INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: asthma; glutaminase; ammonia; lung; breath condensate

Several features of asthmaincluding cough, epithelial cell damage, eosinophil necrosis, impaired ciliary motility, abnormal mucous production, and increased nitric oxide generationmay be initiated by exogenous exposure of the airway epithelium to acid in vitro and in vivo (1). We have recently reported that the pH of the exhaled water of patients during acute exacerbations of asthma is over two log orders lower than that of control subjects (3), raising the possibility that endogenous airway acids could contribute to asthma pathophysiology. Several cellular and inorganic processes may form and release acids in the airway, particularly in the setting of inflammation (6). However, the mechanisms by which this acid is buffered to maintain airway pH homeostasis are incompletely characterized.

The proximal renal tubule responds to an acid load by upregulation of the kidney isoform of glutaminase (KGA), an enzyme that buffers urinary acid by net formation of the base, ammonia* (pK_a = 9.3) from glutamine (10, 11). A splice-variant isoform of glutaminase, termed GAC, has been identified in human lung homogenates (12), but the cellular source is not known. We hypothesized that glutaminase may be expressed in the airway epithelium and have a role in pH homeostasis. To test this hypothesis, we studied the expression and activity of glutaminase both in human airway epithelial cells in culture and in humans with and without asthma. Here, we report that (1) glutaminase is transcribed in normal human bronchial epithelial cells (NHBE), small airway epithelial cells (SAEC), and the Type II alveolar cell line A549; (2) airway epithelial cells produce ammonia stoichiometrically from glutamine; (3) these cells respond to acidic challenge by increasing production of ammonia, which neutralizes acid and enhances survival; (4) interferon (IFN)- and tumor necrosis factor (TNF)- inhibit this epithelial ammonia production; (5) patients with acute exacerbations of asthma have substantially less ammonia in their exhaled breath condensates than do control subjects; and (6) glutaminase protein is expressed in the human airway epithelium in vivo, and expression may be decreased in asthma. Because acidification and ammonia deprivation have antimicrobial effects (13, 14), we speculate that inhibition of glutaminase in the airway epithelium may be an important lung defense mechanism. In subjects with asthma, however, this mechanism may contribute to the pathophysiology of airway injury.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Human Airway Epithelial Cell Culture

A549 (ATCC, Manassas, VA), normal human bronchial epithelial cells (NHBE, Clonetics, Walkersville, MD), and small airway epithelial cells (SAEC, Clonetics) were incubated in bicarbonate-buffered Ham's F12K or bicarbonate-free CMRL media or in modified bicarbonate-buffered glutamine-controlled SAGM medium (Clonetics) and studied at confluence prior to third passage. Selected wells were incubated with TNF- and IFN- (100 units/ml each) and/or the glutaminase competitive antagonist 5-diazo-6-oxo-L-norleucine (DON, 2-4 mM) at various pH values and concentrations of glutamine or glutamate (0.5-10 mM). Assays of cell viability were performed by direct addition of Trypan Blue to cell wells.

Chemical Methods

Unless otherwise noted, reagents were obtained from Sigma (St. Louis, MO). Ammonia concentration was determined spectrophotometrically (15). pH was measured with a Corning microelectrode (Corning, NY) attached to an Orion 520A pH Meter (Orion, Beverly, MA). Deaeration/decarbonation of breath condensate specimens was achieved by bubbling with argon (350 ml/min) as previously reported (3).

Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) for Glutaminase mRNA

Total RNA was isolated from cells following the RNeasy protocol (Qiagen, Valencia, CA). Aliquots were mixed with KGA- or GAC-specific nested primers constructed according to Elgadi and coworkers (12). The PCR products were analyzed by gel electrophoresis.

