IL-13 induced increases in nitrite levels are primarily driven by increases in inducible nitric oxide synthase as compared with effects on arginases in human primary bronchial epithelial cells

Summary

Background

Exhaled nitric oxide is increased in asthma, but the mechanisms controlling its production, including the effects of T-helper type 2 (Th2) cytokines, are poorly understood. In mouse and submerged human epithelial cells, Th2 cytokines inhibit expression of inducible nitric oxide synthase (iNOS). Arginases have been proposed to contribute to asthma pathogenesis by limiting the arginine substrate available to NOS enzymes, but expression of any of these enzymes has not been extensively studied in primary human cells.

Objectives

We hypothesized that primary human airway epithelial cells in air–liquid interface (ALI) culture would increase iNOS expression and activity in response to IL-13, while decreasing arginase expression.

Methods

iNOS and arginase mRNA (real-time PCR) and protein expression (Western blot and immunofluorescence) as well as iNOS activity (nitrite levels) were measured in ALI epithelial cells cultured from bronchial brushings of normal and asthmatic subjects following IL-13 stimulation.

Results

IL-13 up-regulated iNOS mRNA primarily at a transcriptional level in epithelial cells. iNOS protein and activity also increased, arginase1 protein expression decreased while arginase 2 expression did not change. The changes in iNOS protein correlated strongly with changes in nitrites, and inclusion of arginase (1 or 2) did not substantially change the relationship. Interestingly, iNOS mRNA and protein were not correlated.

Conclusions

These results contrast with many previous results to confirm that Th2 stimuli enhance iNOS expression and activity. While arginase 1 protein decreases in response to IL-13, neither arginase appears to substantially impact nitrite levels in this system.

Introduction

Exhaled nitric oxide (Fe_NO), a highly reactive gaseous mediator involved in both pro- and anti-inflammatory physiologic processes, is a relatively specific biomarker for human asthma [1–5]. Although nitric oxide (NO) is produced by 3 enzymatic isoforms [neural (nNOS), endothelial (eNOS) and inducible (iNOS)], iNOS is believed to contribute the bulk of the NO measured in human airways, due to its expression in asthmatic airway epithelial cells [6]. However, as T-helper type 2 (Th2) cytokines have generally been reported to decrease expression of iNOS [7–10], the mechanisms leading to up-regulation of iNOS and Fe_NO in asthma are not clear.

Studies from animal models, immortal cell lines or immature (submerged) human airway epithelial cells systems have reported Th1 and/or innate cytokines as the most potent inducers of iNOS [7–11]. In contrast, in these same systems, Th2 cytokines particularly IL-4 and IL-13, inhibit iNOS expression and nitrite production, or modestly increase it in combination with IFN-γ [7, 9–11]. Recently, IL-13 was reported to variably induce iNOS mRNA and protein expression in 2 of 3 normal human bronchial epithelial cell lines in air–liquid interface (ALI) culture, but no mechanisms for the increase were reported [12].

Arginase enzymes (1 and 2), which convert l-arginine to l-ornithine and urea may modulate NOS enzyme activity by controlling cell arginine levels, the substrate for NOS enzymes [13]. Although arginase 1 expression was originally described to be liver specific, recent studies suggest more widespread distribution, including airway epithelial cells [14, 15]. Arginase 2 has not been previously reported in airway epithelial cells. Increased arginase activity could modify asthma pathogenesis by limiting arginine levels which would lower NO levels, but enhance ornithine production. Increased ornithine levels could lead to greater airway remodeling through increases in proline synthesis [16, 17]. In addition, single nucleotide polymorphisms (SNPs) in both arginase 1 and 2 have been associated with increased risk of allergies and asthma [18].

Given the increase in Fe_NO in asthma, we hypothesized that Th2/IL-13 stimulation of well-differentiated primary human epithelial cells would up-regulate iNOS expression and activity and that this effect would involve both transcriptional and post-transcriptional mechanisms. Additionally, we hypothesized that IL-13 would decrease expression and activity of arginase enzymes and that the balance in expression of these enzymes would predict overall cell nitrite levels.

