Collagen I from human skin, Pronase, DNase I, SBTI, accutase, and IgG from human serum were purchased from Sigma-Aldrich; PneumaCult™-EX PLUS, PneumaCult™-ALI were from STEMCELL Technologies; EASYstrainer Cell Sieves (100 and 40 μm), transwell inserts and 96 well plates with cell-repellent surface were from Greiner Bio-One; High Pure RNA Isolation kit, Transcriptor First Strand cDNA synthesis kit, SYBR Green PCR Master Mix were from Roche Diagnostics.

Antibodies for flow cytometry staining were ACE2-AlexaFluor 647 (clone E-11), AXL-AlexaFluor647 (clone B-2), acetylated α-Tubulin-AlexaFluor 488 (clone 6-11B-1), MUC5AC-AlexaFluor 488 (clone 45M1), CC10-AlexaFluor 488 (clone E-11), Cytokeratin 5-AlexaFluor 488 (clone RCK103) were purchased from Santa Cruz Biotechnology; Mouse IgG1-AlexaFluor 488, Mouse IgG1-AlexaFluor 647 were from BioLegend, Mouse IgG2b-FITC was from DAKO; and Live/Dead™ Fixable Blue Dead Cell Stain was from Invitrogen.

Lung tissue and/or bronchus from 11 patients were provided by the Pathology Department (Research Center Borstel, Germany). All patients underwent lung resection and were characterized as lung cancer stages ranging from IA to IIB. These were samples from residual material without a declaration of consent (AZ 12-220) with defined accompanying data. Here, the regulations of the Bio Material Bank Nord are applied (AZ 14-225). The study was conducted in accordance with the Declaration of Helsinki. The use of biomaterial and data for this study was approved by the local ethics committee of the University Lübeck (AZ 17-131). Sex, age, and smoking behavior, and additional information for every individual are provided in Table 1 and Supplementary Table S1.

Tissue was assembled during routine surgical intervention from lung cancer patients. Since it has been reported that there are significant difference in expression of SARS-CoV-2 entry genes between tumor and the adjacent tumor-free tissue (13), epithelial cells were isolated from tumor free tissues. Lung or bronchial tissue was placed in a 100 mm culture dish and rinsed in D-PBS. Tissue was cut in smaller pieces, excess surrounding tissue was dissected out, and bronchial tissue separated. A scalpel was used to cut open the bronchial pieces, and cut the tissue into 2–3 mm pieces, for better exposure to enzyme solution. Bronchial tissue was then transferred to a tube containing 20 ml of RPMI1640, supplemented with 2 mM L-glutamin, 1.4 mg/ml pronase, 0.1 mg/ml DNAse I and placed at 4–8°C for overnight incubation. The next day, FCS was added to 10% to stop the action of the protease, and the tube was inverted several times to mix well and dislodge cells. Bronchial tissue was discarded, and the cell-containing supernatant was transferred through EASYstrainer Cell Sieves (100 and 40 μm) to new tubes and centrifuged 10 min at 600 × g. Epithelial cells were then collected into a new tube and centrifuged after which pellet was resuspended in RPMI1640 to rinse cells and allow them to be counted. Cells were centrifuged again and the pellet was resuspended in expansion medium (PneumaCult EX PLUS™ Complete Medium), transferred to collagen I-coated (3.45 μg/cm²) 100 mm culture plate (150.000–300.000 cells per culture plate), and incubated at 37°C, 5% CO₂. Medium was changed every 3–4 days until cells reached desired confluency. Cells isolated from lung tissue (secondary/tertiary bronchi and bronchioles) were named small airway epithelial cells (SAEC) and cells isolated from main (primary) bronchus were named human bronchial epithelial cells (HBEC) (Supplementary Figure S1). All cells could be expanded to passage 8–11. Most of the experiments were done with cells from passage 3–6.

Primary human airway epithelial cells were expanded and differentiated at ALI in-vitro following the protocol given by STEMCELL Technologies. Cells were expanded on collagen I-coated (3.45 μg/cm²) dishes in expansion medium, and harvested by trypsinization with SBTI neutralization. The apical chambers of 6.5 mm transwell inserts were coated with 5 μg/cm² collagen I, and cells were seeded at 3 × 10⁵ cells/insert in expansion medium, and the basolateral chambers received 500 μL of expansion medium alone. Once the epithelial cells reached confluence, the apical growth media was removed, and the basolateral medium was replaced with PneumaCult™-ALI Maintenance Medium (ALI-MM), initiating day 0 of the ALI mucociliary culture system. Epithelial cultures were allowed to differentiate at ALI in-vitro for 14–21 days at 37°C, 5% CO₂. In the basolateral chambers ALI-MM was changed every other day. Differentiation of the cells was confirmed by immunohistochemistry (Supplementary Figure S2).

