key: cord-0939432-08p8ns2d authors: Chen, H.; Liu, W.; Liu, D.; Zhao, L.; Yu, J. title: SARS-CoV-2 activates lung epithelia cell proinflammatory signaling and leads to immune dysregulation in COVID-19 patients by single-cell sequencing date: 2020-05-13 journal: nan DOI: 10.1101/2020.05.08.20096024 sha: 21472b4330e32d3941eecf8187d80a8b0d19fcd6 doc_id: 939432 cord_uid: 08p8ns2d Objective: The outbreak of Coronavirus disease 2019 (COVID-19) caused by SARS-CoV-2 infection has become a global health emergency. We aim to decipher SARS-CoV-2 infected cell types, the consequent host immune response and their interplay in the lung of COVID-19 patients. Design: We analyzed single-cell RNA sequencing (scRNA-seq) data of lung samples from 17 subjects (6 severe COVID-19 patients, 3 mild patients who recovered and 8 healthy controls). The expression of SARS-CoV-2 receptors (ACE2 and TMPRSS2) was examined among different cell types in the lung. The immune cells infiltration patterns, their gene expression profiles, and the interplay of immune cells and SARS-CoV-2 target cells were further investigated. Results: Compared to healthy controls, the overall ACE2 (receptor of SARS-CoV-2) expression was significantly higher in lung epithelial cells of COVID-19 patients, in particular in ciliated cell, club cell and basal cell. Comparative transcriptome analysis of these lung epithelial cells of COVID-19 patients and healthy controls identified that SARS-CoV-2 infection activated pro-inflammatory signaling including interferon pathway and cytokine signaling. Moreover, we identified dysregulation of immune response in patients with COVID-19. In severe COVID-19 patients, significantly higher neutrophil, but lower T and NK cells in lung were observed along with markedly increased cytokines (CCL2, CCL3, CCL4, CCL7, CCL3L1 and CCL4L2) compared with healthy controls as well as mild patients who recovered. The cytotoxic phenotypes were shown in lung T and NK cells of severe patients as evidenced by enhanced IFN{gamma}, Granulysin, Granzyme B and Perforin expression. Moreover, SARS-CoV-2 infection altered the community interplay of lung epithelial cells and immune cells: the interaction between epithelial cells with macrophage, T and NK cell was stronger, but their interaction with neutrophils was lost in COVID-19 patients compared to healthy controls. Conclusions: SARS-CoV-2 infection activates pro-inflammatory signaling in lung epithelial cells expressing ACE2 and causes dysregulation of immune response to release more pro-inflammatory cytokines. Moreover, SARS-CoV-2 infection breaks the interplay of lung epithelial cells and immune cells. The Coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), poses a tremendous global challenge recently. As of May 7, 2020, a total of 3,820,869 COVID-19 cases and 265,098 COVID-19 deaths have been reported affecting 212 countries and territories, and the number is still growing as a result of human-to-human transmission 1,2 . SARS-CoV-2 belongs to coronaviruses family which are singlestranded and positive-sense RNA viruses characterized by club-like spike on their surface 3 . SARS-CoV-2 binds to the surface expressed proteins, angiotensin-converting enzyme 2 (ACE2), to entry into cells which is similar as SARS-CoV 4-6 . In addition to ACE2, the expression of serine protease TMPRSS2 on target cells is required for activation of viral spike (S) proteins to facilitate viral entry 6 . However, the ACE2-and TMPRSS2-expressing cell types and their expression level in the lung of COVID-19 patients are unclear. Although SARS-CoV-2 could be recognized by the host immune system to mount an antiviral response 5,7 , imbalanced immune responses have been observed in most patients, as exemplified by high neutrophil to lymphocyte ratio 8-12 . Moreover, a large number of severe COVID-19 patients suffered cytokine storm with markedly release of proinflammatory cytokines such as interleukin 6 (IL-6), interleukin 10 (IL-10) and tumor necrosis factor (TNF)-α, leading to the progression of acute respiratory distress syndrome (ARDS) and potentially death 9, 13 . However, it is still unknown how SARS-CoV-2 infection contributes to dysregulated immune response in the lung of COVID-19 patients. 3 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. . https://doi.org/10.1101/2020.05. 08.20096024 doi: medRxiv preprint In this study, we comprehensively evaluated the single cell sequencing data from the lung of 17 subjects (6 severe COVID-19 patients, 3 recovered COVID-19 patients with mild symptoms and 8 healthy donors) to uncovered cell types with ACE2 and TMPRSS2 expression in the lung infected with SARS-CoV-2. We To examine the expression of SARS-CoV-2 entry genes, ACE2 and TMPRSS2, in different cell types of human lung after SARS-CoV-2 infection, single-cell RNA sequencing (scRNA-seq) data of lung bronchoalveolar lavage fluid (BALF) from 3 recovered mild cases and 6 severe cases (GSE145926) 14 , as well as 8 normal lungs (GSE122960) 15 Figure 1C) . Notably, the percentages of ACE2 positive cells among these three types of lung epithelial cells were all significantly higher in BALF samples from either severe or mild COVID-19 patients as compared to those in lung derived from healthy controls ( Figure 1D ). In keeping with this, the ACE2 mRNA expression level was significantly higher in COVID-19 patients compared to healthy controls in club, basal and ciliated cells, respectively ( Figure 1E ). . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 13, 2020. . https://doi.org/10.1101/2020.05.08.20096024 doi: medRxiv preprint However, the correlation between increased ACE2 expression in the lung epithelia cells from COVID-19 patients and SARS-CoV-2 infection needs further in-depth investigation, considering the small sample size in this study and the treatment administrated to these patients. A total of 65 common up-regulated genes and 53 down-regulated genes were identified in these three type of cells after virus infection (adjusted p ≤ 0.01 and |log2Fold change (FC)| ≥ 1) (Figure 1F and 1G) . Gene Set Enrichment Analysis (GSEA) of these candidate genes revealed that SARS-CoV-2 infection induced interferon pathway and cytokine signaling in the lung epithelia cells of COVID-19 patients ( Figure 1F ). On the other hand, SARS-CoV-2 was capable to suppress host protein translation ( Figure 1G ). We further studied the specification of immune cells fates in response to SARS-CoV-2 infection. As shown in Figure 1B patients compared with healthy controls, while it was restored to normal after the patients recovered ( Figure 2B) . Whilst, macrophage number was significantly lower in severe COVID-19 patients compared to healthy controls, but restored in 5 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 13, 2020. . https://doi.org/10.1101/2020.05.08.20096024 doi: medRxiv preprint recovered patients ( Figure 2B ). SARS-CoV-2 infection significantly increased T/NK cells in the lung, but to a lesser extent in severe COVID-19 patients ( Figure 2B ). We then explored the differential gene expression profiling of immune cells in the lung between COVID-19 patients and healthy controls. As shown in Figure 2C , We evaluated the relationship between epithelia cells and immune cells in the lung from healthy to the diseased status ( Figure 3A and 3B) . In severe COVID-19 patients, the strength of predicted strong correlations between lung epithelial cells (club and basal cells) and neutrophils were significantly reduced, but their correlations with macrophage and T/NK cell were markedly increased. As to ciliated cells, the correlation network appeared to be the same between severe COVID-19 patients and healthy control ( Figure S9) . We investigated the function 6 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. . https://doi.org/10.1101/2020.05.08.20096024 doi: medRxiv preprint of genes constituting the abovementioned network by GSEA gene enrichment analysis and found that these genes were enriched in pathways including phagosome, antigen processing, presentation and interferon alpha/beta signaling ( Figure 3C ). ASS1, CXCL8 and HLA-B were selected for further validation as they are among the list of differentially expressed genes (DEGs) in lung epithelia cells after SARS-CoV-2 infection. In both club and basal cells, the expression of ASS1, CXCL8 and HLA-B were all significantly higher in severe COVID-19 patients as compared to healthy controls or mild recovered patients ( Figure 3D ). In supporting this, SARS-CoV-2 infection significantly increased the mRNA expression of ASS1, CXCL8 and HLA-B in human normal bronchial epithelial cells, human lung cancer cell line Calu-3 and A549 overexpressing ACE2 ( Figure 3E ). Our findings suggest that the specific networks between epithelia cells and immune cells were formed in lung after SARS-CoV-2 infection. In this study, we first identified high expressions of ACE2 and TMPRSS2 in patients. In keeping this, no SARS-CoV-2 viral gene expression was detected in peripheral blood mononuclear cell in three SARS-CoV-2 patients 20 . Thus, lung 7 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. (Figure 4) . Single cell RNA sequence (scRNA-Seq) data were retrieved from published resources, including bronchoalveolar lavage fluid (BALF) from 6 severe and 3 moderate COVID-19 patients 14 We re-analyzed the data from a count quantification matrix due to the unavailable per-cell annotation. Cells with mitochondrial gene proportion higher 9 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. . https://doi.org/10.1101/2020.05.08.20096024 doi: medRxiv preprint than 15% were filtered out. For each individual dataset, raw count matrix was first normalized and the top 2 000 most variable genes were chosen. For each cell, we divided the gene counts by the total counts and multiplied by 10 000, followed by natural-log transformation. High variable genes were determined using 10 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. . https://doi.org/10.1101/2020.05.08.20096024 doi: medRxiv preprint The relative amount of each immune cell type (Neutrophil, Macrophages, T/NK cells and B cells) was defined as its proportion in CD45+ cells. Correlation between gene expression and immune composition was measured by Spearman Correlation Coefficient (SCC) and was computed for healthy control and COVID-19 patients, respectively. Those with difference in SCC higher than a threshold (0.9) between healthy control and COVID-19 were chosen as differentially correlated pairs. Functional enrichment analysis was a method aimed to identify classes of molecules (genes or proteins) that were over-represented in a set of pre-defined molecules and predicted its association with disease phenotypes. We performed this method to uncover potential biological function shift under SARS-CoV-2 infection through mapping the molecules into known molecule sets by WebGestalt 29 . Two databases, KEGG and Reactome were used for canonical pathway detection. Significantly enriched functions were chosen if the corresponding adjusted p-value was below a threshold (0.05). Gene expression levels were represented as mean ± SD unless otherwise indicated. To compare the difference among groups, pairwise Wilcoxon rank sum tests was performed. All statistical analysis was conducted under R computing software. We thank Yifei Wang, Jia Yang, Shanshan Gao and Feixue Wang for their comments and suggestions for this study. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. 12 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 13, 2020. Severe Immune cell and others (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20) Epithelial (1, 2, 3, 4, 5) VCAN LILRA5 S100A8 FCN1 APOC1 FABP4 APOE CCL18 CYP27A1 GPNMB TRBC1 CCL5 CD7 TRAC CD3D CD3E IGKC Clinical infectious diseases : an official publication of the Infectious Diseases Society of America Immunopathological characteristics of coronavirus disease 2019 cases in Guangzhou, China. medRxiv, 2020 Clinical and immunological features of severe and moderate coronavirus disease 2019 The landscape of lung bronchoalveolar immune cells in COVID-19 revealed by single-cell RNA sequencing. medRxiv Single-Cell Transcriptomic Analysis of Human Lung