sweep out superfluous mucus containing dirt from the airways. PEDV DEGs matching the terms "Cilia" and "Synaptic cleft" retrieved form the genotyping program VarElect were further examined by consulting functional information in the NCBI Gene and GeneCards reports in order to evaluate their association with these processes (see supplementary file 3, sheet "Cilia and Synaptic cleft"). Based on this analysis we identified genes in the set of PEDV DEGs which can i) negatively regulate cell adhesion (RND1 and SEMA5A), ii) inhibit formation of cilia (kinases MAK1 and CDK20, highly up-regulated at 6h), and iii) regulate cytoskeleton rearrangements that facilitate axon growth and growth and stabilization of dendritic spines (F2RL2, Regulation of genes involved in histamine/biogenic amines (see above) and formation of cilia/clefts suggested that gene expression related to this intersynaptic signaling between immune cells could be affected in response to infection with PEDV (Fig 4) . In particular, the highly down-regulated gene IL4I1 (30-fold at 4h p.i.) is of interest (see also above). IL4I1 is believed to be secreted from APC's (e.g. DC's) in the immune cleft formed with T cells (31: Molinier-Frenkel et al. 2019). The mechanism how IL4I1 transmits its signal to T cells is not completely understood. It could bind to a receptor that concentrates this amino oxidase in the cleft, resulting in elevated H2O2 and ammonia production, phenylalanine depletion and phenylpyruvate production in the cleft space. These alteration in the concentration of these chemicals in the cleft are sensed by the T cell. The paralog of IL4I1, amino oxidase MAOA (down-regulated 11-fold by PEDV) could also play a similar role in this signaling process. Remarkable was also the strong down-regulation of genes coding for the Olfactory Receptor Family 2 Subfamily A Member 14 (OR2A14; 17-fold at 4h) and Anoctamin 2 (ANO2, alias CaCC;14-fold at 4h). ANO2 is a calcium-activated chloride channel imbedded in the basal membranes of neurons that harbor apical membrane receptors like OR2A14 that sense odorants. By importing chloride ions into the cytosol ANO2 contributes to the depolarization of these neurons (https://www.kegg.jp/kegg-bin/show_pathway?hsa04740+57101). Loss of smell and taste is one of the first noticeable symptoms of COVID-19. Genetic defects in the ANO2 gene are associated with Von Willebrand disease, a bleeding disorder due to defective platelet aggregation ( . Also disease incidence in adult males is significantly higher than in females of the same age.

To assess whether PEDV-DEGs relate to these pathological symptoms, the genotyping program VarElect was used to identify genes matching the terms "ARDS", "Cardiomyopathy", "Obesity (Diabetic)", and "Platelets activation". Detailed information about all matching DEGs is provided in supplementary file 3 in separate sheets for all 4 queried terms. DEGs matching to more than one query term are displayed in Fig. 5 . Remarkable associations of DEGs with these terms are mentioned in sections beneath. . EDN2 and AGT2 are vasoactive peptides and binding of EDN2 and AGT2 to their receptors on granular cells of the juxtaglomerular apparatus in the kidney raises free calcium levels in the cytosol, leading to inhibition of cAMP-mediated secretion of the aspartylprotease renin (REN), the key regulator of renin-angiotensin-aldosterone system (RAAS) (https://www.kegg.jp/kegg-bin/show_pathway?hsa04924+1907). REN converts pre-angiotensinogen (AGT) to the endocrine peptide-hormone AGT1

