Epithelium Depletion in Polarized Human Bronchial Double ...

The Journal of Immunology

Double Mechanism for Apical Tryptophan Depletion in Polarized Human Bronchial Epithelium1 Olga Zegarra-Moran,2* Chiara Folli,* Benedetta Manzari,† Roberto Ravazzolo,*‡ Luigi Varesio,† and Luis J. V. Galietta* Indoleamine 2,3-dioxygenase is an enzyme that catabolizes tryptophan to kynurenine. We investigated the consequences of IDO induction by IFN-␥ in polarized human bronchial epithelium. IDO mRNA expression was undetectable in resting conditions, but strongly induced by IFN-␥. We determined the concentration of tryptophan and kynurenine in the extracellular medium, and we found that apical tryptophan concentration was lower than the basolateral in resting cells. IFN-␥ caused a decrease in tryptophan concentration on both sides of the epithelium. Kynurenine was absent in control conditions, but increased in the basolateral medium after IFN-␥ treatment. The asymmetric distribution of tryptophan and kynurenine suggested the presence of a transepithelial amino acid transport. Uptake experiments with radiolabeled amino acids demonstrated the presence of a Naⴙ-dependent amino acid transporter with broad specificity that was responsible for the tryptophan/kynurenine transport. We confirmed these data by measuring the short-circuit currents elicited by direct application of tryptophan or kynurenine to the apical surface. The rate of amino acid transport was dependent on the transepithelial potential, and we established that in cystic fibrosis epithelia, in which the transepithelial potential is significantly more negative than in noncystic fibrosis epithelia, amino acid uptake was reduced. This work suggests that human airway epithelial cells maintain low apical tryptophan concentrations by two mechanisms, a removal through a Naⴙ-dependent amino acid transporter and an IFN-␥-inducible degradation by IDO. The Journal of Immunology, 2004, 173: 542–549.

T

he periciliary fluid (PCF),3 which covers the airway surface, is very important in the protective role played by the airway epithelium against injuring particles and infectious agents in the breathed air. Optimal thickness and fluidity of PCF are essential to maintain effective cilia beating and to allow mucociliary clearance. PCF volume and ion composition are controlled by the balance between fluid/electrolyte secretion and absorption. The airway epithelium may also modify the composition of PCF by secreting or removing various organic molecules such as proteins, peptides, and amino acids, and influence the innate and adaptive immune response (1, 2). Indeed, epithelial cells directly secrete antibacterial products in the PCF under resting and stimulated conditions (3, 4). During an inflammatory response, airway epithelial cells also produce regulatory soluble factors (e.g., PGs, NO, cytokines, and chemokines) upon direct interaction with pathogens or following stimulation by cytokines secreted by leukocytes (1, 2, 5, 6). The cross talk between airway epithelium and

*Laboratorio di Genetica Molecolare and †Laboratorio di Biologia Molecolare, Istituto Giannina Gaslini, Genoa, Italy; and ‡Dipartimento di Pediatria e Centro de Eccellenza per la Ricerca Biomedica, Universita` di Genova, Genoa, Italy Received for publication September 25, 2003. Accepted for publication April 21, 2004. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This study was supported by Fondo per gli Investimenti della Ricerca Grant to R.R. O.Z.-M. was supported by Fondo per gli Investimenti della Ricerca funds to Dipartimento di Pediatria e Centro de Eccellenza per la Ricerca Biomedica, Universita` di Genova. 2 Address correspondence and reprint requests to Dr. Olga Zegarra-Moran, Laboratorio di Genetica Molecolare, Istituto Giannina Gaslini, L.go G. Gaslini, 5, Genoa 16148, Italy. E-mail address: [email protected] 3 Abbreviations used in this paper: PCF, periciliary fluid; CF, cystic fibrosis; CFTR, CF transmembrane conductance regulator; ENaC, epithelial Na⫹ channel; Isc, shortcircuit current.

Copyright © 2004 by The American Association of Immunologists, Inc.

other types of cells is important in the local resistance to infectious agents. IFN-␥ is secreted by CD4⫹ Th1 cells, NK cells, and a subset of CD8⫹ T cells (7), and it is a potent modulator of airway epithelium functions. IFN-␥ alone or in combination with other cytokines induces several responses, including the expression of ICAM-1 (8), NO synthase 2 (9), cyclooxygenase (10), and chemokines (11, 12). We obtained evidence supporting that IFN-␥ can change PCF properties. We used a differentiated model of airway epithelium in which bronchial epithelial cells are allowed to polarize on a permeable support, and we found that IFN-␥ changes the activity of the epithelial Na⫹ channel (ENaC) and of Ca2⫹-activated Cl⫺ channels in a way that would favor fluid secretion vs absorption (13). To further investigate the effect of IFN-␥ on airway epithelium functions, we recently analyzed differential gene expression by using microarrays, and we found that, among others, IDO was strongly up-regulated by IFN-␥ (our unpublished results). IDO is a mammalian enzyme that catabolyzes the essential amino acid tryptophan by oxidizing its pyrrole moiety. It represents the ratelimiting enzyme of the kynurenine pathway, the major tryptophan catabolic pathway in mammals (14). IDO is ubiquitous in nonhepatic tissue, where it is inducible by inflammatory mediators, including IL-12, IL-18 (15), bacterial LPS (16), and IFNs (17). IDO up-regulation by IFN-␥ is associated to expression of antimicrobial response (18, 19) designed to limit the proliferation of invading pathogens by depleting the essential amino acid tryptophan from the intra- and extracellular environments. Furthermore, IDO may also contribute to modulation of leukocyte function (20). A role for IDO in T cell tolerance has been reported, and the suppressor activity of macrophages and dendritic cells on T cell proliferation in vitro may involve reduction of tryptophan concentrations in the medium via IDO-dependent mechanisms (21–23). Tryptophan-derived catabolites, in particular L-kynurenine, seem to be responsible at least in part for the inhibition of T and NK cell proliferation 0022-1767/04/$02.00

