Introduction

The small airway epithelium (SAE), comprised of basal, intermediate, club, mucous and ciliated cells, plays a central role in the pathogenesis of chronic obstructive pulmonary disease (COPD), idiopathic pulmonary fibrosis (IPF), cystic fibrosis (CF) and most lung cancers [1,2,3,4,5,6]. In humans, the basal cells (BC) function as the SAE stem/progenitor for the intermediate and differentiated club, mucous and ciliated cells [7,8,9,21]. At least 20 metaphase spreads were analyzed. All metaphases were fully karyotyped.

Real time PCR

cDNA was synthesized using TaqMan Reverse Transcriptase Reaction kit (Thermo-Fisher). All reactions were run on an Applied Biosystems Sequence Detection System 7500. Relative expression levels were determined using the △Ct method with 18S ribosomal RNA as the endogenous control. Premade TaqMan Gene Expression Assays (Thermo-Fisher) were used as listed in Additional file 1: Table S1.

Microarray and RNAseq

Trachea, large airway epithelium (LAE, 3rd -4th order bronchi) and small airway epithelium (SAE, 10th to 12th order bronchi) were collected from normal individuals via flexible bronchoscopy as described previously [13, 17, 20, 23]. Following RNA extraction and sample quality assessment, genome-wide expression assessment was performed using microarray (Human Genome U133 Plus 2.0 arrays, Affymetrix, Santa Clara, CA). The raw microarray data of the 9 matched trachea, large and small airway epithelium, and the SAE from 36 COPD smokers and 60 healthy nonsmokers have been published (GSE64614, GSE76327). The cluster analysis of cell line donor’s SAE with the 9 matched trachea, large and small airway epithelium was performed in Partek Genomics Suite (Partek, St. Louis, MO). Genes differentially expressed between the 9 paired trachea and SAE (fold changes > 2 fold, Benjamini-Hochberg corrected p < 0.05) were used as filter to generate heat map. For RNA-Seq (Illumina flow cells, San Diego, CA) [18], samples of passage 47 hSABCi-NS1.1 cell at basal stage (right before ALI culture) and ALI-d28 were used. The data are publicly available in the NCBI Gene Expression Omnibus (GEO accession number: GSE124265).

Lentivirus-mediated SPDEF expression

The methods for lentivirus production were as previously described [22, 23]. Lenti-SPDEF-IRES-GFP (expressing SAM Pointed Domain Containing ETS Transcription Factor) and lenti-IRES-GFP (control) plasmids were generated using the PLKO lentiviral vector backbone (Thermo Scientific). The replication deficient lentiviruses were generated in 293A cells using compatible packaging vectors. Both viruses were pseudo-typed with the VSV-G envelope and concentrated with PEG-it™ Virus Precipitation Solution (System Biosciences, Mountain View, CA) using the manufacturer’s protocol. For infection of the cells, recombinant lentiviruses were added to basal cells immediately prior to plating on Transwell inserts by combining lentiviruses with suspended basal cells at a multiplicity of infection sufficient to produce > 90% positive GFP staining. The cells and virus were deposited in the apical chamber of the transwell insert at a concentration of 1.5 × 105 cells/100 μl/well (designated ALI-day − 2; i.e., two days before removal of the apical medium). The following day, the infectious medium was removed, and the ALI culturing protocol continued as described above. The infection efficiency was confirmed by monitoring GFP positive staining.

Statistical analysis

The two-tailed Student’s t-test was used to compare gene expression in both in vivo and in vitro data. In all analyses, a p value less than 0.05 was deemed significant.

Results

Generation of hSABCi-NS1.1

Based on our previous published sub-dataset [20], small airway epithelium has a different gene expression pattern than matched-tracheal and large airway epithelium from healthy nonsmokers (Fig. 1). For example, expression of SFTPB (surfactant protein), LTF (secretory cell gene) and small airway development-associated transcription factors GATA6 and SOX9 [24,25,26,27] are enriched in the small airway epithelium (Fig. 1). To ensure that the small airway epithelium recovered from the donor had typical SAE transcriptome, unsupervised clustering was carried out on the SAE transcriptome of the donor to compare with the previous small, large and trachea epithelium dataset. As expected, the microarray data of the donor clustered with the SAE samples when differential expression gene list of trachea vs small was assessed.

