Antioxidant imbalance in the lungs of cystic fibrosis transmembrane conductance regulator protein mutant mice

Leonard W. Velsor, Anna van Heeckeren, Brian J. Day


Recent studies suggest that the cystic fibrosis (CF) transmembrane conductance regulator (CFTR) protein modulates epithelial reduced glutathione (GSH) transport and when defective creates an antioxidant imbalance. To test whether the CFTR protein modulates lung antioxidant defenses in vivo, epithelial lining fluid (ELF) and lung tissue from CFTR knockout (CFTR-KO) and wild-type (WT) mice were compared for GSH content and the activities of glutathione reductase, glutathione peroxidase, and γ-glutamyltransferase. In the CFTR-KO mice, the ELF concentration of GSH was decreased (51%) compared with that in WT mice. The concentration of GSH in the lung tissue of CFTR-KO mice, however, was not significantly different from that in WT mice. The activities of glutathione reductase and glutathione peroxidase in the lung tissue of CFTR-KO mice were significantly increased compared with those in WT mice (48 and 28%, respectively). Tissue lipid and DNA oxidation were evaluated by measurement of thiobarbituric acid-reactive substances and 8-hydroxy-2′-deoxyguanosine, respectively. The levels of thiobarbituric acid-reactive substances and 8-hydroxy-2′-deoxyguanosine in the lung tissue of CFTR-KO mice were significantly increased compared with those in WT mice. These data support our hypothesis that a mutation in the CFTR gene can affect the antioxidant defenses in the lung and may contribute to the exaggerated inflammatory response observed in CF.

  • epithelial lining fluid
  • glutathione
  • oxidative stress

cystic fibrosis (CF) is a lethal autosomal recessive disorder associated with mutations in the gene encoding for the CF transmembrane conductance regulator (CFTR) protein. In the general population, CF has an overall incidence of ∼1:3,000, with large differences among ethnic groups (27). CFTR is a 168-kDa integral membrane protein primarily expressed in the epithelia of the lung, pancreas, sweat glands, and vas deferens (27). CFTR couples ATP hydrolysis with the transport of Cl and possibly other large anions across apical cell membranes to maintain the composition of secretions on the epithelial surfaces (26, 27). Although CF has serious clinical implications for the gastrointestinal and genital tracts, pulmonary disease is the primary cause of death in 90% of CF patients (27).

The progressive obstructive lung disease associated with CF is maintained by recurrent episodes of infection, predominantly byPseudomonas aeruginosa (10). A current view is that the inflammatory responses associated with the persistent infection drive an injury and repair process that leads to pulmonary fibrosis, airway obstruction, and, ultimately, respiratory failure (20). Tissue injury is a direct consequence of the oxidative environment created by the inflammatory response. Inflammatory cell-derived oxidants and proteases react with critical cellular biomolecules (i.e., lipids, DNA, and proteins) and lead to cell necrosis. Chronic repair processes lead to fibrosis and the progressive deterioration of pulmonary function (20).

In vitro studies suggest that CFTR may be involved in maintaining the antioxidant homeostasis of the pulmonary epithelial lining fluid (ELF), and a mutation in CFTR may impair lung antioxidant defenses, thereby making the CF lung more susceptible to oxidative stress and fibrosis (16, 25). Analysis of bronchoalveolar lavage fluid (BALF) from adult CF patients demonstrated that these patients have lower concentrations of ELF reduced glutathione (GSH) than normal control subjects (35). Whether this decrease is a direct consequence of the CFTR mutation or an artifact of an underlying infection cannot be ascertained; however, these observations fuel speculation that the CFTR mutation might indirectly attenuate lung antioxidant defenses. Recently, experiments using pulmonary epithelial cell lines demonstrated that cells containing defective CFTR secreted less GSH than control cells containing functional CFTR and that transfection of these cells with functional CFTR restored GSH secretion to that of control cells (16). These data provide support for the hypothesis that the CFTR protein is involved in antioxidant homeostasis of the ELF.

