Basal Cells Contribute to Innate Immunity of the Airway Epithelium through Production of the Antimicrobial Protein RNase 7

Mucosal epithelial cells are an essential component of the host defense barrier against microbial pathogens. In the lung, a pseudostratified layer of airway epithelial cells provides this host defense via a passive barrier function but also active innate immune defense mechanisms (1, 2). These mechanisms are primarily executed by airway epithelial luminal cells (LCs), mainly composed of ciliated and secretory cells, which facilitate for instance mucociliary clearance, defense mediated by periciliary tethered mucins, production of innate immune mediators upon pathogen recognition, and Ig transcytosis toward the airway surface liquid (3–6). In contrast to LCs, it has been largely unexplored whether airway epithelial basal cells (BCs) have an innate immune function. BCs are characterized as epithelial progenitor cells and defined based on the expression of p63 and cytokeratin 5 (KRT5) (7–9). BCs display a low turnover at normal homeostatic conditions but rapidly expand and subsequently differentiate toward other epithelial phenotypes upon injury to restore the airway epithelial barrier. The basolateral localized BCs are shielded from direct microbial contact by the overlying LCs in intact airway epithelium. However, airway epithelial injury caused by disruption of the epithelial barrier because of the loss of cell–cell contact or shedding of LCs after extensive injury may increase microbial exposure. This may enhance the vulnerability of BCs and underlying tissues to respiratory infections. Recently, Byer et al. (10) demonstrated selective expression of the proinflammatory mediator IL-33 by a subpopulation of BCs in patients with chronic obstructive pulmonary disease. On the basis of this observation, we speculate that BCs might display other unique innate immune responses, which act as a second line of airway epithelial host defense in addition to the activity of LCs.

Antimicrobial proteins and peptides (AMPs) are important innate immune mediators involved in airway epithelial host defense (11). AMPs display direct antimicrobial activity against bacteria, viruses, and fungi but also exhibit immunomodulatory and wound healing properties (12, 13). The antimicrobial RNase 7 is a secreted RNase A family member displaying antimicrobial activity toward a range of pathogens, including the respiratory pathogen Pseudomonas aeruginosa (14–16). Previous studies exploring the regulation of RNase 7 expression in skin, urinary tract, and ocular surface epithelial cells, demonstrated increased RNase 7 expression upon stimulation with microbes, microbial compounds, and proinflammatory cytokines (17–19). Moreover, mechanical injury and UV radiation induce expression of RNase 7 in keratinocytes (20, 21), demonstrating that epithelial injury is a potent inducer of RNase 7 expression. Although RNase 7 expression has been detected in cultured airway epithelial cells (14), it is currently unknown how RNase 7 expression is regulated in these cells.

In this paper, we provide evidence that BCs are the principal cell producing RNase 7 in cultured human primary bronchial epithelial cells (PBEC) upon exposure to the respiratory pathogen nontypeable Haemophilus influenzae (NTHi) and epithelial barrier disruption caused by cigarette smoke exposure. Submerged cultured undifferentiated PBEC (S-PBEC), primarily consisting of BCs, but not mucociliary differentiated air–liquid interface–cultured PBEC (ALI-PBEC), expressed RNase 7 upon stimulation with UV-inactivated NTHi. Transient epithelial injury induced by whole cigarette smoke coincides with epidermal growth factor receptor (EGFR)-dependent expression of RNase 7 in ALI-PBEC, and this expression is specifically observed in BCs. These findings demonstrate a novel innate immune defense function of BCs, which act as a second line of airway epithelial defense induced upon injury.

