Low expression of EXOSC2 protects against clinical COVID-19 and impedes SARS-CoV-2 replication

Abstract

New therapeutic targets are a valuable resource in the struggle to reduce the morbidity and mortality associated with the COVID-19 pandemic, caused by the SARS-CoV-2 virus. Genome-wide association studies (GWAS) have identified risk loci, but some loci are associated with co-morbidities and are not specific to host-virus interactions. Here, we identify and experimentally validate a link between reduced expression of EXOSC2 and reduced SARS-CoV-2 replication. EXOSC2 was one of 332 host proteins examined, all of which interact directly with SARS-CoV-2 proteins; EXOSC2 interacts with Nsp8 which forms part of the viral RNA polymerase. Lung-specific eQTLs were identified from GTEx (v7) for each of the 332 host proteins. Aggregating COVID-19 GWAS statistics for gene-specific eQTLs revealed an association between increased expression of EXOSC2 and higher risk of clinical COVID-19 which survived stringent multiple testing correction. EXOSC2 is a component of the RNA exosome and indeed, LC-MS/MS analysis of protein pulldowns demonstrated an interaction between the SARS-CoV-2 RNA polymerase and the majority of human RNA exosome components. CRISPR/Cas9 introduction of nonsense mutations within EXOSC2 in Calu-3 cells reduced EXOSC2 protein expression, impeded SARS-CoV-2 replication and upregulated oligoadenylate synthase ( OAS) genes, which have been linked to a successful immune response against SARS-CoV-2. Reduced EXOSC2 expression did not reduce cellular viability. OAS gene expression changes occurred independent of infection and in the absence of significant upregulation of other interferon-stimulated genes (ISGs). Targeted depletion or functional inhibition of EXOSC2 may be a safe and effective strategy to protect at-risk individuals against clinical COVID-19.

Figure 1:. Unbiased screen of host proteins identified as high confidence interacting partners of SARS-CoV-2 proteins links RNA exosome components to risk of clinical COVID-19.

(a) Schematic of the study design. Known host-viral interactions were screened for disease-association by combining lung-specific eQTLs with a GWAS for COVID-19 symptoms. Identification of a positive correlation between EXOSC2 expression and increased severity of COVID-19 led to further study of interactions between the SARS-CoV-2 polymerase and the entire human RNA exosome by AP-MS. Finally, CRISPR editing of EXOSC2 within human lung cells and subsequent infection with SARS-CoV-2 facilitated validation of the relationship between EXOSC2 expression and viral replication and interrogation of the underlying biological mechanism. (b) Lung eQTLs were used to group genetic variants according to their effect on expression of 332 host genes encoding proteins which interact with viral proteins. Only expression of EXOSC2 was significantly associated with clinical risk of COVID-19 after Bonferroni multiple testing (red line). (c-d) Lung eQTLs were used to group genetic variants according to their effect on expression of all genes encoding components of the RNA exosome. Expression levels of EXOSC7, EXOSC9 and EXOSC2 were significantly linked to clinical COVID-19 and in each case higher expression was associated with higher risk of infection. p=0.05 is indicated by a red dashed line.

Figure 2:. AP-MS analysis confirms the interaction of the SARS-CoV-2 RNA polymerase with EXOSC2 and the majority of components of the host RNA exosome.

Replicate affinity purifications of HEK293T cells expressing Strep-Nsp8 and untagged Nsp7 and control purifications (mock-transfected) were analysed by label-free quantitative mass spectrometry. (a) Volcano plot of Strep-Nsp8 pulldowns from cells co-expressing Nsp7 compared to mock-transfected cells. Data points in red are proteins significantly enriched in Strep-Nsp8 pulldowns with a permutation-based FDR 0.05. (b) RNA exosome complex proteins within the set of enriched proteins are labelled.

Figure 3:. Reduced expression of EXOSC2 in Calu-3 cells is not toxic and leads to reduced viral replication.

(a) Calu-3 cells were targeted with the indicated sgRNAs and cell viability was analysed by MTT assay. Data for unedited control cells were set to 100%. (b-f) Calu-3 cells targeted with sgRNAs and subsequently reconstituted with EXOSC2 as indicated were infected with SARS-CoV-2 (MOI=1) for 17 hours. As a negative control, cells infected with virus were exposed to a neutralising antibody. (b) Viral titres in supernatant samples were analysed by TCID50 assay. (c-d). Viral RNA levels were measured by absolute RT-qPCR quantification of N1 and N2 SARS-CoV-2 genomic RNA. (e) Viral genomic reads as a proportion of total RNA-sequencing reads. (f) Viral genomic RNA-sequencing reads mapped across the SARS-CoV-2 genome by normalised read-depth; colours represent distinct viral transcripts. Data are from three independent biological repeats. In panels (a-e), individual data points are shown with mean and standard error. Significance was tested by paired t-test and p values are indicated.

Figure 4:. Transcriptomic analysis confirmed the inflammatory response to SARS-CoV-2 infection of Calu-3 cells and identified upregulation of OAS genes in the context of reduced EXOSC2 expression.

RNA for sequencing was extracted from Calu-3 cells in the presence and absence of CRISPR editing with sgRNA targeted against EXOSC2; with and without infection with SARS-CoV-2 (MOI=1) at 17h; three biological replicates were obtained for all conditions. (a) First and second principal components for total gene expression across all sequenced samples. Samples include WT unedited Calu-3 cells and EXOSC2 edited Calu-3 cells; +/− indicates the presence/absence of SARS-CoV-2 infection. (b) Heatmap representation of genes upregulated in WT cells in the presence of SARS-CoV-2 infection. A darker colour indicates higher expression. (c) Volcano plot to compare gene expression in uninfected Calu-3 cells with and without CRISPR editing of EXOSC2. Dotted lines represent fold change of +/− 2 and a Bonferroni multiple testing threshold for p-value by genewise exact test. (d) Heatmap representation of 397 ISGs (Schoggins et al. 2011) across all sequenced samples. (e-f) Normalised expression of OAS1 (d) and OAS3 (e) in all four conditions.