Methods

Bronchoscopy protocol with epithelial cell brushing

Bronchoscopy with endobronchial epithelial brushing was performed as previously described [20]. Asthmatic subjects all met the American Thoracic Society criteria for asthma. Mild-Moderate asthma subjects had a forced expiratory volume in 1 s (FEV₁) of ⩾80% predicted and used inhaled β-agonists alone or in conjunction with low dose inhaled corticosteroids (CS). Severe asthmatic subjects were referred to National Jewish Medical and Research Center or the University of Pittsburgh Medical Center for refractory asthma with continuous oral CS or frequent bursts, frequent hospitalizations and/or emergency room visits, evidence for ongoing severe airflow limitation and continuous symptoms [21]. No subject currently smoked or had a history of smoking > 5 pack years. The study was approved by the National Jewish and the University of Pittsburgh Institutional Review Boards and all subjects gave informed consent.

Air–liquid interface culture system

Freshly harvested epithelial cells were placed directly into 10 mL of ice-cold phosphate-buffered saline, centrifuged, washed, and resuspended in 1 mL of serum-free, hormonally supplemented bronchial epithelial cell basal medium (BEBM) with insulin, transferrin, triodothyronine, epinephrine, bovine hypothalamus extract, retinoic acid and human epithelial growth factor (LONZA, Basel, Switzerland) containing 50 μg/mL gentamicin and 50 μg/mL amphotericin). 9 × 10⁴ cells were seeded into 60 mm tissue-culture dishes coated with rat-tail type I collagen (BD Discovery Labware, Bedford, MA, USA), and cultured at 37 °C in a 5% CO₂ environment. When 80% confluent, the cells were passed onto collagen-coated polyester 12-well Transwell inserts (pore size 0.4 μm) at 4 × 10⁴ cells/cm² with BEBM/Dulbecco’s modified Eagle’s medium (DMEM) now supplemented with insulin (4 μg/mL), transferrin (5 μg/mL), hydrocortisone (0.5 μg/mL), epinephrine (0.5 μg/mL), bovine hypothalamus extract (52 μg/mL), bovine serum albumin (5 μg/mL), retinoic acid (30 ng/mL) and human epidermal growth factor (10 ng/mL). When confluent, they were shifted to ALI culture by removing all but 50 μL of the apical medium, and maintaining 1200 μL in the lower chamber. The media (with or without IL-13) was changed every other day. In the majority of studies except time course studies, the ALI state was maintained for 7–10 days, as previous studies have demonstrated this time is required for the mucociliary differentiation [22].

Quantitative real-time polymerase chain reaction

Epithelial expression of iNOS, arginase1 and 2 mRNA was determined by reverse transcription (RT), followed by real-time quantitative PCR as described previously [23]. The primers and probes, labeled with 5′-reporter dye 6-carboxy fluorescein (FAM) and 3′-quencher dye 6-carboxy N, N, N′,N′tetramethyl-rhodamine (TAMRA), were all purchased from Applied Biosystems (Assays on Demand, Foster City, CA, USA). (Catalogue: iNOS, Hs00167248_m1, arginase 2, Hs00265750_m1, arginase 1, Hs00163660_m1). VIC-labeled ready-for-use human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) primers and probe were obtained from Applied Biosystems (Genbank accession no: NM-002046, Catalogue : 4310884E). Real-time PCR was performed on the ABI Prism 7700 or 7900 sequence detection system (Applied Biosystems, Foster City, CA, USA). mRNAs of interest were indexed to GAPDH using the formula $1/2^{\Delta C_{\text{t}}}\times 1000$.