Human PBEC were isolated from tumor-free resected lung tissue from anonymous donors by enzymatic digestion as reported previously (22). Mucociliary-differentiated ALI-PBEC were cultured as previously described (23), using semipermeable Transwell membranes with a 0.4-μm pore size (Corning Costar, Cambridge, MA). Transwells were coated with a mixture of 30 μg/ml PureCol (Advanced BioMatrix, San Diego, CA), 10 μg/ml BSA (Sigma-Aldrich, St. Louis, MO), and 10 μg/ml fibronectin (isolated from human plasma) diluted in PBS, at 37°C, 5% CO₂ for 2–24 h. PBEC (passage 2) were seeded on coated Transwells at a density of 40,000 cells and cultured in submerged condition using a 1:1 mixture of DMEM (Life Technologies, Bleiswijk, the Netherlands) and bronchial epithelial growth medium (Lonza, Verviers, Belgium) (B/D medium) with supplementation of BEGM BulletKit singlequots (0.4% [w/v] bovine pituitary extract, 1 μM hydrocortisone, 0.5 ng/ml human hEGF, 0.5 μg/ml epinephrine, 10 μg/ml transferrin, 5 μg/ml insulin, T3, and 0.1 ng/ml retinoic acid) (Lonza), 1 mM HEPES (Lonza), 1 mg/ml BSA (Sigma-Aldrich), 100 U/ml penicillin and 100 μg/ml streptomycin (Lonza), and additional supplementation of 15 ng/ml retinoic acid (Sigma-Aldrich) to induce mucociliary differentiation. After reaching confluence (∼4–5 d), cells were cultured air exposed for an additional period of 2 wk before use. Undifferentiated S-PBEC were cultured in 12-well tissue culture plates in B/D medium including supplements as earlier described but without additional retinoic acid supplementation. Upon reaching 80–90% confluence, S-PBEC were used for further experiments.

NTHi strain D1 (24) was cultured on chocolate agar plates (bioMérieux, Craponne, France) at 5% CO₂ and 37°C overnight. A selected colony was resuspended in 10 ml Tryptone soya broth + X- and V-factor (Mediaproducts BV, Groningen, the Netherlands) and incubated at 37°C while shaking overnight. Next, 1 ml of the overnight broth containing NTHi was transferred into new 10 ml Tryptone soya broth + X- and V-factor and incubated for 3 h at 37°C while shaking. Afterward, the bacteria-containing broth was centrifuged for 10 min at 3000 rpm, and the pellet was washed with PBS. The bacterial pellet was resuspended in PBS after an additional washing step, and the concentration was adjusted to an OD₆₀₀ 1 (1 × 10⁹ CFU/ml). The NTHi was subsequently inactivated by UV exposure for 2 h.

ALI-PBEC were exposed to cigarette smoke using 3R4F reference cigarettes (University of Kentucky, Lexington, KY) in a whole smoke exposure model adapted from Beisswenger et al. (25) as depicted in Fig. 2A. ALI-PBEC were placed into modified hypoxic chambers (Billups Rothenberg, Del Mar, CA), further referred to as exposure chambers, localized inside an incubator at 37°C and 5% CO₂. Next mainstream smoke derived from one cigarette or air as negative control was infused inside respectively a “Smoke” and “Air” exposure chamber by a mechanical pump using a continuous flow of 1 l/minute for a period of 4–5 min. A ventilator inside the exposure chambers equally distributed and mixed the infused air/smoke with humidified warm air derived from the incubator to prevent inflammatory responses by PBEC as reported previously (25). After exposure, residual smoke inside the exposure chamber was removed for a period of 10 min by flushing the chambers with air derived from the incubator. The amount of smoke infused inside the exposure chamber was measured by determining the weight difference of a filter placed between the pump and exposure chamber before and after exposure. Approximately 2 mg cigarette smoke–derived particles were deposited on the filter as determined by measuring the filter of different exposures. After smoke or air exposure, the culture medium of ALI-PBEC was refreshed, and cells were incubated at 37°C and 5% CO₂ for the indicated period of times. UV-inactivated NTHi or Pam₃CSK₄ (InvivoGen, San Diego, CA) was administered in a volume of 100 μl in PBS at the apical surface of ALI-PBEC or added to the culture medium when stimulating S-PBEC. In the experiment examining cell signaling pathways and the effect of rTGF-α, ALI-PBEC were starved for 24 h before the experiment with supplemented B/D medium earlier described, without 0.4% [w/v] bovine pituitary extract, 0.5 ng/ml human hEGF, and 1 mg/ml BSA. To examine cell signaling pathways, the following chemical inhibitors were applied: AG1478, anti-EGFR neutralizing Ab, and U0125 (all Calbiochem, Darmstadt, Germany). Inhibitors were added into the basal culture medium and preincubated for 1 h prior to smoke exposure. Stimulation of ALI-PBEC with rTGF-α (PeproTech, Rocky Hill, NJ) was done by adding this growth factor to the culture medium. In confocal imaging and FACS experiments, 5 μg/ml brefeldin A (Sigma-Aldrich) was added in the culture medium to prevent protein secretion.

