Transcellular thiocyanate transport by human airway epithelia

Abstract

Human airway mucosa synthesizes and secretes lactoperoxidase (LPO). As H₂O₂ and thiocyanate (SCN⁻) are also present, a functional LPO antibacterial defence system exists in the airways. SCN⁻ concentrations in several epithelial secretions are higher than in serum, although the mechanisms of transepithelial transport and accumulation in these secretions are unknown. To examine SCN⁻ accumulation in secretions, human airway epithelial cells, re-differentiated at the air–liquid interface, were used in open-circuit conditions. [¹⁴C]SCN⁻, in the basolateral medium, was transported across the epithelium and concentrated tenfold at the apical surface. Measurement of the transepithelial potential showed that the basolateral compartment was positive relative to the apical surface (13.7 ± 1.8 mV) and therefore unfavourable for passive movement of SCN⁻. Transport was dependent on basolateral [SCN⁻] and saturable (K_m,app = 69 ± 25 μm); was inhibited by increased apical [SCN⁻]; and was dependent on the presence of basolateral Na⁺. Perchlorate (K_i,app = 0.6 ± 0.05 μm) and iodide (K_i,app = 9 ± 8 μm) in the basolateral medium reversibly inhibited transport, but furosemide did not. Iodide was also transported (K_m,app= 111 ± 69 μm). RT-PCR and immunohistochemistry confirmed expression of Na⁺−I⁻ symporter (NIS) in the airways. SCN⁻ transport was insensitive to apical disulphonic acid Cl⁻ channel blockers, but sensitive to apical glibenclamide and arylaminobenzoates. Forskolin and dibutyryl cAMP increased transport. These data suggest SCN⁻ transport may occur through basolateral NIS-mediated SCN⁻ concentration inside cells, followed by release through an apical channel, perhaps cystic fibrosis transmembrane conductance regulator.

The lactoperoxidase (LPO) antibiotic system is synthesized by epithelia of the mammary, salivary, lacrimal and airway glands and it provides protection against infection. To function in epithelial host defence, LPO uses H₂O₂ to oxidize an anion, for example thiocyanate (SCN⁻) or I⁻, forming OSCN⁻ or OI⁻, which are antibacterial, antifungal and antiviral (for review, Reiter & Perraudin, 1991; Thomas et al. 1991). To maintain adequate enzyme activity, the epithelia must provide sufficient substrate concentrations. Since plasma levels of the used anions are low, diffusion rates are most likely to be insufficient in secretions of these glands to maintain significant LPO activity. Thus, epithelia must concentrate the anion from the blood compartment to secretions to provide sufficient substrate for the enzyme. In the case of saliva, SCN⁻ is normally present at 0.5–3 mm (Dacre & Tabershaw, 1970; Schultz et al. 1996). Airway secretions have recently been shown to contain 0.4 mm SCN⁻ (Wijkstrom-Frei et al. 2003). These values are more than tenfold higher than serum values (0.05 mm, Lundquist et al. 1995). As SCN⁻ is not made by epithelial cells in sufficient amounts to account for these secreted amounts of SCN⁻, active transport seems to be responsible for SCN⁻ movement from the blood compartment to the epithelial surface. However, studies of the mechanism responsible for transport and concentration of SCN⁻ have not been reported for salivary, lacrimal and airway glands, which are the major sites of active LPO.

Most secretions that contain SCN⁻ also contain I⁻. I⁻ is present at 10- to 20-fold higher concentrations in these secretions compared to serum (Brown-Grant, 1961), and I⁻ is known to be transported across the thyroid epithelium in a Na⁺−I⁻ symporter (NIS)-dependent fashion (De La Vieja et al. 2000). I⁻ is used by thyroid peroxidase for the synthesis of thyroid hormones, but it can also be utilized by other haeme peroxidases such as lactoperoxidase. Mammary, salivary and gastric glands express NIS, and I⁻ transport has been demonstrated in these tissues as well as in the kidney. NIS transcripts and protein have been shown to be present in the lungs (Wapnir et al. 2003), but there are no reports of functional NIS in the airway. SCN⁻ is well known as a competitive inhibitor of I⁻ transport by NIS, and NIS is known to transport SCN⁻ in the thyroid gland (for review, De La Vieja et al. 2000). However, the function of NIS has not been experimentally linked with SCN⁻ transport in these extrathyroidal tissues.

LPO enzymic activity has been shown to be a component of host defence in airways (Wijkstrom-Frei et al. 2003), and thus transport of SCN⁻ across the airway epithelium might play a role in maintaining airway sterility. The regulation of SCN⁻ secretion by epithelia provides a possible mechanism to manage peroxidase-mediated host defence activity on the luminal surface. For these reasons, we studied whether airway epithelia actively transport SCN⁻, and we characterized the properties of SCN⁻ transport with regard to possible molecular components at both the basolateral (BL) and apical membranes.

Methods

Materials

Unless otherwise stated all materials were obtained from Sigma Chemical Company (St Louis, MO, USA).

