ENaC, iNOS, mucins expression and wound healing in cystic fibrosis airway epithelial and submucosal cells

Overview

Abstract:
We compared airway epithelial cell models relevant for cystic fibrosis (CF): 16HBE cells with endogenous wild-type cystic fibrosis transmembrane conductance regulator (CFTR), CFBE cells with mutated ΔF508-CFTR, corrected CFBE cells overexpressing CFTR, CFSME (CF submucosal) and Calu-3 (non-CF submucosal) cells with respect to the epithelial sodium channel (ENaC), inducible NO synthase (iNOS) and mucins (MUC) (studied by quantitative Real-Time-Polymerase Chain Reaction, qRT-PCR and Western blot), and wound healing.

CFBE cells had significantly more expression of β- and γ-ENaC mRNA and of β-ENaC protein than 16HBE cells. Compared to corrected CFBE cells, CFBE cells had increased mRNA expression of all ENaC subunits and β-ENaC protein. For ENaC, the CFSME/Calu-3 mRNA ratio was very low and contradictory to the ENaC upregulation in CF cells. CFBE cells showed decreased expression of iNOS at both mRNA and protein levels compared to 16HBE cells and only at the mRNA level compared to corrected CFBE cells. CFSME cells showed expression of iNOS whereas Calu-3 cells did not. Higher expression of MUC2 and MUC5B was found in corrected CFBE cells compared to CFBE cells. Wound healing in CFBE cells was delayed compared to corrected CFBE cells, but not to 16HBE cells, and in CFSME cells compared to Calu-3 cells.

Our data suggest CFSME as an inappropriate CF cell model for Calu-3 cells, and provide partial support for the theory that the differences (in ENaC, iNOS and wound healing) between these cell lines are associated to the presence of CFTR in the bronchial airway epithelial cells.

Keywords:
CFTR; CFTR inh-172; ENaC; iNOS; mucin; wound healing

Introduction

Cystic fibrosis (CF) is an inborn genetic disease due to a mutation in the gene coding for cystic fibrosis transmembrane conductance regulator (CFTR) (Anderson et al., 1991). CFTR is a cAMP-activated chloride (Cl⁻) channel critically involved in the regulation of epithelial surface fluid volume and composition. The mutation in CFTR results in abnormally viscous secretions that can obstruct the airway lumen. The chronic airway disease, including airway obstruction, appears as the main symptom in CF and the most commonly observed structural abnormality is ‘bronchiectasis’ with repeated cycles of epithelial healing. There are many differences between CF and normal airways. One possibility is that since CF is a monogenic disease, all these differences have a common origin in the mutation in CFTR and that there is a simple relationship between CFTR and the observed differences in other parameters. It has, however, also been suggested that the mutation in CFTR may have wide-spread effects on other cellular processes (Mehta, 2005), including signalling processes, and that CFTR would act as a hub in the cellular metabolism (Kunzelmann and Mehta, 2013). In that case, the cellular effects of the mutation in CFTR would be complex. This might result in the varied clinical symptoms that affect CF patients.

Another ion channel that has been extensively discussed in relation to CF pathology is the epithelial sodium channel (ENaC), responsible for Na⁺ transport. ENaC is composed of a number of subunits, initially three (α-, β- and γ-subunits) (Canessa et al., 1994) were described, but an additional δ-subunit has also been identified (Ji et al., 2006; Giraldez et al., 2012; Ji et al., 2012). The interaction of CFTR with ENaC is the most extensively studied regulatory interaction. A direct physical interaction between CFTR and ENaC has been shown (Berdiev et al., 2007). Many studies have shown ENaC downregulation by CFTR (Stutts et al., 1995; Letz and Korbmacher, 1997; Kunzelmann and Schreiber, 1999), but more recently this downregulation of ENaC by CFTR has been questioned (Collawn et al., 2012; Livraghi-Butrico et al., 2013). Hence, the relation between CFTR and ENaC is unclear, and the question whether CF is (mainly) due to reduced Cl⁻ secretion or to hyperabsorption of Na⁺ is still not answered (Kunzelmann and Schreiber, 2012).

