Angiotensin-converting enzyme 2 in peripheral lung club cells modulates the susceptibility to SARS-CoV-2 in chronic obstructive pulmonary disease.

Accumulating evidence has confirmed that chronic obstructive pulmonary disease (COPD) is a risk factor for development of severe pathological changes in the peripheral lungs of patients with COVID-19. However, the underlying molecular mechanisms remain unclear. Because bronchiolar club cells are crucial for maintaining small airway homeostasis, we sought to explore whether the altered susceptibility to SARS-CoV-2 infection of the club cells might have contributed to the severe COVID-19 pneumonia in COPD patients. Our investigation on the quantity and distribution patterns of angiotensin-converting enzyme 2 (ACE2) in airway epithelium via immunofluorescence staining revealed that the mean fluorescence intensity of the ACE2-positive epithelial cells was significantly higher in club cells than those in other epithelial cells (including ciliated cells, basal cells, goblet cells, neuroendocrine cells, and alveolar type 2 cells). Compared with nonsmokers, the median percentage of club cells in bronchiolar epithelium and ACE2-positive club cells was significantly higher in COPD patients. In vitro, SARS-CoV-2 infection (at a multiplicity of infection of 1.0) of primary small airway epithelial cells, cultured on air-liquid interface, confirmed a higher percentage of infected ACE2-positive club cells in COPD patients than in nonsmokers. Our findings have indicated the role of club cells in modulating the pathogenesis of SARS-CoV-2-related severe pneumonia and the poor clinical outcomes, which may help physicians to formulate a novel therapeutic strategy for COVID-19 patients with coexisting COPD.

INTRODUCTION

Peripheral airways contain an abundance of terminal bronchioles, giving rise to respiratory bronchioles that divide into the ducts that extend to the alveoli to facilitate gas exchange. Club cells are secretory cells that are highly prevalent in bronchioles and are crucial for maintaining small airway homeostasis via promoting self-renewal, mucociliary clearance, and secretion of club cell 10-kDa protein (CC10), a member of the secretoglobin family with anti-inflammatory properties (1, 2). Recently, single-cell RNA sequencing for three-dimensional human distal lung culture revealed that angiotensin-converting enzyme (ACE2), a membrane-bound protein and host receptor for the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is predominantly localized within the club cells and alveolar type 2 (AT2) cells (3, 4). Club cells have also been further identified as the cell population targeted by SARS-CoV-2 in animal studies (5, 6), but the roles of club cells have not been fully elucidated.

Comorbid chronic obstructive pulmonary disease (COPD) has been associated with more severe pneumonia in patients with COVID-19, possibly because of the increased ACE2 expression (7–10). However, the association between the characteristics of airways epithelial cells and the severity of SARS-CoV-2 infection remains unclear. We hypothesized that the dysregulated ACE2 expression in club cells might have predisposed to more severe SARS-CoV-2 infection in the lung periphery, where the pathological changes mainly take place in COPD patients with COVID-19. To this end, we sought to investigate the distribution patterns and quantity of ACE2 via immunofluorescence (IF) staining with specific endothelial cells (ECs) (ciliated cells, club cells, basal cells, goblet cells, neuroendocrine cells, and AT2 cells) markers in the bronchiolar and alveolar epithelium. We further performed air-liquid interface (ALI) culture of the primary human small airway epithelial cells (hSAECs) to explore SARS-CoV-2 cell tropism. Our results provided novel insights into the roles of club cells in SARS-CoV-2 infection and might shed light on the development of therapeutic interventions for COVID-19 patients with coexisting COPD.

MATERIALS AND METHODS

See details in the Supplemental Data (all Supplemental material is available at https://doi.org/10.6084/m9.figshare.18070460.v1).

Subject Recruitment and Sample Processing

Studyprotocol approval was obtained from the institutional review boards of The First Affiliated Hospital of Guangzhou Medical University. All study participants provided written informed consent. A total of 82 nonsmokers and 35 current or former smokers were recruited from the subjects without airflow limitation who underwent lung resection for solitary peripheral carcinoma. We also recruited 23 current or former smokers with severe to very severe COPD, 18 of whom were lung transplant recipients and 5 of whom underwent lung resection for solitary peripheral carcinoma. Among patients with carcinoma, all biopsies were randomly obtained from the peripheral tissues at least 1 cm away from the tumor.