For the generation of spheroids 2–3 × 10⁶ cells were resuspended in 10 ml ALI-MM and seeded onto 96 well plates with cell-repellent surface (100 μl cells/well). Plates were centrifuged for 1 min at 450 × g and incubated at 37°C, 5% CO₂ for 4–7 days. The spheroids were featured by ciliated cells on the apical, outer side (Supplementary Figure S3).

For flow cytometry ALI or 3D cultured cells were used. ALI cultured cells were detached from transwell membranes using accutase, and 3D cultured cells were carefully resuspended. The surface markers of cells were analyzed for the presence of ACE2, AXL, and acetylated α-Tubulin using AlexaFluor 488- or AlexaFluor 647 conjugated monoclonal antibodies (all diluted 1/100 in D-PBS/Live-Dead stain (1:1000), supplemented with 60 μg/ml human IgG (Sigma Aldrich) at 8°C for 30 min. After that, cells were washed in FACS buffer (D-PBS/0.1% BSA), fixed in 4% PFA at room temperature (RT) for 15 min and permeabilized in 0.2% Triton X-100 at RT for 20 min. Intracellular markers MUC5AC, CC10 and cytokeratin 5 were analyzed using AlexaFluor 488 conjugated monoclonal antibodies (all diluted 1/100 in FACS buffer, supplemented with 60 μg/ml human IgG) at 8°C for overnight. The cells were washed and analyzed by flow cytometry (LSRII, BD Biosciences), and post-acquisition analysis was carried out using FCSExpress 7 (De Novo Software).

RNA was extracted from cell lysates of primary airway epithelial cells (either expanded in submerged cultures or differentiated in ALI cultures), by using the High Pure RNA Isolation kit, and reverse transcription of 0.5–1 μg of total RNA was performed using the Transcriptor First Strand cDNA synthesis kit. Amplifications of target and HPRT genes were performed using SYBR Green PCR Master Mix, following the manufacturer's instructions, in a LightCycler 480 System (Roche Diagnostics). Data analyses were done using the LightCycler 480 relative quantification software with data normalized to the mRNA level of HPRT housekeeping gene. The sequences of the primer used were synthetized by Metabion and include: ACE2 forward: AACTACCCGGAGGGCATAG and reverse: CTGGGATGTCCGGTCATATT; AXL forward: AACCAGGACGACTCCATCC and reverse: AGCTCTGACCTCGTGCAGAT; CTSL forward: GGGAGGGCAGTTGAGGAC and reverse: GCAAGGATGAGTGTAGGATTCA; TMPRSS2 forward: CGCTGGCCTACTCTGGAA and reverse: CTGAGGAGTCGCACTCTATCC; HPRT forward: TGACCTTGATTTATTTTGCATACC and reverse: CGAGCAAGACGTTCAGTCCT.

For all statistical analyses, the GraphPad Prism 5 software package was used (GraphPad Software). For quantitative data, statistical significance between groups was determined by paired t test, unpaired t test, one-way ANOVA followed by Tukey's Multiple Comparison Test, or Kruskal-Wallis test followed by Dunn's Multiple Comparison Test. Pearson's correlation was used to measures the statistical relationship between two continuous variables. For patients whose SAEC were used in multiple cultures, mean values of gene expression levels were used for the correlation analysis. A p values of 0.05 or lower were considered statistically significant.

As summarized in Table 1, tumor free lung tissues from 11 patients with lung cancer was studied, including 2 with squamous cell carcinoma, 8 with adenocarcinoma and 1 with non-small cell lung cancer (NSCLC; not otherwise specified). The 11 patients were 8 former smokers and 3 current smokers, 7 men and 4 women. The mean age of the patients was 72 years and ranged from 57 to 83 years. SAEC were isolated from all 11 patients, whereas HBEC were isolated from 4 of them. The isolated primary airway epithelial cells were passaged and then cultured in submerged or ALI cultures to examine the expression of SARS-CoV-2 entry genes.

First, we compared HBEC and SAEC with respect to the expression of SARS-CoV-2 entry genes. HBEC and SAEC isolated from same donors were cultured under both, submerged and ALI conditions, and the expression levels of ACE2, TMPRSS2, CSTL, and AXL were determined and normalized to the expression of housekeeping gene HPRT. As shown in Figure 1A, none of the four genes were significant differentially expressed when compared between submerged cultured HBEC and SAEC. When cells were differentiated under in ALI conditions, no significant difference was observed between HBEC and SAEC in expression levels of ACE2, TMPRSS2, and CTSL. However, expression of AXL was significantly higher in HBEC than in SAEC (0.98 ± 0.28 vs. 0.54 ± 0.23, p = 0.0168) (Figure 1B). Therefore, with the exception of AXL expression under ALI culture conditions, in-vitro differentiated HBEC and SAEC expressed comparable levels of SARS-CoV-2 entry genes.

