Elevated FiO₂ increases SARS-CoV-2 co-receptor expression in respiratory tract epithelium

Abstract

The severity of coronavirus disease 2019 (COVID-19) is linked to an increasing number of risk factors, including exogenous (environmental) stimuli such as air pollution, nicotine, and cigarette smoke. These three factors increase the expression of angiotensin I converting enzyme 2 (ACE2), a key receptor involved in the entry of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)—the etiological agent of COVID-19—into respiratory tract epithelial cells. Patients with severe COVID-19 are managed with oxygen support, as are at-risk individuals with chronic lung disease. To date, no study has examined whether an increased fraction of inspired oxygen (FiO₂) may affect the expression of SARS-CoV-2 entry receptors and co-receptors, including ACE2 and the transmembrane serine proteases TMPRSS1, TMPRSS2, and TMPRSS11D. To address this, steady-state mRNA levels for genes encoding these SARS-CoV-2 receptors were assessed in the lungs of mouse pups chronically exposed to elevated FiO₂, and in the lungs of preterm-born human infants chronically managed with an elevated FiO₂. These two scenarios served as models of chronic elevated FiO₂ exposure. Additionally, SARS-CoV-2 receptor expression was assessed in primary human nasal, tracheal, esophageal, bronchial, and alveolar epithelial cells, as well as primary mouse alveolar type II cells exposed to elevated oxygen concentrations. While gene expression of ACE2 was unaffected, gene and protein expression of TMPRSS11D was consistently upregulated by exposure to an elevated FiO₂. These data highlight the need for further studies that examine the relative contribution of the various viral co-receptors on the infection cycle, and point to oxygen supplementation as a potential risk factor for COVID-19.

INTRODUCTION

The world currently finds itself in the grip of the coronavirus disease 2019 (COVID-19) pandemic caused by infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (18). A number of risk factors for severe COVID-19 have been identified, including advanced age, as well as comorbidities including hypertension, obesity, malignancy, and chronic lung disease (11). Risk factors could also include stimuli that modify susceptibility to SARS-CoV-2 infection, for example, by modulating the expression of receptors that mediate viral entry into cells. Endogenous physiological modulators of viral entry receptors such as estrogen (16) and interferon (19) have recently been identified. However, exogenous (environmental) modulators of virus receptor expression may represent hitherto unrecognized risk factors for COVID-19. Indeed, air pollution (3) as well as cigarette smoke (8) and nicotine (7) have recently been proposed to increase the risk of severe COVID-19 by driving expression of the SARS-CoV-2 entry receptor angiotensin I converting enzyme 2 (ACE2) (5).

COVID-19 patients with acute respiratory failure are managed with supplemental oxygen (18), and elderly patients and infants born preterm with restrictive and obstructive lung disease may use home oxygen therapy. This raises the question of whether an elevated FiO₂ might alter the expression of the virus entry receptor ACE2, as well as the virus co-receptors that facilitate cellular entry of SARS coronaviruses, including the serine peptidases TMPRSS1, TMPRSS2, and TMPRSS11D (4, 10). The full spectrum of SARS-CoV-2 receptors has not been definitively clarified, and a growing number of candidate receptors continue to be profiled in the epithelium of the entire respiratory tract (1, 17). To answer the question of whether an elevated FiO₂ might impact the regulation of virus receptor genes, the expression of SARS-CoV-2 entry receptors and co-receptors was assessed in lung tissue from preterm-born infants with bronchopulmonary dysplasia (BPD) managed chronically with elevated FiO₂, or from mouse pups exposed to elevated FiO₂ for protracted periods in an animal model of BPD. It is critical to emphasize that this study is neither a BPD study, nor a study on addressing COVID-19 in infants (9). Rather, the clinical and experimental BPD models served purely as models of chronic supplemental oxygen exposure. This report should not be misunderstood as a pediatric study. Newborn mice were employed in the present study because newborn mice tolerate protracted periods of exposure to high FiO₂, whereas adult mice do not and rapidly succumb to lethal acute lung injury. The expression of entry receptors and co-receptors was also assessed in adult mouse and adult human epithelial cells originating from different regions of the respiratory tract between the nose and the alveoli, as well as the esophagus, which were exposed in vitro to elevated oxygen concentrations.

MATERIALS AND METHODS

Mouse tissue.

In an experimental animal model of BPD [Institutional Animal Care and Use Committee (IACUC) approval for RNA profiling of mouse lung tissues under approval references B2/277 and B2/351]. Newborn mouse pups inhaled an FiO₂ of 0.21 or 0.85 from the day of birth to post-natal day (P)14 (12).