To test whether a defect in CFTR alters the constitutive lung antioxidant defenses in vivo, ELF and lung tissue from CFTR knockout (CFTR-KO) and wild-type (WT) mice were compared. In the ELF, the concentration of GSH was significantly decreased in the CFTR-KO mice, whereas tissue concentrations of GSH were similar. In the CFTR-KO lung, the activities of glutathione reductase (GR) and glutathione peroxidase (GPx) were increased, whereas the activity of γ-glutamyltransferase (γ-GT) was unchanged. Two indicators of oxidative stress, thiobarbituric acid reactive substances (TBARS) and 8-hydroxy-2-deoxyguanosine (8-OHdG), were also increased in the CFTR-KO lung tissue.



The CFTR-KO mice used in this study were congenic B6.129P2-Cftrtm1Unc, which possess the S489X mutation in CFTR that renders the CFTR protein nonfunctional (9,40). Heterozygous breeding pairs were either originally obtained from Jackson Laboratories (Bar Harbor, ME) or a kind gift from Sandra Gendler (Mayo Clinic Scottsdale, Scottsdale, AZ). In this study, the CFTR-KO mice, homozygous for the S489X mutation, were compared with their WT littermates. Genotypes for each mouse were determined by PCR with DNA isolated from tail clips as described in DNA extraction and genotyping by PCR. CFTR-KO mice were maintained on a liquid elemental diet (Peptamen, Nestle, Glendale, CA). WT mice received the regular solid mouse chow (Teklad 9F sterilizable rodent diet 8760, Harland, Madison, WI) or the liquid elemental diet (14). Autoclaved tap water in bottles with sipper tubes was provided ad libitum, and all mice were housed on inedible corncob bedding to prevent the intestinal obstruction associated with the CFTR mutation (14).

DNA extraction and genotyping by PCR.

DNA for genotyping was isolated from tail clips. Approximately 0.5 cm of tissue was hydrolyzed in 50 μl of 0.2 N NaOH and incubated at 75°C for 15–20 min. The samples were then neutralized with 200 μl of 40 mM Tris (pH 7.5), and the debris was removed by centrifuging for 30 s.

Murine CFTR DNA from 129/Sv, C57BL/6J, A/J, BALB/cJ, DBA2/J, and C3H/HeJ mice was sequenced with a modification of previously published methods (40). In the CFTR sequence from 129/Sv mice, single nucleotide differences in exons 14a and 17a alter theRsaI and AluI restriction sites, respectively. Although both exons amplify very efficiently, exon 14a is closer to the actual CFTR mutations. With a single PCR assay, WT and CFTR-KO mice can be easily distinguished by restriction of the PCR product and size separation on agarose.

A PCR master mix [1.5 μl of 10× PCR buffer, 0.6 μl of primer, 0.2 μl of 5 U/ml of Taq polymerase (GIBCO BRL, Life Technologies, Grand Island, NY), 1.2 ml of deoxynucleotide triphosphates, 8.6 ml of water, and 0.3 ml of magnesium chloride per 2 μl of tail sample] was used for genotyping. The PCR primers forRsaI, mCFEx14a5′ (GAG TGT TTT CTT GAT GAT GTG) and mCFEx14a3′ (ACC TCA ACC AGA AAA ACC AG), were obtained from Integrated DNA Technologies (Coralville, IA) at a concentration of 20 nM each. Restriction buffer was made with 2 μl of 10 U/ml of RsaI and 2 μl of 10× RsaI buffer (10 mM MgCl2, 10 mM bis-Tris propane-HCl, pH 7.0, and 1 mM dithiothreitol) in 2.8 μl of water per sample. PCR products were separated on an agarose gel (1% agarose and 2% NuSieve) in Tris-acetate-EDTA buffer. Bands at 109 and 21 bp were identified as WT, and an undigested 130-bp band was identified as 129/Sv-derived CFTR mutants.

Isolation of BALF and lung tissue.

The ELF was sampled for antioxidant concentrations with BALF and normalized for dilution with the urea method (8). The mice were killed by carbon dioxide anoxia followed by exsanguination by direct cardiac puncture. BALF was collected with three individual 1-ml aliquots of sterile phosphate-buffered saline (PBS), pH 7.4. The three aliquots were pooled, acidified with 5% metaphosphoric acid, and centrifuged (4,000 g for 5 min at 4°C) to remove cells; the supernatant was retained and stored at −80°C for subsequent analyses. The right and left lungs were then removed and snap-frozen in liquid nitrogen. The lungs were then ground into a fine powder with a liquid nitrogen-cooled mortar and pestle and then stored at −70°C until analysis. Aliquots of the ground tissue were carefully removed under liquid nitrogen as needed for subsequent analyses.