The Journal of Immunology (24). Interestingly, L-kynurenine inhibits proliferation only of activated cells, leaving resting cells able to respond to a subsequent stimulus. Evidence that IDO-dependent T cell inhibition may operate also in vivo has been presented. An important consequence of IDO activity has been identified in the placenta, where tryptophan degradation seems to play an important role in the prevention of the allogeneic fetus rejection by maternal T cells (25). To our knowledge, there is only one report on IDO gene expression induction by IFN-␥ in the airway epithelium (26). However, there is no indication of the consequences of IDO activation on a polarized epithelium. We aimed at elucidating the effects of IFN-␥ on IDO expression and tryptophan metabolism on a polarized preparation of human bronchial epithelial cells. Our purpose was to determine whether tryptophan is transported through the polarized epithelial barrier and whether IDO up-regulation depletes tryptophan from apical and/or basolateral sides of epithelium. Our results show that a Na⫹-dependent amino acid transporter is responsible for the unbalanced distribution of tryptophan across the polarized epithelium that accumulates in the basolateral side. IFN-␥, by activating IDO, causes an additional depletion of tryptophan at the apical side to almost undetectable levels.

Materials and Methods Cell culture Human bronchial epithelial cells were cultured, as previously described (13, 27). Briefly, cells were detached from bronchi after overnight incubation with protease XIV. Cells were obtained from lung transplants (cystic fibrosis (CF) patients) or lung resections (non-CF patients). After detachment, cells were grown on culture flasks for three to six passages in serumfree LHC basal medium/RPMI 1640 medium with 2 mM L-glutamine, 100 U/ml penicillin, and 100 ␮g/ml streptomycin. Cells used in this work were obtained from three non-CF and from two CF donors. We found no significant differences among donors; therefore, data from different individuals were pooled. CF cells were homozygous for ␦F508 mutation. To obtain polarized monolayers, cells were plated at high density (5 ⫻ 10⫺5 cells/ cm2) on permeable supports (Snapwells; Corning-Costar, Cambridge, MA). The medium was DMEM/Ham’s F12 (1:1) and contained 2% FBS, 2 mM L-glutamine, 100 U/ml penicillin, 100 ␮g/ml streptomycin, and various hormones and supplements (27, 28). Cells were maintained at 37°C in a humidified incubator in an atmosphere containing 5% CO2, and the experiments were done 8 –11 days after plating. At this time, the epithelial resistance was 1182 ⫾ 27 ⍀/cm2 (n ⫽ 52) in control cells, and 3212 ⫾ 108 ⍀/cm2 (n ⫽ 31) in IFN-␥-treated cells.

Conventional RT-PCR Total RNA was extracted from cultured epithelia with TRIzol. Reverse transcription on 1 ␮g of total RNA was performed with a commercial kit (BD Clontech, Palo Alto, CA), following the manufacturer’s instruction. PCR was done using Amplitaq polymers and a reaction kit (PerkinElmer, Wellesley, MA). PCR cycles included denaturation at 95°C (40 s), primer annealing at 60°C (30 s), and extension at 72°C (30 s). Specific sequences for amplification of ␤2-microglobulin were 5⬘-GCGCTACTCTCTCTT TCTGG-3⬘ and 5⬘-TCCAATCCAAATGCGCCATC-3⬘ (sense and antisense strand, respectively). PCR using these amplimers yields a 380-bp product. IDO-specific sequences were amplified by using the following sequences:5⬘-CAAAGGTCATGGAGATGTCC-3⬘and5⬘-CCACCAATAG AGAGACCAGG-3⬘ as the sense and the antisense primer, respectively. PCR product was 240 bp length. Number of PCR cycles (reported in Results) was chosen to keep the reaction in the linear amplification phase.

Quantitative real-time RT-PCR Real-time PCR amplification reaction was done in a total volume of 25 ␮l containing 5 ␮l of cDNA sample, 12.5 ␮l of SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA), and specific primers. The final concentration of primers was 7.5 pM. Primers and probes were designed according to manufacturer’s guidelines. An ABI Prism 7700 sequence detection system (Applied Biosystems) with SYBR Green fluorescence was used for the assays. Cycling conditions were: 10-min hot start at 95°C, followed by 35 cycles of denaturation step at 95°C for 40 s, an annealing step at 63°C for 30 s, and an extension step at 72°C for 30 s. Each sample was run in triplicate. ␤2-microglobulin was used as housekeeping mRNA

543 to normalize IDO transcript abundance. Data were analyzed by using Sequence Detector Systems version 2.0 software (Applied Biosystems). IDO reverse and forward primers were the same used for RT-PCR. Reverse and forward primers for ␤2-microglobulin were 5⬘-TCCAATCCAAATGCG GCATC and 5⬘-GCGCTACTCTCTCTTTCTGG, respectively.

Ion-pairing reverse-phase HPLC Apical and basolateral medium bathing bronchial monolayers were harvested 0, 24, and 48 h from the beginning of the experiment and stored frozen. The sample for HPLC analysis was prepared, as described (29). Briefly, tissue culture supernatants were adjusted to a final concentration of 2% perchloric acid, kept on ice for 30 min, and centrifuged 5 min at 13,000 ⫻ g, and the pellet of precipitated proteins was discarded. The perchloric acid-soluble fraction was neutralized by adjusting to 0.4 M potassium hydroxide; the clear supernatant was filtered through a 0.45-␮m filter (Minispike; Waters, Milford, MA) and analyzed by HPLC. The HPLC separation module was a Waters model 2690 controlled by Millenium 32 data system equipped with a Symmetry Shield RP18 column (Waters). Tryptophan was measured with a fluorescence detector (Waters 474 scanning fluorescence detector) at an excitation wavelength of 285 nm and an emission wavelength of 360 nm. Kynurenine was detected at a wavelength of 360 nm using a UV detector. A diode-array detector was used for the identification of peaks. Tryptophan and kynurenine were identified by comparing the retention times and spectral data (obtained by diode-array detection) with the standards. The HPLC conditions were previously described (29). Briefly, the sample was injected in the column equilibrated in a running buffer composed by 2% methanol, 1 mM tetrabutylammonium hydrogen sulfate, and 30 mM phosphate buffer, pH 8.0, and eluted isocratically for 15 min.