Human tissue.

Harvesting of human lung tissue from BPD patients [x̄ birth wt 706 ± 222 g; 6 male/4 female; x̄ duration FiO₂ >0.5, 29 day (range 5–96 days)] and term infant controls (x̄ birth wt 2,033 ± 1,356 g; 2 male/6 female; no oxygen supplementation) without lung disease was approved by the Institutional Review Board of the University of Rochester (2). Written informed consent was obtained from guardians.

Human and mouse lung epithelial cells.

Mouse alveolar epithelial type II cells were isolated from female C57BL/6J mice (3–6 mo of age) as described previously (14). Human airways and lung epithelial cells were purchased from Cell Biologics (Planegg, Germany): tracheal (H-6033), esophageal (H-6046), bronchial (H-6033B), and alveolar (H-6053). Human nasal epithelial cells were purchased from PromoCell (C-12620). The sexes of the donors of the human respiratory tract cells were not disclosed by the vendor. Cells were not passaged, and were cultured per manufacturer’s recommendations, and maintained on air-liquid interfaces in a 21% or 85% O₂ atmosphere (14).

Real-time RT-PCR profiling of gene expression.

The abundance of mRNA transcripts was screened by real-time RT-PCR (14) using the primer pairs (5′→3′): human ACE2 (ACAGTCCACACTTGCCCAAAT, TGAGAGCACTGAAGACCCATT), TMPRSS1 (CCTGCCCCTCACAGAATACAT, GACATCATTGCTGATTATGGGGA), TMPRSS2 (CAAGTGCTCCAACTCTGGGAT, AACACACCGATTCTCGTCCTC), TMPRSS11D (CACAGTTCCAGAGCTAAGGCA, ACTCCACCTTGAGGTACTCCA), with POLR2A as reference gene (GCGGAATGGAAGCACGTTAAT, CCCAGCACAAAACACTCCTC); and mouse Ace2 (GCAGATGGCTACAACTATAACCG, CCTCCTCACATAGGCATGAAGA), Tmprss1 (TACCTTCCCTTTCGAGACCCT, CCATAGAACTGTGTGTTACCCCA), Tmprss2 (GCTGTCTTGCTTTGGAGGTTC, GACAATGTGCTACCCCGTCA), Tmprss11d (GGTACAGCTCCGTAACTCGTG, GCTGGGAGACATACCCTATGGAT), with Polr2a as reference gene (CTAAGGGGCAGCCAAAGAAAC, CATTCAGCATACAACTCTAGGC). Data are presented as difference in cycle threshold (Ct), ΔCt, given by Ct_(Polr2a) – Ct_{(gene of interest)}. Fold-change in mRNA abundance was given by fold-change = 2^|ΔΔCt|.

Immunoblotting.

Protein extracts were prepared as described previously (15), from the neonate mouse lungs and the human tracheal epithelial cells described above. Proteins were resolved on a 12% reducing SDS-PAGE gel, electroblotted to a nitrocellulose membrane. TMPRSS11D was detected with a rabbit anti-human TMPRSS11D (ab127031, Abcam, Cambridge, UK; 1:1,500) which detected both the human and mouse antigen. Two proteins were employed to document loading equivalence: β-actin and β-tubulin. Two different loading controls were employed because β-actin has a similar molecular mass to TMPRSS11D, and loading equivalence was documented on the same membrane. Thus, the TMPRSS11D signal may have affected the β-actin signal. β-Actin was detected with a rabbit anti-human β-actin antibody (4967S, Cell Signaling Technology, Frankfurt am Main, Germany; 1:1,000), and β-tubulin was detected with a rabbit anti-human β-tubulin antibody (ab6046, Abcam, Cambridge, UK; 1:1,000). Immune complexes were detected with a horseradish peroxidase-conjugated goat anti-rabbit IgG (H+L) (31460, ThermoFisher, Waltham, MA; 1:3,000), and visualized using SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher, Waltham, MA).

RESULTS AND DISCUSSION

Expression of SARS-CoV-2 receptors in experimental and clinical BPD.