Fig. 1
figure 1

Typical small airway transcriptome features of the cell line donor’s small airway epithelium (SAE). Data shown is the unsupervised cluster analysis of microarray data from the cell line donor’s small airway epithelium with data from previously published-microarray datasets that include 9 matched-trachea, large airway and small airway epithelium samples. Genes differentially expressed between the paired trachea and SAE (fold changes > 2 fold, Benjamini-Hochberg corrected p < 0.05) were selected to generate the plot. Examples of SAE-enriched genes (GATA6, SOX9, LTF and SFTPB) are indicated. The donor’s SAE clusters with the reference SAE transcriptome, distinct from the large airway and trachea epithelium

Full size image

After retro-hTERT genetic modification, the SABC were resistant to puromycin selection (Fig. 2a). The resulting cell population was a mixed cell population termed as hSABCi-NS1. A single cell clone was isolated from hSABCi-NS1 (termed as hSABCi-NS1.1) (Fig. 2b). The heterogeneous morphology is likely because these cells were at different phases of the cell [22]. The hSABCi-NS1.1 clone survived and was expanded for more than 1 year in vitro (Fig. 2c). The cells entered consistent growth after 200 days. The population doubling levels between passage 5 and passage 55 were ~ 200 (Fig. 2c). As additional quality control, one vial of passage 46 hSABCi-NS1.1 cells, frozen in liquid nitrogen for > 100 days, was re-cultured and compared with a vial of passage 1 parental primary cells under the same culture conditions at the same time. Even after > 110 additional population doubling levels, the re-cultured passage 46 hSABCi-NS1.1 cells had no signs of replicative senescence (Fig. 2d, Additional file 1: Figure S1). The percentage of live cells over each passage in the re-cultured cells was always > 80% (Fig. 2e). The doubling time of hSABCi-NS1.1 cells during exponential growth of passage 52 was approximately 16 h (Fig. 2f). Based on morphology, the cells remain healthy in the late passages.

Fig. 2
figure 2

Isolation and growth of the hSABCi-NS1.1 immortalized small airway basal cells from a healthy nonsmoker subject. a Morphology of cells following drug selection, day 9. Left - parental cells without TERT, plus puromycin selection; right - hTERT infected cells with puromycin selection (mixed clones, hSABCi-NS1). Bar = 100 μm. b Morphology of the single cell-derived cell clone hSABCi-NS1.1 on day 6, passage 0. Single-cell clones were generated by culturing limited-numbers of cells of hSABCi-NS1 cell in 10 cm plate and tracked under the microscope from day 2. Bar = 100 μm. c Growth of hSABCi-NS1.1. Cell population doubling levels were quantified from passage 5. Each dot represents one passage in T25 flask. d Growth of hSABCi-NS1.1 after recovery from cryopreservation. Passage 46 of hSABCi-NS1.1 cells (red dots) which had been frozen in the liquid nitrogen for 131 days, was re-cultured to test the consistence of growth. The population doubling levels were re-counted from the day of re-culturing. Passage 1 parental pre-immortalized primary cells were cultured at the same time as control (blue triangles). e % live cells of the re-cultured cells. The % live cells of the re-cultured hSABCi-NS1.1 cells were counted during each passage. Only adherent cells were counted. Red – hSABCi-NS1.1, passage 46; blue – parental primary cells pre-TERT, passage 2. f Growth of the re-cultured hSABCi-NS1.1 cells at passage 52. Cell doubling time between day 2 and day 3 after seeding was 16 h. The rate of cell doubling was reduced after day 5 when the cells reached confluence

Full size image

Karyoty** analysis of 23 hSABCi-NS1.1 cells from passage 50 cell showed that all of the cells hade an abnormal male karyotype with trisomy 8 in all cells analyzed. Additionally, 22 (96%) cells contained abnormalities of chromosome 19 (resulting in net gain of 19q and/or 19p) and 15 (65%) cells contained trisomy 20.

Basal/stem cell markers

Consistent with the source of the hSABCi-NS1.1 cell line, immunostaining showed that the parental cells prior to immortalization expressed KRT5 (Additional file 1: Figure S2). hSABCi-NS1.1 cells were positive for basal cells markers, TP63 and KRT5, at passage 6 (Fig. 3a). The positive staining was maintained at passage 49 (Fig. 3b). TaqMan PCR confirmed that hSABCi-NS1.1 (passage 8 to passage 54) had similar mRNA expression levels of TP63 compared with primary small airway basal cells (Fig. 3c).