[3H]Lysine and [3H]tryptophan transport For uptake experiment on permeable supports, the apical and basolateral culture medium was removed and replaced with 0.5 and 2 ml of Krebs solution, respectively. The apical medium contained either 100 ␮M lysine and 1 ␮Ci/ml [3H]lysine or 50 ␮M tryptophan and 1 ␮Ci/ml [3H]tryptophan. Aliquots of 20 and 1000 ␮l were removed every 5 min from apical and basolateral medium, respectively. After each sampling, the filter was moved to another well containing fresh basolateral solution. Experiments were done at 37°C. The radioactivity in the samples was determined by liquid scintillation counting to calculate the amount of lysine or tryptophan remaining in the apical medium at each time, and the efflux of radioactivity at the basolateral side.

Ussing chamber experiments After 8 –11 days in culture, permeable supports with human bronchial epithelial monolayers were mounted in a vertical diffusion chamber (CorningCostar). The apical and basolateral chambers were filled with Krebs bicarbonate solution that contained (in mM): 126 NaCl, 0.38 KH2PO4, 2.13 K2HPO4, 1 MgSO4, 1 CaCl2, 24 NaHCO3, and 10 glucose. The solution was bubbled with 5% CO2-95% air. Experiments were done at 37°C. The transepithelial potential difference was clamped at the desired value with a voltage clamp amplifier, connected to the apical and basolateral chambers via Ag-AgCl electrodes and agar bridges. For convention, reported voltages are referred taking the basolateral side as ground. Potential difference and fluid resistance between potential-sensing electrodes were compensated.

Chemicals LHC9 medium was prepared from LHC basal medium (Biofluids, Rockville, MD) with the addition of supplements, as previously described (27, 28). Triiodothyronine and retinoic acid were from Biofluids. [3H]Tryptophan and [3H]lysine were from Amersham Biosciences (Amersham, Buckinghamshire, U.K.). All other chemicals were from Sigma-Aldrich (St. Louis, MO), except DMEM, RPMI 1640, Ham’s F12, and serum that were from Euroclone (Paignton, Devon, U.K.).

Statistics Data are presented as representative traces or as means ⫾ SEM. Significance was assessed using Student’s t test for unpaired groups of data.

544

TRYPTOPHAN DEPLETION IN HUMAN BRONCHIAL EPITHELIUM

Results

IFN-␥ induces the expression of IDO on bronchial epithelial cells To evaluate IDO expression, we used a semiquantitative RT-PCR on cDNA samples generated from control and IFN-␥-treated human bronchial epithelia grown on permeable supports. Untreated epithelia did not show an IDO signal after 25 amplification cycles using either diluted or undiluted cDNA. On the contrary, a clear amplification product was detected from the cDNA of cells incubated with IFN-␥ even at very low concentrations of the template (Fig. 1). The housekeeping gene, ␤2-microglobulin, was similarly expressed on both cell preparations. IDO expression could be demonstrated in untreated cells only after 32 PCR cycles using the undiluted template. To quantify the IDO response to IFN-␥, we determined the relative amount of mRNA using a real-time RTPCR (Fig. 1B). We found that IDO mRNA increased at ⬃60- or 40-fold after 24 or 48 h of IFN-␥ stimulation, respectively. These results indicate that IDO mRNA is expressed at very low levels in resting bronchial epithelial cells, and that its expression is strongly up-regulated by treatment with IFN-␥.

basolateral side, because it is effective only from this side of the monolayer (our unpublished results). We measured tryptophan and kynurenine concentration in the apical and basolateral fluid of cultured bronchial epithelia. Initial tryptophan concentration in our culture medium was 36 ⫾ 3 ␮M. When the medium was recovered from untreated cell cultures after a 24-h incubation, we found an asymmetrical distribution of tryptophan levels. Taking into consideration the volumes of the culture medium on both sides of epithelium (see Materials and Methods), the tryptophan changed from 18 ⫾ 1.5 nmol to 6 ⫾ 1 nmol on the apical side, and from 72 ⫾ 6 nmol to 84 ⫾ 4 nmol on the basolateral side (Fig. 2, A and B). After 48 h, tryptophan levels were even lower on the apical (4 ⫾ 1.5 nmol) and did not change on the basolateral side (81 ⫾ 6 nmol).

Extracellular tryptophan levels under resting and IFN-␥stimulated conditions The effect of IFN-␥ on extracellular tryptophan concentrations was investigated on human bronchial epithelial cells grown on permeable supports as a polarized epithelium. IFN-␥ was applied to the

FIGURE 1. IDO expression on bronchial epithelial cells. Semiquantitative PCR (A) using serial dilutions of cDNA from control and IFN-␥treated human bronchial epithelial cells. Dilutions 1/2, 1/4, 1/8, and 1/16 indicate that 0.5, 0.25, 0.125, and 0.0625 ␮g of RNA were used for the PCR. Treated cells were incubated for 48 h with 1000 U/ml IFN-␥. ␤2microglobulin was used as housekeeping gene. In control conditions, epithelia did not show an IDO signal after 25 amplification cycles even using undiluted cDNA (1 ␮g). In contrast, a stable amplification product was found from the cDNA of cells incubated with IFN-␥ still after 16-fold dilution of the template. B, A quantitative determination of IDO mRNA was conducted using real-time RT-PCR. Calculation of IDO abundance after normalization to ␤2-microglobulin showed changes due to IFN-␥ treatment. Each bar is the mean of six measurements on different cell preparations. Error bars represent SEM.

FIGURE 2. Time course of apical and basolateral tryptophan and kynurenine levels in control and in IFN-␥-treated epithelia. IFN-␥ caused a reduction of apical (A) and basolateral (B) tryptophan (TRP), and an increase of apical (C) and basolateral (D) kynurenine (KYN). Each point is the mean of 9 –12 experiments ⫾ SEM. The effect of IFN-␥ was partially prevented by norharmane (n ⫽ 5). The IDO inhibitor norharmane (50 ␮M) was added to the apical and basolateral medium at the same time as IFN-␥. Asterisks indicate statistically significant differences with IFN-␥ alone. ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01. Time course of kynurenine level on both sides of the epithelia. After exogenous application of 15 nmol of apical kynurenine, we found that it decreases on the apical (E) and increases on the basolateral (F) side of the monolayer (n ⫽ 3).