Cell culture

Human airways were from organ donors whose lungs were not to be used for transplant; they were obtained from the Life Alliance Organ Recovery Agency of the University of Miami. Local IRB-approved written consents for use of these tissues for research were obtained by the Life Alliance Organ Recovery Agency and conformed to the standards set by the Declaration of Helsinki. Airway epithelial cells were isolated, grown and re-differentiated at an air–liquid interface (ALI) on either 6.5 mm or 24 mm T-clear filters (Costar Corning, Corning, NY, USA) having a porosity of either 0.4 or 3 μm (only for permeabilization experiments) and coated with human placental collagen type IV as previously described (Bernacki et al. 1999; Nlend et al. 2002).

Transport experiments

All cultures had a resistivity of ≥300 Ω cm. All experiments were performed under open-circuit conditions.

Time course of SCN⁻ accumulation in the apical compartment

To determine the time course of SCN⁻ accumulation from initial addition of [¹⁴C]SCN⁻ to the BL medium to the point where [¹⁴C]SCN⁻ concentration was near steady state, [¹⁴C]SCN⁻ (85 μm, 50 mCi mmol⁻¹, Amersham Pharmacia, Piscataway, NJ, USA) was added to the BL culture medium of fully re-differentiated ALI cultures, and Dulbecco's phosphate-buffered saline (PBS, pH 7.4, Gibco, NY, USA) was added to the apical surface of the culture. Unlabelled SCN⁻ was not added to any of the media, thus the specific activity of [¹⁴C]SCN⁻ was that supplied by the manufacturer. Samples of the apical PBS were collected at different times and transported SCN⁻ was determined by liquid scintillation counting.

In experiments using cultures that were permeabilized basolaterally with α-toxin from Staphylococcus aureus, ALI cultures were differentiated on 24 mm inserts with 3 μm pores. BL and apical compartments were washed with PBS and then 140 mm potassium gluconate, 0.33 mm CaCl₂, 10 mm NaCl, 20 mm Hepes, pH 7.2, containing S. aureusα-toxin (10 000 U ml⁻¹) was added to the BL compartment for 30 min at 37 °C in the absence of carbon dioxide. Permeabilization was apparent by the near absence of ciliary beating. The BL compartment was briefly rinsed with same buffer lacking α-toxin, but containing 5 mm MgATP, 50 U ml⁻¹ creatine phosphokinase, 10 mm creatine phosphate and [¹⁴C]SCN⁻. Addition of ATP and the ATP regeneration system to the BL compartment restored ciliary beating, suggesting the BL membrane was permeabilized as reported previously by others (Ostedgaard et al. 1992; Illek et al. 1999). After addition of [¹⁴C]SCN⁻ to the BL compartment, appearance of [¹⁴C]SCN⁻ at the apical surface was then measured by sampling small amounts of apical PBS followed by scintillation counting. Similar treatment of the apical surface did not decrease ciliary beating supporting the idea that α-toxin was not able to permeabilize the apical membrane (Ostedgaard et al. 1992).

Unidirectional transport

To measure unidirectional transport of SCN⁻ from the BL to apical compartments in the absence of accumulated apical SCN⁻, non-permeabilized cultures were washed with PBS and incubated in BL medium containing different concentrations of [¹⁴C]SCN⁻ (50 mCi mmol⁻¹) until the system approached steady state with regard to apical [¹⁴C]SCN⁻ concentration, as determined from time course experiments described in the previous section. To measure I⁻ transport, unlabelled I⁻ was used. Transport of these anions was then determined by rapidly washing the apical surfaces of the cultures three times with PBS to remove anion accumulated on the apical surface. Following the third wash, additional aliquots (e.g. 500 μl per 24 mm filter) were placed on the apical surface for sequential 2 min incubations at 37 °C in humidified 5% CO₂ until a steady flow of anions to the apical surface was obtained. Removal of washes after 2 min and replacement with fresh PBS prevent accumulation of anion in the apical compartment. BL medium was sampled after the last wash. [¹⁴C]SCN⁻ in BL medium and apical washes was determined by liquid scintillation counting. All of the collected [¹⁴C]SCN⁻ was soluble following 10% trichloroacetic acid (TCA) precipitation, showing that radiolabel was not covalently attached to protein.

A colourimetric I⁻ assay (O'Kennedy et al. 1989) based on the Sandell–Kolthoff reaction was used to measure I⁻ transport. A standard curve ranging from 0.01 to 0.6 μm KI was used. Reagent addition was precisely timed and transmittance at 414 nm was measured in a microplate reader (SpectraMax, Molecular Devices, Sunnyvale, CA, USA) at 10 s intervals for 15 min after addition of ceric ammonium sulphate. The time points giving a linear range of transmittance with respect to standards (usually 5 min) were used to determine I⁻ concentrations in samples.

To determine the transport of SCN⁻ from the apical to BL compartment (back transport), apical surfaces of cultures were incubated with PBS containing [¹⁴C]SCN⁻, and BL medium was removed at 2 min intervals for liquid scintillation counting.

Apparent K_m values were estimated by non-linear regression fitting to the Michaelis–Menten equation: V = V_maxS/(S +K_m). Inhibition data were fit to V = V_maxS/{[(1 +i/K_i)S ]+K_m}−L, using the K_m,app determined from the experiments described. L corrects for small amounts of non-inhibitable SCN⁻ transport.