To validate the specificity of ACE2 antibodies and perform positive control staining, we recruited samples from patients who underwent kidney (n = 6) or gallbladder resection (n = 4). To detect ACE2 expression in nasal and bronchial ECs, we obtained the inferior turbinate from patients undergoing septoplasty due to anatomic variation (n = 6) and bronchi from patients undergoing lung resection for solitary peripheral carcinoma (n = 9), respectively.

Immunofluorescence Staining and Immunohistochemistry Assays

Tissue sections and the membranes of Transwell inserts were used to perform IF staining. Three-micrometer-thick tissue sections were dewaxed in xylene, rehydrated in graded alcohols, and rinsed in distilled water. Sections were then subject to heat-induced antigen retrieval in Tris-EDTA buffer (pH 9.0) at 95°C for 15 min in microwave oven and cooled at room temperature. To confirm successful staining and to validate the specificity and sensitivity of ACE2 antibodies, three monoclonal antibodies (mAbs; including MAB933, sc-390851, and 66699-1) and three polyclonal antibodies (pAbs; including HPA000288, ab15348, and 21115-1-AP) were used for evaluating ACE2 protein expression (Supplemental Fig. S1). Protein expression of ACE2 (1:150, anti-rabbit, HPA000288, Sigma) was analyzed by using ImageJ software through calculating the mean fluorescence intensity (MFI) of the positively stained cells in different organs (11). The information of primary antibodies is listed in Supplemental Table S1.

For immunohistochemistry assays, staining of ACE2 (1:150, anti-rabbit, HPA000288, Sigma) was performed using a modified horseradish peroxidase (HRP) technique with Dako EnVision+ System-HRP (Dako A/S, Denmark).

The staining specificity was confirmed by utilizing the isotype-matched immunoglobulin controls including rabbit polyclonal IgG antibody (ab37415, abcam), rabbit monoclonal IgG antibody (ab125938, abcam), mouse monoclonal IgG1 antibody (ab170190, abcam), mouse monoclonal IgG2a antibody (ab18415, abcam), and mouse monoclonal IgG2b antibody (ab18469, abcam).

Protein Expression Profiling of Bronchiolar Epithelium and Alveoli

To evaluate the bronchiolar areas, we enumerated 200 ECs in 5 randomly selected high-power fields (HPFs). We calculated the percentage of ACE2+ cells by performing staining with specific markers, including α-tubulin+/FOXJ1+ for ciliated cells, KRT5+ for basal cells, CC10+ for club cells, chromogranin A+ for neuroendocrine cells, and MUC5AC+ for goblet cells. The percentage was calculated by dividing the number of the positively stained cells by 200 ECs and multiplied by 100%, respectively. To evaluate the alveolar areas, we also calculated the percentage of ACE2+ cells in AT2 cells (labeled with SPC+) in five randomly selected HPFs.

Primary Cell Culture and SARS-CoV-2 Infection

Primary hSAECs were isolated from freshly resected peripheral lung tissues, and human bronchiolar epithelial progenitor cells were cultured in the cell culture system, with the cell medium being changed every 3 days. The cells were cultured in air-liquid interface (ALI) condition for ∼4 wk until fully differentiated (12, 13). The cells of ALI culture on the apical chamber of Transwell inserts were infected with SARS-CoV-2 viral inoculum at a multiplicity of infection (MOI) of 0.1 or 1.0 (GenBank: MT123290) (14). The inoculated plates were incubated for 2 h at 37°C with 5% CO2. At the end of incubation, the inoculum was removed from the apical chamber. At 24, 48, and 72 h postinfection, the viral RNA was extracted from the cell supernatant by using the EZ-press Viral RNA Purification Kit (EZB-VRN1, EZBioscience, Roseville, MN). The copy number of SARS-CoV-2 N gene was determined with the use of novel coronavirus (2019-nCoV) nucleic acid diagnostic kit (PCR-Fluorescence Probing) (Daan Gene, Guangzhou, China) according to the manufacturer’s protocol. We found that as compared with the MOI of 1.0, MOI of 0.1 led to a lower degree of viral replication at each time point, but both of the MOIs can generate replicative infection in our ALI model (Supplemental Fig. S2).