The Journal of Immunology Kynurenine was absent from the medium of bronchial epithelial cells under resting conditions (Fig. 2, C and D), thus confirming that basal IDO expression is very low and suggesting that the decrease of apical tryptophan in untreated cells is not due to catabolism. After IFN-␥ treatment, tryptophan decreased on both sides of cultured epithelia, disappearing completely from the apical side in 48 h and decreasing to 20 ⫾ 3 nmol on the basolateral side. In parallel, kynurenine increased dramatically in the basolateral medium to ⬃50 nmol, whereas kynurenine levels on the apical side were modest (2 ⫾ 0.5 nmol). Experiments were done to establish whether changes in tryptophan and kynurenine levels were indeed mediated by IDO activity. As shown in Fig. 2, 50 ␮M norharmane, a concentration near the IC50 for IDO inhibition (30), partially prevented the tryptophan depletion and the production of kynurenine induced by IFN-␥. We hypothesized that the asymmetrical kynurenine distribution could be due to the presence of a specific transepithelial transport. According to this hypothesis, kynurenine would be produced by IDO-mediated tryptophan degradation and released by diffusion through apical and basolateral membrane. Subsequently, apical kynurenine would be removed by an apical membrane transporter. To test this hypothesis, we added 15 nmol of kynurenine to the apical culture medium. Kynurenine was rapidly removed from the mucosal surface and transported to the basolateral side of the monolayers, thus producing an inverted asymmetric distribution (Fig. 2, E and F).

545 the epithelia at t ⫽ 0. The amount of lysine that disappeared from the apical chamber in 20 min was 10.1 ⫾ 1.4 nmol (n ⫽ 10; Fig. 3). A total of 100 ␮M tryptophan or 100 ␮M kynurenine added to the apical solution at the beginning of the experiment reduced significantly the lysine uptake, suggesting that both compounds compete with lysine for the transporter. The uptake in Na⫹-free solution was almost zero, as expected for a Na⫹-dependent transporter. The transport of tryptophan and kynurenine is electrogenic and Na⫹ dependent One characteristic of the apical amino acid transporter in cultured bronchial epithelial cells is the ability to generate a transepithelial electrical current (28). After clamping epithelia at 0 mV and blocking ENaC with amiloride, we measured the short-circuit currents (Isc) in response to amino acids added to the apical chamber. As previously shown, apical lysine generates an Isc in a dose-dependent way in airway epithelial cells (Fig. 4, A and E). Interestingly,

Tryptophan and kynurenine are actively removed from the apical surface Our data indicated that tryptophan and kynurenine asymmetrical distributions under resting and stimulated conditions could be due to a transepithelial transport. We reported previously that the apical membrane of cultured bronchial epithelia has an electrogenic amino acid transporter that couples Na⫹ influx to the uptake of neutral and cationic amino acids (28). Given the broad selectivity, we hypothesized that this transporter could be also responsible for the apical removal of kynurenine as well as of tryptophan. First, we tested this possibility indirectly by measuring lysine uptake. A total of 100 ␮M lysine (50 nmol) was applied to the apical side of

FIGURE 3. Time course of lysine uptake. Lysine uptake measured alone (F) or in the presence of 100 ␮M tryptophan (‚) or kynurenine (䡺). In 䡺, the lysine uptake in a Na⫹-free solution. Cells were grown for 8 –11 days on permeable supports. Before experiments, the apical and basolateral culture medium was removed and replaced with 0.5 and 2 ml of Krebs solution. A total of 100 ␮M lysine (1 ␮Ci/ml [3H]lysine) was applied to the apical side of the epithelia at t ⫽ 0, and every 5 min thereafter, aliquots of 20 and 1000 ␮l were removed from apical and basolateral medium, respectively. The radioactivity in the samples was determined by liquid scintillation counting to calculate the lysine uptake. Each point is the mean of 4 –10 experiments ⫾ SEM.

FIGURE 4. Ussing chamber experiments. The transepithelial potential difference was short circuited at 0 mV. Isc elicited by increasing concentrations of lysine (A), kynurenine (B), and tryptophan (C). The effect of lysine and tryptophan/kynurenine was not additive, suggesting that they compete for the same amino acid transporter. As expected, in Na⫹-free solution, no Isc increase was observed after kynurenine or lysine addition (D). E and F, Dose-response relationships of tryptophan, kynurenine, and lysine. The points were fitted to a Michaelis-Menten function. The maximal current obtained with tryptophan was 1.3 ⫾ 0.3 ␮A/cm2, about onehalf that of kynurenine (2.6 ⫾ 0.7 ␮A/cm2), and one-third that of lysine (3.9 ⫾ 0.6 ␮A/cm2). The half-effective concentration was instead 12.1 ⫾ 2.9, 13 ⫾ 3.4, and 79.3 ⫾ 5.8 ␮M for tryptophan, kynurenine, and lysine, respectively. No differences on these parameters were observed on IFN␥-treated epithelia (F, open symbols).