RT-PCR

RNA was extracted from normal human bronchial epithelial (NHBE)-ALI cultures using Trisolv (InVitrogen, NY, USA) and cDNA was obtained using SuperScript First Strand Synthesis System for RT/PCR kit (InVitrogen). NIS-specific oligonucleotide primers were designed according to Ajjan et al. (1998) (sense, 5′CTCCCTGTAACGACTCCAG3′; anti-sense 5′CTATCTCTATTACGGTGC3′). NIS cDNA was amplified by 35 cycles of 1 min at 94 °C, 55 °C and 72 °C, followed by a final 3 min elongation at 72 °C. PCR product was cloned using pGEM-T Easy Vector system (Promega, Madison, WI, USA) and sequenced using the ABI Prism 3100 Genetic Analyser (Applied Biosystems, CA, USA).

Immunolocalization

Tracheal tissue

Immunolocalization in tissue was performed according to Castro et al. (1999b) with slight modifications. Normal human tracheas were fixed in 4% paraformaldehyde and prepared for embedding in paraffin in a Microwave Tissue Processor (Microwave Materials Technologies Inc., Knoxville, TN, USA) according to the manufacturer's protocol using a non-xylene method. Embedding and sectioning were performed by the Histology Laboratory at the University of Miami Hospital and Clinics, Sylvester Comprehensive Cancer Center. Autofluorescence of de-paraffinized sections was reduced by incubating sections in sodium borohydride in PBS (5 mg ml⁻¹). This was followed by quenching of endogenous peroxidase activity for 2 h in 3% H₂O₂ in methanol. Sections were subjected to antigen retrieval by incubating in 10 mm citrate buffer (pH 6) for 15 min at 80 °C followed by blocking for 1 h at room temperature in 5% normal goat serum (NGS, Chemicon, Tamecula, CA, USA) in PBS with 0.05% Tween-20 (PBS-T). Biotin autofluorescence was reduced utilizing Endogenous Biotin Blocking kit according to manufacturer's instructions (Molecular Probes, Eugene, OR, USA). Sections were incubated overnight at 4 °C with either anti-human NIS monoclonal antibody against amino acid residues 469–643 (Castro et al. 1999a), or non-immune mouse IgG (both from Chemicon, Tamecula, CA, USA) as a negative control. Both antibodies were used at 10 μg ml⁻¹ in 1% NGS–PBS-T. Sections were incubated with affinity-purified goat anti-mouse IgG conjugated with horseradish peroxidase at a concentration of 0.02 μg ml⁻¹ (KPL, Gaithersburg, MD, USA) for 45 min at room temperature followed by labelling using Tyramide Signal Amplification Kit (TSA, Molecular Probes) according to the manufacturer's protocol.

Cultures

ALI cultures were fixed in 4% paraformaldehyde and permeabilized with methanol for 10 min. Autofluorescence of fixed ALI cultures was reduced by incubating filters in sodium borohydride in PBS (5 mg ml⁻¹). Cultures were subjected to antigen retrieval by incubating in 10 mm citrate buffer (pH 6) for 15 min at 80 °C followed by blocking for 1 h at room temperature in 100% NGS. Cultures were incubated overnight at 4 °C with either anti-human NIS monoclonal antibody or non-immune mouse IgG as a negative control; both were at a concentration of 10 μg ml⁻¹ in 100% NGS. ALI cultures were incubated with purified goat anti-mouse IgG conjugated with AlexaFluor 555 (Molecular Probes) for 1 h at a concentration of 2 μg ml⁻¹ in 100% NGS. Fluorescent images were obtained using a Zeiss LSM-510 confocal laser scanning microscope in the University of Miami Analytical Imaging Core Facility.

Results

Time course of SCN⁻ accumulation

To assess the ability of human airway epithelial cells to accumulate SCN⁻ in the apical compartment, de-differentiated human tracheobronchial epithelial cells were cultured in a two-chamber system that allowed re-differentiation at the air–liquid interface. Differentiation was evident by the presence of cells with beating cilia and goblet cells that secrete mucus. These cultures maintain a small volume of liquid on the apical surface (Matsui et al. 1998) and do not normally secrete LPO (as measured by enzyme activity and Northern blots, data not shown) that might metabolize transported SCN⁻, thus providing an ideal system to study the transport of this anion.

The accumulation of SCN⁻ from the BL compartment to the apical compartment was first examined by adding [¹⁴C]SCN⁻ to the BL medium and then following the time course of [¹⁴C]SCN⁻ appearance at the apical surface. PBS, not containing SCN⁻, was added to the apical surface and then sampled at different times after addition of isotope and [¹⁴C]SCN⁻ was determined by scintillation counting. After a brief lag, [¹⁴C]SCN⁻ began accumulating at the apical side and to a 10.0 ± 1.1-fold higher concentration (mean ± s.e.m., n = 6) than that found in the BL medium. An example experiment is shown in Fig. 1A. Accumulation of SCN⁻ on the apical side also occurred using culture media in both the apical (instead of PBS) and BL compartments. Addition of [¹⁴C]SCN⁻ to the apical compartment did not result in accumulation in the BL compartment, and the appearance of apically applied [¹⁴C]SCN⁻ in the BL compartment occurred at a much slower rate (see below).

Figure 1. Time course of thiocyanate (SCN⁻) transport by airway epithelial cells in open-circuit conditions.

