EGF shifts human airway basal cell fate toward a smoking-associated airway epithelial phenotype

Abstract

The airway epithelium of smokers acquires pathological phenotypes, including basal cell (BC) and/or goblet cell hyperplasia, squamous metaplasia, structural and functional abnormalities of ciliated cells, decreased number of secretoglobin (SCGB1A1)-expressing secretory cells, and a disordered junctional barrier. In this study, we hypothesized that smoking alters airway epithelial structure through modification of BC function via an EGF receptor (EGFR)-mediated mechanism. Analysis of the airway epithelium revealed that EGFR is enriched in airway BCs, whereas its ligand EGF is induced by smoking in ciliated cells. Exposure of BCs to EGF shifted the BC differentiation program toward the squamous and epithelial-mesenchymal transition-like phenotypes with down-regulation of genes related to ciliogenesis, secretory differentiation, and markedly reduced junctional barrier integrity, mimicking the abnormalities present in the airways of smokers in vivo. These data suggest that activation of EGFR in airway BCs by smoking-induced EGF represents a unique mechanism whereby smoking can alter airway epithelial differentiation and barrier function.

Keywords: airway epithelial barrier; cigarette smoking; progenitor cell.

Fig. 1.

Enrichment of EGFR in airway BCs. (A) Normalized expression of the ErbB family receptors in the complete differentiated LAE of healthy nonsmokers (n = 21) and LAE-derived BCs of healthy nonsmokers (n = 4) based on the microarray analysis. N.S., not significant. *P < 0.05. (B) Localization of EGFR in the human airway epithelium. (Left) EGFR immunohistochemistry of the LAE biopsy samples. (Right, Top) EGFR immunocytochemistry of an airway epithelial brushing sample. (Right, Middle and Bottom) Immunofluorescence colocalization of EGFR and KRT5 in the airway epithelial brushings. (C) Representative immunofluorescence image of BCs cultured from the LAE stained for EGFR and KRT5. (D) Western blot analysis of EGFR protein expression in airway BCs at baseline and in BC-derived airway epithelium generated from BCs after 8 d and 28 d of culture in ALI. GAPDH expression is shown as a loading control. (E) Immunofluorescence analysis of cytopreparations of airway epithelial cells generated from BCs after 28 d of culture in ALI double-stained for EGFR and KRT5; two representative images are shown. (Scale bars: 20 μm.)

Fig. 2.

EGF expression in differentiated airway epithelial cells. (A) Normalized EGF gene expression in the indicated groups based on microarray analysis. (B) EGF expression levels, in reads per kilobase per million (RPKM) mapped reads, in the complete airway epithelium of healthy nonsmokers (n = 4) and healthy smokers (n = 6), based on RNA-Seq analysis. (C) Immunohistochemistry analysis of EGF protein expression in LAE biopsy samples from healthy nonsmokers (Left) and smokers (Right). (Scale bars: 20 μm.) (D) Contribution (%) of different cell populations to EGF-expressing cells detected in the airway epithelium by immunohistochemistry as shown in C and in Fig. S3C. (E) Normalized expression of the EGF gene after stimulation of the ALI-differentiated airway epithelium with CSE (1% and 2%) from the apical side every other day for 14 d compared with unstimulated cells. (F) ELISA-based comparison of EGF levels in the apical and basolateral supernatants of ALI-differentiated airway epithelial cells after stimulation with CSE as described in E.

Fig. 3.

Modulation of the BC phenotype and airway epithelial differentiation by EGF. (A) Western blot analysis showing expression of phosphorylated EGFR (p-EGFR) (Upper) and total EGFR (Lower) in airway BCs stimulated with EGF for 5, 15, or 30 min vs. unstimulated (control) BCs. (B) Representative micrographs of airway BCs after 48 h of stimulation with 10 ng/mL EGF (Right) vs. unstimulated control BCs (Left). (Scale bars: 20 μm.) (C) EGF modulation of the normalized expression of genes related to various aspects of airway epithelial differentiation: CD44, KRT5, KRT6A, KRT6B, KRT14, IVL, SFN, SNAI2, VIM, FOXJ1, DNAI1, MUC5B, and SCGB1A1. The data were generated during 14 d of ALI culture stimulated with 10 ng/mL EGF (n = 4) from the basolateral side vs. unstimulated controls (n = 4). (D and E) Immunofluorescence analysis of cytopreparations (D) and sections (E) of the airway epithelium samples differentiated during 28 d of ALI culture of BCs stimulated with 10 ng/mL EGF (Right) vs. unstimulated controls (Left) for the expression of squamous cell markers IVL or KRT14 and mesenchymal/EMT marker VIM. Each sample was costained for the BC marker KRT5. (Scale bars: 20 μm.) (F) Frequency of IVL⁺, VIM⁺, and KRT5⁻ cells in the airway epithelium samples differentiated during 28 d of ALI from EGF-treated vs. untreated control BCs, as shown in D and E and in Figs. S8 and S9.

Fig. 4.

EGF-treated BCs generate a more permeable airway epithelial barrier. (A) EGF modulation of the expression of genes related to the AJC and epithelial polarity: CDH1, TJP1, TJP3, CLDN3, CLDN8, OCLN, PARD3, PARD6B, and PTEN. The data were generated during 14 d of ALI culture of BCs stimulated with 10 ng/mL EGF (n = 4) vs. unstimulated controls (n = 4). (B) Rt of the airway epithelium generated from BCs stimulated with EGF (10 ng/mL) vs. unstimulated control cells at different time points of the ALI culture. (C) Rt of the airway epithelium generated at day 14 of ALI culture from BCs treated with media only (control), DMSO (vehicle for AG1478), EGFR inhibitor AG1478 (10 µm), and EGF (10 ng/mL) with or without a 1-h pretreatment with AG1478 (10 µm) every second day starting on day 0 of ALI. (D) Mean fluorescence intensity detected in the basolateral chamber of ALI at different time points after apical application of FITC-dextran to the airway epithelium samples differentiated in ALI from EGF-treated or control BCs. In B–D, asterisks indicate statistically significant differences at given time points between EGF-stimulated and control groups: *P < 0.05; **P < 0.01; ***P < 0.005.