546


apical kynurenine also generates a current, and the same effect was induced by tryptophan (Fig. 4, B and C). In agreement with the hypothesis of a common transporter for these three amino acids, saturating concentrations of tryptophan or kynurenine largely prevented lysine induced current. In agreement with uptake experiments and with the assumption of a common Na⫹-dependent transporter, kynurenine, tryptophan, and lysine were ineffective under Na⫹-free conditions (Fig. 4D). Dose-response relationships for tryptophan and kynurenine in control and IFN-␥-treated epithelia were similar (Fig. 4F and Table I), thus suggesting that IFN-␥ treatment does not modify transporter expression or activity. The amino acid transport depends on transepithelial potential Given the electrogenic activity of the apical amino acid transporter, we anticipated that the response to amino acids had to be affected by transepithelial electrical potential. We measured lysine- and tryptophan-induced Isc at ⫹30, 0, and ⫺30 mV. The fit of dose-response relationships to a Michaelis-Menten function (Fig. 5) yielded similar half-maximal activations at the three voltages (77.8 ⫾ 15.9, 79.8 ⫾ 11.1, and 96.3 ⫾ 21.9 ␮M for lysine, and 6.72 ⫾ 3.15, 14.24 ⫾ 4.01, and 10.06 ⫾ 2.11 ␮M for tryptophan; n ⫽ 4 –5 at each condition). In contrast, the maximal activity was statistically different: 13.5 ⫾ 1.8, 6.1 ⫾ 0.8, and 3.5 ⫾ 0.5 ␮A for lysine, and 2.18 ⫾ 0.25, 1.76 ⫾ 0.2, and 1.11 ⫾ 0.11 ␮A for tryptophan at ⫹30, 0, and ⫺30 mV, respectively. Amino acid uptake in CF epithelia Airway epithelia from CF patients have altered transepithelial potential. Given the voltage sensitivity of amino acid transport, we decided to measure the lysine uptake on cells obtained from CF subjects. We found that the amount of lysine that disappeared from the apical chamber in 20 min was 6.6 ⫾ 0.3 nmol (n ⫽ 4; Fig. 6A). We measured also lysine uptake on non-CF epithelia in the presence of 400 ␮M glibenclamide, a known inhibitor of CF transmembrane conductance regulator (CFTR), to mimic the CF phenotype. We found that in this condition the uptake of lysine was also reduced to ⬃3.6 ⫾ 0.6 nmol (n ⫽ 4). Lysine uptake in CF epithelia and in glibenclamide-treated non-CF epithelia was significantly lower than in untreated non-CF epithelia ( p ⬍ 0.05 and p ⬍ 0.01, respectively). To have a direct evidence of the movement of tryptophan, we measured also the uptake of this amino acid on CF and non-CF epithelia. As shown on Fig. 6B also, the tryptophan uptake was significantly lower in CF than in non-CF epithelia (2.5 ⫾ 1.3 and 6.6 ⫾ 0.3 nmol, respectively). The presence throughout the experiment of either glibenclamide or of the recently identified CFTR-specific inhibitor CFTRInh-172 (31, 32) caused a significant reduction of tryptophan uptake on non-CF cells to 4.5 ⫾ 0.6 and 3.4 ⫾ 1.5 nmol. We modified the apical membrane potential also by blocking the ENaC with amiloride. This caused a significant increase of tryptophan uptake to 8.6 ⫾

Table I. Kinetic parameters of amino acid-dependent currenta Amino Acid L-Lysine L-Tryptophan L-Tryptophan L-Kynurenine L-Kynurenine a

(IFN-␥) (IFN-␥)

n

Km (␮M)

Imaxb (␮A/cm2)

10 5 4 8 4

79.3 ⫾ 5.8 12.1 ⫾ 2.9 9.6 ⫾ 2.3 13 ⫾ 3.4 11.1 ⫾ 6.1

3.9 ⫾ 0.6 1.3 ⫾ 0.3 1.1 ⫾ 0.4 2.6 ⫾ 0.7 2.3 ⫾ 0.8

Parameters obtained fitting dose-response relationships to Michaelis-Menten equation. n is the number of experiments. Values obtained on epithelia clamped at 0 mV. b Imax, maximal current.

FIGURE 5. Dose-response relationships at different transmembrane potentials. The transepithelial potential difference was short circuited at different clamping potentials, ⫹30, 0, and ⫺30 mV. The points were fitted to a Michaelis-Menten function. Although half-maximal activations at the three voltages were very similar for each amino acid (⬃85 ␮M for lysine and 10 ␮M for tryptophan), the maximal activity was clearly dependent on the transepithelial potential. Each point is the mean of four to five different experiments, and vertical bars are SEM.

1.5 nmol on CF epithelia, while it was without effect on non-CF epithelia.

Discussion

IFN-␥ induces a variety of antimicrobial mechanisms, including induction of IDO and catabolism of tryptophan. The oxidative cleavage of the essential amino acid tryptophan by IDO (14) may be associated to antimicrobial defense (18, 19). In fact, IDO is expressed in cells infected with a variety of intracellular pathogens such as Toxoplasma, Chlamydia, and viruses, and it has also been found to inhibit the growth of extracellular bacteria such as group B streptococci (19). IDO effectively restricts the growth of tryptophan-dependent pathogens by starving them for an amino acid. We have studied the functional effects of IFN-␥-dependent IDO induction in human bronchial epithelia. We hypothesized that IDO up-regulation in airway epithelium could be particularly important to deplete tryptophan in the PCF. Our results first demonstrate that IDO is poorly expressed in resting epithelia. This result is supported by the finding that kynurenine is absent in the culture medium despite relatively high concentrations of tryptophan. Our data also show that tryptophan is removed from the mucosal surface in resting conditions and transported to the basolateral side of the epithelium. IFN-␥ treatment strongly induces IDO expression. In this condition, we found IDO responsible for a further decrease in tryptophan concentration and for kynurenine production. Surprisingly, kynurenine increased only on the serosal side of the monolayer. We hypothesized that kynurenine produced by IDO-mediated tryptophan degradation diffuses through the cell membrane to

The Journal of Immunology

FIGURE 6. Amino acid uptake in CF and non-CF bronchial epithelial cells. A, Time course of lysine uptake in CF (E) and non-CF (F) cells, or in non-CF cells in the presence of 400 ␮M glibenclamide (䡺). Each point is the mean of 4 –10 experiments ⫾ SEM. B, Tryptophan uptake in CF and non-CF (WT) epithelia measured at 20 min in the absence or presence of 10 ␮M amiloride (amil), 20 ␮M CFTRInh-172 (Inh-172), 400 ␮M glibenclamide (glib), or in a Krebs solution without Na⫹ (Na-free). Each bar is the mean of 4 – 6 experiments ⫾ SEM. Asterisks indicate that values are statistically different from the respective control value: ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01. As for experiments on Fig. 3, cells were grown for 8 –11 days on permeable supports. Before experiments, the apical and basolateral culture medium was replaced with Krebs solution, and 100 ␮M lysine or 50 ␮M tryptophan (1 ␮Ci/ml [3H]lysine or [3H]tryptophan) was applied to the apical side of the epithelia. Aliquots of 20 and 1000 ␮l were removed from apical and basolateral medium every 5 min, and the radioactivity in the samples was determined by liquid scintillation counting to calculate the lysine/tryptophan uptake.