Lactate dehydrogenase (LDH) release in the apical surface liquid, collected by washing the epithelial cultures apical with 100 μl PBS, and culture medium was assessed using a cytotoxicity detection kit (Roche, Basel, Switzerland), according to the manufacturer’s protocol. One hundred microliters of 0.1% (v/v) Triton X-100 in PBS was applied on the apical surface as positive control. The transepithelial electrical resistance (TEER) was measured using the MilliCell-ERS (Millipore, Bedford, MA). Cell permeability was determined by the FITC–dextran permeability assay. In this assay, 1 mg/ml FITC–dextran (4 kDa; Sigma-Aldrich) diluted in PBS was added to the apical side and incubated for 2 h. Next, 100 μl of the basal culture medium was transferred into solid black 96-well plates (Costar-Corning), and the fluorescence was measured with an excitation/emission wavelength of 490/521 nm using a Wallac Victor 2 Microtiter Plate Reader (PerkinElmer, Waltham, MA).

Total RNA was isolated using the miRNeasy Mini Kit (Qiagen, Venlo, the Netherlands) following the manufacturer’s protocol, and RNA quantities were measured using the Nanodrop ND-1000 UV-visible (UV-Vis) spectrophotometer (Nanodrop Technologies, Wilmington, DE). cDNA synthesis was performed by RT-PCR of 1 μg RNA reverse mixed with oligo (deoxythymidine) primers (Qiagen) and Moloney murine leukemia virus polymerase (Promega, Leiden, the Netherlands) at 37°C. Quantitative real-time PCR (qPCR) was performed using IQ Sybr green supermix (Bio-Rad) and a CFX-384 real-time PCR detection system (Bio-Rad). Reactions were performed using the primers indicated in Table I. The housekeeping genes RPL13A and ATP5B were selected using the “Genorm method” (Genorm; Primer Design, Southampton, U.K.). Bio-Rad CFX manager 3.0 software (Bio-Rad) was used to calculate arbitrary gene expression by using the standard curve method.

Protein secretion by ALI-PBEC was determined in the apical surface liquid and in the basal medium. The secretion of IL-8 was determined by ELISA following the manufacturer’s protocol (Sanquin, Amsterdam, the Netherlands). RNase 7 secretion was assessed by ELISA as described previously (16). The OD values were measured with a microplate reader (Bio-Rad).

The basal and luminal fraction of ALI-PBEC were isolated according to a method adapted from Jakiela et al. (26). Cultured ALI-PBEC were washed twice with calcium-free PBS. Next, calcium-free MEM (Life Technologies) was applied to the apical side (700 μl) and the basolateral compartment to break down the integrity of the intercellular junctions. This was monitored by TEER measurements, which reached background levels after ∼15–20 min. Next, the apical medium was substituted by 700 μl Trypsin Versene (Lonza) and incubated for 7–10 min until luminal cells started to detach. The detached luminal fraction was collected, and the apical surface was subsequently washed twice gently with 700 μl DMEM supplemented with 5% heat-inactivated FCS (Life Technologies). The cells that were collected by this washing step were added to the luminal fraction, which was then centrifuged for 7 min at 1200 rpm. After removal of the supernatant, mRNA was isolated from the cell pellet (luminal fraction) and the epithelial cells that remained attached to the inserts (basal fraction).

For immunofluorescence staining, ALI-PBEC were fixed in 4% paraformaldehyde, permeabilized in 0.3% (v/v) Triton X-100 in PBS, and blocked in blocking solution consisting of 5% (w/v) BSA and 0.3% (v/v) Triton X in PBS. All incubation steps were 30 min at room temperature and contained washing steps with PBS in-between. Membranes containing ALI-PBEC were detached from the Transwell using a razorblade and transferred to 24-well tissues culture plates for Ab staining. The primary Abs (mouse mAb against RNase 7 and rabbit mAb against p63 [Abcam, Cambridge, U.K.]) were diluted in blocking solution and added to the cells for 1 h at room temperature. The secondary Abs goat anti-mouse Alexa Fluor 488, goat anti-rabbit Alexa Fluor 568 (both Invitrogen), and DAPI, diluted in blocking buffer, were added to the cells for 30 min at room temperature in the dark. Next, the PBEC-containing membranes were placed on coverslips and mounted in Vectashield Hard Set Mounting Medium (Vector Labs, Burlingame, CA). Confocal images were taken using a Leica TCS SP5 confocal inverted microscope with HCX PL APO CS 40× 1.25–0.75 Oil objective (Leica Microsystems, Wetzlar, Germany) and processed using the Leica Application Suite Advanced Fluorescence software (Leica Microsystems).