(https://www.kegg.jp/kegg-bin/show_pathway?hsa04614+5972). AGT1 is further cleaved to variants with specific endocrine activity by Angiotensin I Converting Enzymes (e.g. ACE and ACE2: ). On the surface of bronchial epithelial cells ACE2 was identified as entry receptor for SARS-CoV-1 and -2 (43: Hoffmann et al. 2020). The octamer peptide AGT2 stimulates secretion of the mineralocorticoid hormone aldosterone by the adrenal glands. Aldosterone, and the AGT and EDN peptide hormones regulate an array of physiological processes in the body, e.g. vascular smooth muscle contraction, blood pressure, fluid and electrolyte homeostasis (44: Agapitov et al. 2002) . All processes that are important for proper functioning of the vascular system, heart muscles and kidneys. CTF1 is directly involved in the pathology of numerous cardiovascular diseases by promoting cardiac myocyte hypertrophy (41: Wollert et al. 1996) , which may lead to the onset of heart-diseases like "hypertrophic cardiomyopathy" or "dilated cardiomyopathy", and eventually, to (lethal) heart failure. PEDV induced a strong down-or up-regulation of several other genes directly involved in the function of cardiomyocytes. Sodium voltage-gated channel subunit 4 (SCN4B) was strongly upregulated (15-fold) and MYLK2 (see above), Citrate Synthase (CS; down-regulation 24-fold) and Ankyrin Repeat Domain 1 (ANKRD1) were strongly down-regulated. Down-regulation of CS may reduce oxidative capacity in cardiomyocytes. Gene expression of ANKRD1 was down-regulated 12-fold in response to MRV infection at 4 h.p.i, but reverted to a 24-fold up-regulation 2h later. ANKRD1 is a putative transcription factor involved in regulation of gene expression in hypertrophic myocytes (https://www.wikipathways.org/index.php/Pathway:WP516) Regulation of cytosolic calcium levels in the cytosol of cardiomyocytes, e.g. by binding of COL1A1 and COL2A (up-regulated by PEDV, see above) to integrin subunit alpha on the surface of cardiomyocytes or after import of calcium ions mediated by calcium voltage-gated channels (e.g. by CACNA1H; down-regulated 5-fold at 4h by PEDV) may also trigger myocyte hypertrophy. PEDV strongly down-regulated gene expression of a potassium voltage-gated channel (KCNQ4;16-fold). This in contrast to a strong up-regulation of the sodium symporter SCN4B. For the potassium channel CACNA1H (alias Kv7.4) it was reported that this channel regulates the membrane potential and Ca 2+ permeability of mitochondria located in the vicinity the sarcoplasmic reticulum in rat cardiomyocytes (45: Testai et al. 2016). All three above mentioned ion channels are also involved in the process of excitation, contraction/relaxation and repolarization of cardiac myocytes. MRV also downregulated gene expression 3 to 4-fold for the potassium (KCNQ4) and calcium channel CACNA1H, but did not increased expression of the sodium symporter gene SCN4B or orthologs of this gene.

Chronic hypertension and heart disease/failure is a complication frequently observed in obese/diabetic patients. In accordance with this, 16 out of the 65 PEDV DEGs matching the term "Obesity" also matched with the term "Cardiomyopathy" (see Additional remarkable PEDV-DEGs.

Highly up-or down-regulated PEDV-DEGs not mentioned in the text, and to our opinion interesting with regard to coronavirus infection, are briefly described in Table  2 . Among these DEGs several genes coding for transcription factors and genes transcribed in antisense RNA's that inhibit translation of their coding counterparts.

For more functional information about these DEGs we refer to the weblinks provided in supplementary file 3, sheet "PEDV KEY DEGs". Table 2 . Remarkable PEDV-DEGS not mentioned in the text.

In this report we measured the transcriptional response of Vero cells shortly after infection with the coronavirus PEDV. The function of the responding host genes and the biological processes in which they act were studied in detail by us to find plausible relations to COVID-19 pathology. Because of the differences in genomic organization and expression of viral proteins between SARS-CoV-2 and PEDV, we paid less attention to couple the response of specific host genes to the function of specific coronavirus proteins. We were able to infect the majority of Vero cells (>50%) with PEDV and MRV synchronically. This resulted in a unique set of highly up-and down-regulated DEGs for PEDV. Not more than 14% of the 266 PEDV-DEGs (n=37) were similar to MRV-DEGS (total of 727 MRV-DEGs). In contrast to MRV, we observed no typical response of antiviral genes and related cytokine/chemokine genes in Vero cells within 6 h.p.i. for PEDV. For MRV these processes started already before 4 h.p.i.. We have to notice that PEDV replication started 2h later than MRV3 replication, which could in part be the reason for not detecting transcriptional regulation of specific cytokine, chemokine and antiviral genes for PEDV. Longer incubation times than 6h were not planned in the original design of our experiments and would have resulted in a set of PEDV-DEGs dominated by genes involved in syncytia forming and apoptotic/necrotic cell death. Nevertheless, at 6h replication of PEDV RNA was detected by RT-PCR, indicating that dsRNA was present in the cells and could have be sensed by cytosolic pattern recognition receptors of the RIG-I-like family to initiate an antiviral and related cytokine/chemokine response. Similar as observed in another study with PEDV and Vero cells, and in analogy with SARS-CoV-1 and -2, we observed a high up-regulation of the transmembrane serine protease gene (TMPRSS13) that acts as a co-factor in the infection process of cells ( We observed a reduced expression of EEF1A1, as part of a transcription factor-complex that binds and activates the promotor of IFNγ, and of the cytokine EDA and its receptor (EDARADD) involved in activation of canonical-NFKB transcription of antiviral cytokine-chemokine genes like CXCL8 (alias IL8) an CXCL10. This reduced expression of EEF1A1 and EDA and its receptor may play a crucial role in delaying or downgrading an IFN-mediated antiviral and cytokine/chemokine response in our Vero cell system. The elevated transcription of many cytokine and chemokine genes in Vero cells by MRV suggests that replication of this RNA virus in epithelial cells induces secretion of these immune effectors (more information about the genes/processes that responded to MRV infection will be published elsewhere). PEDV, and also the MRV3 strain we used both replicate in enterocytes lined up in the intestinal mucosal layer.