GSH assay in BALF and lung tissue.

To minimize GSH loss, BALF was acidified with 5% metaphosphoric acid (150 μl/ml), cooled on ice, and centrifuged (10,000 g for 10 min at 4°C) to remove precipitated proteins. To determine GSH concentrations in lung tissue, ∼20 mg of the ground tissue were dissolved in 600 μl of PBS; this solution was then acidified with 5% metaphosphoric acid, cooled, and centrifuged to remove precipitated proteins. GSH concentrations in the concentrated BALF and tissue homogenates were then determined spectrophotometrically with a commercially available assay that forms a chromogen with GSH (GSH-400, Oxis International, Portland, OR). Determination of the GSH-to-oxidized glutathione (GSSG) ratio in ELF was calculated from the difference in GSH concentrations in the BALF treated with 0.5 mU/ml of GR and 10 μM NADPH (28).

Serum and BALF urea concentrations.

To determine actual ELF concentrations of soluble antioxidants from BALF, a dilution factor was derived from the difference between BALF and serum urea concentrations. It is assumed that the urea concentrations in the vascular and ELF compartments are equivalent because urea is freely diffusible (8). A dilution factor is thereby obtained by dividing the serum urea concentration by the BALF concentration. ELF concentrations are then calculated by multiplying the BALF concentration by the dilution factor. Urea concentrations in the samples were determined with a commercially available reagent (Sigma Diagnostics, St. Louis, MO).

Lung GR activity.

To determine GR activity in the lung, 10–25 mg of ground lung tissue were dissolved in 800 μl of cold homogenization buffer (50 mM potassium phosphate and 1 mM EDTA, pH 7.5) and centrifuged (8,500g for 10 min at 4°C), and the supernatant was retained for analysis. GR activity in the lung homogenate was determined spectrophotometrically (340 nm) from the rate of NADPH consumption by GR in the reduction of GSSG with a commercially available kit (Oxis International). GR activity is expressed as units per milligram of sample protein (Coomassie Plus, Pierce, Rockford, IL). The pellets from these homogenates were utilized for the determination of γ-GT activity (see Lung γ-GT activity).

Lung GPx activity.

To determine GPx activity in the lung, 10–35 mg of ground lung tissue were dissolved in 1.0 ml of cold homogenization buffer (50 mM Tris · HCl, 5 mM EDTA, and 1 mM 2-mercaptoethanol, pH 7.5) and centrifuged (7,500 g for 15 min at 4°C), and the supernatant was retained for analysis. The GPx activity in the homogenate was determined with a commercially available kit (Oxis International) to which t-butyl hydroperoxide was added as a GPx substrate to generate GSSG. The rate of NADPH consumption by GR in the subsequent reduction of GSSG was used to calculate GPx activity. GPx activity is expressed as units per milligram of sample protein.

Lung γ-GT activity.

To determine lung γ-GT activity, the pellets obtained from the homogenates for GR analysis were resuspended in homogenization buffer (100 mM Tris · HCl, 10 mM serine, and 0.1 mM EDTA, pH 7.6) and centrifuged (5,500 g for 10 min at 4°C), and the supernatant was retained for γ-GT analysis. A 200-μl aliquot of the supernatant was mixed with 1.0 ml of reagent solution (3.2 mM γ-glutamyl-3-carboxy-4-nitroanilide, 110 mM glycine-glycine, and 110 mM Tris · HCl, pH 8.3). γ-GT activity was calculated from the rate of 3-carboxy-4-nitroaniline production, which absorbs at 405 nm (11, 37). γ-GT activity was normalized to the supernatant protein concentration (Coomassie Plus, Pierce).


Oxidation of tissue lipids produces various aldehydes that can be measured colorimetrically by their reaction with thiobarbituric acid (29). Approximately 25 mg of ground lung tissue were dissolved in 50 mM phosphate buffer (pH 7.4) containing 1 mM butylated hydroxytoluene. An aliquot of the sample was then acidified with an equal volume of phosphoric acid. Thiobarbituric acid (0.1 M) was added, and the mixture was heated to 90°C for 45 min. The chromogen in the sample was extracted with n-butanol, and the absorbance at 535 nm was measured with a plate reader (SpectraMax 340PC, Molecular Devices, Sunnyvale, CA). TBARS in the samples were calculated from a standard curve and normalized for protein content.