It also remains to be elucidated whether mucus hypersecretion in CF is linked to the mutation in CFTR. The main clinical manifestation of CF is accumulation of mucus on epithelial surfaces such as respiratory epithelia (Boucher, 2002; Nadel et al., 2010). MUC2, MUC5AC and MUC5B are gel-forming secretory mucins (Hovenberg et al., 1997; Wickstrom et al., 1998). The hypotheses that CFTR expression is linked to the hyperexpression of mucins fall into two categories. Some hypotheses directly link CFTR to mucus hyperproduction in CF (Zhang et al., 1995; Xia et al., 2005), while others state that mucus hyperproduction is due to downstream effects of the CFTR mutation (Kreda et al., 2012). The CFTR mutation results in low levels of exhaled nitric oxide (NO) and reduced inducible NO synthase (iNOS) expression in CF patients (Kelley and Drumm, 1998; Moeller et al., 2006) and makes the airway more vulnerable to inflammation and infection. Infection not only increases MUC2 (Li et al., 1997) MUC5AC and MUC5B (Rose and Voynow, 2006) mRNA expression in CF airways, but also results in increased bacterial colonisation leading to loss of cell membrane permeability and selectivity, leading to airway epithelial injury (Voynow et al., 2005). The structural integrity of the airway epithelium is important for its proper functioning as a physical barrier and as a regulator of ion and water transport into and out of cells. The airway epithelium is capable of repair by cell migration and proliferation (Zahm et al., 1997). CFTR has been shown to play a role in wound repair (Trinh et al., 2008; Schiller et al., 2010) and such data were partly based on the use of CFTR inhibitor-172 which was described by others to have unspecific effects (Kelly et al., 2010; Itani et al., 2011).

Together, these data suggest that it would be of interest to test the hypothesis that the presence of CFTR affects the expression of ENaC, iNOS, mucins and also the rate of wound healing. For such a study, in vivo measurements appear unsuitable, since fluid transport in the airways of CF patients is likely to be affected, among other factors, by bacterial infection. Hence, cultured airway epithelial cells might be a more suitable system, also because in addition to a CF cell line (homozygous for the ΔF508-CFTR mutation), and a cell line of non-CF (‘healthy’) origin (Cozens et al., 1994), a ‘corrected’ CF cell line is available in which wild-type CFTR was transfected and is overexpressed (Illek et al., 2008). This would allow testing of the hypothesis that differences between CF airway epithelia and healthy airway epithelia would be due to the amount of (functional) CFTR in the cell membrane (no or very little in CF cells, normal in ‘healthy’ control cell, and excessive in ‘corrected CF’ cells). Such a test would not be possible in vivo. In addition, we have used another experimental system, airway submucosal gland cells. For the submucosal gland cells, non-CF Calu-3 cells and CFSME cells (compound heterozygotes for the ΔF508-CFTR and the 2QX mutation) (da Paula et al., 2005) were used. The CFSMEo- cell line is assumed to be a good CF cell model showing very low CFTR mRNA and no CFTR protein expression (da Paula et al., 2005), while Calu-3 cells express very high levels of CFTR at the mRNA and protein levels (Farinha et al., 2004; MacVinish et al., 2007). Such a gradient of CFTR also exists in the surface epithelial cells where CFTR expression shows the following gradient: corrected CFBE > 16HBE > CFBE cells (Farinha et al., 2004; Illek et al., 2008).

The purpose of the present study hence was to compare a CF bronchial cell model to their two non-CF counter parts (one with wild type endogenous CFTR, and the other with CFTR-corrected isogenic cells) to see whether there is a straightforward link between the presence of wild-type CFTR or mutated CFTR and other parameters of potential physiological and clinical relevance, such as the expression of ENaC and iNOS, mucins and epithelial repair. Although some studies regarding ENaC, iNOS, MUC (2, 5AC and 5B), and wound healing, have been performed on CF-cells and non-CF cells (Ismailov et al., 1996; Berdiev et al., 2007; Schiller et al., 2010; Rubenstein et al., 2011; Trinh et al., 2012) neither the set of non-CF, CF and corrected CF surface epithelial cells, nor the submucosal gland cells have been the subject of such a comparison. Since a CF model for Calu-3 cells is not available, comparisons based on the previous findings would be helpful to determine if CFSME cells are appropriate CF models for Calu-3 cells.

If there is a straightforward relation between CFTR and changes in the above parameters, we expect that (1) the differences between the corrected CFBE cells overexpressing wild-type CFTR, and the CFBE cells with mutant (and largely non-functional) ΔF508-CFTR should be same as the differences between 16HBE expressing endogenous CFTR and CFBE cells, (2) the differences between CFBE cells and the normal human bronchial epithelial cells should at least be partially mimicked by exposure of the normal human bronchial epithelial cells to the CFTR inhibitor and (3) the differences between the CFSME (CF submucosal gland cells) and Calu-3 cells (their non-CF counterparts) should be similar to the differences between CF and non-CF bronchial epithelial cells.