In the intestinal and bronchial epithelial layer, microfold (M) cells are imbedded between these lined up epithelial cells. M-cells sense and engulf foreign pathogens/antigens from the lumen to present them to residing immune cells.

According to pathway analysis, most T cell related immune genes regulated in response to MRV infection were part of the pathways "T-helper 17 (Th17) differentiation/activation" and "IL17 signaling". Antigen presentation by M cells to Th17 cells in the submucosal layers stimulate secretion of different types of IL17 cytokines (IL17A-D, IL25 and Il17F) resulting in activation of different types of innate immune cells and T cells, including Th1/Th2 cells. Dysregulation of this pathogen-induced IL17 response may disturb the balance between Th1/Th2-cell mediated immune responses, resulting in excessive inflammation, damage to the epithelial layer and on the long term, to autoimmune reactions. TF RORC (or specific isoforms of this TF, see above) plays a pivotal in controlling IL17 expression and secretion by Th17 cells. PEDV strongly up-regulated gene expression of RORC in Vero cells whereas MRV down-regulated expression of this TF. This difference in regulation of TF RORC suggests that virus-induced activation or suppression of IL17 secretion by Th17 cells in submucosal layers of airway epithelium can be an important mechanism to dysregulate the activation of T cell responses. Therefore, TF RORC might be a potential target for drug treatment/development. Drugs affecting expression of RORC, like the fluorinated steroid "Dexamethasone" and the synthetic tetracycline derivative "Doxycycline" (http://ctdbase.org/basicQuery.go?bqCat=gene&bq=RAR+related+orphan+receptor +C) are already under investigation in relation to SARS-CoV-2 pathogenesis. Transcriptional regulation of a set of genes coding for enzymes involved in biogenic amine metabolism was unique for PEDV, and not observed for MRV. Most associations of these PEDV-DEGs were found with histamine, a compound produced by mast cells and basophils, and released by these cells in response to allergens and pathogen-induced inflammation. The 10-fold up-regulation of the enzyme AOC1 suggests that histamine is converted to imidazole-acetaldehyde (see Fig. 3 ).

However, without data of intra-and extracellular concentrations of the chemicals this remains speculative. Recent reports indicated that submucosal mast cells in the lungs were triggered by SARS-CoV-2 infection to release pro-inflammatory cytokines (IL1, IL6 and TNF-alpha) . MDSCs infiltrate these tumors and inflamed tissues to suppress the local activity of specific immune cells. Therefore, the role of infiltrating MDSCs at inflammatory sites in the lungs of COVID-19 patients, as part of an SARS-CoV-2 immune-evading strategy, and the role of IL4I1 in this process, is worthwhile to investigate in more detail. Expression of genes that promote or inhibit the formation and motility function of cilia were time-dependently regulated by PEDV. The 20-fold up-regulation of FEZ1 gene expression at 4h (20-fold) descended within 2h to a moderate 6-fold upregulation. This descend occurred simultaneously with elevation of gene expression of the kinases MAK and CDK20, both involved in inhibition of cilia formation. Because PEDV efficiently replicates in enterocytes that carry ciliated membrane protrusions (microvilli) on their luminal surface, regulation of these genes could be related to structural changes in the cytoskeleton of cells imposed by virus replication (e.g. syncytia forming in PEDV infected Vero cultures). Likewise, SARS-CoV-2 replication could also impose structural changes in cilia protruding from the surface of upperairway epithelial cells (nose, trachea) and bronchi. Based on our results we cannot pinpoint a specific process in which these cilia-regulating genes act. Possible processes can either be formation of an immune cleft, a virological cleft to promote more effective infection of neighboring cells, or cytoskeleton rearrangements to support virus replication, morphogenesis and budding from cells. Interestingly, two recent studies revealed a high level of expression of the SARS-CoV-2 entry receptor ACE2 and its co-receptor TMPRSS2 in ciliated airway epithelial cells ( . These processes deserve more attention, and may also be considered as possible target-processes for interference with drugs.