both apical and basolateral compartments. Then apical kynurenine is rapidly removed by the same amino acid transporter that is responsible for the clearance of tryptophan from the apical surface in resting conditions. A few years ago, we described the characteristics of a broad specificity Na⫹-dependent amino acid transport in the apical membrane of bronchial epithelial cells (28). This activity might be mediated by the recently cloned hATB0,⫹ (33), which is highly ex-

547 pressed in the lung, in particular in ciliated epithelial cells of the trachea and bronchioles (34). The involvement of this amino acid transporter in the asymmetric distribution of tryptophan and of kynurenine was established in uptake experiments using radiolabeled amino acids. We found that tryptophan and kynurenine inhibited the uptake of lysine, a well-known substrate of this transporter. We interpreted this effect as competition between tryptophan/kynurenine and lysine for the same transporter. This behavior was then confirmed by measuring the Isc elicited by direct application of these amino acids to the epithelium apical surface. Both substrates elicited a Na⫹-dependent current, and this effect was not additive with that of lysine. Comparison of the doseresponse relationships showed that the maximal current obtained with tryptophan was about one-half that of kynurenine and onethird that of lysine (see Fig. 4, E and F). Interestingly, the apparent Km for tryptophan and kynurenine were lower than that for lysine (see Table I) and for other good substrates of the transporter such as arginine and alanine (28), thus indicating that tryptophan and kynurenine are also physiological substrates. In IFN-␥-treated cells, the maximal current and the half-effective concentration of tryptophan and kynurenine were similar to those of control cells, indicating that IFN-␥ does not modify the transporter expression or activity. An important finding was that the rate of amino acid uptake by the apical transporter strongly depended on transepithelial potential, as expected from its electrogenic nature. Indeed, the maximal Isc for lysine and tryptophan at ⫹30 mV was 13.5 and 2.2 ␮A, while it was 3.5 and 1.1 ␮A, significantly lower, at ⫺30 mV. This is not surprising because more positive potentials impose a higher driving force for apical-to-basolateral Na⫹ movement through the epithelium that might drive faster amino acid uptake. Conversely, at more negative transepithelial potential, this driving force is reduced. Given these results, we speculate that apical membrane potential changes might affect the rate of amino acid uptake. CF is a disease in which the mutation of a gene encoding an apical Cl⫺ channel, the CFTR, causes the transepithelial potential difference to be significantly more negative than in non-CF subjects (35). There are indications that in CF the sodium absorption through ENaC is up-regulated (36), and this could be the main reason for the more negative value of the transmembrane potential. In a previous work, we measured the activity of the apical amino acid transporter in CF and non-CF epithelia in short-circuit conditions, i.e., with transepithelial voltage clamped at 0 mV (28). We found that the half-effective concentration and the maximal current evoked by arginine on epithelia obtained from CF patients were similar to those of non-CF cells, indicating that CFTR is not directly involved in the modulation of the transporter. However, we could not exclude an indirect effect mediated by electrochemical coupling. Therefore, in this study, we measured lysine and tryptophan uptake on CF cells in conditions in which the transepithelial potential is not clamped. Our results demonstrate that, in open circuit conditions, CF epithelia exhibit a reduced lysine and tryptophan uptake compared with non-CF cells. We hypothesized that this result is due to an altered transepithelial voltage in CF. Actually, the lack of functional CFTR could make the voltage more negative by two mechanisms that could happen at the same time. In the absence of CFTR protein, the ENaC would be hyperactive, as reported previously (35–37). Furthermore, reduced Cl⫺ permeability will result in the lack of a shunt conductance for Na⫹ transport. In both cases, the outcome would be hyperpolarization of the epithelium that could decrease voltage-dependent amino acid transport. We tested this hypothesis using selective inhibitors of ENaC and CFTR. The block of ENaC should increase amino acid

548


transport by making the transepithelial voltage less negative. Interestingly, the treatment with the ENaC blocker amiloride restores the amino acid uptake on CF to levels seen in non-CF epithelia. In contrast, we treated non-CF epithelia with two CFTR inhibitors to mimic the CF phenotype. We found that in these conditions the non-CF epithelium do behave as a CF tissue displaying a reduced amino acid uptake (see Fig. 6). Our conclusion is that membrane voltage changes really influenced the apical amino acid uptake in airway epithelium, and that reduced uptake may occur in CF in vivo as a result of CFTR-impaired activity. The degradation of basolateral tryptophan after IFN-␥-dependent IDO induction requires tryptophan uptake into the cell. We have not studied the basolateral transport system(s) responsible for this uptake. However, we can exclude that tryptophan moves through the same type of Na⫹-dependent transporter found in the apical membrane, because we previously showed that application of basic amino acids on the basolateral side did not cause a change of the Isc (28), indicating that this transporter is not expressed in the basolateral membrane. Tryptophan is normally transported into the cells by the aromatic amino acid system T (38), by branchedchain and aromatic amino acid transporter system L (39, 40), and by system y⫹ L (39). We previously found that amino acid efflux from the basolateral membrane is sensitive to transstimulation or amino acid exchange (28), a phenomenon characteristic of systems y⫹ and system L; therefore, it is possible to speculate that one or both of these systems might be responsible for tryptophan uptake through the basolateral membrane. The same transporters could be responsible for IFN-␥-dependent kynurenine secretion across the basolateral membrane, which would be facilitated by the exchange with tryptophan from the culture medium. Amino acids are critical for cell survival. In most cases, bacteria would rather use amino acids from their environment than make them de novo. Indeed, in nutrient-rich conditions, as is the case of the mammalian lower intestine, bacterial growth may not be energy limited because they may obtain a substantial fraction of their amino acids from their environments rather than through biosynthesis. Amino acid synthesis is a high cost process from the energetic point of view. The metabolism provides precursor metabolites for synthesis of the 20 aa incorporated into proteins, and thus, energy is lost by redirecting chemical intermediates from fueling reactions. Additional energy is required to convert precursor metabolites to amino acids. The range of costs of biosynthesis varies from 11 ATP equivalents per molecule of glycine, alanine, and serine to ⬎70 ATP per molecule of tryptophan. Tryptophan is the most energetically expensive amino acid to synthesize. In conclusion, our work provides the first evidence that human airway epithelial cells maintain low apical tryptophan and kynurenine concentrations by two important mechanisms: First, under resting conditions, there is a dramatic removal through a Na⫹dependent amino acid transporter. This transporter seems to be able to remove also kynurenine, and this could be important because tryptophan catabolites may suppress the proliferation of T and NK cells (20, 24). Second, in the presence of IFN-␥, the induction of IDO decreased further the tryptophan level. Because tryptophan synthesis is energetically very expensive for bacteria, its depletion from mucosal surface could be fundamental in the innate resistance of airways to bacterial colonization. A defective amino acid transport in CF patients, due to abnormal transepithelial potential, could be an important factor favoring bacterial colonization.

