Neutrophil elastase decreases SARS-CoV-2 spike protein binding to human bronchial epithelia by clipping ACE-2 ectodomain from the epithelial surface

Abstract

Patients with cystic fibrosis (CF) have decreased severity of severe acute respiratory syndrome-like coronavirus-2 (SARS-CoV-2) infections, but the underlying cause is unknown. Patients with CF have high levels of neutrophil elastase (NE) in the airway. We examined whether respiratory epithelial angiotensin-converting enzyme 2 (ACE-2), the receptor for the SARS-CoV-2 spike protein, is a proteolytic target of NE. Soluble ACE-2 levels were quantified by ELISA in airway secretions and serum from patients with and without CF, the association between soluble ACE-2 and NE activity levels was evaluated in CF sputum. We determined that NE activity was directly correlated with increased ACE-2 in CF sputum. Additionally, primary human bronchial epithelial (HBE) cells, exposed to NE or control vehicle, were evaluated by Western analysis for the release of cleaved ACE-2 ectodomain fragment into conditioned media, flow cytometry for the loss of cell surface ACE-2, its impact on SARS-CoV-2 spike protein binding. We found that NE treatment released ACE-2 ectodomain fragment from HBE and decreased spike protein binding to HBE. Furthermore, we performed NE treatment of recombinant ACE-2-Fc-tagged protein in vitro to assess whether NE was sufficient to cleave recombinant ACE-2-Fc protein. Proteomic analysis identified specific NE cleavage sites in the ACE-2 ectodomain that would result in loss of the putative N-terminal spike-binding domain. Collectively, data support that NE plays a disruptive role in SARS-CoV-2 infection by catalyzing ACE-2 ectodomain shedding from the airway epithelia. This mechanism may reduce SARS-CoV-2 virus binding to respiratory epithelial cells and decrease the severity of COVID19 infection.

Keywords: ACE-2; SARS-CoV-2; airway epithelial cell; ectodomain cleavage; mass spectrometry; neutrophil; proteinase; spike protein; virus entry.

Figure 1

Quantitation of ACE-2 levels in plasma and airway mucus from subjects with and without CF and association of ACE-2 levels in CF sputum with neutrophil elastase activity. Blood samples collected from subjects with or without CF were processed for plasma collection, aliquoted, and stored at −80 °C until further use. Frozen sputa from patients with CF or mucus from endotracheal tubes (ETTs) of healthy adult patients were mixed with NS containing 10% Sputolysin at 1:1 (sputum/mucus [mg]: volume [μl]) (37 °C, 15 min), and sputum/mucus supernatants were collected by centrifugation. Plasma and sputum and mucus supernatants were analyzed for sACE-2 by ELISA. Data are summarized as (mean ± SEM) for plasma samples (20 non-CF and 16 CF) (A) and sputum/mucus samples (12 non-CF and 14 CF) (B). Statistical comparisons were made using Mann–Whitney U test. Soluble ACE-2 levels were significantly increased in CF sputum compared to non-CF, ∗p = 0.0146. For analysis of sputum NE activity compared to sputum ACE-2 levels, sputum supernatants prepared with DNase-1 (0.3 mg/ml) for 2 h at 37 °C were used. Scatter dot plot showed a linear correlation between ACE-2 and NE activity levels in CF sputum supernatants (n = 20, r² = 0.24, p = 0.03) (C). ACE-2, angiotensin-converting enzyme 2; CF, cystic fibrosis; NE, neutrophil elastase; NS, normal saline; sACE-2, soluble ACE-2.

Figure 2

NE treatment of HBE and release of ACE-2 ectodomain fragment in conditioned media. HBE cells cultured on plastic (A) or at air–liquid interface (B) were treated with vehicle control or NE (200 or 500 nM) for 1 h or 2 h. Conditioned media (CM) was collected and concentrated to enrich ACE-2 ectodomain fragment release. Concentrated media (22 μl) proteins were separated on a 4 to 20% SDS-PAGE and probed for ACE-2 fragment following transfer to nitrocellulose. Blots were probed with primary rabbit anti-ACE-2 antibody that recognizes the N-terminal ACE-2 ectodomain of the protein (1:1000 dilution, O/N in 5% milk); secondary horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG antibody (1:5000 dilution) and the immunoreactive protein complexes were detected using lightning ultrachemiluminescence substrate. A, immunoblot showing the release of ACE-2 fragment in the conditioned media from undifferentiated NHBE cells cultured on plastic. B, immunoblot of apical conditioned media from differentiated NHBE cells cultured at ALI. Positive control (PC) was Caco2 whole cell lysates (catalog number ab3950, Abcam). Western blots shown were representative of n = 4 undifferentiated and n = 2 differentiated NHBE donor cells. ACE-2, angiotensin-converting enzyme 2; IgG, immunoglobulin G; HBE, human bronchial epithelial; NE, neutrophil elastase.