ALI-PBEC were detached from the Transwell by trypsinization using soft trypsin, consisting of 0.3 mg/ml trypsin 250 (BD Biosciences, San Jose, CA), 0.1 mg/ml EDTA, 1 mg/ml glucose (both BDH Chemicals, Poole, U.K.), 100 U/ml penicillin, and 100 μg/ml streptomycin in PBS (pH 7.45). Trypsinization was stopped using 1.1 mg/ml soy bean trypsin inhibitor (Sigma-Aldrich) in keratinocyte serum-free medium (Life Technologies). Cells were washed once in cold PBS and stained with live/dead stain aqua (Invitrogen). Then, cells were washed in PBS, permeabilized using 0.5% (v/v) saponin (Sigma-Aldrich) in PBS for 5 min on ice, and fixed with 3.9% (v/v) formaldehyde (Merck) in PBS for 15 min at room temperature. Next, primary Abs (mouse mAb against RNase 7 [Abcam, Cambridge, U.K.] and rabbit polyclonal Ab against KRT5 [Santa Cruz Biotechnology, Santa Cruz, CA]) were added, followed by secondary Ab treatment (goat anti-mouse Alexa Fluor 488 and goat-anti-rabbit Alexa Fluor 647 [both Invitrogen]), both diluted in 0.5% saponin supplemented with human FcR binding inhibitor (eBioscience, San Diego, CA). Both primary and secondary Abs were incubated for 30 min at room temperature. Cells were measured on a FACSCanto II (BD Biosciences), and analysis was performed using FlowJo (version 7.6.5) software (Tree Star, Ashland, OR). Gates were placed according to unstained and single stains with secondary Abs (Supplemental Fig. 1).

SDS-PAGE gel electrophoresis and Western blotting was performed as described previously (27). In short, PBEC were washed with ice-cold washing buffer consisting of 5 mM Tris (pH 6.4), 100 mM NaCl, 1 mM CaCl₂, and 1 mM MgCl₂, and cell lysates were prepared in lysis buffer consisting of 0.5% (v/v) Triton X-100, 0.1 M Tris-HCl (pH 7.4), 100 mM NaCl, 1 mM MgCl₂, 1 mM CaCl₂ 1 mM Na₃VO₄, and protease inhibitor mixture (Roche) and subsequently suspended in SDS-PAGE sample buffer consisting of 4% (w/v) SDS, 10% (v/v) 2-ME, 20% (v/v) glycerol, 0.5 M Tris (pH 6.8), and 0.003% (w/v) Bromophenol blue. SDS-gel electrophoresis was conducted using 10–12.5% glycine-based gels, and afterward, separated proteins were transferred to polyvinylidene difluoride membranes. Membranes were blocked with blocking buffer consisting of 5% (w/v) skim milk (Sigma-Aldrich) in PBS/0.05% (v/v) Tween 20 and incubated with primary Abs (rabbit polyclonal Ab ERK1/2, p-ERK1/2, p-EGFR, and mouse mAb β-actin [all Cell Signaling Technology, Leiden, the Netherlands], mouse mAb EGFR [BD Biosciences], and mouse mAb RNAse 7 [Abcam]) in 5% (w/v) BSA TBS/0.05% (v/v) Tween 20 overnight at 4°C. Afterward, membranes were incubated with secondary HRP-labeled Ab (Sigma-Aldrich) in blocking buffer for 1 h, and membranes were subsequently developed with ECL substrate (ThermoScientific, Rockford, IL). Results were analyzed by densitometry using Adobe Photoshop (Adobe Systems, San Jose, CA).

Graphs were made and statistical analysis was performed in GraphPad PRISM 6.0 (GraphPad Software, La Jolla, CA). Data are shown as means ± SEM. Differences were considered significantly different with p < 0.05 explored by (un)paired Student t test or one/two-way ANOVA as appropriate.