A, PBS was added to the surfaces of two cultures followed by the addition of 85 μm [¹⁴C]SCN⁻ to the basolateral (BL) medium. The [SCN⁻] in the apical compartment (•) was measured as a function of time by sampling small aliquots of the apical PBS for liquid scintillation counting (LSC). The decrease in [SCN⁻] of the BL compartment is also plotted (▪). Mean values of the two filters are shown. In these intact cultures, after 240 min, the apical [SCN⁻] was tenfold higher than the BL [SCN⁻] despite an unfavourable electrical potential. B, the BL membranes of another culture were permeabilized with α-toxin as described in Methods, and 100 μm [¹⁴C]SCN⁻ was added to the BL compartment. The apical compartment [SCN⁻] (•) increased to, but did not exceed, the level of the BL compartment [SCN⁻] (▪). C, the rates of appearance of SCN⁻ in the apical compartments of the experiments in A (•) and B (▪) are compared.

The transepithelial potential of the re-differentiated cultures before initiation of experiments was −13.7 ± 1.8 mV (mean ± s.e.m., n = 15), i.e. the BL compartment was positive with respect to the apical compartment. Within 10 min after adding PBS to the apical surface, this potential dropped to −7.6 ± 0.7 mV (mean ± s.e.m, n = 15) and then increased back to baseline over the following 2 h. Thus, the electrical potential was unfavourable for movement of SCN⁻ from the BL to apical compartment and the observed tenfold accumulation of SCN⁻ could not be attributed to electrical potential differences.

In identical experiments using separate cultures permeabilized with α-toxin in the BL compartment, apical [SCN⁻] rapidly approached the BL [SCN⁻] but did not increase above that level (Fig. 1B). Permeabilization of the BL membrane was apparent by a dramatic decrease in ciliary beating that was restored upon addition of ATP and an ATP regeneration system in the BL compartment. Apically applied α-toxin did not alter ciliary beating, consistent with other reports that α-toxin does not permeabilize the apical surface. Thus, an intact BL membrane was apparently needed for concentration of SCN⁻ at the apical surface of the ALI cultures.

The difference in accumulation of SCN⁻ by the intact and permeabilized cultures is reflected in the rate at which SCN⁻ appeared in the apical compartments (Fig. 1C). The rate of SCN⁻ appearance in intact cells increased, presumably as the cellular compartment filled, and then decreased over several hours as the apical [¹⁴C]SCN⁻ concentration increased and the system approached the steady state. In contrast, in the culture permeabilized with α-toxin in the BL compartment, the rate of SCN⁻ appearance reached a maximum after about 30 min and then declined to near zero for the remainder of the experiment. The ¹⁴C collected from the apical surface was soluble in 10% TCA indicating that the radiolabelled SCN⁻ had not been bound to protein.

SCN⁻ transport rates

To quantify transport rates, cultures were instead pre-incubated with [¹⁴C]SCN⁻ in the BL medium until apical [¹⁴C]SCN⁻ approached a steady value. The rate of appearance was then followed by adding and removing aliquots of PBS at the apical surface at 2 min intervals and determining the amount of ¹⁴C in the collected PBS as a measure of apical [SCN⁻]. These short washes insured that apical [SCN⁻] remained very low (< 1 μm) during transport measurements, thus transport was not inhibited by accumulation of SCN⁻ in the apical compartment as suggested by Fig. 1.

The first wash showed a high level of isotope, suggesting that SCN⁻ had accumulated in the small volume of apical surface liquid present during the pre-incubation period. Later washes contained a lower constant amount of isotope suggesting a constant rate of [¹⁴C]SCN⁻ appearance (Fig. 2A). This constant rate obtained in the later washes increased with increasing BL [¹⁴C]SCN⁻ concentration. Experiments were carried out on triplicate filters at each different BL [SCN⁻] used. The experiments used cultures from two different donors and 4 different days (n = 52 cultures). The rate was concentration dependent and appeared to be saturable at higher BL concentrations (Fig. 2B). Fitting the data to the Michaelis–Menten equation predicted an apparent K_m of 69 ± 25 μm and an apparent V_max of 24 ± 3 nmol h⁻¹ cm⁻² (fit coefficient ± s.d.) for transepithelial transport of SCN⁻. These experiments were conducted under open-circuit conditions, thus these constants are not corrected for any effects of electrical potentials. In addition, these values do not distinguish the different contributions of different subcellular compartments and surface membranes and thus are apparent values for the entire transepithelial transport process under the experimental conditions.

Figure 2. Concentration dependence of SCN⁻ transport by airway epithelial cultures.

Cultures, re-differentiated at the air–liquid interface (ALI), were incubated with [¹⁴C]SCN⁻ in the BL medium under open-circuit conditions. After incubation, the apical surfaces were washed at 2 min intervals with PBS. [¹⁴C]SCN⁻ in washes was determined by LSC of triplicate samples at each time point. A, the basal transport rates of single cultures incubated with 31 (•), 92 (▪), 153 (▴), and 306 (▾) μM SCN⁻ in the BL medium (concentrations determined from the specific radioactivity of BL medium). B, mean (±s.e.m.) basal transport rates from triplicate cultures at each BL [SCN⁻]. Cultures in B were obtained from two different lungs and assayed on 3 separate days, n = 51 cultures. Fitting the data to the Michaelis–Menten equation by non-linear regression analysis gave apparent K_m and V_max for the transport process of 69 ± 25 μm and 24 ± 3 nmol h⁻¹ cm⁻², respectively (fit coefficient ± s.d.) under open-circuit conditions.