In addition, the infected cells were harvested for assessing N-protein expression by using SARS-CoV antibodies via IF staining and the mRNA expression levels of type I (IFN-α and IFN-β) and type III interferon (IFN-λ1) and inflammatory markers (TNF-α, IFN-γ, IL-6, IL-8, andIL-13) by performing PCR assays.

Scanning Electron Microscopy and Transmission Electron Microscopy

Tissue specimens and ALI culture samples were prepared according to a standard protocol for scanning electron microscopy (SEM) or transmission electron microscopy (TEM). The surface of epithelium was examined with FEI Quanta 250 FEG (SEM) or Hitachi JEM-1400 PLUS (TEM).

RNA Extraction and Quantitative Real-Time PCR

Total RNA was extracted from frozen lung tissues in RNA later using the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) and reverse transcribed into cDNA using Maxima Reverse Transcriptase Kit (Thermo Fisher Scientific, Waltham, MA) according to manufacturer’s protocol. The mRNA expression levels of ACE2, dACE2, TMPRSS2, FURIN, basigin (BSG), cathepsin L, integrin subunit alphav (ITGAV), IFN-α, IFN-β, IFN-λ1, TNF-α, IFN-γ, IL-6, IL-8, and IL-13 were assessed with quantitative real-time PCR. The details of primer sequences are presented in Supplemental Table S2. The relative gene expression was calculated by using the comparative 2−ΔCt method, which was normalized against the housekeeping gene normalized against glyceraldehyde 3-phosphate dehydrogenase.

Statistical Analysis

Statistical analyses were conducted with SPSS21.0 software (IBM, Chicago, IL) and GraphPad Prism 6 (GraphPad Software, La Jolla, CA). The Kolmogorov-Smirnov tests and Shapiro-Wilk tests revealed that the data were not normally distributed. The Mann-Whitney two-sided nonparametric test was used as appropriate to compare the continuous variables between two groups. Correlation analysis was performed with Spearman’s model. P < 0.05 was deemed statistically significant for all analyses.

RESULTS

Subject Characteristics

The clinical characteristics of nonsmokers, smokers, and COPD patients are summarized in Table 1. COPD patients had significantly higher pack-years of smoking than smokers without airflow limitation (P < 0.001). Most smokers and COPD patients were male and older, compared with nonsmokers (both P < 0.001). Of the 23 COPD patients, 12 were users of inhaled corticosteroids (ICS).

Table 1.

Baseline characteristics of study participants

	Nonsmokers	Smokers without Airflow Obstruction	Patients with COPD	P Value
Subjects, n	82	35	23
Sex (M/F), n	28/54	33/2	23/0	P < 0.01
Age, yr	47.9 ± 11.0	59.7 ± 10.9	64.2 ± 5.6	P < 0.01
Body mass index, kg/m2	22.36 ± 3.19	22.85 ± 4.63	20.90 ± 2.90	P < 0.01
Smoking history, pack-year		32.9 ± 21.0	58.9 ± 38.5	P < 0.01
Smoking status (Current/Former)*		25/10	19/4	P = 0.37
ICS users (Y/N)	0/82	0/35	11/12	P < 0.01
FEV1, %predicted	104.8 ± 12.0	100.1 ± 13.3	38.1 ± 16.8	P < 0.01
FEV1/FVC, %	84.5 ± 6.6	79.0 ± 5.8	49.4 ± 16.2	P < 0.01
GOLD classification of severity of COPD, n (%)
Mild			0 (0)
Moderate			5 (21.74)
Severe			7 (30.43)
Very severe			11 (47.83)

Values are means ± SE. COPD, chronic obstructive pulmonary disease; F, female; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; GOLD, Global Initiative for Chronic Obstructive Lung Disease; ICS, inhaled corticosteroids; M, male; N, no; Y, yes. *Current smoking indicates that the participant reported smoking for the 1 yr before the baseline visit.

