Influenza virus infection increases ACE2 expression and shedding in human small airway epithelial cells

Abstract

Background: Patients with coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) demonstrate high rates of co-infection with respiratory viruses, including influenza A (IAV), suggesting pathogenic interactions.

Methods: We investigated how IAV may increase the risk of COVID-19 lung disease, focusing on the receptor angiotensin-converting enzyme (ACE)2 and the protease TMPRSS2, which cooperate in the intracellular uptake of SARS-CoV-2.

Results: We found, using single-cell RNA sequencing of distal human nondiseased lung homogenates, that at baseline, ACE2 is minimally expressed in basal, goblet, ciliated and secretory epithelial cells populating small airways. We focused on human small airway epithelial cells (SAECs), central to the pathogenesis of lung injury following viral infections. Primary SAECs from nondiseased donor lungs apically infected (at the air-liquid interface) with IAV (up to 3×10⁵ pfu; ∼1 multiplicity of infection) markedly (eight-fold) boosted the expression of ACE2, paralleling that of STAT1, a transcription factor activated by viruses. IAV increased the apparent electrophoretic mobility of intracellular ACE2 and generated an ACE2 fragment (90 kDa) in apical secretions, suggesting cleavage of this receptor. In addition, IAV increased the expression of two proteases known to cleave ACE2, sheddase ADAM17 (TACE) and TMPRSS2 and increased the TMPRSS2 zymogen and its mature fragments, implicating proteolytic autoactivation.

Conclusion: These results indicate that IAV amplifies the expression of molecules necessary for SARS-CoV-2 infection of the distal lung. Furthermore, post-translational changes in ACE2 by IAV may increase vulnerability to lung injury such as acute respiratory distress syndrome during viral co-infections. These findings support efforts in the prevention and treatment of influenza infections during the COVID-19 pandemic.

FIGURE 1

ACE2 and TMPRSS2 mRNA expression in human lungs. a) Uniform manifold approximation and projection projections (UMAP) of the single-cell RNA data obtained from human donor lungs from individuals without lung disease (n=3). Note colour-coded populations of specific cell types, as noted, including airway epithelial cell types identified through unsupervised clustering. b–d) UMAP and violin plots of normalised expression of SCGB1A1, a marker of b) airway epithelial cells, c) ACE2 and d) TMPRSS2 in the identified cell type; colocalisation with SCB1A1 is indicated by arrows. ATI: alveolar type I cells; ATII: alveolar type II cells.

FIGURE 2

Effect of influenza A virus (IAV) infection on ACE2 transcription in small airway epithelial cells (SAECs). Relative levels of a) STAT1 and b) ACE2 mRNA measured using reverse transcription quantitative PCR using 18S rRNA as control, expressed as log₂ of 2^−ΔΔCT, following infection of SAECs at the air—liquid interface with IAV low dose (H1N1 pdm09 virus, 3×10² pfu·transwell⁻¹, 72 h) or high dose (H1N1 pdm09 virus, 3×10⁵ pfu·transwell⁻¹, 48 h). Each data point represents an independent experiment from n=6 and n=4 (low and high dose, respectively). Data are presented as mean±SEM; one-way ANOVA—Sidak multiple comparison test. c) Correlation by linear regression between STAT1 and ACE2 expression levels in SAECs; each data point represents an independent IAV-infected or -uninfected experimental condition. Pearson correlation coefficient r=0.78; R²=0.61, p<0.0001.

FIGURE 3

Effect of influenza A virus (IAV) on angiotensin-converting enzyme (ACE)2 protein levels and shedding in small airway epithelial cells (SAECs). a) Validation of ACE2 antibody (#AF933) using immunoblotting of recombinant human (rh)ACE2 protein and whole-cell lysate (1.6 μg) of human embryonic kidney (HEK)293T cells induced to overexpress human ACE2 (HEK^ACE2). b,c) Intracellular ACE2 protein in SAECs infected at the air—liquid interface with IAV (+) (H1N1 pdm09 virus, 3×10⁵ pfu·transwell⁻¹, 48 h) compared to uninfected cells (−), with indicated pre-exposure (+) to e-cigarette vapour (e-cig); b) ACE2 was detected by immunoblotting with the polyclonal antibody #AF933 and c) quantified by densitometry after normalisation to actin levels used as loading control. Cell lysates obtained from distinct donors (#1–3) are noted. d,e) Released ACE2 protein in apical supernatants (normalised by volume) of SAECs infected with IAV (+) at the air—liquid interface, with the indicated pre-exposure (+) to e-cig; ACE2 d) detected by immunoblotting with #AF933 antibody and e) quantified by densitometry. Supernatants obtained from cells from distinct donors (#1–4) are noted. f) Released ACE2 protein in apical supernatants from SAECs infected while submerged in culture media with lower dose (LD; 0.5×10⁵ pfu·transwell⁻¹) or higher dose (HD; 1.0×10⁵ pfu·transwell⁻¹) IAV for 48 h; ACE2 was detected by immunoblotting (with antibody # 21115-1-ap). Cell lysate (1.6 μg) from HEK^ACE2 was used as control. Graphs show individual data points from independent experiments, mean±SEM; t-test.

FIGURE 4

Effect of influenza A virus (IAV) on ADAM17 and TMPRSS2 levels in small airway epithelial cells (SAECs). a–d) Expression levels of indicated proteases measured by reverse transcription (RT) quantitative PCR in SAECs uninfected or infected at the air—liquid interface with IAV at low dose (H1N1 pdm09 virus, 3×10² pfu·transwell⁻¹, 72 h) or high dose (H1N1 pdm09, 3×10⁵ pfu·transwell⁻¹, 48 h); each data point represents an independent experimental condition. a) Correlation between STAT1 and ADAM17 expression levels in SAECs by linear regression; Pearson correlation coefficient R²=0.59, p<0.0001; b) ADAM17 mRNA levels; c) correlation between STAT1 and TMPRSS2 expression levels in SAECs by linear regression; Pearson correlation coefficient R²=0.70, p<0.0001; d) TMPRSS2 mRNA levels. e) TMPRSS2 protein (zymogen and mature forms) in SAECs infected with IAV (+) compared to uninfected cells (−), with indicated pre-exposure (+) to e-cigarette vapour (e-cig) detected by Western blotting; cell lysates obtained from distinct donors (#1–3) are noted; f) relative changes in zymogen, mature and total TMPRSS2 levels induced by IAV were quantified by densitometry relative to uninfected controls, using actin loading control for normalisation. Individual data points indicate independent experiments; mean±SEM, Wilcoxon signed rank test. Pie charts indicating relative levels of mature forms of total TMPRSS2 protein.

FIGURE 5

Schematic showing the interpretation of results in the context of putative severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) co-infection and resultant lung injury. IAV: influenza A virus; ALI: acute lung injury; ACE2: angiotensin-converting enzyme 2.