Back transport was measured by an identical procedure except cultures were pre-incubated with [¹⁴C]SCN⁻ in the apical PBS, and the BL surface was washed with culture medium at 2 min intervals. The SCN⁻ transport rate from the apical to BL compartment was less than 0.05 that of the BL to apical transport rate at the same [SCN⁻].

Na⁺−I⁻ symporter

Since SCN⁻ can be actively transported by the NIS in thyroid gland and since NIS has also been detected in salivary gland, mammary gland and stomach (Ajjan et al. 1998; Spitzweg et al. 1998, 1999), all organs that transport SCN⁻, airway epithelia were probed for the presence of NIS mRNA and protein. RT-PCR using total RNA isolated from re-differentiated airway epithelial cultures showed a band of the expected size (Fig. 3). Sequence analysis of this band showed that it was 100% identical to NIS nucleotide sequence (GenBank accession no. U66088) suggesting that NIS is expressed in airway epithelia.

Figure 3. RT-PCR detects Na⁺−I⁻ symporter (NIS) message in airway epithelial cells.

Total RNA from re-differentiated airway epithelial cells was isolated and reverse transcribed. PCR was performed using human NIS-specific oligonucleotide primers (lanes 1 and 3) and GAPDH primers (lanes 2 and 4). Bands of the expected size (346 bp) for human NIS (lane 1) and GAPDH (lane 2) were seen in RNA prepared from ALI cultures. Control PCR reactions without reverse transcriptase (lanes 3 and 4) and without primers (lanes 5 and 6) showed no amplified products. The NIS band was sequenced and was identical to the expected portion of human NIS (GenBank accession no. U66088). Markers (lane M) were HaeIII-digested ΦX174.

Immunolocalization (Fig. 4) with a monoclonal antibody against recombinant human NIS (Castro et al. 1999a) confirmed expression of the protein in human trachea. Immunoreactivity was mainly associated with submucosal gland cells (Fig. 4A and B) at the basal portion of the acinar cells (Fig. 4C and D). In contrast to NIS localization in salivary gland (Jhiang et al. 1998), NIS in airway submucosal glands was not restricted to ducts. A smaller amount of product was also specifically associated with superficial epithelial cells (data not shown). Since cultures studied here were derived mainly from surface epithelia but also contained cells with glandular phenotype (Gray et al. 1996; Yoon et al. 1997), we examined the cultures directly for the expression of NIS. Confocal images of immunostained epithelial cell cultures re-differentiated on filters showed that approximately half of the cells expressed NIS (Fig. 4G and H), apparently in the lateral membrane of the cells (circular profiles, Fig. 4H). Phase contrast images suggested that NIS-expressing cells were not ciliated (data not shown). Together these data strongly suggest that NIS is expressed in large human airways and that it may be the molecule responsible for SCN⁻ transport at the basolateral membrane.

Figure 4. Immunolocalization of NIS in human airway mucosa and cultured airway epithelia.

Sections of human trachea (A–F), and ALI cultures of human airway epithelia (G–J) were incubated with monoclonal antibody against human NIS (A, B, C, D, G and H) or non-immune mouse IgG (E, F, I and J), followed by anti-mouse IgG coupled to horseradish peroxidase (A–F) or AlexaFluor 555 (G–J). Tyramide was used to localize HRP-coupled antibody (A–F). NIS in human trachea was primarily seen in submucosal glands (B) and localized to the basal surface of gland acinar cells (D, corresponding to region outlined by a rectangle in A). Arrowheads point toward lateral cell borders and the asterisks mark the small lumen in these acini. NIS was also present in about half of the cells in ALI cultures (H). A, B, E and F, bar = 100 μm; C and D, bar = 20 μm; G–J, bar = 10 μm.

Biochemical and pharmacological properties of NIS activity in thyroid gland are well characterized (De La Vieja et al. 2000), and thus we assessed the airway epithelial SCN⁻ transport with respect to the known properties of NIS. To measure the requirement of Na⁺ for SCN⁻ transport, cultures were incubated with [¹⁴C]SCN⁻ containing normal BL culture medium. After establishing a baseline transport rate, the medium was replaced with [¹⁴C]SCN⁻ containing Hank's balanced salt solution in which choline was substituted for Na⁺, and SCN⁻ transport was monitored for changes in transport (Fig. 5). Replacement of Na⁺ by choline in the BL medium blocked [¹⁴C]SCN⁻ transport. Transport could be restored by transferring the cultures to Na⁺-containing medium.

Figure 5. SCN⁻ transport is Na⁺ dependent.