Identification and Quantification of ACE2+ ECs in the Whole Airways

To explore the distribution and characteristics of ACE2 protein within the airways, we examined its expression in the ECs of inferior turbinate (Fig. 1), bronchi (Fig. 1), bronchioles (Fig. 1), and alveoli (Fig. 1) of the lung and calculated the mean fluorescence intensity (MFI) of the positively stained area. We found significantly higher MFI of positively stained ECs in bronchioles than those in the inferior turbinate, bronchi, and alveoli (all P < 0.001) (Fig. 1). We then examined the association between ACE2 expression and the expression of bronchiolar ECs markers (CC10, α-tubulin, KRT5, MUC5AC, and chromogranin A) in nonsmokers (n = 70) (Fig. 1). Within the bronchiolar epithelium, we found that ACE2 was expressed in 48.4% (2115/4368) of club cells, 0.3% (22/7239) of ciliated cells, 1.7% (42/2466) of basal cells, 1.1% (2/184) of goblet cells, and 0% (0/77) of neuroendocrine cells (Fig. 1).

Figure. 1.

Identification and quantification of the angiotensin-converting enzyme 2-positive (ACE2+) epithelial cell populations in the peripheral airways. Representative hematoxylin-eosin images and immunofluorescent images of the ACE2 protein expression (white arrow) in epithelial areas derived from the human inferior turbinate (A), bronchi (B), bronchioles (C), and alveoli (D), respectively are shown. Significantly higher mean fluorescence intensity (MFI) of ACE2 expression was detected in the bronchioles than in the inferior turbinate, bronchi, and alveoli (E). Within the bronchiolar epithelium, ACE2 protein was ubiquitously expressed in club cells but rarely expressed in ciliated cells, basal cells, goblet cells and neuroendocrine cells (F). The positive control staining of ACE2 protein was performed in the kidney and gallbladder (G), and the MFI of ACE2 was consistent with that of the bronchiole (H). The negative control staining in the bronchiole was demonstrated (I). DAPI labeling of nuclei is in blue. CC10, club cell 10-kDa protein; KRT5, keratin 5; MUC5AC, mucin 5AC; PNEC, pulmonary neuroendocrine cells. ****P < 0.0001.

We also examined the association between the expression of ACE2 and ECs markers in pseudostratified columnar epithelium from the inferior turbinate (n = 6) and bronchi (n = 9). In nasal and bronchial epithelium, we found that ACE2 was commonly expressed in basal cells (64.8% and 70.2%), infrequently expressed in ciliated cells (7.1% and 8.5%), and completely absent in goblet (0% and 0%) and neuroendocrine cells (0% and 0%). No significant difference of ACE2 expression was found between the nasal and bronchial epithelium (all P > 0.05) (Supplemental Fig. S3, A–F).

Altered ACE2 Expression among ECs and CC10+ Cells in Smokers and COPD Patients

We next performed IF staining to enumerate the club cells within the bronchiole in nonsmokers, smokers, and COPD patients. As shown in Fig. 2, , the median (the 1st, 3rd quartile) percentage of club cells was significantly higher in patients with COPD [38.0% (32.5%, 44.0%)] than in nonsmokers [30.7% (27.8%, 35.6%)] and smokers [30.0% (26.4%, 34.6%)] (both P < 0.05) but was comparable between nonsmokers and smokers (P > 0.05).

Figure 2.

Angiotensin-converting enzyme 2-positive (ACE2+) club cell count within the bronchiolar epithelium from smokers and comorbid chronic obstructive pulmonary disease (COPD) patients. Scanning electron microscopy images indicated an increased percentage of club cells (green) in bronchiole from COPD patients compared with nonsmokers and smokers (A–C and G). Compared with nonsmokers, the percentage of ACE2+ endothelial cells (ECs) and percentage ACE2+ club cells were both significantly increased in the bronchiolar epithelium of smokers and COPD patients (D–F, H, and I). ACE2+ EC count was significantly increased in COPD patients compared with smokers (H). In addition, the mRNA expression levels of ACE2 were significantly increased both in smokers and COPD patients (J). DAPI labeling of nuclei is in blue. CC10, club cell 10-kDa protein.