HPLC analysis for 8-OHdG in lung DNA.

DNA from mouse lung tissue was obtained by a chloroform-isoamyl alcohol extraction of proteinase K-digested lung homogenates (38). The purified DNA was then hydrolyzed to nucleosides with nuclease P1 and alkaline phosphatase. Samples were analyzed for 8-OHdG and 2′-deoxyguanosine (2-dG) by HPLC coupled with coulometric electrochemical and ultraviolet detection (CoulArray model 5600, ESA, Chelmford, MA), respectively (38). Sample analysis was done with a 4.6 × 150-mm, C-18 reverse-phase column (YMCbasic, YMC, Wilmington, NC) with a mobile phase of 100 mM sodium acetate in 5% methanol at pH 5.2 (38). 2-dG was detected by ultraviolet light at 265 nm, whereas 8-OHdG was detected electrochemically with electrode potentials of 285, 365 and 435 mV. Under these conditions, 2-dG and 8-OHdG had retention times of ∼7.4 and 9.5 min, respectively. Nucleoside concentrations were calculated from standard curves generated daily with freshly prepared standards.

Statistical analysis.

Data are presented as means ± SE. Unless noted otherwise, each experimental group contained five samples, with each sample measured in duplicate. Data were subsequently analyzed for significant differences with t-tests, with significance attained whenP ≤ 0.05.


GSH concentrations in the ELF and lung tissue.

The ELF GSH and GSSG concentrations in WT and CFTR-KO mice were calculated from their respective concentrations in BALF (Fig.1 A). Because BALF is a manyfold dilution of the actual ELF, serum and BALF urea concentrations were measured to determine the ELF dilution factor. No differences in BALF or serum urea concentrations between the WT and CFTR-KO mice were observed (data not shown). ELF GSH concentration in the WT mice was 512 ± 63 μM and in agreement with previously published data (7). The CFTR-KO mice demonstrated a significant decrease (51%) in ELF GSH, with a mean concentration of 244 ± 59 μM (Fig. 1 A). There was also a significant decrease (60%) in the ELF GSSG concentration in the CFTR-KO mice (Fig. 1 A). In contrast to the decrease in ELF GSH concentration, the concentration of GSH in the lungs of CFTR-KO and WT mice was not significantly different (Fig. 2 A). Although the CFTR-KO mice were maintained on a liquid diet, this diet did not adversely affect the GSH concentration in the ELF of WT mice compared with that in WT mice maintained on the regular solid diet (data not shown).

Fig. 1.

Reduced (GSH) and oxidized glutathione (GSSG) concentrations (A) and GSH-to-GSSG (GSH/GSSG) ratios (B) in epithelial lining fluid (ELF) of wild-type (WT) and cystic fibrosis transmembrane conductance regulator knockout (CFTR-KO) mice (n ≥ 5). ELF concentrations in WT and CFTR-KO mice were 512 ± 63 and 249 ± 59 μM, respectively, for GSH and 106 ± 15 and 41 ± 18 μM, respectively, for GSSG. Both GSH and GSSG concentrations were significantly decreased in CFTR-KO mice (* P = 0.015 and 0.024, respectively). GSH/GSSG ratios in WT and CFTR-KO mice (8.6 ± 1.7 and 13.4 ± 2.6, respectively; n ≥ 8) were not significantly different (P = 0.12).

Fig. 2.

A: concentration of GSH in the lung tissue normalized to sample weight. Although there was a small increase in the GSH content in CFTR-KO mouse lungs compared with WT lungs (0.31 ± 0.03 and 0.38 ± 0.03 nmol/mg lung, respectively;n ≥ 6), the increase was not significant (P = 0.13). B: GSH/GSSG ratios in the WT (22.1 ± 3.9) and CFTR-KO (25.8 ± 1.8) mouse lungs (n ≥ 6) were also not significantly different (P = 0.46).

Activities of lung antioxidant enzymes.