Immunohistochemistry

Formalin-fixed sections (3-5 µm) from open lung biopsy specimens were blocked and incubated (4° C) with rabbit polyclonal antibody to rat glutaminase (graciously supplied by Dr. Norman P. Curthoys, Department of Biochemistry, Colorado State University) (16). Glutaminase isoforms are highly conserved between rat and human (12), and the specificity of this antibody had been previously described (17). After washing, sections were incubated with or without goat anti-rabbit biotinylated secondary antibody, followed by incubation with avidin-biotinylated horseradish peroxidase macromolecular complex solution. Glutaminase staining was visualized using 3,3'-diaminobenzidine substrate (Sigma).

Precipitation of Exhaled Breath Fluid

Nonsmoking, otherwise healthy human subjects with acute exacerbations of asthma were recruited from the clinics and inpatient wards. Subjects performed quiet tidal breathing for 5 min, exhaling into a plastic RTube disposable breath condensation collection device (Respiratory Research, Inc., Charlottesville, VA). Approximately 1 ml of fluid was collected during the procedure. Collected samples were frozen at 4 to 20° C. This study was reviewed and approved by the institutional human investigation committee, and patients provided informed consent.

Western Blotting

Proteins collected from epithelial cell lysates were separated by SDS- PAGE, transferred to a nitrocellulose membrane, and probed with rabbit polyclonal anti-rat glutaminase antibody. Glutaminase was visualized after addition of horseradish peroxidase-conjugated goat anti-rabbit antibody with subsequent enhanced chemiluminescence (Amersham).

Statistics

Data are presented as median and range or, if parametrically distributed, as arithmetic or geometric (pH) mean ± standard error of the mean. Differences between subjects with asthma and control subjects and the subgroups of treated incubated cells were analyzed by repeated measures ANOVA or ANOVA on ranks with appropriate pairwise comparisons (18). Differences were considered significant at p values < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airway Epithelial Cells Transcribe Message for Two Isoforms of Glutaminase

RT-PCR was performed using two sets of nested primers for GAC and KGA on RNA isolated from NHBE, SAEC, and A549 cells. Agarose gel electrophoresis of the resulting cDNA product revealed the presence of appropriately located bands at 480 bp (both the GAC and nested KGA cDNA product), 640 bp (KGA), and 260 bp (GAC nested) (Figure 1). Nonspecific bands were not detected.

View larger version (141K): [in this window] [in a new window]   Figure 1.   Airway epithelial cells transcribe glutaminase message. NHBE cells were harvested and RNA isolated. RT-PCR was performed using two sets of nested primers for GAC and KGA. The band at 680 (Lane 5) represents the cDNA product of KGA-specific primers. Bands at 464 are the RT-PCR product of nested KGA primers (Lane 3), and separately, GAC primers (Lane 4). Nested primers for GAC yielded the expected 248-bp product (Lane 2). Nonspecific bands were not detected. There was no evidence of DNA contamination of isolated RNA, as evidenced by lack of any bands in one lane (Lane 7) with primers but no reverse transcriptase. Lane 1 is a 100-bp DNA ladder. Lane 6 is the RT-PCR product of GAPDH-specific primers. Data are representative of three independent experiments and are identical with results from SAEC cells and A549 cells.

Airway Epithelial Cells in Vitro Convert Glutamine to Ammonia

All epithelial cell lines (A549, NHBE, and SAEC) produced ammonia stoichiometrically from glutamine (Figure 2A-2C). Addition of the glutaminase antagonist DON caused a pronounced decrease in ammonia production (p < 0.001, n = 3) in all three cell types. Similarly, replacement of glutamine with glutamate caused a prominent decrease in ammoniagenesis (Figure 2D). In contrast, medium without cells (n = 5, each medium type, each glutamine concentration) produced very low levels of ammonia from spontaneous decomposition of glutamine (data not shown).

