SARS-CoV-2 spike protein induces endothelial inflammation via ACE2 independently of viral replication

Abstract

COVID-19, caused by SARS-CoV-2, is a respiratory disease associated with inflammation and endotheliitis. Mechanisms underling inflammatory processes are unclear, but angiotensin converting enzyme 2 (ACE2), the receptor which binds the spike protein of SARS-CoV-2 may be important. Here we investigated whether spike protein binding to ACE2 induces inflammation in endothelial cells and determined the role of ACE2 in this process. Human endothelial cells were exposed to SARS-CoV-2 spike protein, S1 subunit (rS1p) and pro-inflammatory signaling and inflammatory mediators assessed. ACE2 was modulated pharmacologically and by siRNA. Endothelial cells were also exposed to SARS-CoV-2. rSP1 increased production of IL-6, MCP-1, ICAM-1 and PAI-1, and induced NFkB activation via ACE2 in endothelial cells. rS1p increased microparticle formation, a functional marker of endothelial injury. ACE2 interacting proteins involved in inflammation and RNA biology were identified in rS1p-treated cells. Neither ACE2 expression nor ACE2 enzymatic function were affected by rSP1. Endothelial cells exposed to SARS-CoV-2 virus did not exhibit viral replication. We demonstrate that rSP1 induces endothelial inflammation via ACE2 through processes that are independent of ACE2 enzymatic activity and viral replication. We define a novel role for ACE2 in COVID-19- associated endotheliitis.

Figure 1

rS1p induces endothelial cell inflammation and injury. Proteomic analysis of rS1p-treated endothelial cells. (A) The figure represents 1150 hMEC protein interactome created by STRING. Each circular dot (blue, pink) represents the node (protein) and the grey lines the edges. (B) represents the heat map of differentially expressed proteins after 24 h of recombinant S1 spike protein (rS1p). Shades of blue represent the z-score intensities (scale below the heat map). Grey shade represents missing intensities. Human endothelial cells (hMEC) were stimulated with rS1p (0.66 μg/mL) for 5 h and 24 h for assessment of IL-6 (C) and MCP-1 (D) mRNA expression; IL-6 (E) and MCP-1 (F) production. Data are expressed as ± SEM; *p < 0.05 control (Ctl) (non-stimulated cells) vs. rS1p stimulated cells after student’s t-test.

Figure 2

SARS-CoV-2 does not replicate in human endothelial cells but pseudovirus induces cell inflammation. SARS-CoV2 England-02 virus replication in hMECs. VeroE6 (A) and hMECs (B) were exposed to SARS-CoV-2 England-02 virus and virus plaques were measured after 72 h for VeroE6 cells and after 24 h, 48 h and 72 h for hMECs after incubation. The dotted line represents the lower limit of detection at 50 pfu/mL (n = 4–5). Data are expressed as ± SEM; *p < 0.05 negative vs. SARS-CoV-2 pseudotype virus exposed cells after 1-way ANOVA followed by Tukey’s post-hoc test.

Figure 3

Endothelial cell microparticle formation and proliferation in response to rSP1. Human endothelial cells were stimulated with rS1p (0.66 μg/mL) for 5 h and 24 h for assessment of functional changes. Endothelial microparticle formation after 24 h exposure to rSP1 (A). Endothelial cell proliferation in response to rSP1 expressed as the cumulative area under the curve (AUC) (B) and the cell index/minutes of treatment with rS1p from xCELLigence studies (C). Data are expressed as ± SEM; *p < 0.05 control (Ctl) (non-stimulated cells) vs. rS1p stimulated cells after student’s t-test.

Figure 4

rS1p pro-inflammatory effects in human endothelial cells are mediated via ACE2. Human endothelial cells (hMEC) were stimulated with rS1p (0.66 μg/mL) for 5 h and 24 h for assessment of ACE2 mRNA levels (A), protein expression (B) and activity (C) in the presence or absence of MLN-4760 (440 pmol), an ACE2 inhibitor. IL-6 (D) and MCP-1 (E) production; and ICAM-1 (F) and PAI-1 (G) protein expression were also assessed after rS1p stimulation in the presence or absence of MLN-4760 (n = 5–20). Data are expressed as ± SEM; *p < 0.05 control (Ctl) (non-stimulated cells) vs. rS1p stimulated cells; ^†p < 0.05 rS1p stimulated cells vs. rS1p stimulated cells treated with MLN-4760 after 1-way ANOVA followed by Tukey’s post-hoc test. Original blots are presented in Supplementary Figs. S13 and S14.

Figure 5

ACE2 siRNA blocks rS1p pro-inflammatory effects in human endothelial cells. (A) ACE2 siRNA efficiency. Human endothelial cells (hMEC) were stimulated with rS1p (0.66 μg/mL) for 24 h for assessment of IL-6 (B) and MCP-1 (C) production, ICAM-1 (D) and PAI-1 (E) protein expression, and microparticles formation (F) in the presence or absence of ACE2 siRNA (n = 9). Data are expressed as ± SEM; *p < 0.05 control (Ctl) (non-stimulated cells) vs. rS1p stimulated cells; ^†p < 0.05 rS1p stimulated cells vs. rS1p stimulated cells treated with ACE2 siRNA after student’s t-test (A) or 1-way ANOVA followed by Tukey’s post-hoc test. Original blots are presented in Supplementary Figs. S15, S16 and S17.

Figure 6

ACE2 siRNA blocks pro-inflammatory signalling activated by rS1p in human endothelial cells. Human endothelial cells (hMEC) were stimulated with rS1p (0.66 μg/mL) for 10 min for assessment of ERK1/2 (A), NFκB (B) and eNOS (C) phosphorylation in the presence or absence of ACE2 siRNA (n = 9). Human embryonic kidney (HEK) cells expressing (HEK293-ACE2) or not (HEK293) ACE2 (D) were stimulated with rS1p (0.66 μg/mL) for 24 for assessment of IL-8 production (E). Data are expressed as ± SEM; *p < 0.05 control (Ctl) (non-stimulated cells) vs. rS1p stimulated cells or HEK293 vs. HEK293-ACE2 non-stimulated cells; ^†p < 0.05 rS1p stimulated cells vs. rS1p stimulated cells treated with ACE2 siRNA or rS1p stimulated HEK293 cells vs. rS1p stimulated HEK293-ACE2 cells; ^‡p < 0.05 ACE siRNA vs. rS1p stimulated cells treated with ACE2 siRNA after 1-way ANOVA followed by Tukey’s post-hoc test. Original blots are presented in Supplementary Figs. S18, S19, S20 and S21.