Exposure of newborn mouse pups to FiO₂ 0.85 increased lung expression of Tmprss11d 10.5-fold after 2 days, and Tmprss11d expression remained elevated, up to 40-fold at 10 days (Fig. 1A). By 14 days, Ace2, Tmprss1, and Tmprss2 expression was increased 1.5-, 2-, and 1.5-fold, respectively (Fig. 1A). The increased protein expression of TMPRSS11D in the lungs of mice exposed to FiO₂ 0.85 was validated by immunoblot (see Fig. 3A). In BPD patients that were chronically managed with FiO₂ > 0.5, no changes in ACE2, TMPRSS1, and TMPRSS2 expression were noted. However, a 2.5-fold increase in TMPRSS11D expression was noted, compared with term infants without oxygen supplementation (Fig. 1B).

Fig. 1.

Steady-state levels of lung mRNA transcripts encoding severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) entry receptors and co-receptors in experimental animals and clinical subjects chronically exposed to elevated FiO₂. A: steady-state mRNA levels were determined by real-time RT-PCR for Ace2, Tmprss1, Tmprss2, and Tmprss11d in lung homogenate cDNA from C57BL/6J mouse pups (n = 5 animals per group) at post-natal day (P)2, 3, 5, 10, and 14, exposed either to room air (FiO₂ 0.21; yellow bars) or hyperoxia (FiO₂ 0.85; blue bars) for the first 14 days of post-natal life. B: steady-state mRNA levels were similarly determined for ACE2, TMPRSS1, TMPRSS2, and TMPRSS11D in lung homogenate cDNA from infants without (Ctrl, n = 8 patients; yellow bars) or with bronchopulmonary dysplasia (BPD; n = 10 patients; blue bars). Data reflect mean ΔCt ± SD. Pairwise comparisons were made between the 21% O₂ and 85% O₂ groups by unpaired Student’s t test (A), and between the Ctrl and BPD groups by Mann-Whitney U test (B). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ACE2, angiotensin I converting enzyme 2; POLR2A/Polr2a, RNA polymerase II subunit A; TMPRSS, transmembrane serine protease.

Fig. 3.

Steady-state TMPRSS11D protein expression levels in human tracheal epithelial cells and in the lungs of mice or exposed elevated oxygen levels. A: newborn mouse pups were exposed to either an FiO₂ of 0.21 or 0.85 for the first 14 days of postnatal life, after which lungs were harvested for protein isolation. B: primary human tracheal epithelial cells were cultured on an air-liquid interface and exposed to a 21% O₂ or an 85% O₂ incubator headspace environment, and processed for protein isolation. Protein extracts were resolved on a 12% reducing SDS-PAGE gel, electroblotted to nitrocellulose, and TMPRSS11D was detected by immunoblot. Blots were reprobed for both β-actin (which has the same formula mass as TMPRSS11D), and β-tubulin, which served as controls for loading equivalence. The uncropped blots are provided in Supplemental Fig. S1 (https://doi.org/10.6084/m9.figshare.12855341.v1). TMPRSS, transmembrane serine protease.

Impact of elevated oxygen levels on SARS-CoV-2 receptor expression.

A pilot range-finding study was used to identify the minimum duration of oxygen exposure that was required to induce perturbations to co-receptor gene expression (Fig. 2A). Exposure of mouse alveolar epithelial type II cells to 85% O₂ in vitro increased Tmprss11d expression 4.4-fold by 48 h (Fig. 2A) without affecting Ace2, Tmprss1, or Tmprss2 expression. In human airways, lung and gut epithelial cells, TMPRSS2 expression was consistently increased in tracheal (2-fold), esophageal (1.7-fold), bronchial (1.5-fold), and alveolar (2-fold) epithelial cells after 48 h exposure to 85% O₂ (Fig. 2B). TMPRSS11D expression was increased (4-fold) in tracheal epithelial cells. However, no increase in the protein expression of TMPRSS11D was detected in tracheal epithelial cells after a 48-h 85% O₂ exposure (Fig. 3B see also Supplemental Fig. S1; https://doi.org/10.6084/m9.figshare.12855341.v1). These data suggest that while a 48-h hyperoxia-exposure period is sufficient to increase TMPRSS11D mRNA transcript levels, an impact of hyperoxia on steady-state protein expression levels may require a longer oxygen-exposure period. There was no impact of 85% O₂ exposure on the expression of ACE2 or TMPRSS1 (Fig. 2B).

Fig. 2.

