Systematic analysis of SARS-CoV-2 infection of an ACE2-negative human airway cell

Abstract

Established in vitro models for SARS-CoV-2 infection are limited and include cell lines of non-human origin and those engineered to overexpress ACE2, the cognate host cell receptor. We identified human H522 lung adenocarcinoma cells as naturally permissive to SARS-CoV-2 infection despite complete absence of ACE2. Infection of H522 cells required the SARS-CoV-2 spike protein, though in contrast to ACE2-dependent models, spike alone was not sufficient for H522 infection. Temporally resolved transcriptomic and proteomic profiling revealed alterations in cell cycle and the antiviral host cell response, including MDA5-dependent activation of type-I interferon signaling. Focused chemical screens point to important roles for clathrin-mediated endocytosis and endosomal cathepsins in SARS-CoV-2 infection of H522 cells. These findings imply the utilization of an alternative SARS-CoV-2 host cell receptor which may impact tropism of SARS-CoV-2 and consequently human disease pathogenesis.

Figure 1.. The H522 cell line is null for ACE2 expression and is permissive to SARS-CoV-2 infection.

A, Normalized RNA-seq reads were aligned to the GRCh38 and Vervet-African green monkey genomes and quantified with Salmon (v1.3.0). The read counts for ACE2, TMPRSS2, FURIN, CTSB, CTSL, and NRP1 are given for the indicated cell lines. See also Figure S1 and Table S1. B, qRT-PCR for ACE2 and TMPRSS2 expression normalized to 1μg input RNA for each cell line. Cercopithecus aethiops specific primers against TMPRSS2 were used for the Vero E6 samples. Each bar represents mean, error bars indicate SEM (n=3). C, Immunoblot showing ACE2 expression across 10 lung and upper airway cancer cell lines and Vero E6 cells (representative of n=3). ACE2 expression was quantified using Licor Image Studio software in which ACE2 levels were normalized to β-ACTIN, set relative to Vero E6, and are indicated below the immunoblots. D, qRT-PCR for cell-associated SARS-CoV-2 RNA at 4 and 72 hpi at MOI=0.015 or 0.15. MOIs were determined by titration on Vero E6 cells. Error bars represent SEM (n=3). * indicates p<0.05 where significance was determined using two-way ANOVA and the Šidák correction for multiple comparisons. E, qRT-PCR for cell-associated SARS-CoV-2 RNA in H522 cells across various time points and MOIs. Error bars represent SEM (n=2). F, qRT-PCR for SARS-CoV-2 RNA in the supernatant of H522 cells across various time points and MOIs. Error bars represent SEM (n=2). G, Percent of SARS-CoV-2 infected H522 and Vero E6 cells determined by FACS for Nucleocapsid positive cells across various time points and MOIs. Error bars represent SEM (n=2). H, Plaque assays on H522 and Vero cells using two viral dilutions (10⁻² and 10⁻¹). Data are representative of three independent experiments. I, Representative images of H522 cells infected with SARS-CoV-2 at MOI=1. H522 cells were fixed and stained for SARS-CoV-2 RNA (green) by RNAScope reagents and Nucleocapsid (N) protein (red) at 4 and 96hpi and imaged by confocal microscopy (representative of n=2). See also Figure S2. J, Representative images using transmission electron microscopy (TEM) on Vero E6 and H522 cells infected with SARS-CoV-2 (MOI=0.1 pfu/cell and 24 hpi for Vero, MOI: 1 pfu/cell and 96 hpi for H522).

Figure 2.. The SARS-CoV-2 S protein is necessary but not sufficient for viral entry in the H522 cell line.

A, Representative immunoblot showing ACE2 expression and Vinculin as the loading control in Vero E6, H522, H522-ACE2, basal HBEC, and basal HBEC-ACE2 cells. B, Viruses were pre-treated with increasing concentrations of S neutralizing antibody for 1 h and then cells were infected with SARS-CoV-2 at MOI=0.1 in the presence of the S neutralizing antibody. Cell-associated SARS-CoV-2 RNA was detected by qRT-PCR at 24 hpi and was normalized to mock treated (n=3). *** indicates p<0.001 where significance was determined using two-way ANOVA and the Dunnett correction for multiple comparisons. C, SARS-CoV-2 viruses were pre-treated with increasing amounts of soluble ACE2-Fc for 1 h and then cells were infected with SARS-CoV-2 at MOI=0.1 in the presence of ACE2-Fc. Cell-associated SARS-CoV-2 RNA was detected by qRT-PCR at 24 hpi and was normalized to mock treated (n=3). ** indicates p<0.01 and *** indicates p<0.001 where significance was determined using two-way ANOVA and the Dunnett correction for multiple comparisons. D, Representative images of cells infected with VSV-SARS-CoV-2-S_Δ21 at 0 and 8hpi using an Incucyte® S3 Live Cell Analysis System (n=3). Percent GFP positive cells seeded in triplicate were quantified over time with the shaded grey region indicating standard deviation.