Acknowledgments We are grateful to Dr. E. Caci and R. Cusano for assistance in preparing the RT-PCR and real-time RT-PCR.

References 1. Martin, L. D., L. G. Rochelle, B. M. Fischer, T. M. Krunkosky, and K. B. Adler. 1997. Airway epithelium as an effector of inflammation: molecular regulation of secondary mediators. Eur. Respir. J. 10:2139. 2. Thompson, A. B., R. A. Robbins, D. J. Romberger, J. H. Sisson, J. R. Spurzem, H. Teschler, and S. I. Rennard. 1995. Immunological functions of the pulmonary epithelium. Eur. Respir. J. 8:127. 3. Goldman, M. J., G. M. Anderson, E. D. Stolzenberg, U. P. Kari, M. Zasloff, and J. M. Wilson. 1997. Human ␤-defensin-1 is a salt-sensitive antibiotic in lung that is inactivated in cystic fibrosis. Cell 88:553. 4. McCray, P. B. J., and L. Bentley. 1997. Human airway epithelia express a ␤-defensin. Am. J. Respir. Cell Mol. Biol. 16:343. 5. Black, H. R., J. R. Yankaskas, L. G. Johnson, and T. L. Noah. 1998. Interleukin-8 production by cystic fibrosis nasal epithelial cells after tumor necrosis factor-␣ and respiratory syncytial virus stimulation. Am. J. Respir. Cell Mol. Biol. 19:210. 6. DiMango, E., H. J. Zar, R. Bryan, and A. Prince. 1995. Diverse Pseudomonas aeruginosa gene products stimulate respiratory epithelial cells to produce interleukin-8. J. Clin. Invest. 96:2204. 7. Boehm, U., T. Klamp, M. Groot, and J. C. Howard. 1997. Cellular responses to interferon-␥. Annu. Rev. Immunol. 15:749. 8. Look, D. C., S. R. Rapp, B. T. Keller, and M. J. Holtzman. 1992. Selective induction of intercellular adhesion molecule-1 by interferon-␥ in human airway epithelial cells. Am. J. Physiol. 263:L79. 9. Asano, K., C. B. Chee, B. Gaston, C. M. Lilly, C. Gerard, J. M. Drazen, and J. S. Stamler. 1994. Constitutive and inducible nitric oxide synthase gene expression, regulation, and activity in human lung epithelial cells. Proc. Natl. Acad. Sci. USA 91:10089. 10. Watkins, D. N., M. J. Garlepp, and P. J. Thompson. 1997. Regulation of the inducible cyclo-oxygenase pathway in human cultured airway epithelial (A549) cells by nitric oxide. Br. J. Pharmacol. 121:1482. 11. Sauty, A., M. Dziejman, R. A. Taha, A. S. Iarossi, K. Neote, E. A. Garcia-Zepeda, Q. Hamid, and A. D. Luster. 1999. The T cell-specific CXC chemokines IP-10, Mig, and I-TAC are expressed by activated human bronchial epithelial cells. J. Immunol. 162:3549. 12. Fujimoto, K., T. Imaizumi, H. Yoshida, S. Takanashi, K. Okumura, and K. Satoh. 2001. Interferon-␥ stimulates fractalkine expression in human bronchial epithelial cells and regulates mononuclear cell adherence. Am. J. Respir. Cell Mol. Biol. 25:233. 13. Galietta, L. J. V., C. Folli, C. Marchetti, L. Romano, D. Carpani, M. Conese, and O. Zegarra-Moran. 2000. Modification of transepithelial ion transport in human cultured bronchial epithelial cells by interferon-␥. Am. J. Physiol. 278:L1186. 14. Shimizu, T., S. Nomiyama, F. Hirata, and O. Hayaishi. 1978. Indoleamine 2,3dioxygenase. J. Biol. Chem. 253:4700. 15. Liebau, C., A. W. Baltzer, S. Schmidt, C. Roesel, C. Karreman, J. B. Prisack, H. Bojar, and H. Merk. 2002. Interleukin-12 and interleukin-18 induce indoleamine 2,3-dioxygenase (IDO) activity in human osteosarcoma cell lines independently from interferon-␥. Anticancer Res. 22:931. 16. Fujigaki, S., K. Saito, K. Sekikawa, S. Tone, O. Takikawa, H. Fujii, H. Wada, A. Noma, and M. Seishima. 2001. Lipopolysaccharide induction of indoleamine 2,3-dioxygenase is mediated dominantly by an IFN-␥-independent mechanism. Eur. J. Immunol. 31:2313. 17. Carlin, J. M., Y. Ozaki, G. I. Byrne, R. R. Brown, and E. C. Borden. 1989. Interferons and indoleamine 2,3-dioxygenase: role in antimicrobial and antitumor effects. Experientia 45:535. 18. Taylor, M. W., and G. S. Feng. 1991. Relationship between interferon-␥, indoleamine 2,3-dioxygenase, and tryptophan catabolism. FASEB J. 5:2516. 19. Thomas, S. R., and R. Stocker. 1999. Redox reactions related to indoleamine 2,3-dioxygenase and tryptophan metabolism along the kynurenine pathway. Redox Rep. 4:199. 20. Melillo, G., G. W. Cox, D. Radzioch, and L. Varesio. 1993. Picolinic acid, a catabolite of L-tryptophan, is a costimulus for the induction of reactive nitrogen intermediate production in murine macrophages. J. Immunol. 150:4031. 21. Munn, D. H., E. Shafizadeh, J. T. Attwood, I. Bondarev, A. Pashine, and A. L. Mellor. 1999. Inhibition of T cell proliferation by macrophage tryptophan catabolism. J. Exp. Med. 189:1363. 22. Munn, D. H., M. D. Sharma, J. R. Lee, K. G. Jhaver, T. S. Johnson, D. B. Keskin, B. Marshall, P. Chandler, S. J. Antonia, R. Burgess, et al. 2002. Potential regulatory function of human dendritic cells expressing indoleamine 2,3-dioxygenase. Science 297:1867. 23. Hwu, P., M. X. Du, R. Lapointe, M. Do, M. W. Taylor, and H. A. Young. 2000. Indoleamine 2,3-dioxygenase production by human dendritic cells results in the inhibition of T cell proliferation. J. Immunol. 164:3596. 24. Frumento, G., R. Rotondo, M. Tonetti, G. Damonte, U. Benatti, and G. B. Ferrara. 2002. Tryptophan-derived catabolites are responsible for inhibition of T and natural killer cell proliferation induced by indoleamine 2,3-dioxygenase. J. Exp. Med. 196:459. 25. Munn, D. H., M. Zhou, J. T. Attwood, I. Bondarev, S. J. Conway, B. Marshall, C. Brown, and A. L. Mellor. 1998. Prevention of allogeneic fetal rejection by tryptophan catabolism. Science 281:1191. 26. Van Wissen, M., M. Snoek, B. Smids, H. M. Jansen, and R. Lutter. 2002. IFN-␥ amplifies IL-6 and IL-8 responses by airway epithelial-like cells via indoleamine 2,3-dioxygenase. J. Immunol. 169:7039. 27. Galietta, L. J. V., S. Lantero, A. Gazzolo, O. Sacco, L. Romano, G. A. Rossi, and O. Zegarra-Moran. 1998. An improved method to obtain highly differentiated monolayers of human bronchial epithelial cells. In Vitro Cell. Dev. Biol. Anim. 34:478.