Fig. 3
figure 3

Maintenance of basal cell features in the immortalized hSABCi-NS1.1 cells. a Basal cell markers staining at passage 6. Left, TP63 (green); middle, KRT5 (red); right, merged image. Bar = 50 μm. b Basal cell marker staining at passage 49. Left, TP63 (red); middle, KRT5 (green); right, merged image. Bar = 20 μm. Note that some nuclei appear negative for TP63 because of image contrast. c Expression of TP63 mRNA. Gene expression was assessed by TaqMan PCR. Each circle is one passage of hSABCi-NS1.1 cells. Primary small airway basal cells from 3 healthy nonsmokers at passage 2 were used as controls (blue triangle on right)

Full size image

Differentiation capacity

To assess whether the hSABCi-NS1.1 cells still retain the potential to form tight junctions, the cells were cultured on ALI. Measurement of trans-epithelial electrical resistance showed that the resistances were consistently > 200 Ohms x cm2 after ALI-day 14 (Fig. 4a). Immunofluorescence staining revealed that there was sharp and continued tight junction protein 1 (TJP1) staining between the cells on the apical side (Fig. 4b).

Fig. 4
figure 4

Cell junction formation of the immortalized hSABCi-NS1.1 cells on air-liquid interface (ALI) culture. a Trans-epithelial electrical resistance (TEER). Each color represents one independent experiment (n = 7 in total, passage 47 to 56 cells). Each dot represents average TEER of 3 ALI wells between day 0 and day 35 of ALI culture. b Junction protein staining of hSABCi-NS1.1 on ALI culture. Shown is tight junction protein 1 (TJP1) staining (green) of passage 47 cells on ALI-day 35. Bar = 20 μm. This composite image was created from a stack of individual images acquired at different depths in the monolayer (z-stack covering 40 μm). As a result, the merged image includes cell junction staining at the apical surface overlaid on the image of the epithelial cell nuclei. The difference in depth of the two structures accounts for difference in alignment of nuclei and apical tight junctions

Full size image

To assess the differentiation capacity of hSABCi-NS1.1 cells on ALI, the ALI cells were stained with markers for SAE-associated lineages (Fig. 5a-e), including ciliated (ARL13B+ motile cilia), club (SCGB1A1+ organelles), mucous (MUC5AC+ or MUC5B+ organelles), ionocytes (FOXI1+ nuclei), neuroendocrine (CHGA+ organelles), and less-defined alveolar-like cells (SFTPA+ or SFTPB+ organelles). ARL13B staining showed a characteristic apical collection of curvilinear structures while SCGB1A1, MUC5AC, MUC5B, SFTPA, SFTPB, and CHGA staining occurred in collections of intracellular puncta typical of secretory vesicles. The transcription factor, FOXI1, co-localized with nuclei consistent with its role in regulation of gene expression. Ciliated cell and club cell staining were non-overlap** while MUC5AC and MUC5B were partially overlap** as expected.

Fig. 5
figure 5

Differentiation capacity of immortalized hSABCi-NS1.1 cells on ALI culture. Passage 47 to passage 56 of hSABCi-NS1.1 cells were assessed on ALI day 14 to 35 for airway epithelial cell lineage markers. a Ciliated and club cells of passage 49 cells, ALI-day 21. Red-ARL13B; Green-SCGB1A1. Inset, higher magnification, showing cilia and SCGB1A1+ granules. Bar = 20 μm. b Mucous cells in passage 49 cells, ALI-day 16. Red-MUC5AC; green-MUC5B; yellow-co-localization (arrowheads). Bar = 20 μm. c Surfactant protein expressing cells, passage 49, ALI-day 35. Left, red-SFTPA; right, red-SFTPB. In both fields, nuclei were stained with DAPI (blue). d Ionocyte, passage 47, ALI-day 28. Red-FOXI1. Nuclei were stained with DAPI (blue). e Neuroendocrine cells, passage 49 cells, ALI-day 35. Red-CHGA. Nuclei were stained with DAPI (blue). Bar = 10 μm

Full size image

RNAseq was used to confirm the differentiation capacities of hSABCi-NS1.1 on ALI. Compared to the undifferentiated basal cell stage, ALI-day 28 hSABCi-NS1.1 had down-regulation of basal cell genes (KRT5, KRT14, TP63, KRT15), with up-regulation of genes typical of ciliated cells (FOXJ1, DNAI1, SPAG6, TEKT1), secretory cells (SCGB1A1, SCGB3A1, MUC5B, LYZ), alveolar-like cells (SFTPA1, SFTPA2, SFTPB and SFTPD), neuroendocrine cells (CHGB and SCG5), ionocyte cells (FOXI1, ATP6V1B1 and CFTR; Table 1). All of these markers have been detected in small airway epithelium observed by brushing during fiberoptic bronchoscopy using RNAseq [17]. By functional term enrichment analysis, “cilium” and “immunity” were among the top 3 categories in the top 1500 up-regulated genes (Table 2), while “ribosomal protein” was the top 1 functional category associated with the top 1500 down-regulated genes (Table 2), consistent with the fact that basal cells contain abundant ribosomes [12]. In addition, many Notch pathway genes showed more than 1.5-fold changes during differentiation, e.g., NOTCH1 (− 2.1-fold), NOTCH2 (2.7-fold), NOTCH4 (1.5-fold), JAG2 (− 2.9-fold), HES1 (− 6.4-fold), HES2 (− 8.6-fold), HES4 (− 1.6-fold), HES6 (3.0-fold), HEY1 (1.8-fold) and HEYL (2.9-fold). The data is consistent with the knowledge that the dynamic changes of Notch pathway are involved in the commitment of airway epithelial lineages.