Figure 3.. The H522 cell line is permissive to SARS-CoV-2 infection independent of ACE2 expression.

A, Cells were pre-treated with 20 μg/ml of the indicated blocking antibodies for 1 h and then infected with SARS-CoV-2 at MOI=0.1 in the presence of the blocking antibodies. Cell-associated SARS-CoV-2 RNA was detected by qRT-PCR at 72 hpi (n=3). *** indicates p<0.001 where significance was determined using two-way ANOVA and the Dunnett correction for multiple comparisons. B, Polyclonal populations of H522 and Calu-3 (ACE2^+/+ and ACE2^−/−) cells were infected with SARS-CoV-2 virus and cell-associated SARS-CoV-2 RNA was detected by qRT-PCR 4 and 72 hpi (n=8). Error bars indicate the SEM. *** indicates p<0.001 where significance was determined using two-way ANOVA and the Tukey correction for multiple comparisons. See also Figure S3. C, Polyclonal populations of H522 and Calu-3 (ACE2^+/+ and ACE2^−/−) cells were pre-treated with 20 μg/ml of the indicated blocking antibodies for 1 h and then infected with SARS-CoV-2 at MOI=0.1 in the presence of the blocking antibodies. Cell-associated SARS-CoV-2 RNA was detected by qRT-PCR 72 hpi (n=3). Error bars indicate the SEM. *** indicates p<0.001 where significance was determined using two-way ANOVA and the Tukey correction for multiple comparisons. See also Figure S3. D, Monoclonal populations from H522 ACE2^+/+ (6 clones), ACE2^−/− (2 clones), and ACE2^+/− (1 clone) were infected with SARS-CoV-2 at MOI=0.1 and cell-associated SARS-CoV-2 RNA was detected by qRT-PCR 4 and 72 hpi (n≥3). Error bars indicate the SEM. See also Figure S3.

Figure 4.. H522 infection by SARS-CoV-2 is dependent on clathrin-mediated endocytosis and endosomal cathepsins.

A, H522 cells were pre-treated with increasing concentrations of bafilomycin A, SGC-AAK1–1, E64D, apilimod, or camostat mesylate for 1 h and then infected with SARS-CoV-2 at MOI=1 in the presence of the inhibitors. Cell-associated SARS-CoV-2 RNA was detected by qRT-PCR 24 hpi and normalized to DMSO treated cells (n≥3). See also Figure S4. B, Immunoblot showing pAP2M1 (T156), AP2M1, and AAK1 levels in H522 cells infected with SARS-CoV-2 over time (representative of n=2). pAP2M1 (T156) levels were normalized to total AP2M1 and set relative to the 4 hours mock control. Quantification was performed using the Licor Image Studio software and values are indicated below the immunoblots. C, Basal HBECs from 5 different donors were pre-treated with increasing concentrations of SGC-AAK1–1 for 2 h and then infected with SARS-CoV-2 in the presence of the inhibitor. Cell-associated SARS-CoV-2 RNA was detected by qRT-PCR 72 hpi and normalized to DMSO treated cells.

Figure 5.. H522 transcriptome response to SARS-CoV-2 infection.