The Journal of Immunology 28. Galietta, L. J. V., L. Musante, L. Romio, U. Caruso, A. Fantasia, A. Gazzolo, L. Romano, O. Sacco, G. A. Rossi, L. Varesio, and O. Zegarra-Moran. 1998b. An electrogenic amino acid transporter in the apical membrane of cultured human bronchial epithelial cells. Am. J. Physiol. 275:L917. 29. Dazzi, C., G. Candiano, S. Massazza, A. Ponzetto, and L. Varesio. 2001. New high-performance liquid chromatographic method for the detection of picolinic acid in biological fluids. J. Chromatogr. B Biomed. Sci. Appl. 751:61. 30. Saito, K., C. Y. Chen, M. Masana, J. S. Crowley, S. P. Markey, and M. P. Heyes.1993. 4-Chloro-3-hydroxyanthranilate, 6-chlorotryptophan and norharmane attenuate quinolinic acid formation by interferon-␥-stimulated monocytes (THP-1 cells). Biochem. J. 291:11. 31. Taddei, A., C. Folli, O. Zegarra-Moran, P. Fanen, A. S. Verkman, and L. J. Galietta. 2004. Altered channel gating mechanism for CFTR inhibition by a high-affinity thiazolidinone blocker. FEBS Lett. 558:52. 32. Ma, T., J. R. Thiagarajah, H. Yang, N. D. Sonawane, C. Folli, L. J. Galietta, and A. S. Verkman. 2002b. Thiazolidinone CFTR inhibitor identified by highthroughput screening blocks cholera toxin-induced intestinal fluid secretion. J. Clin. Invest. 110:1599. 33. Sloan, J. L., and S. Mager. 1999. Cloning and functional expression of a human Na⫹ and Cl⫺-dependent neutral and cationic amino acid transporter B0⫹. J. Biol. Chem. 274:23740.

549 34. Sloan, J. L., B. R. Grubb, and S. Mager. 2003. Expression of the amino acid transporter ATB0⫹ in lung: possible role in lumenal protein removal. Am. J. Physiol. 284:L39. 35. Grubb, B. R., and R. C. Boucher. 1999. Pathophysiology of gene-targeted mouse models for cystic fibrosis. Physiol. Rev. 79:S193. 36. Widdicombe, J. H. 2002. Regulation of the depth and composition of airway surface liquid. J. Anat. 201:313. 37. Grubb, B. R., R. N. Vick, and R. C. Boucher. 1994. Hyperabsorption of Na⫹ and raised Ca2⫹-mediated Cl⫺ secretion in nasal epithelia of CF mice. Am. J. Physiol. 266:C1478. 38. Kim, D. K., Y. Kanai, A. Chairoungdua, H. Matsuo, S. H. Cha, and H. Endou. 2001. Expression cloning of a Na⫹-independent aromatic amino acid transporter with structural similarity to H⫹/monocarboxylate transporters. J. Biol. Chem. 276:17221. 39. Castagna, M., C. Shayakul, D. Trotti, V. F. Sacchi, W. R. Harvey, and M. A. Hediger. 1997. Molecular characteristics of mammalian and insect amino acid transporters: implications for amino acid homeostasis. J. Exp. Biol. 200:269. 40. Kudo, Y., and C. A. Boyd. 2001. Characterization of L-tryptophan transporters in human placenta: a comparison of brush border and basal membrane vesicles. J. Physiol. 531:405.