Western blots for inducible nitric oxide synthase and Arginase

Twenty-five μg protein samples were resolved on 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred and immuno-probed with rabbit polyclonal antibody for iNOS (1 : 500, BD Transduction Laboratories) and arginase 1 and 2 (1 : 500, Santa Cruz Biotechnology). Horseradish peroxidase (HRP) conjugated secondary antibody (F(ab′)2) (1 : 4000) was followed by ECL plus Western blot detection reagent (Amersham, Buckinghamshare, UK) and chemiluminescence detected by Luminocent image analyzer, LAS-3000 (Fuji Film, Tokyo, Japan). Blots were stripped and then re-probed for β-actin (1 : 5000) (Sigma, St Louis, MO, USA) or GAPDH (1 : 2000, NOVUS, Littleton, CO, USA). Densitometry was performed using the Multi Gage software (Fuji Film) and the protein of interest was indexed to β-actin or GAPDH. To evaluate the presence of the fully dimerized iNOS protein, Western blot was performed without dithiopthreital (DTT) in the lysates. The positive control for iNOS protein was submerged epithelial cells stimulated with a mix of 50 ng/mL (each) of IL-1β, TNF-α and IFN-γ for 24 h as previously described [10]. Positive controls for aringase 1 and 2 protein were rat liver and kidney lystates [24].

Immunofluorescence

Transwell membranes were fixed in acetone and embedded in glycol-methacrylate resin. Serial 2-μm sections were immunostained using rabbit polyclonal antibody against iNOS (1 : 50) (Oxford Biomedical Research, Oxford, MI, USA) and fluorescein-labeled secondary antibody (Vector Laboratories Inc Burlingame, CA, USA). Nuclei were stained with (4′,6-diamidino-2-phenylindole) (DAPI) dye and the negative control consisted of secondary antibody only.

Nitrite assay

Nitrite assays were performed on 75 μL of media which had been added to the upper wells, incubated for 30 min and removed. This was done to standardize the upper volume as evaporation and absorption led to variable volumes after 48 h. Nitrite in the upper supernatants was quantified by using a colorimetric assay based on the Griess reaction [25]. Briefly, 10 μL of the supernatants were added to 100 μL of Griess reagent (1% sulfanilamide, 0.1% naphthylethylenemide dihydrochrolide in 5% phosphoric acid) of a 96-well plate. Absorbance was determined at 540 nm. The concentration of ${\text{NO}}_2^{-}$ was calculated by comparison to a standard curve of known nitrite concentrations (5 to 50 μM).

MessengerRNA stability

mRNA stability was analyzed by adding 5 μg/mL actinomycin D (Act D) (Sigma) to cells stimulated with IL-13 or IFN–γ (R&D systems, Minneapolis, MN, USA). As extremely low levels of iNOS mRNA are present in unstimulated cells, a comparison of the T_1/2 (half-life) of IL-13 stimulated iNOS mRNA vs. iNOS mRNA from cells treated with media alone was not possible. Therefore, a comparison was made between the mRNA stability of iNOS following stimulation with IL-13 vs. IFN-γ. Cells were cultured for 6 days with media alone. On the 6^th day, cells were treated with IL-13 or IFN-γ (100 ng/mL). After 12 h, actinomycin (Act) D was added and cells were harvested at 0, 1, 6, and 12 h. Total RNA was extracted for quantification of iNOS mRNA using real-time PCR, and the amount of mRNA compared at each time point for the two conditions. Additionally, the levels of iNOS mRNA were determined after 24 h stimulation of either IL-13 or IFN-γ.

In vitro polymerase chain reaction-based nuclear run-on assay

Nuclear run-on assay was performed on nuclei (CelLytic nuclear extraction kit; Sigma) isolated from epithelial cells stimulated with media alone or IL-13 for 7 days according to previously described methods [26–28] The nuclei were stored in 200 μL of storage buffer [50 mm Tris HCl, pH 8.3, 40% glycerol (vol/vol), 5 mm MgCl₂, and 0.1 mm EDTA] split into two aliquots and incubated at 37 °C in 20% glycerol, 30 mm Tris HCl, 2.5 mm MgCl₂, 150 mm KCl, 1 mm DTT, and 40 U of RNAse inhibitor (Applied Biosystems). ATP, CTP, GTP, and UTP (0.5 mm each) were added to one aliquot (+NTPs), while no NTPs were added to the second aliquot (−NTPs). After 30 min incubation, RNA was isolated, DNAse I digestion performed, and RT and real-time PCR products analyzed as described to determine the quantity of newly formed mRNA, calculated as follows:

$$\text{Relative}\text{mRNA}=2^{-\left\{\text{iNOS}{C}_t\right[(+\text{NTPs})-(-\text{NTPs})]-\text{GAPDH}{C}_t[(+\text{NTPs})-(-\text{NYPs})\left]\right\}}$$

Statistics

Variables were checked for normality of distribution. When possible, data were log transformed to achieve linearity (changes in nitrites, iNOS and arginase 1 and 2 protein). All other data were not normalized by log transformation and were analyzed using nonparametric tests. Paired t-tests (or the non-parametric equivalent, the Wilcoxon test) were performed to compare basal and stimulated responses. Pairwise correlation analysis was performed to evaluate relationships between changes in iNOS, arginases and their ratios with fold changes in nitrites. For all comparisons, P-values <0.05 were considered significant.