We first investigated whether the common respiratory pathogen NTHi induced RNase 7 expression in cultured PBEC. Therefore, undifferentiated S-PBEC were exposed to UV-inactivated NTHi for 6 and 24 h, and RNase 7 expression was assessed by qPCR. The results showed that NTHi induced RNase 7 mRNA expression by S-PBEC after 24 h but not 6 h of stimulation (Fig. 1A). In contrast, mRNA expression of the well-defined bacterial-induced antimicrobial peptide human β-defensin-2 (hBD-2) was increased by NTHi both at 6 and 24 h. As shown by others (28), we observed that undifferentiated S-PBEC primarily consist of p63⁺ BCs (data not shown). Because mucociliary-differentiated ALI-PBEC are more comparable to the pseudostratified airway epithelium in vivo (29), we subsequently compared RNase 7 mRNA expression in ALI-PBEC and S-PBEC. Interestingly, ALI-PBEC did not display an increase in RNase 7 mRNA in response to various concentrations of NTHi after 24 h of stimulation, whereas dose-dependent induction was observed in S-PBEC (Fig. 1B). In contrast, a dose-dependent NTHi-induced expression of hBD-2 was observed in both ALI- and S-PBEC. Further evaluation of the expression of other innate immune mediators demonstrated that NTHi induced mRNA expression of IL-8, IL-6, CCL20, and lipocalin 2 (LCN2) in both S-PBEC and ALI-PBEC (Fig. 1C). To explore baseline RNase 7 expression during mucociliary differentiation, mRNA levels of RNase 7 were measured in S-PBEC and in ALI-PBEC at 1 and 2 wk after initial air exposure. Compared with S-PBEC, no difference in mRNA levels was observed in ALI-PBEC cultured 1 wk in air exposed conditions, whereas after 2 wk a significant decrease in RNase 7 mRNA was observed (Fig. 1D). As expected, airway epithelial cell differentiation furthermore resulted in a decreased mRNA expression of the BC marker p63 and increased expression of ciliated and goblet cell markers, respectively, FOXJ1 and MUC5AC. These findings suggest, that in contrast to other examined innate immune mediators, RNase 7 is selectively produced by airway BCs but not by mucociliary-differentiated airway epithelial cells at baseline conditions and upon NTHi stimulation.

BCs display enhanced activity upon injury, and it has been demonstrated that injury induces RNase 7 expression in keratinocytes (7, 20, 21). Therefore, we hypothesized that injury of ALI-PBEC can reinitiate expression of RNase 7 by remaining undifferentiated BCs. Cigarette smoke is a well-known inhaled toxic mixture that induces airway epithelial injury (30, 31). We therefore examined cigarette smoke–induced injury in ALI-PBEC using a whole cigarette smoke exposure model (Fig. 2A). First, the cytotoxic effect of smoke exposure was determined by a LDH release assay (Fig. 2B). No significant increase in LDH release was detected in the basal medium of smoke-exposed cells compared with air controls. However, a significant increase in LDH release was observed in the apical surface wash. Despite this cytotoxic effect, no significant difference was observed in the TEER of air- and smoke-exposed cultures after 24 h (Fig. 2C). In contrast, a significant decrease in TEER was observed in the first 2 h after smoke exposure, indicating transient injury. Further evaluation of the epithelial barrier function by FITC–dextran permeability also confirmed increased barrier disruption in the first 2 h after smoke exposure, whereas no difference in permeability was detected 24 h after exposure (Fig. 2D). To examine whether transient epithelial damage induced by cigarette smoke affected the expression of RNase 7, mRNA expression levels were measured at different time periods after exposure. A time-dependent increase in RNase 7 mRNA expression was detected after smoke exposure, which peaked after 3 h and remained significantly different compared with air exposed cells after 6 h (Fig. 3A). Similar to increased RNase 7 mRNA expression, protein expression was increased in lysates of ALI-PBEC at 6, 12, and 24 h after smoke exposure (Fig. 3B, 3C). Moreover, increased RNase 7 secretion was measured in the apical surface liquid but not in the basolateral culture medium 24 h after exposure, indicating polarized secretion of RNase 7 (Fig. 3D). We further examined the expression of other mediators besides RNase 7 to delineate the innate immune response induced by cigarette smoke. In contrast to RNase 7, cigarette smoke did not increase mRNA expression of the AMPs hBD-2, hBD-3, LCN2, and Psoriasin/S100A7 (data not shown). However, smoke did induce mRNA expression of IL-8 and IL-6 (Fig. 3E), in line with previous observations by others (25). Furthermore, cigarette smoke exposure increased mRNA expression of CCL20, which previously was reported to be induced in airway epithelial cells upon disruption of the airway epithelial barrier integrity (32). In line with mRNA expression, we also detected cigarette smoke–induced IL-8 protein secretion in the basal medium and apical surface liquid (Fig. 3F). In summary, we demonstrated that transient injury induced by cigarette smoke exposure is paired with increased expression and polarized secretion of RNase 7 by ALI-PBEC. This response is part of a distinct innate immune response including the expression of IL-8, IL-6, and CCL20, but lacks induction of other AMPs that were investigated.