View larger version (31K): [in this window] [in a new window]   Figure 2.   Ammoniagenesis by airway epithelial cells is dependent on glutamine availability. Confluent A549 cells (A) in Ham's F12K medium, (B) in bicarbonate-free CMRL medium, and (C ) SAEC cells and NHBE cells (not shown) in modified SAGM medium in six-well culture plates were incubated with various concentrations of glutamine at initial pH of 7.8-8.0. Supernatants were assayed every 24 h for ammonia. Glutamine-dependent ammonia production was evident in all cell lines. (D) Addition of 4 mM of the glutaminase competitive antagonist DON resulted in 42% reduction of A549 ammonia production when incubated with 4 mM glutamine (p < 0.001). Replacement of GLN with glutamate prevented ammonia production. Addition of 3 mM DON resulted in a prominent decrease of ammonia production in NHBE and SAEC cells incubated with 3 mM glutamine (data not shown). Control medium without cells produced only trace amounts of ammonia. Results presented are means of three to six concurrent experiments for each cell type, and are representative of multiple separate experiments. Data are presented as mean ± SE. Except time 0, all data points plotted represent significantly different means.

Glutaminase Expression Is Upregulated by Acidic Media

A549 cells incubated in CMRL media supplemented with 3 mM glutamine and challenged with initial pH 5.6 revealed 25-96% increased levels of glutaminase proteins compared with control cells incubated at normal pH, as quantified by densitometry after Western blot with rabbit anti-rat glutaminase antiserum (Figure 3). Three bands were visualizedconsistent with the results of previous investigators revealing limited proteolytic digestion of two splice variant isoforms of glutaminase found in the lung.

View larger version (54K): [in this window] [in a new window]   Figure 3.   Glutaminase protein is enhanced by low pH media. Western blotting for glutaminase reveals approximately 25-96% increase in glutaminase after > 48 h in cells treated with initial media at pH 5.6 compared with normally treated cells. The three bands represent partial proteolytic digests of two isoforms of glutaminase (KGA and GAC). Lane 1 represents an equal amount of protein from homogenized rat kidney, revealing the rat KGA isoform.

Ammonia Production by Airway Epithelial Cells in Vitro Is Upregulated by Acidic Stress, Neutralizes Acidic Cell Culture Media, and Enhances Cell Survival

A549 cells in bicarbonate-free CMRL medium at initial pH values of 5.8 were supplemented with 0, 4, and 10 mM glutamine; pH and ammonia production were monitored for 72 h. Control wells without cells produced minimal ammonia, with no change in pH. The medium supernatant from cells incubated at initially acidic pH was partially or completely neutralized depending on the initial concentration of glutamine and resultant ammonia production (n = 3, p < 0.01) (Figure 4A). In general, cells survived poorly at low pH, with cell death evidenced by both inability to exclude trypan blue and sloughing of the cells from the plating surface. However, the cells in some wells incubated at low pH responded to the stress by a pronounced upregulation of ammonia production (n = 3 replicates of two independent experiments, p < 0.001 compared with normal pH) (Figure 4B) and concomitant neutralization of pH (p < 0.01) (Figure 4C). The cell populations that were able to accomplish this upregulation also had enhanced survival at 72 h (97 ± 0.33%)similar to control wells at neutral pH (97 ± 0.6%)when compared with nonresponding cells (51 ± 2.4%; p < 0.001) (Figure 4D).

View larger version (25K): [in this window] [in a new window]   Figure 4.   Ammonia production alkalinizes cell medium supernatant. A549 cells were incubated in a CO2-free atmosphere in bicarbonate-free CMRL medium supplemented with 0, 4, or 10 mM glutamine at initial pH of 5.8. Control wells without cells produced minimal ammonia, without pH alteration over time (not shown). (A) The medium of cells incubated initially at pH 5.8 gradually neutralized, with rate controlled by availability of glutamine. (B) During some experiments, acid-stressed cells responded with an accelerated increase in ammonia production (n = 6 responding wells, 6 non- responders) that resulted in (C ) more rapid neutralization of pH and (D) enhanced survival when compared with non-responders (p < 0.01).