Results

Subjects

Thirteen normal, seven mild-moderate and 15 severe asthma subjects underwent bronchoscopic airway brushing to collect primary epithelial cells, and were included here (Table 1). Not all subject’s cells were used for every experiment. As expected, FEV₁ (% predicted) was different between groups (P<0.0001). In this small sample size and including atopic normal controls, Fe_NO was marginally higher comparing all asthmatic subjects to the normal controls (P = 0.06) but there was no difference in Fe_NO by severity (P = 0.15).

Table 1.

Subject characteristics for studies

	Gender (M/F)	Age (year)	FEV1 (%)*	ICS (μg)*	Oral GC (mg)*	FeNO (p.p.b)†
Normal Control	6/7	37(22–59)	102 ± 3	0	0	28 ± 10 (n = 8)
Mild-Moderate	3/4	40(22–64)	86 ± 8	178 ± 141	0	65 ± 24 (n = 5)
Severe	8/7	42(22–58)	46 ± 3	977 ± 189	18 ± 6	73 ± 21 (n = 13)
P-values	NS	NS	< 0.0001	N/A	N/A	0.15

FEV₁, forced expiratory volume in 1 s; FeNO, exhaled nitric oxide.

Mean ± SEM, ICS: inhaled corticosteroid GC: glucocorticoid.

Interleukin-13 increases inducible nitric oxide synthase expression and activity in bronchial epithelial cells

Messenger RNA.

Primary airway epithelial cells (n = 33) from asthmatic and normal control subjects were cultured for 10 days in the presence of IL-13 (10 ng/mL). IL-13 consistently increased expression of iNOS mRNA (Fig. 1a, P<0.0001).

Fig. 1.

(a) IL-13 (10 ng/mL for 10 days) significantly increased inducible nitric oxide synthase (iNOS) mRNA (n = 33, P<0.0001) compared with cells stimulated with media alone. (b) IL-13 (10 ng/mL for 10 days) significantly increased iNOS protein levels (n = 32, P < 0.0001). Representative Western blots are shown above the graph. *Positive control = submerged primary human epithelial cells stimulated with IL-1β, TNF-α and INF-γ (50 ng/mL each) for 24 h.

Protein.

Primary airway epithelial cells (n = 32) from asthmatic and normal control subjects were cultured for 10 days in the presence of IL-13 (10 ng/mL). Similar to mRNA, IL-13 up-regulated iNOS protein expression (Fig. 1b, P<0.0001). To evaluate the distribution of iNOS protein in the epithelial ALI system with and without IL-13 stimulation (10 ng/mL for 10 days), the Transwell membranes were evaluated by immunofluorescence for iNOS protein. The protein was located primarily in the apical cells, often in association with goblet cells (Fig. 2). There was no correlation between iNOS mRNA and protein in IL-13 stimulated group (ρ = −0.03, P = 0.87, n = 32).

Fig. 2.

Representative immunofluorescence of bronchial epithelial inducible nitric oxide synthase (iNOS) protein from five independent subjects. Original magnification is × 200. (a) Secondary antibody alone with DAPI stain on unstimulated cells at day 10 of ALI. (b) Basal expression of iNOS in unstimulated cells at day 10. (c) Secondary antibody alone for IL-13-stimulated cells at day 10. (d) IL-13 induced expression of iNOS after 10 days. The green and blue colors represent iNOS antibody and nuclear DAPI staining, respectively.

Enzyme activity.