ALI cultures, incubated with 70 μm [¹⁴C]SCN⁻ in the BL medium, were washed normally to determine baseline transport rates (0–20 min, not shown). Two cultures with similar transport rates were transferred to BL Hank's balanced salt solution (HBSS) containing 70 μm [¹⁴C]SCN⁻, and baseline transport was re-determined with ten additional apical washes (20–40 min). BL HBSS was then exchanged for HBSS (•) or Na⁺-free HBSS (▪), both containing 70 μm [¹⁴C]SCN⁻, and washing was continued (40–80 min). Cultures were then transferred back to BL culture medium containing 70 μm [¹⁴C]SCN⁻ and washed at the pre-determined intervals (80–95 min). Transport rates were determined by LSC of triplicate samples from the apical washes. Standard errors typically fell within the plotted point. Replacement of Na⁺ by choline reversibly blocked the transport of SCN⁻.

To rule out the possibility that the Na⁺−K⁺−2CL⁻ cotransporter was responsible for SCN⁻ transport in ALI cultures, transport was measured in the presence of 100 μm furosemide (Fig. 6A). No alteration in SCN⁻ transport was observed; however, transport was inhibited by inclusion of 100 μm perchlorate, a known inhibitor of the NIS (Fig. 6A). Perchlorate inhibition was reversible (Fig. 6B) and concentration dependent (Fig. 7). The K_i,app was estimated by non-linear regression fitting of the data to the Michaelis–Menten equation to be 0.6 ± 0.05 μm (fit coefficient ± s.d.), a value near the previously reported K_i values (1–2 μm) in other tissues (Dohan et al. 2003).

Figure 6. SCN⁻ transport is resistant to furosemide but sensitive to perchlorate and iodide.

A, three ALI cultures, incubated with 10 μm [¹⁴C]SCN⁻ in the BL medium, were apically washed with PBS at 2 min intervals. After five washes, cultures were transferred to BL medium containing [¹⁴C]SCN⁻ alone (•) or BL medium with 100 μm furosemide (▪) or 100 μm perchlorate (▴) and washing with PBS continued (12–48 min). Furosemide had no effect, while perchlorate inhibited transport. B, three ALI cultures, incubated with 10 μm [¹⁴C]SCN⁻ in the BL medium, were apically washed with PBS at 2 min intervals. After eight washes, cultures were transferred to BL medium containing [¹⁴C]SCN⁻ alone (•) or BL medium with 100 μm potassium iodide (▪) or 10 μm perchlorate (▴) and washing with PBS continued (18–38 min). Cultures were then transferred to normal BL medium without inhibitors, but containing [¹⁴C]SCN⁻, for additional washes (40–70 min). Transport rates were determined by LSC of triplicate samples from the apical washes. Standard errors typically fell within the plotted point. Furosemide had no effect while perchlorate and I⁻ reversibly inhibited transport.

Figure 7. Perchlorate is an inhibitor of SCN⁻ transport.

Airway epithelial cell cultures were incubated with [¹⁴C]SCN⁻ containing BL medium and then various concentrations of perchlorate were added to the BL medium. Transport at each perchlorate concentration was determined with triplicate cultures and values are means ± s.e.m. Perchlorate inhibited SCN⁻ transport in a concentration-dependent manner. Transport rates were fit by non-linear regression to Michaelis–Menten kinetics for a competitive inhibitor and predicted an apparent K_i = 0.6 ± 0.05 μm (fit coefficient ± s.d.).

If NIS was responsible for uptake of SCN⁻ into airway epithelia, I⁻ should compete with SCN⁻ and should be transported. Addition of I⁻ to the BL medium reversibly inhibited SCN⁻ transport (Fig. 6B) and competed with SCN⁻ for transport across the airway epithelial cell cultures in a concentration-dependent manner (Fig. 8B) with a K_i,app = 9 ± 8 μm. To measure I⁻ transport, varying concentrations of non-radioactive I⁻ were included in the BL medium, and the appearance of I⁻ at the apical surface was determined by a chemical assay (O'Kennedy et al. 1989). I⁻ was effectively transported across airway epithelial cultures with a K_m,app of 111 ± 69 μm (Fig. 8A), close to that observed in thyroid (20–36 μm, De La Vieja et al. 2000).

Figure 8. I⁻ transport by airway epithelial cultures.

A, airway epithelial cells were incubated with BL medium containing various concentrations of I⁻. Transport of I⁻ was measured by colourimetric assay of apical washes from three individual cultures at each BL [I⁻] in two separate experiments, and plotted values are means ± s.e.m. Transport rates were fit by non-linear regression to the Michaelis–Menten equation and predicted an apparent K_m = 111 ± 69 μm and apparent V_max = 37 ± 16 nmol h⁻¹ cm⁻² (fit coefficient ± s.d.). B, airway epithelial cell cultures were incubated with 70 μm [¹⁴C]SCN⁻-containing BL medium and then various concentrations of I⁻ were added to the BL medium. Transport at each [I⁻] was determined with triplicate cultures and means ± s.e. were plotted. I⁻ inhibited SCN⁻ transport in a concentration-dependent manner. Transport rates were fit by non-linear regression to Michaelis–Menten kinetics for a competitive inhibitor and predicted an apparent K_i = 9 ± 8 μm (fit coefficient ± s.d.).

Thus, NIS is expressed in airway mucosa and NIS activity is similar to that in thyroid gland. Therefore, the observed properties of SCN⁻ and I⁻ transport in ALI cultures suggest that NIS is likely to be responsible for BL transport of these anions, at least in this experimental system.