A, Experimental design of transcriptomics experiments. H522 cells were infected with SARS-CoV-2 at MOI 1.0, 0.25, 0.06, or 0.015 and harvested after 4, 24, 48, 72, and 96 h. Mock-infected cells were harvested after 4 h. All conditions were performed in duplicate. B, Relative expression of SARS-CoV-2 RNA vs. H. sapiens RNA from H522 (n=2). C, Principle component analysis of highly expressed genes from MOIs 0.25 and 1 across all time points. D, Volcano plot of gene expression changes comparing mock infection to 96 hours post infection of MOIs=0.25 and 1. Select changes in IFN response genes (purple) and SARS-CoV-2 genes (salmon) are highlighted. See also Table S2. E, Hierarchical clustering of differentially expressed genes (DEGs) after infection. Genes were filtered for an absolute log₂ fold change >2 and adjusted p-value < 0.005 at any time point. F, Log₂ fold changes of DEGs as grouped by clustering. The colored lines represent quantification of an individual gene whereas the solid black represents the cluster mean. G, Hypergeometric enrichment analysis of biological gene sets in the identified gene clusters (D-E). See also Table S3. H, Rank-based gene set enrichment analysis. Gene sets were queried if identified by hypergeometric analysis RNA seq (5F) or proteomics data (6E). Display indicated p-adjusted < 0.05. N.E.S. = normalized enrichment score.

Figure 6.. H522 infection with SARS-CoV-2 results in proteome changes within the type I IFN, cell cycle, and DNA replication pathways.

A, Experimental design of proteomics experiments. H522 cells were infected with SARS-CoV-2 at MOI=1 and harvested after 4, 12, 24, 48, 72, and 96 h. Mock-infected cells were harvested after 4 and 96 h. Peptides labeled with TMT10 reagents were analyzed by liquid chromatography-mass spectrometry. B, Principal component analysis of whole cell proteomics of H522 cells infected with SARS-CoV-2 across a 4-day time course (n=3). C, Quantification of total ion intensities for each identified SARS-CoV-2 protein over time and normalized to the 4 h mock control. The shaded grey regions represent SEM. D, Volcano plot of protein abundance at 96hpi compared to the 96 h mock control. See also Table S4 E, Differentially expressed proteins from ‘D’ were clustered based on z-score. F, Quantification of total ion intensities normalized to the 4 h mock control for each protein across the 7 identified clusters in ‘D’. The colored lines represent quantification of an individual protein whereas the solid black and dashed black lines represent the mean of infected and mock samples, respectively. G, Hypergeometric enrichment analysis from three different databases for each individual cluster in ‘D’ (Hallmark, Reactome, Gene Ontology). The color of the circle represents significance (q-value), whereas the size of the circle indicates the percentage of the cluster represented in the pathway. See also Table S5. H, Distribution of Pearson’s correlation coefficient between a gene’s transcript and protein log₂ fold change over 4 h mock for all proteins and differentially expressed proteins. Correlations used the matching time points of 4, 24, 48, 72, 96 hpi. I, Rank-based gene set enrichment analysis. Differentially expressed proteins were ranked by their correlation to transcript levels. J, Protein complexes of differentially expressed H522 and SARS-CoV-2 proteins associated with DNA replication and cell cycle checkpoint. Complexes and functions were extracted from the CORUM database. The colors correspond to the whole cell proteomic clusters identified in ‘D’. See also Figure S5. K, Protein interaction network of differentially expressed H522 and SARS-CoV-2 proteins associated with the IFN response. Interactions were determined from the BioGRID Multi-Validated Datasets. Interferon related functions were extracted from GO terms in MSigDB. The colors correspond to the whole cell proteomic clusters identified in ‘D’. See also Figure S5.

Figure 7.. MDA5 mediates the IFN response to SARS-CoV-2 infection.

A, Immunoblot depicting the IFN response in H522 cells infected with SARS-CoV-2 over time (representative of n=2). β-actin represents the loading control. B, ISG mRNA levels was detected by qRT-PCR in H522 cells infected with SARS-CoV-2 96 hpi. H522 cells were either mock transfected or transfected with a non-targeting (NT) siRNA 24 h prior to infection. C, ISG mRNA levels was detected by qRT-PCR in H522 cells infected with SARS-CoV-2 96 hpi. H522 cells were transfected with a non-targeting (NT) siRNA or a panel of siRNAs targeting genes involved in viral sensing 24 h prior to infection. * indicates p<0.05, ** indicates p<0.01 and *** indicates p<0.001 where significance was determined using two-way ANOVA and the Dunnett correction for multiple comparisons. See also Figure S6. D¸ qRT-PCR for cell-associated SARS-CoV-2 RNA in H522 cells 96hpi. H522 cells were transfected with a non-targeting (NT) siRNA or a panel of siRNAs targeting genes involved in viral sensing 24 hours prior to infection.