Enzyme activity and directionality were determined by measuring nitrite levels in apical and basal supernatants from cells treated with media alone and with IL-13 (10 ng/mL) for 10 days (n = 32). IL-13 consistently increased the levels of nitrite in the apical supernatants, Interestingly, although IL-13 increased nitrite levels in normal control (P = 0.0009) and asthmatic subjects (P<0.01), the increase and the absolute nitrite levels in the normal controls were significantly greater than those observed in asthmatic subjects (P<0.05 for change, P<0.02 for absolute level following IL-13) (Fig. 3). Most lower chamber nitrite levels were undetectable, perhaps due to the sensitivity of the assay and the more dilute media. The percent increase in nitrite levels following IL-13 stimulation was correlated with the percent increase in iNOS protein expression (r = 0.38, P = 0.04 n = 31).

Fig. 3.

IL-13 (10 ng/mL for 10 days) increased nitrite levels in the apical supernatants. IL-13 increased nitrite levels in normal control (P = 0.0009) and asthmatic subjects (P<0.01). The increase and the absolute nitrite levels in the normal controls were significantly greater than those observed in asthmatic subjects (P<0.02). Changes in nitrite values were significantly greater in normal controls compared with asthmatics (P<0.05).

Time course and dose response.

Epithelial cells from three normal and one mild asthma subjects were stimulated with IL-13 in doses ranging from 0.5 to 20 ng/mL for varying times (1, 4, 7 and 10 days). In addition, a short time course (0–24 h) was performed on an additional four normal and five asthmatic subjects. The increases in iNOS mRNA and protein followed a dose and time-dependent relationship, generally peaking between 10 and 20 ng/mL of IL-13 and between 7 and 10 days of stimulation (Figs 4a–e). However, increases in mRNA were consistently seen as early as 5 h after stimulation (baseline 0.02 ± 0.01, 5 h is 1.3 ± 0.5 (iNOS mRNA relative to GAPDH, P = 0.0002, n = 9). To confirm that IL-13 increased the functional homodimeric enzyme, four samples were analyzed on non-denatured gels. With IL-13 stimulation, the homodimer was visible at the expected ~250 kDa molecular weight band (Fig. 4c). The ~250 kDa fully dimerized protein increased in dose and time dependent manner to IL-13, while the non-dimerized band at 130 kDa decreased.

Fig. 4.

(a) Dose response: IL-13 (0.5–20 ng/mL) added to the cultures. IL-13 increased inducible nitric oxide synthase (iNOS) mRNA and peaks at 10 ng/mL (n = 4, P<0.01 over all). (b) Time course study: IL-13 (10 ng/mL) added from 1 to 10 days before harvest. IL-13 increased iNOS mRNA which peaked at 7 days (n = 4, P<0.01 over all). In small inset, increases in mRNA were consistently seen as early as 5 h after stimulation (baseline 0.02 ± 0.01, 5 h is 1.3 ± 0.5 [iNOS mRNA relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH), P = 0.0002, (n = 9)]. (c) Representative Western blots for iNOS protein (denatured gels) (upper blots). IL-13 increased iNOS protein in dose and time dependent manner. Non-denatured gels (lower blots) demonstrating IL-13 induced dose and time dependent increases in the higher molecular weight iNOS homodimer (~250 kDa), with concomitant decreases in the monomeric form (130 kDa). (d) Quantitative densitometric analysis of the dose dependent IL-13 induced increase in iNOS protein. (P = 0.02). (e) Quantitative densitometric analysis of the time dependent IL-13 induced increase in iNOS protein (P = 0.03).

Inducible nitric oxide synthase mRNA stability in the presence of interleukin-13

iNOS mRNA stability was analyzed as the percent mRNA remaining over time following addition of Act D to IL-13 and IFN-γ stimulated cells (Fig. 5a). The rate of decline in iNOS mRNA levels with stimulation with IL-13 or IFN-γ was not significantly different with the T_1/2 around 5 h for both. Of note, the absolute increases in iNOS mRNA levels were compared after IL-13 (10 ng/mL) or IFN-γ (100 ng/mL stimulation for 24 h. IL-13 induced significantly higher amounts of iNOS mRNA than IFN-γ [(7.2 ± 3.6 (n = 16) vs. 0.9 ± 0.3 (n = 7)] relative units, P = 0.02) suggesting that additional mechanisms beyond mRNA stabilization were likely to be involved.