Having revealed a greater proportion of club cells in COPD patients than in nonsmokers and smokers (Fig. 2, ), we then explored the expression of ACE2 in bronchiolar ECs (Fig. 2, ). In nonsmokers, the median (the 1st, 3rd quartile) percentage of ACE2+ ECs was 14.0% (9.0%, 20.3%), which was significantly lower than that in smokers [21.0% (14.0%, 31.0%)] and COPD patients [29.0% (19.0%, 39.0%)] (all P < 0.05) (Fig. 2). We next determined the percentage of ACE2+ club cells. As shown in Fig. 2, the percentage of ACE2+ club cells was significantly higher in smokers (median: 81.8% vs. 49.6%, P < 0.05) and COPD patients (median: 77.1% vs. 49.6%, P < 0.05) than in nonsmokers but was comparable between smokers and COPD patients (P = 0.39).

We further explored whether the levels of ACE2 expression (including the expression in ACE2+ ECs and ACE2+ club cells) could be explained by the key clinical characteristics (including age, sex, and body mass index) of nonsmokers and the smoking status of smokers (Supplemental Fig. S4). We found that the levels of ACE2 expression did not correlate with most of these clinical characteristics, except that the percentage of ACE2+ ECs and ACE2+ club cells correlated negatively with age (Supplemental Fig. S4, E and F). Among COPD patients, the median (the 1st, 3rd quartile) percentage of club cells in bronchiolar epithelium cells (ECs) was 33.7% (29.4%, 39.9%) in ICS users, which was significantly lower than that of ICS nonusers [40.6% (38.0%, 46.4%)] (P = 0.009, Supplemental Fig. S5A). However, the median percentage of ACE2+ ECs and ACE2+ club cells was comparable between ICS users and ICS nonusers in COPD patients, respectively (both P > 0.05, Supplemental Fig. S5, B and C). Furthermore, the median percentage of ACE2+ ECs, ACE2+ club cells, and club cells in ECs was comparable between current smokers and ex-smokers of COPD patients, respectively (Supplemental Fig. S5, D–F).

We also specifically evaluated ACE2 expression by costaining with SPC, which revealed no remarkable difference in AT2 cell count and in the median percentage of ACE2+ AT2 cells between nonsmokers, smokers, and COPD patients (Supplemental Fig. S6).

We next performed quantitative real-time PCR on peripheral lung tissues derived from nonsmokers (n = 72), smokers (n = 31), and COPD patients (n = 11). Compare with nonsmokers, ACE2 mRNA levels were markedly higher in COPD patients (P < 0.05) (Fig. 2). The mRNA level of dACE2, a novel truncated isoform of ACE2, was markedly higher in COPD patients than in nonsmokers and in smokers, but was comparable between nonsmokers and smokers (Supplemental Fig. S7A). Furthermore, the mRNA expression levels of BSG, cathepsin L, and ITGAV were consistently reduced in COPD patients as compared with nonsmokers, while the expression levels of TMPRSS2 and FURIN were not significantly dysregulated (Supplemental Fig. S7, B–F).

Characterization of hSAECs after ALI Culture and SARS-CoV-2 Infection

Next, we cultured primary hSAECs derived from nonsmokers (n = 4) and COPD patients (n = 4) on ALI conditions, followed by infection with SARS-CoV-2 at a multiplicity of infection (MOI) of 1.0 for 1, 2, and 3 days, respectively (Fig. 3). Before SARS-CoV-2 infection, ACE2 was expressed in 3.5% of hSAECs in nonsmokers, which was significantly lower than in COPD patients (8.4% of hSAECs) (Fig. 3, and ). We then examined the ALI culture by performing IF staining for SARS-CoV-2 nucleocapsid protein (NP) and ACE2 to demonstrate viral replication. We identified a subset of NP (green) that colocalized with ACE2 (red) (Fig. 3, and ). There was a trend toward an increase in the percentage of NP+ hSAECs, a decrease in the number of ACE2+ hSAECs postinfection (Fig. 3), and an increase in NP+ hSAECs in COPD patients, compared with that in nonsmokers (Fig. 3).