Similar experiments were performed in 20 mM bicarbonate-containing Ham's F12K medium. This media required preincubation and pH adjustment in a 5% CO₂ atmosphere before use. In the absence of cells, bicarbonate-buffered medium acidified with HCl to pH 5.4 spontaneously increased pH (pH = 6.46 ± 0 at 48 h) while producing minimal ammonia (< 150 µM). The presence of A549 cells in medium at initial pH 5.4 greatly augmented both ammonia production at 48 h (2,747 ± 43 µM, p < 0.001 compared with medium, n = 3) and speed of neutralization of pH (7.5 ± 0.03 at 48 h; p < 0.001). In comparison, cells treated initially with pH 7.8 produced much less ammonia (1,585 ± 50 µM; p < 0.001 compared with pH 5.4 treated cells, n = 3) while alkalinizing the media more modestly (pH = 8.5 ± 0.0 at 48 h; n = 3). The ability to further alkalinize medium starting from above neutral is present only with the bicarbonate-buffered medium consistent with the hypothesized roles of HCO₃ and ammonia in airway epithelial pH control.

Glutaminase Activity in Airway Epithelial Cells Is Downregulated by IFN- and TNF-

SAEC cells at confluence in six-well culture clusters in SAGM media (pH 7.5) supplemented with 3 mM glutamine were treated with 100 units/ml of IFN- and 100 units/ml of TNF-. Ammonia production by cytokine-treated cells was significantly less than that by untreated controls at 24, 48, and 72 h. For example, at 48 h, treated cells produced 371 ± 52 nmol of ammonia compared with control cells that produced 846 ± 49 nmols (n = 3 replicates each; p < 0.001) (Figure 5).

View larger version (18K): [in this window] [in a new window]   Figure 5.   IFN- and TNF- downregulate ammonia production in vitro. SAEC cells were grown in supplemented SAGM medium with 3 mM GLN in the presence and absence of 100 units/ml of both IFN- and TNF-. Cytokines significantly decreased SAEC ammonia production at all time points (n = 3 each, p < 0.001).

Ammonia Levels Are Low in Fluid Condensed from the Expired Air of Subjects with Acute Exacerbations of Asthma

Exhaled breath condensate was collected from 32 human subjects, and ammonia was assayed as noted in METHODS. Median breath condensate ammonia concentration was 10-fold lower in specimens from acutely ill subjects with asthma (30 µM [range: 0-233], n = 18, age 23 ± 2.5 yr) than from control subjects (327 µM [14-1,220], n = 24, age 24 ± 2.4 yr, p < 0.001) with a positive logarithmic relationship to same sample pH measurements (r = 0.58, p < 0.001) (Figure 6A).

View larger version (12K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 6.   Ammonia concentration in exhaled breath condensate of patients with acute asthma is low. Exhaled breath condensate was collected from patients admitted to the hospital or seen in the emergency department with acute exacerbations of asthma or from healthy volunteers. (A) The median ammonia concentration in subjects with asthma (30; range, 0-233; n = 18, age 23 ± 2.5 yr) was 10-fold lower than in control subjects (327; range, 14-1220; age 24 ± 2.4 yr) by Mann-Whitney rank-sum testing (p < 0.001). (B) Ammonia concentrations were correlated with logarithmically transformed H+ concentrations (pH) (r = 0.55; p < 0.001). Though loss of ammonia production is permissive for acidification, it is clearly not the only determinant.

Immunohistochemistry

Open lung biopsy specimens from a subject with severe asthma (kindly provided by Dr. Serpil Erzurum) were compared with histologically normal lung tissue obtained from a steroid-naive subject without asthma and a subject who had been treated with systemic corticosteroids. Tissues were immunostained as described in METHODS. There was strong immunoreactivity for glutaminase in all specimens, particularly near the apical surface of the epithelium. Epithelial staining was dramatically less, however, in the asthmatic tissue. Primary and secondary antibodies alone revealed no staining in any specimen (Figure 7).