To determine whether cigarette smoke–induced RNase 7 was specifically expressed by BCs in ALI-PBEC, we analyzed mRNA expression in the basal and luminal cell fraction of ALI-PBEC that were separated based on a previous described method (26). To validate the nature of the isolated fractions, we initially examined the mRNA expression of the BC markers p63 and KRT5. Both p63 and KRT5 mRNA levels were significantly higher in the basal fraction compared with the luminal fraction in smoke- and air-exposed cultures (Fig. 4A). Interestingly, we observed a significant decrease in p63 and KRT5 mRNA levels in the basal fraction of smoke-exposed cultures. At later time points, these levels returned to baseline levels (unpublished data), suggesting that the effect of smoke exposure on these basal cell markers is only transient. In contrast, mRNA expression of RNase 7 was significantly increased 3 h after smoke exposure in the basal fraction, but not in the luminal fraction of ALI-PBEC, whereas smoke-induced mRNA expression of IL-8 was observed in both the basal and the luminal fraction. Examination of RNase 7 protein expression was subsequently determined by assessing colocalization of RNase 7 with p63 using immunofluorescence confocal imaging. Confocal imaging of the apical side of ALI-PBEC lacking p63⁺ cells, demonstrated that RNase 7 expression could not be observed in both air- and smoke-exposed ALI-PBEC after 6 h (Fig. 4B,). However, RNase 7 colocalized with p63⁺ cells at the basal side of ALI-PBEC, which was more abundant in cigarette smoke–exposed cultures than air controls. In these confocal studies, we did not observe a decrease in p63 protein levels or the number of p63⁺ cells in smoke-exposed cultures compared with air controls, contrasting the observed decreased mRNA expression of p63. To quantify RNase 7 expression in BCs, we performed flow cytometric analysis examining RNase 7 expression in KRT5⁺ cells in ALI-PBEC, 6 h after air/smoke exposure (Fig. 4C). Cigarette smoke significantly increased RNase 7 in the KRT5⁺ fraction compared with air exposed cultures (Fig. 4D), whereas no effect was observed in the KRT5^- fraction. Similar to p63 protein detection by confocal imaging, we did not observed a difference in the ratio of KRT5^- and KRT5⁺ cells upon smoke exposure (data not shown). Supporting our hypothesis, we observed that smoke exposure induced RNase 7 expression in ALI-PBEC selectively in BCs.

We further evaluated the additional effect of bacterial stimulation on smoke-induced RNase 7 expression. After smoke exposure, ALI-PBEC were stimulated with either NTHi or the TLR2 ligand Pam₃CSK₄. Consistent with previous studies examining the effect of whole cigarette smoke on bacterial induced expression of IL-8 and hBD-2 (33), cigarette smoke inhibited NTHi- and Pam₃CSK₄-induced hBD-2 expression while increasing smoke-induced IL-8 mRNA expression (Fig. 5). In line with the observation in Fig. 1B, NTHi did not induce RNase 7 mRNA expression at 3, 12, and 24 h after exposure (Fig. 5A, 5B). Moreover, NTHi did not further increase smoke-induced RNase 7 mRNA levels. Similar to NTHi, TLR2 activation with Pam₃CSK₄ did not affect RNase 7 mRNA expression after 3 h of stimulation (Fig. 5C).

Airway epithelial BCs are characterized by selective expression of the EGFR, a receptor that is activated upon disruption of the epithelial barrier function by cigarette smoke (30, 34). Previous studies reported that cigarette smoke extract induced expression of IL-8 in airway epithelial cells, which required activation of EGFR and the downstream MAPK/ERKs 1/2 (ERK1/2) (35, 36). Therefore, we studied whether EGFR and ERK1/2 signaling were required in cigarette smoke–induced expression of RNase 7. Our results clearly showed that similar to cigarette smoke extract, whole cigarette smoke exposure increased phosphorylation of EGFR and ERK1/2 in ALI-PBEC 15 min after exposure (Fig. 6A, 6B). Inhibition of EGFR using a neutralizing Ab, the EGFR tyrosine kinase inhibitor AG1478, and the ERK1/2-activating kinases MEK1/2 inhibitor U0126, attenuated activation of ERK1/2, demonstrating that indeed smoke activates ERK1/2 in an EGFR-dependent manner. Pharmacological inhibition of EGFR with AG1478 and inhibition of the ERK1/2-activating kinases MEK1/2 with U0126 caused a significant decrease in smoke-induced mRNA expression of RNase 7 (Fig. 6C). Similar to RNase 7, and in line with previous studies, mRNA expression of IL-8 and CCL20 was also impaired upon inhibition of EGFR and MEK/ERK1/2, while a trend was observed for IL-6 expression. To further support the conclusion that EGFR activation mediates RNase 7 expression, we examined the effect of direct receptor activation by stimulating ALI-PBEC with TGF-α. Indeed, in line with the effect of cigarette smoke, 3 h stimulation with TGF-α significantly enhanced RNase 7 mRNA expression and furthermore increased IL-8 expression; the increase in CCL20 did not reach statistical significance (Fig. 6D). In conclusion, we observed that cigarette smoke–induced expression of RNase 7 requires activation of the BC-specific receptor EGFR and the downstream ERK1/2 signaling pathway.

In this study, we examined the expression of the antimicrobial RNase 7 in cultured airway epithelial cells. We demonstrated that undifferentiated S-PBEC, primarily consisting of BCs, but not mucociliary-differentiated ALI-PBEC, expressed RNase 7 upon stimulation with the respiratory pathogen NTHi. This was in marked contrast to other examined innate immune mediators that were induced by NTHi in both S-PBEC and ALI-PBEC. Furthermore, differentiated ALI-PBEC displayed a decreased baseline RNase 7 expression compared with S-PBEC after 2 wk of culturing in air-exposed conditions. Cigarette smoke exposure, which caused transient disruption of the epithelial barrier function, reinitiated RNase 7 expression in the remaining undifferentiated BCs in ALI-PBEC, which was demonstrated using confocal microscopy and isolation of luminal and basal cell fractions. Moreover, we showed that this induction required activation of EGFR and the downstream ERK1/2 signaling pathway and furthermore demonstrated that direct EGFR activation by TGF-α also induced RNase 7 expression. In contrast to the innate immune mediators IL-8 and hBD-2, bacterial costimulation did not further affect smoke-induced RNase 7 expression in ALI-PBEC. This data suggest that expression of RNase 7 is a unique innate host defense property of BCs that is induced in conditions of airway epithelial injury.

To our knowledge, we demonstrate for the first time that BCs display a phenotype-distinct antimicrobial response by producing RNase 7. The epithelial phenotype–specific expression of AMPs has been well described in multiple mucosal tissues (37). For instance, intestinal Paneth cells selectively produce human α-defensin 5 and 6, which are not produced by enterocytes (38). In the lung, phenotype-specific expression of AMPs has been observed in submucosal gland serous cells, which produce lactoferrin and lysozyme (39, 40) and also in alveolar cells producing antimicrobial surfactant protein A and D (41). Phenotype-specific expression of RNase 7 has been reported in the skin and urinary tract, demonstrating specific expression by respectively suprabasal keratinocytes and α- and β-intercalated cells (16, 18). In contrast to these epithelial cells, phenotype-distinct expression of AMPs by airway epithelial BCs has not yet been reported.

Similar to previous in vitro observations demonstrating BC-specific expression of IL-33 and EGFR (10, 34), we observed decreased expression of RNase 7 after mucociliary differentiation of PBEC. Consistent with this observation, it was shown in a previous transcriptome analysis that RNase 7 transcripts were increased in BCs but not in differentiated airway epithelial cells (28). Another transcriptome analysis demonstrated enhanced mRNA expression of RNase 1, 4, and 6 in differentiated airway epithelial cells compared with undifferentiated cultures (42), suggesting that RNase 7 production by BCs is unique among other secreted RNases. Interestingly, we did not observe a difference in RNase 7 mRNA expression after 1 wk of mucociliary differentiation of ALI-PBEC. As the amount of BCs declines during the first week of differentiation, we reasoned that the attenuation of RNase 7 expression after 2 wk of differentiation was not caused by a decrease in the number of BCs in cultured ALI-PBEC but because of another mechanism. Because undifferentiated ALI-PBEC initially display a low epithelial barrier function that progresses during differentiation, we hypothesized that RNase 7 expression by BCs in ALI-PBEC is attenuated because of an enhanced airway epithelial barrier function. Our finding that cigarette smoke can reinitiate RNase 7 expression by BCs in ALI-PBEC supported this hypothesis. Previous data demonstrate that E-cadherin–mediated cell–cell contacts inhibit EGFR signaling in airway epithelial cells (43). Because we observed that cigarette smoke decreases epithelial barrier function, which has been shown to be related to interference with the function of E-cadherin (31), and because smoke-induced RNase 7 expression involved EGFR activation, we reason that smoke-induced barrier disruption contributes to RNase 7 expression. In skin host defense, it is well established that injury enhances EGFR-dependent innate immune responses by keratinocytes that provide protection against infections during wound repair (44, 45). Also, in airway epithelial cells, it has been shown that EGFR activation promotes innate immune responses that are depending on the MAPK/ERK1/2 signaling pathway, which also facilitates epithelial wound healing (36, 46). These studies primarily demonstrate EGFR-dependent expression of proinflammatory innate immune mediators such as IL-8, IL-6, and CCL20, which were also induced by cigarette smoke in our model. However, to our knowledge, this is the first study reporting EGFR-dependent AMP expression in airway epithelial cells. Although smoke-induced RNase 7 expression was specifically detected in BCs, we observed secretion of RNase 7 in the apical surface liquid. This may be the result of polarized secretion to the apical side, which can be mediated in part by transepithelial projections between the lateral intracellular spaces formed by BCs (47). These projections might explain the apical secretion of RNase 7. However, this requires further investigation.

Besides decreased baseline RNase 7 expression in ALI-PBEC, we furthermore demonstrated that in contrast to S-PBEC, ALI-PBEC did not induce RNase 7 expression after stimulation with NTHi. Also, additional stimulation with NTHi or the TLR2 ligand Pam3CSK4 did not further enhance cigarette smoke–induced RNase 7 expression, whereas expression of IL-8 was further increased. It can be speculated that the inability of bacterial stimuli to induce RNase 7 expression in BCs of ALI-PBEC is caused by a lack of direct microbial contact of BCs because of apical application of the NTHi and Pam₃CSK₄. Similar to the pseudostratified airway epithelium in vivo, undifferentiated BCs in cultured ALI-PBEC are superimposed by differentiated LCs. In intact airway epithelium, this mechanism might protect BCs from direct microbial contact and subsequently prevents RNase 7 expression. Another explanation of the absence of bacterial induced RNase 7 expression in BCs of ALI-PBEC might be that pattern recognition receptors such as TLR2 are localized at the basolateral site of BCs because of cell polarization. Finally, we cannot formally exclude the possibility that, in contrast to UV-inactivated bacteria, life-NTHi may promote RNase 7 expression in basal cells in vivo (e.g., as a result of epithelial barrier disruption).

In contrast to RNase 7, NTHi-induced hBD-2 in airway epithelial cells was independent of the epithelial differentiation status. Furthermore, cigarette smoke exposure impaired bacterial-induced hBD-2 (33). This finding is in line with a decreased antimicrobial defense of airway epithelial cells exposed to cigarette smoke, which provides an explanation for the increased susceptibility of smokers and chronic obstructive pulmonary disease patients toward respiratory infections (33). We argue that the decreased antimicrobial defense caused by smoking results from the attenuation of a large number of AMPs including hBD-2, which cannot be compensated by RNase 7 expression. To further explore the significance of RNase 7 in airway epithelial host defense upon barrier disruption, other environmental factors should be examined that affect the epithelial barrier function but do not attenuate the expression of other AMPs.

On the basis of our observations, we propose a novel model of airway epithelial host defense that includes a second line of defense mediated by airway BCs, which is induced upon injury. In the intact airway epithelium, LCs are the main cells that mediate host defense via, for example, production of AMPs such as hBD-2 and mucociliary clearance (Fig. 7A). LCs furthermore prevent microbial contact of the basolateral localized BCs, thereby preventing RNase 7 expression. However, BCs do produce RNase 7 upon disruption of the epithelial barrier integrity by inhaled substances such as cigarette smoke (Fig. 7B). Moreover, in damaged epithelium lacking LCs, microbial exposure of BCs induces innate immune responses including cell type–specific expression of RNase 7. This way BCs provide additional innate immune defense against microbial infections during the course of epithelial wound repair.

We thank Jerry van der Ploeg (Department of Instrumental Services, Leiden University Medical Center) for technical assistance in developing the whole cigarette smoke exposure model. We also thank Dr. Joop Wiegant, (Department of Molecular Cell Biology, Leiden University Medical Center) for technical expertise using the Leica TCS SP5 Confocal Microscope, and Sanne de Jong (Department of Parasitology, Leiden University Medical Center) for help with the flow cytometry analysis. We furthermore thank Dr. Jamil Aarbiou (Galapagos BV, the Netherlands) for helpful discussion.

The authors have no financial conflicts of interest.