The activities of three GSH-utilizing enzymes in lung tissue of WT and CFTR-KO mice were compared. No differences in the lung protein concentrations were observed in the lung homogenates from WT and CFTR-KO mice (data not shown). The activity of GR, the intracellular enzyme responsible for reducing GSSG to GSH, was significantly increased in CFTR-KO mouse lungs (Fig.3). GR activity increased from 2.5 ± 0.2 U/mg protein in WT mice to 3.8 ± 0.3 U/mg protein in CFTR-KO mice. GPx, the intracellular enzyme that utilizes GSH to reduce peroxides, was also significantly elevated CFTR-KO mouse lungs. GPx activity increased from 338 ± 20 U/mg protein in WT mouse lungs to 431 ± 28 U/mg protein in CFTR-KO mouse lungs (Fig.4). The process of extracellular catabolism of GSH for subsequent cellular processes, including GSH synthesis, is mediated by γ-GT, which cleaves GSH to yield the γ-Glu and Cys-Gly moieties (33). In the CFTR-KO mouse lung, γ-GT was not significantly altered over that in WT lungs (Fig.5).

Fig. 3.

GSH reductase (GR) activity in the lung tissue. Activity of GR in lung tissue was normalized to sample protein concentrations. CFTR-KO mice had significantly greater lung GR activity than WT mice (3.76 ± 0.27 and 2.54 ± 0.19 U/mg protein, respectively; * P = 0.001; n ≥ 12). No differences in lung protein concentrations were observed in these samples (data not shown).

Fig. 4.

GSH peroxidase (GPx) activity in the lung tissue. Activity of GPx in lung tissue was normalized to sample protein concentrations. CFTR-KO mice had significantly more GPx activity in the lung tissue than WT mice (431 ± 28 and 338 ± 20 U/mg protein, respectively; * P = 0.012;n ≥ 10). No differences in lung protein concentrations were observed in these samples (data not shown).

Fig. 5.

γ-Glutamyltransferase (γ-GT) activity in lung tissue. Activity of γ-GT in lung tissue was normalized to sample protein concentration. Lung γ-GT activity in the CFTR-KO mice was not significantly increased compared with that in WT mice (P = 0.159). No differences in lung protein concentrations were observed in these samples (data not shown).

Markers of oxidative stress in the ELF.

To determine whether there was oxidative stress in the ELF compartment, TBARS and GSH-to-GSSG ratios in the BALF were determined. Not surprisingly, TBARS were below the detection limits of the spectrophotometric assay in both the WT and CFTR-KO mice (data not shown). A change in the GSH-to-GSSG ratio can often be an indicator of oxidative stress (2, 28). Because the decreases in GSH and GSSG were relatively proportional between the WT and CFTR-KO mice (Fig.1 A), the GSH-to-GSSG ratios did not differ significantly and suggested the absence of oxidative stress in the ELF of CFTR-KO mice (Fig. 1 B).

Markers of oxidative stress in lung tissue.

Increases in antioxidants and antioxidant enzymes, such as those shown in Markers of oxidative stress in the ELF, often occur in response to an oxidative stress (30, 32, 33, 36, 41). Consequently, mouse lung tissues from CFTR-KO mice were analyzed for more direct indications of oxidative stress. Tissue GSH-to-GSSG ratios were not significantly different between the mice (Fig. 2 B). More direct indicators of oxidative injury, namely TBARS and 8-OHdG, were significantly elevated in CFTR-KO mouse lungs. The concentration of TBARS, a marker for lipid oxidation, was increased over 25% in the lung tissue of the CFTR-KO mice (Fig.6 A). Levels of 8-OHdG, formed from the oxidation of 2-dG in DNA, were also significantly increased in CFTR-KO mouse lungs (Fig. 6 B).

Fig. 6.

Concentration of thiobarbituric acid-reactive substances (TBARS; A) and 8-hydroxy-2′-deoxyguanosine (8-OHdG;B) in lung tissue of CFTR-KO and WT mice. A: level of TBARS was significantly elevated in the lung tissue of CFTR-KO mice compared with WT mice (100.0 ± 4.0 and 126.6 ± 8.0%, respectively; * P = 0.0251). B: ratio of 8-OHdG to every 105 2′-deoxyguanosine (2-dG) in DNA isolated from lung tissue. Levels of 8-OHdG were significantly increased in DNA from CFTR-KO lungs compared with those in WT lungs (5.67 ± 0.94 and 3.72 ± 0.37 8-OHdG per 1052-dG, respectively; * P = 0.0459).