NIS in thyroid gland concentrates I⁻ in epithelial cells, and I⁻ then exits via a channel in the apical membrane. To characterize the apical pathway by which SCN⁻ exits airway epithelial cells, various inhibitors of anion channels were added to apical PBS washes to assess their ability to inhibit transport of SCN⁻. 4,4′-Dinitrostilbene-2,2′-disulphonic acid (DNDS, 100 μm), an inhibitor of anion exchangers and Cl⁻ channels, had no effect on SCN⁻ transport (Fig. 9), nor did 4,4′-diisothiocyanatostilbene-2,2′-disulphonic acid (DIDS, 200 μm) (data not shown). In contrast, the sulphonylurea, glibenclamide (500 μm), an inhibitor of CFTR, blocked SCN⁻ transport, as did the arylaminobenzoate, diphenylamine-2-carboxylic acid (DPC, 1 mm) (Fig. 9). Inhibition by glibenclamide and DPC was reversible on removal of the compounds, indicating that the cells remained viable during treatment (Fig. 9). 5-Nitro-2-(3-phenylpropylamino)-benzoate (NPPB) also reversibly inhibited SCN⁻ transport (data not shown).

Figure 9. SCN⁻ transport is sensitive to apically applied inhibitors of cystic fibrosis transmembrane conductance regulator.

Three ALI cultures, incubated with 90 μm [¹⁴C]SCN⁻ in the BL medium, were rapidly and apically washed three times with PBS and then at 2 min intervals. After ten washes (20 min) to establish baseline transport rate, PBS washing was continued in the presence of either glibenclamide (500 μm; •), diphenylamine-2-carboxylic acid (DPC, 1 mm; ▪) or dinitrostilbene-2,2′-disulphonic acid (DNDS, 100 μm; ▴) (22–42 min). Compounds were then removed and washing continued in PBS to show reversibility (44–58 min). SCN⁻ transport was determined by LSC of triplicate samples from each time point. Standard errors typically fell within the plotted point. DPC and glibenclamide reversibly inhibited SCN⁻ transport, while DNDS had no effect.

These data suggested that cystic fibrosis transmembrane conductance regulator (CFTR) might be the anion channel responsible for SCN⁻ transport and, if so, predicted that increased intracellular cAMP would stimulate transport. Inclusion of forskolin (10 μm) and dibutyryl cAMP (500 μm) in the apical washes was associated with an immediate increase in SCN⁻ transport that was blocked by glibenclamide (500 μm) (Fig. 10A), consistent with a role for CFTR. Glibenclamide inhibited both basal and cAMP stimulated SCN⁻ transport and the inhibition by glibenclamide was again reversible (Fig. 10A).

Figure 10. SCN⁻ transport is stimulated by increased intracellular cAMP.

A, two ALI cultures, incubated with 70 μm [¹⁴C]SCN⁻ in the BL medium, were rapidly and apically washed three times with PBS and then at 2 min intervals. After ten washes (20 min), dibutyryl cAMP and forskolin were added to the next six PBS washes (22–32 min) causing an abrupt increase in transport. Inclusion of glibenclamide, in addition to the dibutyryl 500 μm cAMP and 10 μm forskolin, caused an immediate block of both the stimulated and baseline SCN⁻ transport (34–46 min). Removal of compounds from washes (48–56 min) restored SCN⁻ transport to the pre-cAMP level, demonstrating the reversibility of glibenclamide inhibition. B, two ALI cultures were incubated with 100 μm [¹⁴C]SCN⁻ in the apical PBS, and either normal BL medium (•) or BL medium containing 100 μm perchlorate (▪). The BL compartment was then washed at 2 min intervals with culture medium to measure back transport from the apical to BL compartment. After 16 washes (32 min), dibutyryl cAMP and forskolin were added to the apical PBS (34–60 min) to 500 and 10 μm, respectively. Stimulation of protein kinase A caused an increase in transport in control, but not in the culture containing perchlorate in the BL medium. Transport rates were determined by LSC of triplicate samples from each wash. Standard errors typically fell within the plotted point.

Back transport from the apical to BL compartment was low in comparison to BL to apical transport (Fig. 10B). Inclusion of perchlorate in the BL medium of these cultures had no effect on back transport unless forskolin and dibutryl cAMP were used to stimulate protein kinase A. When forskolin and dibutryl cAMP were added to the apical PBS, back transport was stimulated (Fig. 10B). This suggested that back transport, in the absence of PKA stimulation was primarily not through BL NIS, and may occur via a paracellular route.

Since CFTR-mediated Cl⁻ current is increased in the presence of the ENaC blocker amiloride (Knowles et al. 1983), the effect of amiloride (100 μm) was also tested for its effect on SCN⁻ transport. Inclusion of amiloride in apical washes was not associated with a significant increase in SCN⁻ transport (Fig. 11), perhaps due to the smaller contribution of SCN⁻ to the overall anion flux through CFTR.

Figure 11. SCN⁻ transport is insensitive to apical amiloride.