Figure 3.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infects the angiotensin-converting enzyme 2-positive (ACE2+) cell and induces a loss of ACE2 protein expression in human small airway epithelial cells. Primary human small airway epithelial cells were cultured in the air-liquid interface (ALI) system and then infected with SARS-CoV-2 at an multiplicity of infection (MOI) of 1.0 for 1, 2, and 3 days, respectively (A). Before SARS-CoV-2 infection, ACE2 protein was expressed in club cells (B and C). A subset of SARS-CoV-2 NP (green) was costained with ACE2 (red) in 1-day postinfection (D and E). There was a trend toward an increase in the percentage of nucleocapsid protein-positive (NP+) cells and a decrease in the number of ACE2+ cell postinfection (F), and the number of NP+ cells was increased in COPD patients than in nonsmokers (G). DAPI labeling of nuclei is in blue. CC10, club cell 10-kDa protein.

We further performed IF staining on ALI cultures maintained across different time points to examine the association between NP expression and the expression levels of hSAECs markers (CC10, FOXJ1, and KRT5) after SARS-CoV-2 infection. Consistent with the results from a recent study documenting SARS-CoV-2 infection in human distal lung organoids (4), we found that NP partially colocalized with club cells (CC10+) but barely colocalized with ciliated cells (FOXJ1+) and basal cells (KRT5+) (Fig. 4, ). Finally, we confirmed that club cells were infected with SARS-CoV-2 (virus-like particles appearing like open sacs within the cytoplasm or scattering on the top of club cells) by using both the TEM (Fig. 4) and SEM (Fig. 4, ).

Figure 4.

Club cells are major target of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in human small airway epithelial cells (hSAECs). In primary hSAECs from air-liquid interface (ALI) culture with SARS-CoV-2 infection at an multiplicity of infection (MOI) of 1.0 for 2 days, there was costaining of nucleocapsid protein (NP) with club cell marker club cell 10-kDa protein (CC10; A) but not with cilia cell marker forkhead box J1 (FOXJ1; B) or basal cell marker keratin 5 (KRT; C). DAPI labeling of nuclei is in blue. Transmission electron microscopy revealed some virus-like particles as open sacs in the cytoplasm of club cells in hSAECs (D). Scanning electron microscope showed a normal club cell before SARS-CoV-2 infection (E), the top of 2 club cells that were infected with virus-like particles (F and G), and a club-like cell that was damaged with a large number of virus-like particles (H).

DISCUSSION

Peripheral lung injury is common in patients with COVID-19 and can trigger a cascade of events that lead to acute respiratory distress syndrome. The peripheral lung (which, in our study, was derived from patients who underwent surgical lung resection) is an important tissue for evaluation of the characteristic of ACE2 expression in the human bronchioles and alveoli. By performing IF assays with a relatively large sample size of surgically resected tissues, our study has further demonstrated the detection and cell type-specific expression of ACE2 protein in human bronchiole and alveoli. We also showed that club cells are an important target of SARS-CoV-2 infection, in light of the abundance of ACE2 expression. Recent autopsy studies have revealed hypersecretion of airway mucus in small airways of patients with COVID-19, indicating that secretory cells (e.g., club cells) might have been directly infected by SARS-CoV-2 (15, 16). We found that ACE2 is not detectable in goblet cells from either the inferior turbinate or from the bronchi (Supplemental Fig. S2); therefore, mucus hypersecretion in patients with COVID-19 might be mainly related to the infection and injury to club cells. We have further reported that the expansion of ACE2+ club cells in COPD might be a consequence of exposure to cigarette smoke. Our results provided new insights into the pathogenesis of SARS-CoV-2-related pneumonia, which may help physicians to formulate novel therapeutic strategy for patients with COVID-19 who had coexisting COPD.