This study demonstrates that a mutation in the CFTR gene decreases the concentration of GSH in the ELF, increases the activities of lung antioxidant enzymes, and possibly stimulates oxidative stress. CFTR-KO mice demonstrated a significant decrease in ELF GSH compared with WT mice. Although the lung tissue did not exhibit any alterations in GSH concentration, the CFTR-KO mice exhibited significant increases in GR and GPx activities. Elevated activities of such antioxidant enzymes are often an adaptive response to oxidative stress (30, 32, 33, 36,41). Increased oxidation of lipids and DNA in the lungs of CFTR-KO mice also suggests the presence of an oxidative stress.

Although studies (18, 34, 35) have documented that ELF concentrations of GSH are considerably lower in normal than in adult CF patients, a direct link to CFTR has been difficult to establish. Even in the absence of overt clinical indications of an infection, many CF patients exhibit evidence of an underlying inflammatory process occurring in their lungs (1, 19). This underlying inflammation has limited the use of these data to support a hypothesis that CFTR was directly involved in the transport or regulation of GSH in the ELF. However, two recent in vitro studies (16, 25) have provided strong evidence that CFTR is associated with the transport of GSH. In the first study, Linsdell and Hanrahan (25) demonstrated that in addition to Cl and other anions, GSH could permeate the cell membrane through CFTR in vitro. The subsequent in vitro study (16) used airway cells isolated from a CF patient and demonstrated that cells lacking functional CFTR had significantly less GSH in their apical fluid than normal cells or cells transfected with functional CFTR. Taken together, these two studies provide substantial support of the hypothesis that CFTR directly regulates the concentration of ELF GSH and may subsequently render the CF lung more susceptible to oxidative injury.

Based on this potential link between CFTR and the transport of a critical antioxidant into the ELF compartment, we used CFTR-KO mice to test the hypothesis that the CFTR mutation alters antioxidant defenses in the lung. The CFTR-KO mouse used in this study (B6.129P2-Cftrtm1Unc ) possesses a mutation that renders the CFTR protein nonfunctional (9, 40, 43, 44). The decreased concentration of GSH in the ELF of CFTR-KO mice indicates that this defect in CFTR alters GSH availability in the ELF. Thus the previously reported decrements in the concentration of GSH in BALF may not only be due to the infection-induced oxidative stress but may also be a direct result of impaired GSH transport into the ELF compartment (25, 35).

The pancreatic dysfunction associated with CF leads to inadequate breakdown and absorption of the fat-soluble nutrients such as vitamin E, β-carotene, and selenium and may contribute to the antioxidant imbalances observed in CF (30). In addition to the pulmonary manifestations, CF patients also suffer from gastrointestinal obstructions and malabsorption of nutrients (10, 42). To minimize these problems in the CFTR-KO mice, the mice were housed on inedible bedding and maintained on a liquid diet (14). WT mice maintained on either the regular solid diet or the liquid diet did not exhibit significant differences in ELF GSH concentration. In addition, CFTR-KO mice with gut-corrected CFTR (46) also have a decreased concentration of GSH in their ELF (data not shown). Because there is little effect of diet in WT mice, the observed difference between the WT and CFTR-KO mice is most likely due to the lack of functional CFTR protein in the knockout mice and not attributed to the liquid diet on which the mice were maintained.

As a consequence of the reduced ELF concentration of GSH, the CF lung may be more vulnerable to oxidative stress that results from an infectious agent (35). After invasion of an infectious agent, there is an influx of neutrophils that generate bactericidal oxidants such as superoxide, hydrogen peroxide, and hypochlorous acid. Extracellular antioxidants such as GSH can protect the underlying epithelial cells from these oxidative processes (34, 35,42). Factors contributing to an imbalance between oxidant production and antioxidant defense yield a net oxidative stress, which can cause epithelial injury concomitant to the bacterial killing. Due to epithelial damage and release of intracellular contents, the inflammatory response could be further stimulated and exacerbate the condition (30). Compared with WT mice, CFTR-KO mice infected with Pseudomonas have higher mortality rates and markedly elevated levels of inflammatory mediators (44). In CF patients, lung infections elicit an intense inflammatory response characterized by an influx of neutrophils and the secretion of cytokines (21). Clearing a pulmonary infection is further complicated in CF patients by the thickened characteristics of the ELF mucus (39). GSH is the predominant mucolytic agent in the ELF; its deficiency, due to lack of functional CFTR, probably contributes to thickened ELF mucus in CF patients. The continued presence of the bacteria to drive inflammation coupled with decreased protection by antioxidants favors destruction of tissue. Thus the inherent GSH deficiency caused by impaired epithelial transport coupled with insufficient dietary absorption of vital antioxidants may seriously predispose the CF lung to oxidative injury (34, 35,42).