Fig. 5.

(a) Stability of IL-13-induced iNOS mRNA expression appears to be similar to that induced by IFN-γ. Act D was added after 12 h of stimulation, the RNA harvested at 0, 1, 6 and 12 h. There is no significant difference between IL-13 and IFN-γ at any time points (n = 3 for IL-13 group and n = 7 for IFN-γ group). (b) Nuclear run on assay on cells stimulated with IL-13 (10 ng/mL for 10 days) and treated with NTPs (or not) for 30 min. Data are presented means ± SE. All data are expressed as percentage of baseline (no NTPs added group, n = 5).

Inducible nitric oxide synthase mRNA transcription in response to interleukin-13

To determine whether IL-13 induced transcription of iNOS, nuclear run-on assays were performed in ALI epithelial cells from five subjects (two normal, one mild-moderate and two severe asthmatic subjects). IL-13 increased iNOS mRNA expression without altering GAPDH mRNA, and was therefore normalized by GAPDH. iNOS mRNA was undetectable in unstimulated conditions (C_t values of 38 to 40). In the presence of IL-13, iNOS mRNA was higher (C_t 29–30). In every subject, the addition of nucleotides for 30 min to isolated nuclei from IL-13 stimulated cells further increased the level of iNOS mRNA compared with the nuclei from IL-13 stimulated cells without nucleotides. The averge C_t differences were around 0.7, or an increase of 60% over no NTPs. These results, indexed to the baseline non-nucleotide iNOS values, confirm active transcription at 24 h post-stimulation with IL-13 (as shown in Fig. 5b).

Arginase expression in bronchial epithelial cells

To determine whether IL-13 reciprocally lowered arginase expression, both arginase 1 and 2 mRNA (n = 26) and protein (n = 25) expression were evaluated. Arginase 2 mRNA and protein were constitutively expressed in bronchial epithelial cells but were not altered by IL-13 (P = 0.09 for protein)) (Figs 6a and b). Further, there was no dose or time dependent response to IL-13 (n = 4) (data not shown). In contrast, arginase 1 mRNA was detectable in a small minority (Five of 26) subjects in the media control samples. However, in the five subjects with detectable mRNA, IL-13 decreased arginase 1 mRNA in four of five subjects (0.002 ± 0.001 (0 to 0.02) for control, 0.0004 ± 0.0004 (0 to 0.01) for IL-13. In contrast to mRNA, 20 of 26 subjects expressed low levels of arginase 1 protein and IL-13 decreased the levels further (P = 0.03, n = 25, Figs 7a and b). Changes in arginase 1 protein levels were not correlated with nitrite levels. When the changes in the ratio of iNOS to arginase 1 protein in response to IL-13 were compared with the increases in nitrite levels, the correlation remained, but was not better than that for iNOS alone (ρ = 0.32 P = 0.18, n = 19).

Fig. 6.

Arginase 2 mRNA and protein are constitutively expressed in primary human bronchial cells. (a) IL-13 (10 ng/mL for 10 days) did not change arginase 2 mRNA expression (P = 0.37, n = 26). (b) Arginase 2 Western blots (representative from 3 subjects). Densitometry analysis of arginase 2 protein normalized by β-actin (n = 25). IL-13 did not impact arginase 2 protein expression (P = 0.09). IL-13 marginally increased arginase 2 protein expression. *Positive control for arginase 2 = rat kidney lysate [24].

Fig. 7.

(a) Representative Western blot of arginase 1. IL-13 (10 ng/mL for 10 days) decreased arginase 1 expression. (b) IL-13 decreased arginase 1 protein expression in 19 subjects with detectable baseline levels of arginase 1 (P = 0.03). *Positive control for arginase 1 = rat liver lysate [24].

Comparison of asthmatic with normal epithelial cells

iNOS, arginase 1, 2 mRNA, protein or activity did not differ by subject group or use of corticosteroids under either media or IL-13 conditions. In addition, there were no correlations of any of these factors with Fe_NO. However, both the change and the absolute levels of nitrites induced by IL-13 were higher in normal than asthmatic cells (Fig. 3).