Table 1 Expression of airway epithelial lineage markers in hSABCi-NS1.1 cells during differentiation on ALI culturea
Full size table
Table 2 Gene function annotation enrichment analysis of the top up-regulated and down-regulated genes during differentiation on ALI culturea
Full size table

Interestingly, differentiated hSABCi-NS1.1 cells expressed inflammatory cell genes CD4 (lymphocyte cell marker) and MHC-II genes (CD74, HLA-DRA; dendritic cell markers; Additional file 1: Table S2). For additional confirmation, RNAseq reads were aligned with exons of selected mRNAs. The alignment clearly demonstrated the expression of immunity genes CD4, CD74 and HLA-DRA (Fig. 6a, b and c). Since hSABCi-NS1.1 was single-cell-derived, contamination with inflammatory cells can be excluded. The findings have corroborated our previous observation from single-cell RNAseq data that some airway epithelial cells were positive for inflammatory cell genes, including CD4, CD74 and HLA-DRA [12].

Fig. 6
figure 6

Example of non-inflammatory cell immunity genes in hSABCi-NS1.1 on ALI. The reads from the RNAseq data were aligned to the introns and exons of reference mRNAs of the selected genes. Both basal and ALI-d28 data were plotted. a CD4 alignment; b CD74 alignment; c HLA-DRA alignment

Full size image

Regional identity in the airways

To assess whether differentiation of airway cell lines have regionally enriched features, the transcriptome of hSBACi-NS1.1 cells was compared with our previously-reported large airway cell line BCi-NS1.1 [21] using the same culture condition in basal and ALI culture. TaqMan PCR analysis demonstrated that hSABCi-NS1.1 cells had lower expression levels of basal, club, and ciliated cell genes, while significantly higher expression levels of small airway-associated genes, including SFTPA1, SFTPB, SFTPD and LTF (Fig. 7a). Consistent with this data, there was a trend of lower expression of SOX2 (proximal airway transcription factor) and significantly higher expression levels of small airway development transcription factors (NXK2–1, GATA6, SOX9, HOPX, ID2, ETV5) [24,25,26,27] (Fig. 7b).

Fig. 7
figure 7

Expression of small airway development-relevant transcription factors in immortalized hSABCi-NS1.1 cells. Gene expression in hSABCi-NS1.1 cells (from small airway) was compared with gene expression in BCi-NS1.1 cells (from large airway). Both cell lines (passage 52–55) were cultured on ALI using identical culture conditions for comparison. Large and small airway lineage markers and transcription factors relevant to the development of large airway (SOX2) and small airway (NKX2–1, GATA6, SOX9, HOPX, ID2 and ETV5) were assessed by TaqMan PCR. Data shown is the mean fold-changes ± standard error of n = 3 independent side-by-side comparison experiments. *, p < 0.05; **, p < 0.01; #, consistent changes in all 3 independent experiments, but not significant because of high variability. a Airway cell lineage marker genes. b Transcription factors relevant to the development of large airway and small airway epithelium

Full size image

Recapitulation of SAE-related disease

To test potential applications of the cell line in the studies of SAE diseases, hSABCi-NS1.1 cells were genetically modified by lentivirus to express SPDEF, a key transcription factor controlling secretory cell differentiation [28]. As shown in Fig. 8a, SAE from COPD smokers demonstrated dysregulation of multiple secretory cell genes, including SPDEF. Based on GFP distribution, the infection efficiency of lentivirus in hSABCi-NS1.1 cells was high and maintained during ALI culture (not shown). As a control, SPDEF overexpression in hSABCi-NS1.1 cells on ALI induced remarkable up-regulation of MUC5AC staining (Fig. 8b). TaqMan PCR data demonstrated that upregulation of SPDEF led to increased-expression of MUC5AC, MUC2, TFF3, CEACAM5, MSMB and decreased-expression of SCGB3A1 and LTF. However, no consistent changes in SCGB1A1 or MUC5B were found (Fig. 8c), supporting our previous proposal that a transcription factor network might be involved [29]. Immunostaining confirmed the striking up-regulation of MSMB (Fig. 8d), which has not been identified as a SPDEF-regulated gene in previous studies [28].