The lower concentration of GSH in the ELF of CFTR-KO mice is not due, however, to an inherent oxidative stress present in the ELF. A change in the ratio of GSH to GSSG is often used as an indicator of oxidative stress (2, 28). In this study, the GSSG concentration in CFTR-KO mouse lungs was significantly decreased. The decrease, however, was nearly proportional to the decrease in GSH and did not significantly alter the GSH-to-GSSG ratio from that of the WT mice. Calculating a GSH-to-GSSG ratio from ELF GSH and GSSG concentrations reported in a previous study with rats (5) produces a ratio of 8.6; this is exactly the same value reported here for the WT mice. Therefore, the decreased ELF concentration of GSH in the CFTR-KO mice is not likely due to increased oxidation but rather is a direct result of impaired epithelial transport.

In contrast to the decreased antioxidant capacity of the ELF, the lung tissue demonstrated increases in the antioxidant enzymes GR and GPx. An increase in these enzymes usually occurs as an adaptive response to increased oxidative stress. GR functions to recycle GSH by using NADPH to reduce GSSG back to GSH. Increases in GR activity have been demonstrated under conditions of oxidative stress in many tissues including the lung (24). Under oxidative stress, upregulation of GR can maintain GSH-to-GSSG ratios as an adaptive response. GPx is another enzyme upregulated by oxidative stress that functions to eliminate oxidizing peroxides, including H2O2 and lipid peroxides (3, 24). Together, these enzymes can neutralize injurious oxidants with GSH and replenish GSH through increased GSSG reduction. The upregulation of this enzyme system is indicative of the presence of an increased oxidant burden in the CFTR-KO lung.

γ-GT is an enzyme found on the luminal surfaces of Clara cells and, to a lesser extent, type II cells of the alveolar epithelium (12). γ-GT in conjunction with a dipeptidase serves to recycle GSH by cleaving it into its three amino acids that are readily transported to the intracellular compartment (17). Oxidative stress produces an increase in γ-GT activity in both in vitro and in vivo models (17, 22). Consequently, increases in γ-GT activity are regarded as an indication of oxidative stress. Although the increase in lung γ-GT activity observed in CFTR-KO mice was not significant, the increased activity of γ-GT is not inconsistent with the GR and GPx data that suggest the presence of an oxidative stress.

More direct markers of oxidative injury provide evidence of an inherent oxidative stress in the CFTR-KO mice. In this study, TBARS and 8-OHdG were utilized as markers of oxidative stress in lung tissue. TBARS in CFTR-KO mouse lung tissue were significantly increased compared with those in WT mouse lungs. This result is consistent with a previous study (31) that found increases in plasma TBARS in children with CF. It could be argued, however, that the elevated TBARS levels could be a direct result of ongoing infections or dietary effects. Another study with CF patients without clinical signs of exacerbations, however, demonstrated increases in plasma TBARS, hydroperoxides, and protein carbonyls and erythrocyte GR and superoxide dismutase activities (13). These studies clearly indicate the existence of an oxidative stress in CF and provide support for the increased TBARS observed here. Although the previous studies indicate the presence of an oxidative stress, the fact that the results were obtained from plasma limits identification of the site of the oxidative events. The results reported here, however, extend these previous observations and identify the lung as a potential site of oxidative events. A more sensitive indicator of oxidative stress is the detection of oxidized bases, particularly 8-OHdG, in DNA (15, 38). A previous study (6) demonstrating elevated levels of 8-OHdG in the urine of CF patients also suggests the presence of an oxidative stress. Moreover, no correlation between clinical indicators of lung function and increases in 8-OHdG was identified, which suggests that CF patients have an inherent oxidative stress independent of clinical causes such as infection. In our study, a small but significant increase in 8-OHdG was detected in DNA from CFTR-KO mouse lungs. These data suggest that a mild oxidative stress is occurring in the lung tissue and provide a plausible explanation for the elevation in lung GR and GPx in CFTR-KO mice.