To simulate the in vivo scenario of SARS-CoV-2 infection, we adopted the ALI culture for the in vitro system, unraveling the specific cell types being infected and the dynamic responses. The primary hSAECs obtained from the lung periphery during surgery or biopsy were subsequently differentiated in the ALI culture. This has allowed for our observations within a dynamic model of hSAECs that consisted of club cells. However, due to the lack of lung tissues from COPD patients infected with SARS-CoV-2, we could not clarify whether the increased expression of ACE2 in club cells would be associated with the severity of pneumonia. Apart from club cells, other epithelial cell types (such as bronchial ciliated cells, alveolar type 2 epithelial cells, and immune cells) could also be infected by SARS-CoV-2 (17–19). We believe that the pneumonia in patients with COVID-19 might be the cumulative effects driven by all the infected cells but not a standalone consequence of club cells.

Intriguingly, we noted a downregulation of ACE2 expression after SARS-CoV-2 infection in vitro. The underlying mechanisms are likely multifactorial. First, ACE2 might have been cleaved by transmembrane serine protease 2 of the host cells during the infection with SARS-CoV-2 (20, 21). Second, splicing, translation, and protein trafficking of ACE2 might have been disrupted by SARS-CoV-2 (22). Furthermore, the cell membrane (localized with ACE2) of the infected cells could have been damaged by SARS-CoV-2 infection. Two contrasting functions of ACE2 have been identified: 1) an endogenous counter-regulator of the renin-angiotensin system, and 2) a cellular receptor for both SARS-CoV and SARS-CoV-2 (23). Evidence from multiple models has revealed the protective effects of ACE2 against the lipopolysaccharide-induced acute lung injury (24–26). The binding to and entry of both SARS-CoV and SARS-CoV-2 into human cells could be facilitated by the interaction between receptor-binding domain of the S1 subunit on the viral spike glycoproteins with ACE2 (27). Although ACE2 was upregulated in the lung tissues and serum of patients with acute respiratory distress syndrome regardless of the SARS-CoV-2 infection status (28), it was unclear to what extent the ACE2 would affect lung injury after SARS-CoV-2 infections. Nonetheless, our findings inspire ongoing studies to elucidate the association between ACE2 expression patterns and SARS-CoV-2 infection both in vivo and in vitro.

These findings have been corroborated by the accumulating cross-sectional studies or epidemiological data, which demonstrated that current smokers and COPD patients are at higher risk of developing severe complications of COVID-19 (29–32). Our study showed smoking and comorbid COPD significantly upregulated both the mRNA and protein expression levels of ACE2 in the lung periphery of Chinese populations. This finding was consistent with that of Smith et al. (33), who reported a 42% increase of ACE2 mRNA expression in the lung of cigarette smokers, compared with that of nonsmokers. Cai et al. (34) found that smoking increased pulmonary ACE2 expression by 25%. By modeling the direct effects of cigarette smoke on SARS-CoV-2 infection of the large airway epithelium in ALI culture, Purkayastha and colleagues (35) concluded that acute exposure to cigarette smoke increased the number of the infected and apoptotic cells, prevented the normal airway basal stem cells from repair, and blunted the innate immune responses. Taken together, our findings provided further evidence that smoking could upregulate ACE2 expression in the peripheral airway epithelium and, more specifically, in club cells.

Several studies have reported that interferon (IFN) can modulate the expression of ACE2 in the airway epithelial cells (36–39). To confirm this finding in our SARS-CoV-2 infection ALI culture model of small airway epithelial cells, we have performed additional measurements on the expression of type-I (IFNα and IFNβ) and type-III IFN (IFNλ1) by performing quantitative PCR assays. All three IFNs were upregulated upon SARS-CoV-2 infection at 48 h in our ALI model (Supplemental Fig. S8, A–C). Besides, the expression levels of all these three IFNs correlated positively with the mRNA levels of ACE2, suggesting that ACE2 expression might be induced by the IFNs (Supplemental Fig. S8, D–F). In addition, we have also conducted experiments to detect the expression levels of several inflammatory factors in our infection model by performing qPCR assays. We found that the expression levels of TNF-α, IFN-γ, IL-6, IL-8, and IL-13 were elevated upon SARS-CoV-2 infection in hSAECs of ALI culture from nonsmokers and COPD patients (Supplemental Fig. S9). Notably, the levels of inflammatory cytokines were higher in nonsmokers than in COPD patients, especially at 72 h after infection. This finding indicates that the immune response in hSAECs from COPD patients was partially suppressed, which may also explain the greater viral loads in COPD hSAECs.