Two ALI cultures, incubated with 70 μm [¹⁴C]SCN⁻ in the BL medium, were washed rapidly three times and then at 2 min intervals with PBS (•) or with PBS containing amiloride (▪). After nine washes (18 min), dibutyryl cAMP and forskolin were added to the next five PBS washes (20–28 min) and then, glibenclamide, in addition to the dibutyryl cAMP and forskolin, was added to the final six washes (causing an immediate block of SCN⁻ transport) (30–40 min). Transport rates were determined by LSC of triplicate samples of each wash. Standard errors fell within the plotted point. Inclusion of amiloride did not significantly affect either basal or stimulated SCN⁻ transport.

The pharmacological profile of SCN⁻ transport suggested that CFTR might be involved in transepithelial transport.

Discussion

The data presented here support the hypothesis that NIS may play an important role in moving and concentrating SCN⁻ into airway epithelial cells, and that CFTR may regulate transport into the airway lumen. In a similar fashion, NIS in salivary, mammary, and gastric glands also may be responsible for developing the significant transepithelial [SCN⁻] gradient seen in those tissues, although the apical channel releasing SCN⁻ into the lumen of these tissues remains unidentified.

The data suggested that SCN⁻ transport occurred primarily via a transcellular route in ALI cultures and not a paracellular route. This was because SCN⁻ was concentrated in the apical compartment in opposition to an electrical potential, transport was reversibly blocked by competitive inhibitors of NIS in the BL compartment and inhibitors of CFTR in the apical compartment, and required Na⁺ in the BL medium and was stimulated by treatments to increase intracellular cAMP. Thiocyanate accumulated to values similar to those reported previously for airway secretions (Wijkstrom-Frei et al. 2003), suggesting that the ALI cultures resemble in vivo epithelial SCN⁻ transport. Importantly, these cultures do not make LPO under the conditions used for growth and differentiation. Finally, transported SCN⁻ was not bound to protein by organification reactions.

Apparent kinetic constants estimated for SCN⁻ and I⁻ transport were within the range of published values for NIS in other tissues (Dohan et al. 2003). This was despite the fact that calculations of airway epithelial kinetic constants were performed under open-circuit conditions and did not take into account effects on transport by the apical membrane of the cells. Measurement of transport across intact epithelial layers under open-circuit conditions more closely resembles the physiological circumstances in airways, but in turn prevents accurate measurement of kinetic constants of the NIS in the BL membrane. Despite this, the measured K_m,app for SCN⁻ transport by airway epithelia (70 μm) was between that reported for SCN⁻ transport by thyroid slices (30 μm, Wolff, 1964) and oocytes expressing recombinant NIS (96 μm, Eskandari et al. 1997). In thyroid samples, the K_m values of NIS for SCN⁻ and I⁻ are identical, although in oocytes expressing NIS, K_m values for these anions are threefold different (36 versus 96 μm). The K_m,app and K_i,app for I⁻ were dissimilar (111 versus 9 μm). Although the reasons for this are unclear, a similar difference was reported for recombinant NIS expressed in oocytes (Dohan et al. 2003). In this regard, both SCN⁻ and I⁻ have higher lipid-bilayer permeability than other anions of similar size that could contribute. Our experiments were conducted under open-circuit conditions and measured transport across the intact epithelium and thus are influenced by movement across the apical membrane as well.

The perchlorate K_i,app, using SCN⁻ as a substrate, was comparable to that reported for NIS using I⁻ as a substrate (Wolff, 1964; Eskandari et al. 1997) and thus the selectivity resembled that of NIS, perchlorate > I⁻≈ SCN⁻. These apparent kinetic constants, together with the Na⁺ requirement of transport, support the hypothesis that NIS may be the BL transporter responsible for uptake of SCN⁻. The absence of furosemide inhibition and the inhibition by perchlorate rule out a major contribution by the Na⁺−K⁺−2Cl⁻ cotransporter.

Back transport studies showed low levels of apical to BL SCN⁻ movement despite a transepithelial potential that favoured transport in this direction. Since apical to BL movement was not inhibited by perchlorate unless cells were first stimulated with forskolin and dibutyryl cAMP, this low level of back transport suggested that the paracellular route probably contributed a small amount to the overall movement of the anion in the BL to apical direction. Back transport was increased by stimulation of PKA and this increase was blocked by perchlorate in the BL medium, consistent with the idea that CFTR was involved in regulating movement of SCN⁻ across the apical membrane.

In addition to the kinetic data, RT-PCR and immunolocalization support the presence of NIS in airway mucosa. Interestingly, immunolocalization suggested that NIS expression is heaviest in airway submucosal glands. Both RT-PCR and immunocytochemistry showed that ALI cultures used for these transport studies contained cells that expressed NIS basolaterally.

The observation that SCN⁻ transport responded to apically applied compounds with an inhibition profile similar to CFTR, suggested that CFTR might regulate or possibly be the channel that passes SCN⁻ to the apical surface. Like other Cl⁻ channels, CFTR has been shown to carry SCN⁻ although these studies primarily used SCN⁻ as probe for channel properties (Tabcharani et al. 1993; Linsdell et al. 1997; Linsdell, 2001). The possibility that CFTR regulates SCN⁻ transport in airway epithelia cells raises important questions regarding the effect of impaired CFTR activity on LPO-mediated airway host defence, and other potential effects of a loss of LPO activity on airway homeostasis in cystic fibrosis.