Discussion

The results reported here are the first systematic analysis of cultured primary human airway epithelial cells from asthmatic and normal subjects for the impact of IL-13 on three enzymes believed primarily responsible for the levels of Fe_NO in human airways. These studies demonstrate a profound ability of IL-13 to enhance iNOS mRNA, protein and activity in well-differentiated human epithelial cells, with some indication that the impact of IL-13 is greater in normal, as compared with asthmatic cells. The increase in iNOS appears to involve transcription, but the absence of any correlation between mRNA and protein suggests a complex relationship, between transcription and translation. In contrast, arginase 2 is constitutively expressed and not impacted further by IL-13, whereas arginase 1 protein expression appears to be decreased by IL-13. Finally, although nitrite levels in these cells may be driven primarily by iNOS levels, arginase 1 protein expression appears to play some role as well.

Many previously published reports have suggested Th2 cytokines (IL-4 and IL-13) have (1) no effects on iNOS mRNA/protein, (2) inhibitory effects on iNOS mRNA/protein or (3) stimulatory effects on iNOS mRNA protein, but only in the presence of one or two additional stimuli (IFN-γ, IL-1β) [11, 29–32]. These studies have utilized cell lines (BEAS-2B, A549, J774), submerged human epithelial (lung and colonic) cells, macrophages (primarily rodent) and various mouse models. A previous extensive evaluation of submerged human epithelial cells was unable to show a positive effect of IL-4 on iNOS, although IL-13 was not evaluated [11]. Perhaps some of the strongest in vivo data to show a negative effect of Th2 cytokines on iNOS come from mice deficient in either IL-4 or IL-4 receptor alpha (IL-4Rα) [33]. These mice, when infected with Plasmodium berghei had greatly enhanced iNOS expression compared with wild type mice, leading to the conclusion that the increased iNOS expression was due to the loss of IL-4 and IL-4Rα inhibitory responses.

The results reported here confirm and extend a previous small report of the increase in iNOS mRNA, protein and activity in primary human airway epithelial cells cultured at the ALI [12]. In that previous study, ALI-epithelial cells from two of three ‘normal’ donor lines were demonstrated to respond to IL-13 with an increase in iNOS mRNA protein and activity. In the current study, we extend those findings by evaluating a total of 35 separate cells from normal and asthmatic subjects over a wide range of severity. In almost all human cells studied, IL-13 induced a profound increase in iNOS expression at the mRNA and protein level. Interestingly, despite the consistent increases in both mRNA and protein in response to IL-13, there was no correlation between the mRNA and either protein or nitrite levels. This poor relationship supports the importance of additional regulatory factors such as splice variants previously published in relation to iNOS mRNA [34, 35]. However, the discrepancy does not appear to be due to endothelial NOS expression by these cells, as similar to a previous report [12], eNOS mRNA and protein were not present in these cells under basal or IL-13 stimulated conditions (data not shown). Finally, the clinical relevance of this effect is strongly supported by the recent study which reported that an inhibitor of the action of both IL-4 and IL-13 on the IL-4Rα receptor significantly decrease Fe_NO levels in humans, supporting the clinical relevance of the increased expression of iNOS in response to IL-13 [36].

In contrast to expected results, the increase in iNOS activity (as measured by nitrites in the apical supernatants, and to some degree with iNOS protein) was significantly greater in the normal subjects, as compared with the asthmatic subjects. Reduced nitrite production from airway epithelial cells derived from asthmatic subjects compared with normal has been reported previously [10] albeit with cells submerged in media. In this study, we did not see differences in baseline levels, only after treatment with IL-13. The clinical relevance or the mechanism for this effect remains to be determined, although it was not due to a higher basal production in asthmatic cells (data not shown).