Although the elevated levels of TBARS and 8-OHdG demonstrate the presence of an oxidative stress in the lung tissue of CFTR-KO mice, the GSH-to-GSSG ratio remains unchanged. The GSH-to-GSSG ratio in the CFTR-KO lungs was not significantly different from WT lungs and was only slightly greater than the ratios recently reported (GSH/GSSG ≈ 19) for rat lungs (23). The lack of any change in this parameter suggests the absence of oxidative stress and is in contradiction to the other markers. The most plausible explanation for this divergence is that the increased GR activity compensated for any elevations in GSSG formation and thereby maintained normal GSH-to-GSSG ratios. Thus the oxidative stress indicated by TBARS and 8-OHdG may alter the GSH-to-GSSG ratio by increasing GSSG production, but the increased GR activity negates its detection.

The previous studies demonstrating the presence of an oxidative stress in the absence of clinical indicators of infection suggest that oxidative stress may be inherent to CF. The data presented here also demonstrate the presence of an oxidative stress; however, effects from infection and diet are largely minimized and provide more convincing evidence that a mild inherent lung oxidative stress occurs in CF. This mild oxidative stress itself produces little lung pathology but may sensitize the lung to overtly respond to stimuli such as infections. Although the data presented here provide evidence of oxidative stress in the lung, they do not identify the source of oxidative stress. It is well known that mitochondrial electron transport can be a significant source of cellular oxidants such as superoxide and hydrogen peroxide under both normal and pathological conditions. CF patients demonstrated increased oxygen consumption and increased energy requirements consistent with mitochondrial dysfunction (45). In that study, cells isolated from CF patients demonstrate increased oxygen consumption and activity of NADH oxidase and NADH dehydrogenase in mitochondria. These data provide further evidence that events in the mitochondria may contribute to injurious oxidative events. Although the increased oxidation of 2-dG demonstrated in this study suggests that an oxidative stress is occurring within the cell, it is unclear whether the oxidized DNA is nuclear, mitochondrial, or both. Studies currently underway in our laboratory will delineate the source of the 8-OHdG.

The exact mechanism by which CFTR may mediate oxidative stress in the mitochondria is unclear. Because mitochondria lack GSH-synthesizing enzymes, all intramitochondrial GSH must be transported from the cytosol. Currently, the mechanism by which GSH is transported into the mitochondria has not been identified. Consequently, it is possible that in addition to apical transport of GSH, the CFTR protein may also facilitate GSH transport into the mitochondria directly or through the modulation of proteins involved in its transport. If this is true, then an imbalance in mitochondrial GSH may be present and provide a link between the CFTR protein and mitochondrial oxidative stress.

The decreased concentration of ELF GSH observed in CF patients has led some investigators to propose that restoring GSH concentrations in this compartment may reduce tissue destruction (34). In this study, application of aerosolized GSH increased the BALF concentration of GSH and reduced phorbol 12-myristate 13-acetate-stimulated superoxide release in macrophages from CF patients. In addition to reducing oxidant-mediated epithelial injury, increased ELF GSH concentrations may also limit tissue destruction by preserving the antiproteases that are present in the ELF (4, 21). Given the likelihood of developing severe pulmonary infections, augmentation of the ELF GSH pool is likely to be critical to improving the lifespan of CF patients. The results of this study demonstrate, in an animal model defective in the CFTR protein, decreased GSH concentrations in the ELF and increased oxidation of lung lipids and DNA. These data support the hypothesis that CFTR regulates lung antioxidants and, when defective, may contribute to oxidative stress associated with CF.


We thank Merle Fleischer, Mark Goldstein, Lisa Hogue, Todd Romigh, and Christiaan van Heeckeren for expert technical assistance in these studies and Drs. Frank Accurso and Pamela B. Davis for helpful suggestions. We also thank Tanya Canafax for secretarial support.


  • This work was supported in part by National Heart, Lung, and Blood Institute Grants HL-59602 and HL-31992 (to B. J. Day); National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-27651 (to A. V. Heeckeren); and research and development grants from the Cystic Fibrosis Foundation (to A. V. Heeckeren and B. J. Day).

  • Original submission in response to a special call for papers on “CFTR Trafficking and Signaling in Respiratory Epithelium.”

  • Address for reprint requests and other correspondence: B. J. Day, National Jewish Medical and Research Center, 1400 Jackson St., Rm. K-706, Denver, CO 80206 (E-mail: dayb{at}

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract