Betacoronaviruses SARS-CoV-2 and HCoV-OC43 infections in IGROV-1 cell line require aryl hydrocarbon receptor

Abstract

The emergence of novel betacoronaviruses has posed significant financial and human health burdens, necessitating the development of appropriate tools to combat future outbreaks. In this study, we have characterized a human cell line, IGROV-1, as a robust tool to detect, propagate, and titrate betacoronaviruses SARS-CoV-2 and HCoV-OC43. IGROV-1 cells can be used for serological assays, antiviral drug testing, and isolating SARS-CoV-2 variants from patient samples. Using time-course transcriptomics, we confirmed that IGROV-1 cells exhibit a robust innate immune response upon SARS-CoV-2 infection, recapitulating the response previously observed in primary human nasal epithelial cells. We performed genome-wide CRISPR knockout genetic screens in IGROV-1 cells and identified Aryl hydrocarbon receptor (AHR) as a critical host dependency factor for both SARS-CoV-2 and HCoV-OC43. Using DiMNF, a small molecule inhibitor of AHR, we observed that the drug selectively inhibits HCoV-OC43 infection but not SARS-CoV-2. Transcriptomic analysis in primary normal human bronchial epithelial cells revealed that DiMNF blocks HCoV-OC43 infection via basal activation of innate immune responses. Our findings highlight the potential of IGROV-1 cells as a valuable diagnostic and research tool to combat betacoronavirus diseases.

Keywords: AHR; COVID-19; DiMNF; Genome-scale CRISPR screening; HCoV-OC43; IGROV-1; SARS-CoV-2; betacoronavirus.

Figure 1.

Identification and characterization of IGROV-1 as a susceptible and permissive cell line for SARS-CoV-2 and HCoV-OC43. (A) ACE2 mRNA and protein expression levels in CCLE database (scatter plot), as well as TMPRSS2 protein expression (table). Normalized expression values are used for the plot with cutoffs at log₂ expression value of 1. (B) ACE2 and TMPRSS2 protein levels in IGROV-1, Vero E6, 293T, Huh7 and Huh7.5.1 cell lines. GAPDH levels were used as internal controls. (C) Renilla luciferase signal measurement in IGROV-1, Vero E6, 293FT, and Huh7 cells, after infection with viruses pseudotyped with SARS-CoV-2 S, SARS-CoV S and VSV G proteins. Data are represented as mean ± SEM (n = 2). (D) SARS-CoV-2 replication kinetics in IGROV-1, Huh7 and Vero E6 cells over 72 h, quantified by plaque assay. Data are represented as mean ± SEM (n = 3). (E) Representative plaque assay plates using IGROV-1 or Vero E6 cells to measure SARS-CoV-2 viral titre. Both plates were subjected to the same set of virus inoculums, with 10 to 10,000 fold dilution. (F) SARS-CoV-2 viral inhibition in IGROV-1 cells, upon treatment with a gradient concentration of chloroquine diphosphate or remdesivir. Viral titres were quantified by plaque assay. Data are represented as mean ± SEM (n = 4). (G) HCoV-OC43 propagation kinetics in IGROV-1 cells. Three different MOIs (0.1, 1, 10) were used and genome copies were quantified by real time qPCR. Data are represented as mean ± SEM (n = 3).

Figure 2.

Application of IGROV-1 for serological neutralization assays. (A) Determining the neutralization potency of eight commercial SARS-CoV-2 monoclonal antibodies to inhibit viral infection in IGROV-1 and Vero E6 cells, measured by PRNT₅₀ assay. Monoclonal antibody cocktail #1: AR6949 and AR6959. Cocktail #2: 9A9C9, AR6949 and AR6959. (B) Paired PRNT₅₀ values for convalescent sera tested on IGROV-1 and Vero E6 cells. Correlation is determined by Pearson method.

Figure 3.