Figure 4

The effect of NE treatment of HBE on SARS-CoV-2 spike protein binding detected by flow cytometry. Dose-dependent binding of spike protein to HBE cells was determined by flow cytometry (A and B). Undifferentiated HBE cells–expressing endogenous ACE-2 were incubated with a concentration curve of recombinant SARS-CoV-2 trimeric spike protein with C-terminal His6-tag (1–10 μg/ml) or a concentration curve of His6 peptide alone as control (1–10 μg/ml), for 2 h at 37 °C. In a separate experiment, HBE were treated with NE (200 or 500 nM) or control vehicle for 2 h prior to incubation with SARS-CoV-2 spike protein (10 μg/ml) (C and D). Following incubation, the cells were stained with rabbit anti-His-tag antibody (1:500 dilution), followed by Alexa Fluor 488–conjugated goat anti-rabbit IgG (1:500). Mean fluorescence intensity (MFI) shows spike protein binding to HBE in a dose-dependent manner (A) black bars; however, there was only background binding with no dose-dependent increase in control His6 peptide binding to HBE (A) (red bars). Representative flow cytometry histograms show differences in spike protein binding (black) and His tag binding (red) (B). NE treatment significantly decreased spike protein binding to HBE (C). Representative histogram overlay displayed the inhibition of spike binding to HBE post-NE exposure (D). Data are presented as MFI ± SEM; n = 3 independent experiments with three different donor cells, with total eight replicates per condition. ∗p = 0.02, ∗∗p = 0.01 versus no spike protein (0); ++p = 0.002 versus Ctrl. Statistical analysis was performed by one-way, nonparametric ANOVA (Kruskal–Wallis test) and post hoc comparisons by Wilcoxon Rank Sum test. ACE-2, angiotensin-converting enzyme 2; HBE, human bronchial epithelial; IgG, immunoglobulin G; NE, neutrophil elastase; SARS-CoV-2, severe acute respiratory syndrome-like coronavirus-2.

Figure 3

HBE cell surface ACE-2 expression was determined by flow cytometry, following treatment with NE or control vehicle. Cell surface ACE-2 expression in HBE cells was analyzed by flow cytometry after the cells were exposed to NE at the indicated dose. Undifferentiated HBE cells–expressing endogenous ACE-2 were exposed to NE (200 or 500 nM) or vehicle control at 37 °C for 2 h. Following incubation, the cells were stained with goat anti-ACE-2 antibody (1 μg/ml) followed by phycoerythrin-conjugated donkey anti-goat IgG (1:100). Mean fluorescence intensity (MFI) showing ACE-2 staining in HBE cells was significantly reduced with NE exposure both at 200 and 500 nM compared to untreated cells (A). Background MFI was determined for HBE cells exposed to no primary and no secondary (white bar) or secondary antibody only control (red bar). Representative flow cytometry histograms showing differences in ACE-2 binding to HBE cells with or without NE exposure (color matched) are shown (B). Data are presented as MFI ± SEM; n = 3 independent experiments with three different HBE donor cells, with a total of nine replicates per condition. ∗∗∗p = 0.0004 versus. control. Statistical analysis was performed by one-way, nonparametric ANOVA (Kruskal–Wallis test) and post hoc comparisons by Wilcoxon Rank Sum test. ACE-2, angiotensin-converting enzyme 2; HBE, human bronchial epithelial; IgG, immunoglobulin G; NE, neutrophil elastase.

Figure 5

Recombinant Fc-tagged human ACE-2 and recombinant His6-tagged SARS-CoV-2 trimeric spike protein susceptibility to NE proteinase activity. Recombinant human ACE-2 protein with Fc-tag or SARS-CoV-2 spike protein with His6 Tag were incubated with control vehicle or NE (50, 100 nM for 15 min) for ACE2 protein or NE (50, 100, 200 nM for 15 and 30 min) for SARS-CoV-2 spike protein. Equal amounts of reaction products were resolved on 4 to 20% SDS-PAGE. ACE-2 protein tagged with Fc, cleaved by NE proteinase activity, was confirmed by imperial protein stain (A). The specificity of ACE-2 cleavage (B) or Fc cleavage (C) was further confirmed by probing with primary rabbit mAbs for ACE-2 (1:1000) or mouse mAb raised against human IgG₁ Fc peptide, followed by secondary horseradish peroxidase (HRP)-conjugated goat anti-rabbit or anti-mouse IgG antibody (1:5000). Immunoreactive complexes were developed by chemiluminescence. SARS-CoV-2 spike protein was not cleaved by NE proteinase activity, confirmed by imperial protein stain (D). Arrows indicate full-length and cleaved ACE-2 or Fc fragments following NE treatment. Data shown are representative of 2 to 3 independent experiments. ACE-2, angiotensin-converting enzyme 2; IgG, immunoglobulin G; NE, neutrophil elastase; SARS-CoV-2, severe acute respiratory syndrome-like coronavirus-2.

Figure 6

In-gel LC-MS/MS characterization of ACE-2 cleavage sites by NE. The ACE-2 protein with Fc-tag was treated for 15 min with neutrophil elastase at 0 (Ctrl), 50, 100, and 200 nM, separated on a 4 to 20% PAGE, and stained with highly sensitive colloidal gold total protein stain to detect full-length and NE-cleaved protein fragments (A). A second 4 to 20% PAGE gel was prepared under identical conditions in duplicate for ACE-2-Fc protein treated with NE at 0 (Ctrl) and 50 nM. The gel was stained with imperial protein stain, followed by excision of 16 gel fragments from four rows from regions outlined by the dashed-line grid spanning ∼120 kDa (“row 1”) to ∼65 kDa (“row 4”) and then analyzed by LC-MS/MS following in-gel tryptic digestion, (B). The LC-MS/MS dataset was searched using enzyme specificities for both trypsin (C-term of R, K) and neutrophil elastase (C-term of A, V) which resulted in the identification of six NE-cleavage sites including three in the ACE-2 domain (C). Relative label-free quantities for the six identified NE-cleaved semitryptic peptides are reported as log₂ NE/Ctrl by gel band row (i.e., MW) and amino acid positions (D). Data summarize one experiment with duplicate samples. ACE-2, angiotensin-converting enzyme 2; MS/MS, tandem mass spectra; MW, molecular weight; NE, neutrophil elastase.