Fig. 8
figure 8

Consequences of overexpression of SPDEF on the differentiation of immortalized hSABCi-NS1.1 cells on ALI, as a model of secretory cell hyperplasia in COPD. Based on the knowledge that SPDEF, a transcription factor that induced secretory cell differentiation is up-regulated in the SAE of COPD smokers. hSABCi-NS1.1 cells were genetically modified by lentivirus expressing SPDEF and cultured on ALI to day 14 to test whether the hSABCi-NS1.1 cells can be used as a model of secretory cell hyperplasia. a Secretory cell gene expression in the SAE of 36 COPD smokers compared to 60 healthy nonsmokers in vivo. Five up-regulated and 4 down-regulated secretory cell genes were used as examples. Data shown is relative fold-change of microarray-based gene expression ± standard error. b Staining of a known SPDEF-regulated gene in ALI. MUC5AC was used as an example. Shown are passage 47 cells on ALI-day 21. Red-MUC5AC; blue-nucleus. Top panel, lenti-GFP infected cells, as control; bottom panel, lenti-SPDEF-infected cells. Bar = 50 μm. c TaqMan PCR assessment of lentivirus-SPDEF-induced gene expression changes in hSABCi-NS1.1 cells. The same genes as in panel A were assessed. Data shown is the mean fold-change ± standard error of n = 3 independent experiments using passage 52 to passage 55 cells. d Staining of MSMB as an example of a SPDEF-regulated secretory gene in ALI. Shown are passage 47 cells on ALI-day 21. Red-MSMB; blue-nucleus. Top panel, lenti-GFP infected cells, as control; bottom panel, lenti-SPDEF-infected cells. For a, c *, p < 0.05; **, p < 0.01; ***, p < 0.001; #, consistent changes in all 3 independent experiments, but not significant because of variability

Full size image

Discussion

We have generated and characterized a single-cell-derived normal human small airway basal cell line (hSABCi-NS1.1) with diverse differentiation capacities and regional features. This human SAE cell line can achieve high cell doubling levels and passage numbers. The hSABCi-NS1.1 cell line provided a novel tool to study SAE biology and pathophysiology in an airway region relevant manner. The availability of hSABCi-NS1.1 cells will permit studies of diseases for which the SAE plays a central role. It also can be used to assess the effect of specific mediators on the SAE. As an example, we found transcription factor SPDEF can induce human COPD-relevant changes in hSABCi-NS1.1 cells. Also relevant to small airway biology, hSABCi-NS1.1 can differentiate to ionocytes, a recently identified airway cell type [30, 31], surfactant protein positive cells [13, 17, 20, 32] and neuroendocrine cells [17]. Basal cells are likely the common progenitor for all these cell types, and hSABCi-NS1.1 cells will be a convenient tool to study these cell populations.

The hSABCi-NS1.1 cell line demonstrates regional (small airway epithelium) identity. Along the proximal to distal axis, the airway epithelium has different morphological features [4, 7, 20, 24,25,26,27]. As examples, compared to the tracheal epithelium, small airway epithelium has a thinner epithelial layer, is composed of basal, secretory and ciliated cells in different ratios, and has surfactant protein positive cells [13, 14, 32]. These airway regional features are apparent in the airway epithelium transcriptome, including SAE-specific surfactant protein genes and SAE enrichment of distal airway patterning-associated transcription factors NKX2–1, GATA6, SOX9 and ETV5 [24,25,26,27].

The regional identity is likely directly involved in the pathogenesis of SAE-related diseases. For example, SAE express surfactant protein genes (SFTPB, SFTPA2), genetic variation of which are associated with COPD and IPF [33, 34]. SAE-enriched transcription factor, NKX2–1, is relevant to lung adenocarcinoma [35], a cancer type that is derived from the small airway epithelium. The expression of MUC5B, an IPF risk gene [36], is enriched in SAE [37]. GATA6 is involved in regulating MUC5AC gene expression [38] and can regulate an IPF-associated mesenchymal phenotype [39]. LTF plays important anti-infection roles in CF [40].

Conclusion

In vitro differentiation models of airway epithelial cells have been widely used in studies for airway development, carcinogenesis, environmental insults and infection by pathogens [41, 42]. Importantly, since the SAE plays a central role in many chronic lung disorders, hSABCi-NS1.1 cells should be a valuable tool for drug screening. In this context, hSABCi-NS1.1 cells are available for other investigators by contacting the senior author.