Transcriptional profiling of SARS-CoV-2 infection in IGROV-1 cells. (A) Differentially expressed genes (light blue) and innate immunity genes (red) in IGROV-1 cells at 24, 48 and 72 hpi, compared to 0 (no virus control). (B) Key antiviral pathways as functionally annotated using Gene Set Enrichment Analysis (GSEA). Size of bubble refers to proportion of genes from gene set lying at leading edge. Colours indicate Normalised Enrichment Scores (NES). Only significant pathways are plotted. (C) Heatmap of log normalised counts from all three cell types and both time-points, clustered by genes and libraries. Genes are selected as follows: top 200 differentially expressed genes by fold change for each condition (cell type, hpi) are selected, and a union of gene lists across all conditions is taken, totalling 935 genes. All conditions are performed in triplicates (n = 3).

Figure 4.

Genome-scale CRISPR screens identified AHR as a common host factor for SARS-CoV-2 and HCoV-OC43. (A) Schematic diagram of CRISPR-Cas9 KO screen workflow. (B) Overlapped MAGeCK RRA scores of SARS-CoV-2 (x axis) and HCoV-OC43 screens (y axis). The values depicted are – Log₁₀ of negative selection scores of the gene summary output files. (C) SARS-CoV-2 viral titre in IGROV-1 WT, AHR KO, and AHR KO with cDNA complementation. KO cells were transduced with a lentivirus carrying GFP protein to control for unspecific lentivirus transduction effects. Data are represented as mean ± SEM (n = 3). (D) HCoV-OC43 viral titre in IGROV-1 WT, AHR KO, and AHR KO with cDNA complementation (left), as well as nucleocapsid (N) protein accumulation following infection (right). KO cells were transduced with a lentivirus carrying GFP protein to control for unspecific lentivirus transduction effects. TCID₅₀ data are represented as mean ± SEM (n = 3). GAPDH levels were used as internal controls for the western blots. LOD denotes the “limit of detection.” N.D. denotes “not detected.”

Figure 5.

AHR modulator, DiMNF, inhibited HCoV-OC43 infection but not SARS-CoV-2. (A and B) HCoV-OC43 viral titre inhibition in response to DiMNF treatment in (A) IGROV-1 and (B) NHBE cells (n = 3). Cell cytotoxicity of DiMNF is measured independently from the infection experiments, via CellTiter-Glo assay (n = 4). Data are represented as mean ± SEM. N.D. denotes “not detected.” (C) SARS-CoV-2 viral titre could not be inhibited in response to DiMNF treatment of IGROV-1 cells. Data are represented as mean ± SEM (n = 3). LOD denotes the “limit of detection.”

Figure 6.

DiMNF treatment and HCoV-OC43 infection impact on normal human bronchial epithelial cells. (A) Proportion of all obtained reads getting aligned to human genome and HCoV-OC43 genome, determined by STAR aligner (n = 3). (B) Clustering of cells based on their overall gene expression count profiles, determined by MDS first two dimensions. Replicates from each condition are highlighted with different colours and are unsupervisedly clustered together. HCoV-OC43 infected cells treated with DiMNF are clustered closer to non-infected conditions. Values in parenthesis show the percentage of total variation captured in each of the first two dimensions. (C) Volcano plots of gene expression changes upon DiMNF treatment and HCoV-OC43 infection, individually. Upregulated genes are determined by Log₂(fold-change) > 1 and Log₁₀(FDR)<−1. (D) Biological processes gene ontology (BP) enrichment in the upregulated genes upon DiMNF treatment (left) and HCoV-OC43 infection (right). Over-representation analysis is done via clusterProfiler package in R, with multiple testing p-value adjustment by Benjamini-Hochberg method. Only top 20 statistically significant enriched entries are shown. Uninfected NHBE cells treated with DMSO are denoted as “- DiMNF - OC43.” NHBE cells infected with HCoV-OC43 and treated with DMSO are denoted as “- DiMNF + OC43.” Uninfected NHBE cells treated with DiMNF drug are denoted as “+ DiMNF - OC43.” NHBE cells infected with HCoV-OC43 and treated with DiMNF drug are denoted as “+ DiMNF + OC43.”