In addition to confirming and comparing the response to IL-13 across a wide range of subjects, we determined whether IL-13 preferentially increased mRNA stability or transcription of iNOS mRNA. Previous reports suggested that iNOS mRNA levels are modulated through both transcription and mRNA stability. In human hepatocytes, IL-1β was reported to transcriptionally increase iNOS mRNA but a direct effect on iNOS mRNA stability was not shown [37]. In human submerged airway epithelial cells, the combination of IFN-γ and IL-4 prolonged the iNOS mRNA half-life [11, 38]. The results from this study suggest that IL-13 stimulation of ALI epithelial cells induces similar mRNA stability to that previously reported for IFN-γ. In these cases, the mRNA T_1/2 is short, in the range of 5 h. Whether this half-life is different from the mRNA half-life without IL-13 (or IFN-γ) stimulation cannot be determined due to the very low levels of iNOS mRNA present in the absence of stimulation. However, more importantly, IL-13 also increases the rate of transcription of iNOS mRNA, with a 66.2% increase in mRNA at 30 min with added NTPs, as compared with samples without NTPs. This effect on transcription in isolated nuclei confirms that the impact of IL-13 to increase iNOS mRNA is not due to a differentiation of epithelial cells by IL-13 into ‘goblet cells’ which indirectly produce iNOS through some other pathway. This effect is further supported by the rapid induction of iNOS mRNA in response to IL-13 (within 5 h). Thus, the effect of IL-13 to increase mRNA in human epithelial cells clearly involves transcription, but involvement from some degree of mRNA stabilization cannot be ruled out. Further studies will be required to determine the transcriptional control of iNOS by IL-13 in human epithelial cells.

In addition to evaluating iNOS, we also evaluated the levels of arginase 1 and 2 at baseline and in response to IL-13. Several studies have suggested that levels of these enzymes, in association with iNOS, control the levels of Fe_NO produced in asthmatic airways [39, 40]. This effect is believed to be through the ability of these enzymes to modulate the cellular levels of arginine, the substrate for iNOS. Our study does not allow us to evaluate the relationship of these enzymes to Fe_NO in vivo, but the interaction of iNOS and arginase1 (and 2) to modulate nitrite levels in vitro was evaluated in a wide number of primary human epithelial cell cultures. In vitro, IL-13 primarily controls iNOS, as compared with arginase1 or 2 with levels of iNOS more relevant to the production of nitrites. However, the results reported here also suggest that arginase 1 in response to IL-13 and in contrast to the constitutive expression of arginase 2 has some impact on nitrite levels.

Although arginase 1 expression was originally believed to be specific to liver, recent reports have suggested that arginase1 is expressed in human epithelial cells as well, with some data to suggest it can be increased by stimuli such as nicotine [39]. Arginase 1 was measurable at the protein level in 20 of 25 subject’s cultured epithelial cells and IL-13 significantly diminished the levels. However, the addition of arginase 1 levels to the equations did not improve the correlation with nitrite levels. Unfortunately, the levels of arginase 1 mRNA were very low, making these results somewhat suspect. While an evaluation of arginase 1 activity in this system should be able to confirm (or not) the enzyme’s presence, the competing levels of arginase 2 in these cultures nullify this approach. Therefore, these results must be interpreted with caution. However, in contrast to these results, a recent study reported IL-13 increased arginase 1 expression in a murine model [41]. While further confirmation is warranted, the results reported here suggest that similar to iNOS, IL-13 has opposite effects on arginase 1 in humans as compared with mice.

This is the first report we are aware to measure arginase 2 and compare it with arginase 1 in human cells. Despite the lack of effect of IL-13 on arginase 2, the mRNA and protein levels of arginase 2 were much greater than those for arginase 1 in these primary human bronchial epithelial cells. Further studies of arginase 2 in vivo and in response to non-Th2 stimuli will be required to determine whether it may modulate nitrite production from cells in other settings.

In summary, these results confirm and greatly expand on the observation that IL-13 strongly induces iNOS expression primarily through transcription in primary human airway epithelial cells in an ALI system. iNOS protein levels directly correlate with the activity as measured by nitrite levels. Preliminary evidence supports a difference in the response to IL-13 in normal as compared with asthmatic cells, with greater enzyme activity in normal cells. In this in vitro system, arginase 2 is constitutively expressed and not impacted by IL-13, while the expression and perhaps clinical relevance of arginase 1 appears to be more complex. Further studies are necessary to better understand the regulatory mechanisms of these enzymes and their in vivo contribution to asthma pathogenesis.