The above experiments provide preliminary evidence for a possible influence of cell culture conditions on the expression of SARS-CoV-2 entry genes. Therefore, HBEC cells derived from submerged and ALI cultures were examined for the expression of SARS-CoV-2 entry genes (Figure 2A). Notably, expression of ACE2 in HBEC differentiated under ALI condition was significantly higher than those grown in submerged culture (0.094 ± 0.032 vs. 0.011 ± 0.004, p = 0.0353). In addition, the expression of the other two genes involved in ACE2-mediated viral entry, TMPRSS2 (4.99 ± 1.10 vs. 1.81 ± 0.79, p = 0.0313) and CTSL (0.0082 ± 0.0012 vs. 0.0046 ± 0.0016, p = 0.0018), was also significantly increased in HBEC grown under ALI conditions compared to those grown in submerged culture. Similar findings were observed in SAEC cells, where differentiation under ALI conditions significantly increased the expression of ACE2 (0.118 ± 0.052 vs. 0.010 ± 0.004, p = 0.0313), TMPRSS2 (4.16 ± 0.95 vs. 0.99 ± 0.32, p = 0.0121) and CTSL (0.0061 ± 0.0009 vs. 0.0024 ± 0.0009, p = 0.0349) compared with submerged culture (Figure 2B). In contrast, the expression of AXL was not significantly different between the two cell culture conditions in both HBEC and SAEC.

Because ALI culture is closer to physiological conditions than submerged culture, we focused on the expression of these genes in SAEC differentiated under the former conditions in the following experiments. Because there is evidence that respiratory diseases such as COPD and asthma may affect the expression of SARS-CoV-2 entry genes (14–16), we next examined whether SAEC derived from patients with COPD or asthma differed in gene expression from that of donors without asthma or COPD. As shown in Figure 3, the expression of ACE2 in SAEC isolated from lung cancer patients without respiratory co-morbidities, lung cancer patients with asthma, and lung cancer patients with COPD was 0.041 ± 0.017, 0.057 ± 0.012, and 0.035 ± 0.011, respectively, with no significant differences between groups. No significant difference was also observed in the expression levels of TMPRSS2, CTSL, and AXL among SAEC isolated from the three patient groups, although there was a tendency for higher expression of CTSL and AXL in SAEC isolated from patients with co-morbidity of COPD, but this was not statistically significant.

Given that no significant difference in expression of SARS-CoV-2 entry genes was observed between SAEC from lung cancer patients with or without respiratory complications, we next combined all samples for further analysis. We first determined the effect of smoking by comparing data sets of current smokers and ex-smokers. Expression of ACE2 in SAEC of current smokers (0.038 ± 0.0085) was not significantly different to that of former smokers (0.046 ± 0.010). Furthermore, levels expression of TMPRSS2, CTSL, and AXL in SAEC of current smokers were also comparable to those of former smokers (Figure 4A). When these SAEC were divided into two groups according to sex of the cell donors, only a significant difference was observed between the two subgroups in the expression of TMPRSS2, with female SAEC showing significantly higher expression than male SAEC (Figure 4B), while comparing of the four genes in patients with or without hypertension no significant differences could be observed (Figure 4C). To evaluate the effect of age of cell donors, a correlation analysis was performed. As shown in Figure 4D, only a marginally significant inverse correlation was detected between cell donor age and TMPRSS2 expression, with a Pearson's correlation coefficient of−0.58 (p = 0.0622). Also, no significant correlation was found between the age of cell donors and expression of and the expression of ACE2, CTSL and AXL.

High expression of ACE2, TMPRSS2 and CTSL in SAEC cultured under ALI condition implies that expression SARS-CoV-2 entry genes is dependent on the differentiation of SAEC. SAEC could either be differentiated under ALI conditions, or in 3D culture. Using flow cytometry, we next determined the surface protein expression of ACE2 and AXL on in-vitro differentiated SAEC. As shown in Supplementary Figure S4A, SAEC differentiated under ALI condition consisted of 27.00 ± 3.77% ciliated cells, 5.17 ± 1.26% goblet cells, 29.26 ± 3.28% club cells, and 24.97 ± 2.33% basal cells. Notably, majority of all four types of cell express ACE2 and AXL (Supplementary Figure S4B). Regarding the levels of surface expression of ACE2 and AXL, goblet cells showed highest expression of both ACE2 and AXL, while the expression levels of the other three types of cells were comparable (Supplementary Figure S4C). Further, flow cytometry analysis revealed that the cell composition in the 3D cultured spheroids is comparable to that in SAEC differentiated under ALI condition, with 19.00 ± 2.92% ciliated cells, 3.40 ± 0.43% goblet cells, 14.51 ± 1.59% club cells, and 19.11 ± 2.38% basal cells (Supplementary Figure S4D). Furthermore, similar to SAEC differentiated under ALI condition, majority of all above of four types of cells express both ACE2 and AXL, and goblet cells express higher levels of ACE2 and AXL than other three cells types (Supplementary Figures S4E,F).