View larger version (130K): [in this window] [in a new window]   Figure 7.   Glutaminase protein is expressed in human airway epithelial cells and is upregulated by corticosteroids in vivo. Open lung biopsy specimens from a subject with a severe asthma exacerbation (A) (provided courtesy of Dr. S. Erzurum) are compared with those from a subject without asthma with a histologically normal lung (B) and a subject without asthma treated with systemic corticosteroids (C ). There was more glutaminase immunoreactivity (brown) in the airway epithelium of the control subject (B) than of the subject with asthma (A). Glutaminase staining is further enhanced in the corticosteroid-treated patient (C ). (D) The secondary antibody in the absence of primary antibody had no reactivity; the same was true using primary antibody and no secondary antibody for all tissues (not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Substantial lung injury may result from exposure of the airway to exogenous acid (1, 2, 4, 19), in part because the cytotoxicity of oxygen radicals and nitrogen oxides produced in the airway is dramatically enhanced by protonation (23, 24). Efforts to characterize the mechanisms by which the airway protects itself from acid have focused primarily on the role of ion channels and luminal albumin (25). Recent evidence that endogenous airway acidification may also contribute to the pathophysiology of cystic fibrosis (26) and asthma (3) has led to renewed interest in the regulation of airway pH. However, it has not previously been considered that ammonia production by the airway epithelium could have a role in this process. Here, we demonstrate that the human airway epitheliumlike the renal tubular epitheliumneutralizes acid by converting glutamine to ammonia and expresses glutaminase in vitro and in vivo (Figure 8). These novel observations may have implications for the management of infectious and inflammatory lung diseases.

View larger version (11K): [in this window] [in a new window]   Figure 8.   Model of airway epithelial ammonia production and lining fluid proton buffering. Glutamine is imported into the cell where it is catabolized by glutaminase to release ammonia, glutamate, and a proton. Ammonia diffuses across the luminal membrane to react with available protons, and then is exhaled or pumped back into the epithelium. The proton generated can be exported basolaterally or be consumed in other cellular processes. OONO, peroxynitrite; NO2, nitrite; NH3, ammonia; NH4+, ammonium.

Ammonia is a highly water-soluble base (pK_a of 9.4) found in gas phase in the exhaled breath and alveolar air of normal subjects (28). It is believed to have a role in neutralizing exogenous acids in air pollution (29) and can be formed by pharyngeal bacteria. Our findings demonstrate that it is also produced directly by airway epithelial cells and can buffer both endogenous and exogenous acids (Figures 4 and 6B). Indeed, we find that airway epithelial cellslike renal tubular epithelial cells (10, 11, 30)are capable of upregulating their ammonia production in vitro in response to acid challenge (Figure 4), and that this ammonia serves to alkalinize pH. Even in the presence of high concentrations of HCO₃ in Ham's F12K media (20 mM compared with airway lining fluid levels of < 6 mM [31]), the presence of epithelial cells strongly augmented the neutralization of pH concurrent with ammonia production. Importantly, glutaminase produces ammonia, glutamate, and a proton (H⁺) in the renal epithelium. The H⁺ may be pumped basolaterally by ion exchange. However, because airway epithelial cells neutralize acid in vitro without basolateral exchange, we speculate that subsequent metabolism of glutamate to HCO₃ is important in determining the ultimate fate of luminal acid. Additional studies will be required to determine the relative importance of H⁺ and HCO₃ exchange mechanismsand of glutamate metabolismto the glutaminase pathway in the airway epithelium in vivo.

During acute exacerbations of asthma, exhaled breath condensate ammonia concentrations are low when pH is low (Figure 6) and airway glutaminase expression may be decreased (Figure 7). Although cause and effect cannot be established unequivocally, these observationsin association with the in vitro data presentedsuggest that inhibition of glutaminase expression may contribute to the low breath condensate pH observed in acute asthma (3). Intriguingly, the biochemical and cellular toxicities that result from exposure of airway epithelium to acid reflect the pathophysiological features of asthma to a striking degree. These toxicities include augmented release of eosinophil inflammatory mediators (23, 32), enhanced reactivity of nitrogen and oxygen species (3, 24, 33), decreased ciliary beating (4), increased mucous viscosity (19), epithelial dysfunction and sloughing (2, 20), and augmentation of neurogenic bronchoconstriction and cough (1, 21). The apparent association between impaired ammonia production and asthmatic airway inflammation raises the possibility that interventions that improve buffering of airway pHincluding those that increase the expression of airway epithelial glutaminasemay be helpful in preventing and treating exacerbations of asthma.