Materials and methods

Cell culture

16HBE14o-(wt/wt CFTR) (normal human bronchial epithelial cells, wild-type CFTR), CFBE41o-(ΔF508/ΔF508), (cystic fibrosis bronchial epithelial cells, homozygous for the ΔF508 mutation) and its corresponding plasmid-corrected CFBE41o-pCep4 (CFBE41oc10-6.2wt pCep4), overexpressing wtCFTR) cell line (Cozens et al., 1994; Illek et al., 2008) were used as a model for airway surface epithelial cells. Under the culture conditions used, these cells are polarised. Under other culture conditions (air–liquid interface), the surface epithelial cells can develop into ciliated cells (Kozlova et al., 2006), but in the present study, the cells were cultured in liquid medium only and did not develop cilia. The CFBE41o-cells have in our hands a very low cAMP-induced Cl⁻ efflux and a low CFTR expression compared to the 16HBE14o-cells (Andersson et al., 2002; Servetnyk et al., 2006). The cells were cultured in T75 flasks (Sarstedt, Landskrona, Sweden) in Medium 199 (Invitrogen/Life Technologies, Carlsbad, CA, USA). Calu-3 (submucosal gland with wild type CFTR), and CFSMEo-(CF submucosal gland epithelial cells, ΔF508/2QX) (Cozens et al., 1992) cells as a model for airway submucosal gland cells were grown in Minimum Essential Medium with GlutaMAX™ (Invitrogen/Life Technologies) supplemented with 1% non essentatial amino acids (Sigma–Aldrich, St. Louis, MO, USA). Cells were maintained in the presence of 10% heat inactivated fetal bovine serum (FBS) (Invitrogen/Life Technologies) and 1% penicillin–streptomycin (PEST) at 37°C in a humidified atmosphere of 5% CO₂. Calu-3 cells were obtained from the American Type Culture Collection (ATCC) (Manassas, VA, USA), the other cells were a kind gift of Dr D. Gruenert (San Francisco, CA, USA).

Cell treatments with CFTRinh-172 and forskolin

Cell monolayers were treated with 20 µM CFTR inhibitor-172 ((5-[(4 carboxyphenyl)methylene]-2-thioxo-3-[(3-trifluoromethyl)phenyl-4-thiazolidinone), and 20 µM forskolin (an activator of adenylate cyclase) for 20 h to find the effect of these compounds on the mRNA expression of ENaC, iNOS and MUC (2, 5AC and 5B), and on the epithelial wound healing time.

Quantitative real time-polymerase chain reaction (qRT-PCR)

RNA extraction was performed by a commercially available purification RNeasy^® Plus Micro kit (Qiagen, Hilden, Germany). RNA quantification was carried out by spectrophotometry using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). One microgram of RNA was reverse transcribed to cDNA by a High Capacity Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). For the evaluation of mRNA expression samples were analysed by qRT-PCR, using the TaqMan Gene Expression Assay consisting of unlabeled primers and target specific FAM™ dye labelled TaqMan^® probe, specific for MUC2 (Hs00159374_m1), MUC5AC (Hs01370716_m1), MUC5B (Hs00861588_m1), iNOS (Hs01075529_m1), α-ENaC (Hs00168906_m1), β-ENaC (Hs00165722_m1) and γ-ENaC (Hs00168918_m1). Target amplification was carried out by using Gene Expression Master Mix using an Applied Biosystems Fast 7500 Sequence Detection System. The mRNA expression of the samples was normalised against glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (HS 99 99 99 05_m1). For validation of GAPDH as an internal control, the effect of treatment (with CFTRinh 172 or forskolin) on the expression of the GAPDH was analysed by calculating the fold change by 2 ^-∆C_t, which was always 1.0 (Livak and Schmittgen, 2001). All PCR reagents as well as primer and probes were obtained from Applied Biosystems. Samples were run as singleplex in triplicate wells. We also treated CFBE cells with CFTRinh-172 and forskolin, to serve as a negative control.

Western blotting

Protein expression of ENaC and iNOS in 16HBE, CFBE and corrected CFBE cells was measured by Western blotting. The cell lysate was obtained in cold radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 50 mM Tris, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 1% Triton X 100) with added protease inhibitor cocktail (Sigma–Aldrich). The protein concentration was measured by the DC protein assay (Bio-Rad Laboratories, Hercules, CA, USA). 25µg protein was separated on 7.5% Tris-HCl ready gel (Bio-Rad) under denaturing and reduced conditions. Blotting to polyvinylidene fluoride (PVDF) membrane was performed according to standard techniques. Non-specific binding was blocked with 5% bovine serum albumin (BSA) for 2 h at room temperature.

Primary antibody incubation for the α- (1:1,000), β- (1:100), and γ-ENaC (1:1,000) subunit, and iNOS (1:1,000) was performed overnight with antibodies sc-22237, sc-48428, sc-22245 and sc-8310, respectively. Membranes were incubated with horseradish peroxidase (HRP) coupled secondary antibodies for 1 h at room temperature. We used Donkey anti-goat IgG-HRP (Sc-2020) (1:10,000) for α- and γ- ENaC subunits, goat anti-mouse IgG-HRP (Sc-2005) (1:10,000) for the β-ENaC subunit and goat anti-rabbit IgG-HRP (Sc-2004) (1:20,000) for iNOS. β-tubulin (1:1,000) (sc-9104) was used as a loading control. Precision Plus Protein™ Standards (Bio-Rad) were used as molecular weight markers. As positive controls we used KNRK (Rat K Ras transformed kidney cell) whole cell lysate for α- and β-ENaC and A549 (human lung carcinoma) cell lysate for γ-ENaC, and RAW 264.7 + LPS/IFN-γ (Mouse leukaemic monocyte macrophage stimulated with lipopolysaccharide plus interferon-γ) cell lysate for iNOS according to the manufacturer's instructions.

