Re-analysis of Single Cell Transcriptome Reveals That the NR3C1-CXCL8-Neutrophil Axis Determines the Severity of COVID-19

Abstract

SARS-CoV-2 infection has recently been declared a pandemic. Some patients showing severe symptoms exhibit drastic inflammation and airway damage. In this study, we re-analyzed published scRNA-seq data of COVID-19 patient bronchoalveolar lavage fluid to further classify and compare immunological features according to the patient's disease severity. Patients with severe symptoms showed DNA damage and apoptotic features of epithelial cells. Our results suggested that epithelial damage was associated with neutrophil infiltration. Myeloid cells of severe patients showed higher expression of proinflammatory cytokines and chemokines such as CXCL8. As a result, neutrophils were abundant in lungs of patients from the severe group. Furthermore, recruited neutrophils highly expressed genes related to neutrophil extracellular traps. Neutrophil-mediated inflammation was regulated by glucocorticoid receptor expression and activity. Based on these results, we suggest that severe COVID-19 symptoms may be determined by differential expression of glucocorticoid receptors and neutrophils.

Keywords: BAL; COVID-19; CXCL8; SARS-CoV-2; glucocorticoid; neutrophil; scRNA-seq.

FIGURE 1

Damaged features of epithelial cells from severe COVID-19 patients. (A) The Uniform Manifold Approximation and Projection (UMAP) presentation of 12 major cell clusters in bronchoalveolar lavage fluid scRNA-seq data. (B) The UMAP presentation of cell clusters was divided into healthy, mild, and severe groups. (C) A volcano plot of differentially expressed genes of epithelial cells between the mild and severe groups. The x-axis denotes the average log-scaled fold change of gene expression. The y-axis denotes the log-scaled Bonferroni-adjusted p value. (D) Upregulated pathways of the severe COVID-19 group compared to the mild group. The adjusted p values of all presented pathways are zero. Pathways were compared using the normalized enrichment score (NES). (E) A violin plot of the expression of GADD45B and DDIT3 of epithelial cells from the healthy (H), mild (M) and severe (S) groups. (F) A violin plot of PMAIP3 and BAG3 expression in epithelial cells from the H, M, and S groups. (G) Gene Set Enrichment Analysis of the sets compared between severe and mild groups with the HALLMARK_APOPTOSIS gene set.

FIGURE 2

The inflammatory landscape of myeloid cells from severe COVID-19 patients. (A) A violin plot of TNF, IL1B, IL6, and IL12A expression. (B–D) A feature plot of CXCL1 expression (B), CXCL2 (C), and CXCL8 (D). (E) A violin plot of CXCL8 expression in each cluster. (F) The average expression level of CXCL8 from myeloid cells for patients by groups. The data were analyzed using Student’s t-test. **p < 0.01. Error bars denote the mean ± SEM.

FIGURE 3

Infiltration and activation of neutrophils in severe COVID-19 patients. (A) A feature plot of the expression of FCGR3B to identify neutrophil clusters. (B) Neutrophil counts of the integrated data (left) and the percentage of neutrophils per patient (right) were compared among the patient groups. Statistical analysis was tested using Student’s t-test. (C,D) The top 10 upregulated (C) and downregulated (D) pathways in neutrophils of severe patients compared to mild patients. (E) Differentially expressed genes in neutrophils. (F,G) Violin plots of HIF1A expression (F) and CYBB expression (G) in neutrophils by group. (H) Correlation of PMAIP1 expression between epithelial cells and the percentage of neutrophils in the lungs of patients. The correlations were tested using the one-tailed Spearman’s test. *p < 0.05; ****p < -0.0001. Error bars denote the mean ± SEM.

FIGURE 4

The relationship between glucocorticoid receptors and the CXCL8-neutrophil axis. (A) Expression of NR3C1 of integrated total cells (left) and average expression per patient by group (right). (B) The expression of NR3C1 of integrated myeloid cells (left) and the average expression per patient by group (right). Student’s t-test was used for statistical analyses. (C) The correlation between NR3C1 and CXCL8 in myeloid cells of severe patients. (D) The correlation between CXCL8 in myeloid cells and the percentage of neutrophils. (E) The correlation between NR3C1 in myeloid cells and the percentage of neutrophils. One-tailed Spearman’s test was used for the correlation analyses. *p < 0.05; ***p < 0.001; ****p < 0.0001. Error bars denote the mean ± SEM.