Steady-state levels of mRNA transcripts encoding severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) entry receptors and co-receptors in airways and lung epithelial cells. A: steady-state mRNA levels were determined by real-time RT-PCR for Ace2, Tmprss1, Tmprss2, and Tmprss11d in cDNA prepared from adult C57BL/6J mouse alveolar epithelial type II cells (n = 4 independent cell-cultures per group) incubated on an air-liquid interface under room air (21% O₂: yellow bars) or hyperoxia (85% O₂: blue bars) conditions for 24, 48, or 72 h. B: steady-state mRNA levels were similarly determined for ACE2, TMPRSS1, TMPRSS2, and TMPRSS11D in adult human nasal, tracheal, esophageal, bronchial, and alveolar epithelial cells (n = 4 independent cell-cultures per group) incubated on an air/liquid interface under room air (21% O₂: yellow bars) or hyperoxia (85% O₂: blue bars) conditions for 48 h. Data reflect mean ΔCt ± SD. Pairwise comparisons were made between the 21% O₂ and 85% O₂ groups by unpaired Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. n.d., not detected. ACE2, angiotensin I converting enzyme 2; POLR2A/Polr2a, RNA polymerase II subunit A; TMPRSS, transmembrane serine protease.

Changes in the expression of the SARS-CoV-2 entry receptor ACE2 and the virus co-receptors TMPRSS1, TMPRSS2, and TMPRSS11D are believed to influence the risk of severe COVID-19. This has led to speculation that air pollution and smoking, which upregulate ACE2 expression in the respiratory epithelium, may exacerbate the course of COVID-19 in affected patients. Consistent with other reports that analyzed adult human respiratory tract tissue (6), the analyses reported here reveal that in cultured epithelial cells from trachea, bronchi, and alveoli, TMPRSS2 was more abundant at baseline than was ACE2; and that ACE2 was most abundant in the nasal epithelium compared with more distal respiratory tract epithelium. Furthermore, baseline abundance of TMPRSS2 and ACE2 were comparable between cultured tracheal, bronchial, and alveolar epithelial cells.

The key question posed in this study: “does increased FiO₂ drive SARS-CoV-2 entry receptor or co-receptor expression?” was addressed using lung tissue from patients with BPD, as well as an experimental animal model of BPD. This was neither a BPD nor a pediatric study. Rather, clinical or experimental BPD was employed as a model of supplemental oxygen support with lengthy use of an FiO₂ of >0.5. The data presented here document that elevated FiO₂ increased expression of TMPRSS2 and TMPRSS11D, which are required for efficient viral entry of SARS-CoV-2 (13). TMPRSS11D in particular has emerged in this study as being oxygen-responsive in both mouse and human airways and lungs. These data are potentially important, given that COVID-19 patients with acute respiratory failure are managed with oxygen support, which might increase viral co-receptor expression in the airways and lung epithelium of affected patients, thus accelerating the infection cycle. High-risk patients such as elderly (or young) patients with chronic lung disease in which home oxygen therapy forms part of routine disease management, or preterm born survivors of BPD discharged on home oxygen, may be similarly affected. These data highlight the need for further studies that examine the relative contribution of the various viral co-receptors on the infection cycle, and point to oxygen supplementation as a potential risk factor for COVID-19.

GRANTS

W.S. and R.E.M. were supported by the Max Planck Society; I.V., S.H., W.S., and R.E.M. were supported by the German Center for Lung Research (Deutsches Zentrum für Lungenforschung; DZL) and the German Research Foundation (Deutsche Forschungsgemeinschaft; DFG) through EXC2026 (390649896) and KFO309 (284237345). R.E.M. and W.S. were supported by the DFG through SFB1213 (268555672). R.E.M. was supported by the DFG through Mo 1789/1-1 (160966624) and Mo 1789/4-1 (420759458). S.H. and W.S. were supported by the DFG through SFB-TR84 (114933180).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.M., M.G., F.C., S.K.R., C.N., I.V., S.H., G.P., W.S., and R.E.M. conceived and designed research; D.M., M.G., F.C., S.K.R., and C.N. performed experiments; D.M., M.G., F.C., S.K.R., C.N., I.V., S.H., G.P., W.S., and R.E.M. analyzed data; D.M., M.G., F.C., S.K.R., C.N., I.V., S.H., G.P., W.S., and R.E.M. interpreted results of experiments; D.M., M.G., F.C., S.K.R., and C.N. prepared figures; D.M., M.G., F.C., S.K.R., C.N., and R.E.M. drafted manuscript; D.M., C.N., and R.E.M. edited and revised manuscript; D.M., M.G., F.C., S.K.R., C.N., I.V., S.H., G.P., W.S., and R.E.M. approved final version of manuscript.