Several studies have examined the localization of ACE2 within the respiratory tract but reported inconsistent results (40–42). To address this technical issue, we further investigated the specificity and sensitivity of six different antibodies against human ACE2 based on the topological map (Supplemental Fig. S1M). Our findings showed that PAbs had higher sensitivity than mAbs did in the gallbladder and peripheral lung tissues (Supplemental Fig. S1N). This was probably because of a larger number of B-cell clones responding to a specific epitope, thus changes in epitopes (e.g., glycosylation or denaturation) might not have affected the binding capacity of pAbs. Furthermore, pAbs would be more stable over a broad range of pH and salt concentration than mAbs (43). In addition, SARS-CoV-2 receptor-binding domain reportedly binded to the NH2-terminal peptidase domain of ACE2 (44). Therefore, the HPA000288 polyclonal antibody (Sigma-Aldrich, St. Louis, MO), which targeting 111 NH2-terminal amino acids of ACE2 (45), was selected for further experiments (Supplemental Fig. S1, M and O).

However, there are several limitations of this study. First, most patients underwent resection for solitary peripheral carcinoma or lung transplant, which might have introduced a bias into our results. Ideally, sampling the tissues from the donors with lung transplantation as the control would be preferred. Second, only the relatively severe COPD patients were studied, which limited the external generalizability of the results. Third, we included a wide age range of healthy subjects with the median age lower than that in pathological subjects, although the percentage of ACE2+ ECs and ACE2+ club cells correlated negatively with age in nonsmokers. Further cross-sectional studies are needed to confirm the impact of age on ACE2 expression in club cells in pathophysiological conditions. Fourth, it would be more accurate to report the expression levels for the whole airway of each patient, but it would be challenging to obtain different generations (e.g., trachea, proximal bronchi, distal bronchi) of the whole airway tissue from patients in real-world practice. Finally, our airway epithelium model lacks both resident and infiltrating immune cells (dendritic cells, macrophages, and lymphocytes, etc.) and therefore cannot help us study the impact of immune responses on the SARS-CoV-2 infections.

In conclusion, our study reveals that club cells are a key source of ACE2 in human peripheral airways and helps explain the mechanisms of severe COVID-19 and the poor clinical outcomes in COPD patients infected by SARS-CoV-2.

DATA AVAILABILITY

The data that support the findings of this study will be made available upon reasonable request from the corresponding author.

Supplemental Information, Supplemental Figs. S1–S9, and Supplemental Tables S1 and S2: https://doi.org/10.6084/m9.figshare.18070460.v1.

GRANTS

This work was supported by the National Natural Science Foundation (No. 81870003), Guangdong Natural Science Foundation (No. 2019A1515011634), and Guangzhou Institute for Respiratory Health Open Project (funded by China Evergrande Group) Project No. 2020GIRHHMS09 and 2020GIRHHMS19 (to W.-J.G.); The National Medical Research Council, Singapore (NMRC/CIRG/1458/2016 and MOH-COVID19RF2-0001) (to D.-Y.W.); as well as the Impact and Mechanisms of Physical, Chemical and Biological Interventions on the Development and Outcome of Acute Lung Injury (No. 81490534), National Key Technology R&D Program (No. 2018YFC1311902), and Guangdong Science and Technology Foundation (No. 2019B030316028) (to N.-S.Z).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

J.-C.Z., D.-Y.W., and N.-S.Z. conceived and designed research; Y.P., Z.-N.W., S.-Y.C., A.-R.X., J.S., Z.-Q.Z., X.-T.H., L.-J.C., and J.-J.M. performed experiments; Y.P., Z.-N.W., and Z.-F.F. analyzed data; Y.P., Z.-N.W., S.-Y.C., A.-R.X., Z.-F.F., J.S., Z.-Q.Z., and X.-T.H. interpreted results of experiments; Y.P. prepared figures; Y.P. drafted manuscript; W.-J.G. and N.-S.Z. edited and revised manuscript; D.-Y.W. and N.-S.Z. approved final version of manuscript.