The acid burden delivered to the airway lumen may be substantial. In addition to serving as the site for excretion of carbon dioxide/carbonic acid, the distal airways and alveoli are also exposed to H⁺ release from Type II alveolar cells because of vacuolar (V) H⁺-ATPase activity in lamellar bodies (8). Inflammatory cells such as macrophages and eosinophils also have VATPases in their vesicles (34); chronic inflammation with necrosis of these cells can result in marked airway acidification analogous to that in infected pleural space (35). These and other processes potentially may be buffered in the airway by luminal albumin and, as in the kidney, by HCO₃ excretion, Na⁺/H⁺ exchange and carbonic anhydrase. However, we suggest that a substantial quantity of titratable acid formed in the lung may react with ammonia. The relevance of ammonia to airway pH homeostasis is suggested by the observations that (1) exposure of airway epithelial cells to physiological concentrations of glutamine (36) results in the production of stoichiometric millimolar concentrations of ammonia (Figure 2); (2) low breath condensate pH is associated with low breath condensate ammonia in vivo (Figure 6); and (3) despite the presence of albumin, bicarbonate, and other buffers in the cell culture medium, airway epithelial cell ammonia production robustly alkalinized pH in vitro (Figure 4A-4C).

Figure 6B suggests that ammonia depletion may be necessary, but not exclusively sufficient, for acidification: breath with low ammonia content can have a neutral pH. The presence of multiple buffering systems appears to be particularly important to airway epithelial cells, which tolerate low pH poorly if they fail to increase ammonia production markedly above control levels (Figure 4D). Indeed, cells that did not respond with increased ammonia production demonstrated sloughing from the culture plate analogous to airway epithelial sloughing observed during asthma exacerbations in vivo (37, 38). The factors that regulate this ammonia-based protective response to low pH remain to be clarified, but may be critically relevant to airway inflammation and are the subject of ongoing investigation. It should be noted that there may be several sources of ammonia in the airway. We found that epithelial cell lines derived from the proximal airway, small airways, and alveoli had glutaminase mRNA activity. The extrathoracic airway may also contribute to ammonia production, especially given the role of glutaminase activity in immune cells and the potential for ammonia production by bacteria (39, 40).

Our data suggest that one mechanism involved in the inhibition of pulmonary ammonia excretion in acute asthma may involve exposure of the airway epithelium to inflammatory cytokines. Sarantos and coworkers have shown that glutaminase activity is downregulated by T helper 1-type lymphocyte (Th1)derived cytokines in a fibroblast cell line (41), and we have confirmed this finding in airway epithelial cell lines (Figure 5). In this regard, it is increasingly appreciated that Th1 cytokines (such as those produced in response to infection with rhinovirus) may help to initiate "common-cold"-associated exacerbations of asthma in the allergen-primed, inflammatory cell-rich asthmatic lung (42, 43). These same cytokines are released in response to Mycobacterium tuberculosisan organism that requires exogenous ammonia (14, 44), and is intolerant of acidic conditions (13, 45). It is tempting to speculate that the asthmatic phenotypein which airway ammonia production is decreased and pH fallsmay limit substrate for ammonia-requiring M. tuberculosis while providing an inhospitable acidic environment. This process could be detrimental, however, in the allergen-exposed, eosinophil-infiltrated lung.

In conclusion, we have shown that glutaminase is expressed and active in the human airway epithelium and production of ammonia by airway epithelial cells serves a pH homeostatic roleincreasing protectively in response to acidic stressand is inhibited by inflammatory cytokines. Direct relevance of this pathway to human disease is suggested by the permissive effect of ammonia depletion on exhaled airway fluid acidification during asthma exacerbations. These observations may have broad implications for the management of infectious and inflammatory lung diseases.