All primary and secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

Detection was carried out by the Amersham enhanced chemiluminescence (ECL) advanced Western blotting detection kit (GE Healthcare, Chalfont St. Giles, UK). Images of bands were acquired by ChemiDoc XRS (Bio-Rad).

Epithelial wound healing by scratch assay

To assess the differences in the epithelial wound healing capability of CF and non-CF cell lines, cells were grown to confluence in 6-well plates and starved overnight in serum-free medium before mechanical wounds were made. Three linear scratches were made in each well with a 120 µL pipette tip (Biohit/Sartorius, Helsinki, Finland). The monolayer was washed and maintained in serum-free medium throughout the experiment. The wound healing process was followed in the absence and presence, respectively, of 20 µM CFTRinh-172 (dissolved in dimethylsulfoxide, DMSO), and forskolin (which raises the intracellular levels of cAMP) (dissolved in ethanol). The final concentrations of DMSO and ethanol in the cell culture were 0.08 and 0.04%, respectively, and had no effect on the wound healing (data not shown). Wounds were photographed at 10× magnification in an Olympus (Tokyo, Japan) inverted microscope at 0, 6, 16 and 20 h. Lines were drawn under the plates so that each wound was photographed at the same position at each time point. Images were analysed for the exact quantification of the healed area by the Wimasis image analysis Wimscratch tool (Wimasis, Munich, Germany). Data are presented as a percentage of scratched area at time 0 h. The comparison was based on wounds with similar scratch areas at 0 h.

Cell migration assay

Differences in cell migration among cell lines were characterised and quantitatively measured by using inserts with a polyethylene terephthalate membrane with 8 µm pore size (BD Biosciences, San Jose, CA, USA) kept in 24-well plates (BD Biosciences). The cell monolayer was starved in serum-free medium overnight before the migration assay was performed. Cells were harvested in trypsin/EDTA (0.025/0.01%), and neutralised with trypsin neutralising solution containing 0.05% trypsin inhibitor from soybean and 0.1% BSA (Invitrogen/Life Technologies). 70,000 cells were seeded in each insert in serum-free medium and allowed to migrate towards medium containing 0, 5 or 10% FBS, respectively, for 6 h at 37°C with 5% CO₂. To quantify the migration, cells on the upper surface of the insert were removed using a cotton swab. Migrated cells attached to the lower surface of the insert were fixed with 3.7% formaldehyde for 10 min, permeabilised with methanol and stained with 0.1% crystal violet in 20% ethanol overnight. Inserts were washed with water to remove extra stain, and finally stained migrated cells were counted in three different fields at 20× magnification.

Cell proliferation assay

To determine the proliferative capacities of 16HBE, CFBE and corrected CFBE cells, cells were seeded at a density of 5,000 cells/well in black 96-well plates (Corning Life Sciences, Corning, NY, USA) in serum-free medium. Eight replicates were included for each cell type along with control wells without cells. Cells were incubated at 37°C with 5% CO₂, and the serum-free medium was changed every other day. We used the Cyquant direct cell proliferation assay kit (Invitrogen/Life Technologies), which is based on a live cell-permeant DNA binding fluorescent dye. At the desired time (day 0, day 1, day 2 and day 3) cells were incubated with nucleic acid stain in combination with a background suppression reagent for 60 min at 37°C. The fluorescence intensity was measured at 480/535 nm in the FLUOstar Optima fluorescence plate reader (BMG Labtech, Offenburg, Germany). Fluorescence intensities were normalised to the control wells lacking cells.

Neutral Red uptake viability assay

The Neutral Red uptake assay (Repetto et al., 2008) was performed to determine the effect of CFTRinh-172, and forskolin on cell viability. Cells were cultured in 96-well plates to confluence and monolayers were treated with varying concentrations (5, 10, 15, 20, 30, 40, 50 µM) of these substances for 20 h. After completion of the treatment, cells were washed with Dulbecco's phosphate-buffered saline (DPBS), and incubated with 4 µg/mL Neutral Red dye (3-amino-7-dimethylamino-2-methyl-phenazine hydrochloride) (Sigma–Aldrich) prepared in Medium 199, for 2 h under cell culture conditions. The cells were washed, and the dye was extracted with acidified ethanol, and quantified at 540 nm with a Multiskan Ascent^® plate reader (Thermo Labsystems, Stockholm, Sweden). To determine the cytotoxic effect of the concentrations of DMSO and ethanol used during this study we included solvent controls. The data are presented as percentage of viable cells.

Statistical analysis

Significance of the differences was calculated by the unpaired t-test, and by one-way analysis of variance. Dunnett's multiple comparison test, and Bonferroni's Multiple comparison test (Graphpad Prism^® version 4; San Diego, CA, USA) were used. Statistical significance of differences is indicated as * or # (P < 0.05), ** or ## (P < 0.01), and *** or ### (P < 0.001) as indicated in the legends. The data represent means ± standard error of the mean (S.E.M.).

Results and discussion

Epithelial sodium channel (ENaC)

CFBE cells showed significantly more mRNA expression of the β- and γ-ENaC subunits (but not of the α-subunit) in comparison with 16HBE cells. The mRNA expression of all three ENaC subunits in the CFBE cells was significantly increased in comparison with the corrected CFBE cells. Comparison between corrected CFBE and 16HBE (both non-CF) cells showed significant differences: corrected CFBE showed less α-ENaC, and more β- and γ-ENaC compared to 16HBE cells (Figure 1A). On the other hand, the CF submucosal gland cell line CFSME showed significantly less expression of ENaC subunits, compared to the non-CF Calu-3 cells. CFSME cells showed a decreased expression of α-ENaC (0.19 ± 0.14 fold), β-ENaC (0.02 ± 0.03 fold) and γ-ENaC (0.012 ± 0.3 fold) compared to Calu-3 cells (Figure 1B). Western blotting showed no significant differences in band density in any bronchial epithelial cell type of α- and γ-ENaC, but CFBE cells displayed a small but significant increase in the expression of β-ENaC compared to 16HBE (1.2 ± 0.05 fold), and corrected CFBE cells (1.3 ± 0.03 fold). No significant differences between the surface epithelial cell lines in the protein expression of the α- and γ-ENaC subunits were found (Figures 2A and 2B).

CFBE cells showed increased mRNA expression of all ENaC subunits compared to the CFTR overexpressing corrected CFBE cells, and increased mRNA expression of β- and γ- ENaC compared to the endogenous CFTR containing 16HBE cells, but at the protein level only a small increase in β-ENaC expression was found in CFBE cells in comparison with 16HBE cells. It is possible that more refined methods of evaluating the Western blot results would decrease the discrepancy between the results at the protein level and at the mRNA level. The finding at the mRNA level gave partial support to the notion of a link between the presence/amount of wild type CFTR and downregulation of ENaC mRNA expression. The importance of overexpression of β-ENaC was highlighted in a CF mouse model (Mall et al., 2004). Despite the presence of endogenous CFTR in this β-ENaC overexpressing mouse model, the ratio of ENaC to CFTR was suggested to be too large to suppress the effect of β-ENaC (Collawn et al., 2012). In an attempt to balance this ENaC/CFTR ratio, human CFTR was overexpressed in a β-ENaC overexpressing mouse model, which, however, failed to reverse the CF phenotype (Grubb et al., 2012). An early study claimed that there exists no difference in the expression of the ENaC subunits in the nasal epithelium of CF patients versus healthy controls (Burch et al., 1995), but if overexpression of CFTR in a cell would have suppressed the mRNA expression of ENaC, then we expected decreased expression of these three ENaC subunits in corrected CFBE cells (having overexpressed wild type CFTR) compared to the 16HBE cells (having endogenous wild type CFTR), but such a suppression was shown for the α-ENaC subunit only. In addition, some have claimed that results from mouse lung have shown that regulation of ENaC activity by CFTR is inversely correlated to CFTR levels (Lazrak et al., 2011), while others have disputed this (Collawn et al., 2012).

Evidence that changes in the ENaC mRNA and protein expression level affect the sodium transport in epithelial cells was provided by Bangel et al. (2008) who showed that increased expression of all three ENaC subunits at the mRNA level, and of α- and β-ENaC at the protein level in CF tissue is responsible for the increased sodium absorption through ENaC in CF epithelia (Bangel et al., 2008). The molecular mechanisms leading to the increase in ENaC mRNA transcription are still unknown. If ENaC is inhibited by CFTR, this inhibition would likely be absent in CF airway epithelia, which would lead to overactive ENaC and hence overactive Na⁺ (Sobczak et al., 2009) and water influx, resulting in reduced airway surface liquid (ASL) volume, and dehydrated airway surfaces (Boucher, 2002), but recent studies have presented evidence from a pig model for CF and from primary cultures of and excised human airway epithelia, that there is no increased Na⁺ absorption in CF airway epithelia (Chen et al., 2010; Itani et al., 2011). The significance of ENaC in CF is hence unclear and controversial (Collawn et al., 2012).

For ENaC, the CFSME/Calu-3 mRNA ratio was very low and the opposite of that for the CFBE/corrected CFBE, and the CFBE/16HBE pairs. An association between ENaC overexpression and CFTR mutation was established at the mRNA and/or protein expression levels more clearly in the CFBE/corrected CFBE pair compared to the CFBE/16HBE pair. Na⁺ uptake would probably not be enhanced in CFBE cells because it is questionable whether ENaC is active in the 16HBE and CFBE cell lines (Cozens et al., 1994), but this association might be relevant in vivo where ENaC is active in airway epithelial cells.

Inducible NO synthase (iNOS)

CFBE cells showed significantly less iNOS mRNA expression compared to 16HBE and corrected CFBE cells (Figure 1A). The expression of iNOS protein was significantly lower in CFBE cells compared to 16HBE cells (0.88 ± 0.05 fold). The difference in iNOS protein expression between CFBE and corrected CFBE was not significant (Figures 2A and 2B). A comparison between the CFSME/Calu-3 pair showed mRNA expression of iNOS in CFSME cells while Calu-3 did not show any expression of this enzyme (data not shown).

A decreased expression of iNOS in the CF airway epithelial cells is well established and may increase the susceptibility for bacterial colonisation as iNOS is generally considered to have a positive effect on the anti-bacterial defence of the host (Meng et al., 1998). Studies have indicated an association between mutated CFTR and reduced iNOS expression (Kelley and Drumm, 1998; Steagall et al., 2000). In the present study such an association for iNOS was shown at the mRNA level in both bronchial epithelial cells while an association at the protein expression level was only detected in the CFBE/16HBE pair, but in the CFSME/Calu-3 pair we could not find such an association.

Mucins

We found no measurable mRNA expression of MUC (2, 5AC and 5B) in 16HBE cells (even after converting 2 µg of total RNA into cDNA), but found a significantly higher expression of MUC2 (3.5 ± 0.2 fold) and MUC5B (1.9 ± 0.16 fold) in corrected CFBE cells compared to CFBE cells (Figure 1C). Neither CFBE nor corrected CFBE cells showed any expression of MUC5AC. Expression of MUC2 was only observed in CFSME cells and expression of MUC5AC only in Calu-3 cells. Expression of MUC5B was found neither in CFSME nor in Calu-3 cells (data not shown).

Mucus secretion is abnormal in CF airways and MUC2, MUC 5AC and MUC 5B are the secretory mucins. Contradictory to the hypothesis that mucins (MUC5AC and MUC5B) are overproduced in CF lung secretions (Kirkham et al., 2002), it has been shown that mRNA expression and secretion of MUC5AC and MUC5B were very significantly decreased in CF airway epithelia compared with normal cells (Voynow et al., 1998; Henke et al., 2004; Martinez-Anton et al., 2006; Hajj et al., 2007a) and that their expression increased during a CF-related pulmonary exacerbation (Henke et al., 2004). The hypothesis that CFTR is directly linked to mucus production in CF (Zhang et al., 1995; Xia et al., 2005) is less likely than the hypothesis that increased mucus production is secondary to the mutation in CFTR (Kreda et al., 2012). The downstream consequences of the mutation in CFTR (i.e., dehydrated airway epithelial surface, and chronic inflammation and infection) provide a stimulus for the mucin gene upregulation (Kreda et al., 2012). Similarly, the data in the present study do not explain that the expression levels of mucins are directly related to the presence of wild-type versus mutated CFTR.

Epithelial wound healing, migration and proliferation

The epithelial wound healing time in CF cells was compared with that in its non-CF counterparts by the commonly employed wound healing assay, which is a highly reproducible technique, and provides an estimation of the capacity of cells to repair monolayers after mechanical injury (Trinh et al., 2008; Maille et al., 2011; Trinh et al., 2012). Wound repair was found to be significantly slower in CFBE cells compared to corrected CFBE cells, but the comparison between CFBE and 16HBE cells showed no significant difference in the healing time (Figure 3A). CFSME also showed a significant and a very prominent delay in wound healing time compared to the non-CF Calu-3 cells (Figure 3B). Since wound healing is dependent on cell proliferation and cell migration that are early phases in the repair process, these properties were investigated separately. No differences were found in the rate of cell proliferation between 16HBE, CFBE and corrected CFBE cells (Figure 4). The capacity for cell migration was determined under three conditions (towards serum-free medium (0% FBS), 5% FBS and 10% FBS, for the three surface epithelial cell lines. No significant differences in migration rate between 16HBE, CFBE and corrected CFBE cells depending on the serum concentration were observed. All bronchial epithelial cell lines showed a more pronounced migratory response towards medium to which serum had been added than towards serum-free medium (Figure 5).

The role of CFTR in airway epithelial wound healing has been studied in immortalised and primary cell culture models using CFTR silencing, CFTR correctors, and CFTR inhibitors (Trinh et al., 2008; Schiller et al., 2010; Trinh et al., 2012). We found no significant differences in the rate of wound healing between CFBE cells and 16HBE cells but faster wound healing in corrected CFBE cells, compared to CFBE cells. The results from the CFBE/corrected CFBE pair agree with those of Trinh et al. (2008) and Trinh et al. (2012) who found that their CF cell line (CuFi) showed slower wound healing than their normal airway epithelial cells (NuLi), and this was also the case for primary airway epithelial cells, whereas corrected CF cells showed faster wound healing (Trinh et al., 2008; Trinh et al., 2012), but wound healing data from the CFBE/16HBE pair contradicts these data. Trinh et al. (2008) found no difference in cell proliferation between CuFi and NuLi cells, which agrees with our finding that there are no significant differences in cell proliferation between 16HBE cells, CFBE cells and corrected CFBE cells. However, cell migration was lower in CuFi cells, which could explain the slower wound healing (Trinh et al., 2008). However, data on cell proliferation of CF epithelial cells are controversial: higher cell proliferation was associated with CF (Gallagher and Gottlieb, 2001; Hajj et al., 2007b), while CFTR inhibition reduced cell proliferation (Trinh et al., 2012).

A comparison between Calu-3 cells and CFSME cells that have been considered a good model for CF cells (da Paula et al., 2005) is complicated by the fact that Calu-3 cells are an adenocarcinoma cell line, positive for an activating K-RAS (p.G13D) gene mutation where RAS is known to induce cell migration and proliferation in cancer cells, while CFSME is a non-cancer cell line (Fleming et al., 2005; Morgillo et al., 2011). Therefore, Calu-3 cells showed much faster healing compared to the CFSME cells. Furthermore, MacVinish et al. (2007) (not aiming to study wound healing but rather to produce a Calu-3 cell line with a CF phenotype) reported that CFTR-silenced Calu-3 cells reached confluence more slowly than control Calu-3 cells, while Schiller et al. (2010) (aiming to study wound healing) attributed it to be the cause of delayed wound healing.

Efffect of CFTR inhibition and activation on the mRNA expression of ENaC, iNOS, mucin and epithelial repair time

To study if CFTR inhibition and CFTR activation by CFTR-inh172, and forskolin respectively, can show a CFTR mediated effect on the mRNA expression of ENaC, iNOS, mucins and wound healing, we used 20 µM CFTR-inh172, and 20 µM forskolin. CFTRinh-172 has been has been described as a specific and potent CFTR channel blocker (Tonghui et al., 2002). Forskolin activates adenylyl cyclase, which leads to the synthesis of cAMP and activation of CFTR (Denning et al., 1992). We did not study the effect of CFTRinh-172 and forskolin treatment on the mRNA expression of ENaC, iNOS and mucin in the CFSME/Calu-3 pair due to the very low mRNA expression of ENaC subunits in CFSME cells compared to Calu-3 cells, absence of iNOS expression in Calu-3 cells and also because of the expression of different mucin genes (i.e., MUC2 was expressed in CFSME, but not in Calu-3 cells, while MUC5AC was expressed in Calu-3 but not in CFSME cells), which could not provide support that CFSME is an appropriate CF cell model for Calu-3 cells.

CFBE cells were also treated with CFTR-inh172 to act as a negative control. Treatment with CFTRinh-172 increased mRNA expression of γ- ENaC, iNOS and mucin not only in non-CF (16HBE and/or corrected CFBE) cells but also in CFBE cells (Figures 6A–6C). Similarly, CFTRinh-172 induced a delay in wound healing, not only in non-CF cells (16HBE and corrected CFBE cells) but also in CFBE cells (Figures 7B–7D). Modulation of γ-ENaC, iNOS, MUC (2, and 5B), and delay in wound healing in CFBE cells (where CFTR is non-functional) indicates that this effect may not be related to the ion channel function of CFTR, but most probably is an unspecific effect of the inhibitor. However, previous studies have shown that treatment of non-CF cells with CFTRinh-172 resulted in cells with a characteristic CF proinflammatory phenotype (Perez et al., 2007; Vij et al., 2009), and reduced iNOS expression (Steagall et al., 2000; Gorgun and Eissa, 2009). Similarly, treatment with CFTRinh-172 delayed wound closure of NHBE cells and Calu-3 cells (Schiller et al., 2010) and also another CFTR inhibitor, GlyH101, reduced wound healing rates (Trinh et al., 2012), but none of these studies treated a CF cell model with the CFTR inhibitor as negative control as we used in the present study. A non-specific effect of CFTRinh-172 was also found by Kelly et al. (2010) in other cell types (Kelly et al., 2010).

Activation of CFTR by forskolin had no effect on the mRNA expression of ENaC subunits in 16HBE cells. Corrected CFBE and CFBE cells showed only small increases in the expression of α-ENaC and MUC2 after treatment with forskolin. Forskolin did not affect iNOS expression in any bronchial epithelial cell line (Figures 6A–6C). Forskolin also caused delayed healing in all cell types (Figures7B–7D). The question whether activation of CFTR by forskolin can induce changes in the mRNA expression of ENaC, iNOS, mucin and in wound healing in non-CF cells, remained unanswered because upon treatment with forskolin either bronchial epithelial cells (16HBE) did not show any effect on gene expression or showed an effect not only on non CF (corrected CFBE) cells, but also on CF (CFBE) cells (e.g., α-ENaC, MUC2 and wound healing). That the effects of forskolin on CFBE and corrected CFBE cells are similar, may be explained by the common origin of these cells, as corrected CFBE cells are the isogenic control of CFBE cells, while 16HBE cells have a different origin (originate from a different donor). A controversy exists about whether the activation of CFTR by intracellular cAMP contributes to the regulation of ENaC, and our data cannot resolve this controversy. Activation of CFTR decreased the surface expression of murine ENaC but not of human ENaC in a Xenopusoocyte expression system (Yan et al., 2004). Furthermore, mucin secretion from Calu-3 cells was affected neither by activation nor by inhibition of CFTR (Kreda et al., 2007). Our data show potential limitations for the use of CFTRinh-172 and forskolin and suggest the treatments with these compounds may lead to the controversies if the effects are not evaluated critically.

Neutral Red uptake viability assay

We investigated if the effect of CFTRinh-172 and forskolin treatment on the delay in wound healing is due to its cytotoxic effect on the bronchial epithelial cells. We found that 20 µM CFTRinh-172 treatment for 20 h killed about 30% of these cells, but forskolin did not show any cytotoxic effect (Figure 8). The theory that non-CF cells could be made to look like CFBE cells by addition of CFTRinh-172 is, however, complicated by the obvious cytotoxic effects of this inhibitor. Data regarding specificity and cytoxicity of CFTRinh-172 is controversial. That CFTRinh-172 was cytotoxic was observed by some investigators (Kelly et al., 2010), but not by others (Tonghui et al., 2002; Schiller et al., 2010). The presence of serum in the cell culture medium decreases the efficacy and cytotoxicity of CFTRinh-172 (Kelly et al., 2010) and this could explain the absence of toxic effects of high concentrations of CFTRinh-172 (in the presence of serum in cultured cells), which has been shown several times (Tonghui et al., 2002; Perez et al., 2007). Schiller et al. (2010) observed that of the cells with wild-type CFTR (NHBE and Calu-3 cells in their study), the NHBE cells were more affected by the cytotoxic effect of CFTR inh-172, but attributed this to the solvent, and lowered the inhibitor concentration to 20 µM, the same concentration as used in the present study. Hence, the result of the exposure to the inhibitor could at the start of this study not be predicted with certainty. The cytotoxic effect of the inhibitor, which is in accordance with the findings of Kelly et al. (2010) (who also showed this effect of CFTRinh-172 on bronchial epithelial cells, due to mitochondrial failure) suggests that cytotoxicity must be seriously considered because it may generate effects not directly related to Cl⁻ transport inhibition (Kelly et al., 2010). Though CFTRinh-172 had a cytotoxic effect on bronchial epithelial cells, still the majority of cells remained viable (Figure 8), and could be analysed for mRNA expression.

Conclusions

The results obtained in this study suggest CFSME as an inappropriate CF cell model for Calu-3 cells. The data from bronchial epithelial cells support the theory that the ΔF508 mutation in CFTR is associated with the upregulation of β-, γ-ENaC and downregulation of iNOS at mRNA and/or protein expression level. The data in the present study do not support the notion that the expression levels of mucins are directly related to the presence of mutated CFTR. The notion that wound healing is abnormal and delayed in ΔF508-CFTR was partially confirmed. CFTRinh-172 showed cytotoxic and non-specific effects. Use of CFTRinh-172 on cells in CF research must be seriously questioned because this inhibitor may generate unspecific effects.

Acknowledgments and funding

This study was financially supported by the Swedish Science Research Council, and the Swedish Heart-Lung Foundation.

Author contribution

Rashida Hussain and Hafiz M. Umer performed the experimental work. Rashida Hussain wrote the paper, Godfried M. Roomans and Maria Björkqvist supervised the work.

Abbreviation

CF cystic fibrosis

CFTR cystic fibrosis transmembrane conductance regulator

CFBE CF bronchial epithelial

CFSME cystic fibrosis submucosal epithelial

ENaC epithelial sodium channel

HBE human bronchial epithelial

iNOS inducible nitric oxide synthase

MUC mucin

qRT-PCR Quantitative Real-Time-Polymerase Chain Reaction

RIPA radio immunoprecipitation assay

SDS sodium dodecyl sulphate

Corresponding author: e-mail: rashida.hussain@oru.se