Within our set of PEDV-DEGs, and the biological processes deduced from this set, we found associations with diverse aspects of COVID-19 pathogenesis, i.e. "ARDS", "Cardiomyopathy", "Obesity (Diabetic)", and "Platelets activation". However, it is unknown whether the proteins encoded by these PEDV-DEGS indeed play a role in the biological processes underlying the symptoms and complications observed in hospitalized persons infected with SARS-CoV-2. Nevertheless, a part of these genes/processes may be starting points for further dedicated research. Research to fine tune drug treatment protocols that are already applied for COVID-19 patients, or research that provides new insights for treatments with alternative prophylactic and therapeutic approved drugs. In Table 3 . we summarized the PEDV-DEGs that are, to our opinion, of interest for modulation of the biological processes underlying the pathogenesis of COVID-19. Table 3 . Possible target genes for COVID-19 therapy.

Colophon.

The overwhelming amount of data published recently made it impossible for us to oversee all (novel) facts about the SARS-CoV-2 virus and pathology of the COVID-19 disease. Some important genes and related processes imbedded in our set of PEDV-DEGs may have been overlooked by us. Therefore, we encourage researchers, especially medical, immunological and pharmaceutical specialists, to study this set of DEGs in detail. The users-friendly supplementary file 3 with functional information about DEGs and related biological processes can be down-loaded from the web. By publishing these PEDV data ahead of our complete study, we hope that some of the gene targets and cognate processes we have identified for the coronavirus PEDV will contribute to a better understanding how hospitalized COVID-19 patients can be treated and cured. A more condensed version of this manuscript, focusing on the original goal of our study, will be submitted to a peer-reviewed virological journal soon. 

Risk Factors Associated With Acute Respiratory Distress Syndrome and Death in Patients With Coronavirus Disease

A new coronavirus-like particle associated with diarrhea in swine

Origin, evolution, and genotyping of emergent porcine epidemic diarrhea virus strains in the United States

Porcine epidemic diarrhea virus infection: Etiology, epidemiology, pathogenesis and immunoprophylaxis

A novel pathogenic Mammalian orthoreovirus from diarrheic pigs and Swine blood meal in the United States

Identification of a Novel Reassortant of a Mammalian Orthoreovirus in Faeces of diarrheic pigs in the Netherlands: presentation at the 11th annual meeting of EPIZONE in Paris

High similarity of novel orthoreovirus detected in a child hospitalized with acute gastroenteritis to mammalian orthoreoviruses found in bats in Europe

Pathogenesis of reovirus infections of the central nervous system

Identification and characterization of a new orthoreovirus from patients with acute respiratory infections

A novel reovirus isolated from a patient with acute respiratory disease

ReoV isolated from SARS patients

Cloning and identification of reovirus isolated from specimens of SARS patients

A Bat-Derived Putative Cross-Family Recombinant Coronavirus with a Reovirus Gene

The persistent prevalence and evolution of crossfamily recombinant coronavirus GCCDC1 among a bat population: a two-year followup

Coronavirus genome structure and replication

Comparative tropism, replication kinetics, and cell damage profiling of SARS-CoV-2 and SARS-CoV with implications for clinical manifestations, transmissibility, and laboratory studies of COVID-19: an observational study

SARS-CoV-2 productively infects human gut enterocytes

The genome landscape of the african green monkey kidney-derived vero cell line

Characterization of the interferon regulatory factor 3-mediated antiviral response in a cell line deficient for IFN production

Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19

Full-length genome sequences of porcine epidemic diarrhoea virus strain CV777; Use of NGS to analyse genomic and sub-genomic RNAs

TMPRSS2 and MSPL Facilitate Trypsin-Independent Porcine Epidemic Diarrhea Virus Replication in Vero Cells

DESC1 and MSPL activate influenza A viruses and emerging coronaviruses for host cell entry

Efficient activation of the severe acute respiratory syndrome coronavirus spike protein by the transmembrane protease TMPRSS2

A novel isoform of the orphan receptor RORγt suppresses IL-17 production in human T cells

Structure, function and evolution of the hemagglutinin-esterase proteins of corona-and toroviruses

Functional screen reveals SARS coronavirus nonstructural protein nsp14 as a novel cap N7 methyltransferase

2'-O methylation of the viral mRNA cap evades host restriction by IFIT family members

Human IL4I1 is a secreted Lphenylalanine oxidase expressed by mature dendritic cells that inhibits T-lymphocyte proliferation

The IL4I1 Enzyme: A New Player in the Immunosuppressive Tumor Microenvironment

Avoiding the void: cell-to-cell spread of human viruses

A common 253-kb deletion involving VWF and TMEM16B in German and Italian patients with severe von Willebrand disease type 3

Pathological findings of COVID-19 associated with acute respiratory distress syndrome

Case-Fatality Rate and Characteristics of Patients Dying in Relation to COVID-19 in Italy

The Tn antigen-structural simplicity and biological complexity

Post-translational modifications of coronavirus proteins: roles and function

Emerging WuHan (COVID-19) coronavirus: glycan shield and structure prediction of spike glycoprotein and its interaction with human CD26

Therapies for bleomycin induced lung fibrosis through regulation of TGF-beta1 induced collagen gene expression

COVID-19-induced acute respiratory failure: an exacerbation of organspecific autoimmunity? medRxiv 2020.04

Cardiotrophin-1 activates a distinct form of cardiac muscle cell hypertrophy. Assembly of sarcomeric units in series VIA gp130/leukemia inhibitory factor receptor-dependent pathways

A985G polymorphism of the endothelin-2 gene and atrial fibrillation in patients with hypertrophic cardiomyopathy

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Role of endothelin in cardiovascular disease

Expression and function of Kv7.4 channels in rat cardiac mitochondria: possible targets for cardioprotection

Human coronary arteriolar dilation to adrenomedullin: role of nitric oxide and K(+) channels

Adrenomedullin improves cardiac function and prevents renal damage in streptozotocin-induced diabetic rats

Adrenomedullin: a possible autocrine or paracrine inhibitor of hypertrophy of cardiomyocytes

Histones and heart failure in diabetes

Histone H4 lysine 16 hypoacetylation is associated with defective DNA repair and premature senescence in Zmpste24-deficient mice

Studies of the SIM1 gene in relation to human obesity and obesity-related traits

Neuropeptide B mediates female sexual receptivity in medaka fish, acting in a female-specific but reversible manner

Porcine epidemic diarrhea virus inhibits dsRNA-induced interferon-β production in porcine intestinal epithelial cells by blockade of the RIG-Imediated pathway

Mast cells contribute to coronavirus-induced inflammation: new anti-inflammatory strategy

Mast cell stabilisers, leukotriene antagonists and antihistamines: A rapid review of effectiveness in COVID-19

Lnc-C/EBPβ Modulates Differentiation of MDSCs Through Downregulating IL4i1 With C/EBPβ LIP and WDR5

Identification of discrete tumor-induced myeloid-derived suppressor cell subpopulations with distinct T cellsuppressive activity

SARS-CoV-2 entry factors are highly expressed in nasal epithelial cells together with innate immune genes

SARS-CoV-2 receptor ACE2 and TMPRSS2 are primarily expressed in bronchial transient secretory cells

Mechanisms of innate immune evasion in re-emerging RNA viruses

Supplementary file 3: Interactive Excel file with sortable tables in separate sheets. Please, first read the sheet "read me" for an explanation and instructions for the use of the tables. Excel sheets contain tables with i) PEDV and MRV DEGs extracted from RNAseq data files, ii) functional information about the PEDV-DEGs, iii) GSEA extracted pathways (for MRV and PEDV), GO-terms (only for PEDV) and compound associations (only for PEDV), and iv) associations of PEDV-DEGs with the terms "cytokines-chemokines", "(anti)-viral", "Biogenic amines", "cilia and synaptic clefts", and the disorders "ARDS, "Cardiomyopathy", "Obesity", and "Platelets activation" (A-C-O-P).

The authors declare that they have no competing interests.Authors' contributions.