INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Farmer's lung (FL) is the most common form of hypersensitivity pneumonitis. It is induced by the inhalation of thermophilic actinomycetes and spores of Aspergillus species found in moldy hay. Although a wide range of immunologic and inflammatory phenomena have been observed in this disease, its pathogenesis is not yet completely understood (1, 2). The clinical presentation may be acute, subacute, or chronic, depending on the intensity of antigen exposure. In chronic stages of the disease, a lymphocytic inflammation prevails in the lungs which, however, can also be found in asymptomatic farmers (3). In contrast, acute episodes of FL are associated with flulike symptoms and a respiratory impairment characterized by a loss of VC and single-breath diffusing capacity of carbon monoxide (DL_CO), and a decrease in Pa_O2, typically observed several hours after exposure to moldy hay. These acute episodes of FL are characterized by an influx of neutrophils into the lungs (6, 7) and oxidative activation of these cells (8). As a consequence, oxidative damage to the alveolar epithelium may occur as the initial step in the chronic interstitial lung disease characteristic of hypersensitivity pneumonitis. In this context, we were interested in the antioxidant defense mechanisms of the lungs of patients with FL during acute episodes of the disease. Since glutathione (GSH) is, at least quantitatively, the most important antioxidant of the lung (9), we investigated pulmonary GSH levels after hay exposure in patients with manifest FL and in asymptomatic farmers (AF).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

We investigated a total of 15 symptomatic patients with FL (one male and 14 females, 42 ± 1 yr old [mean ± SEM]). The following criteria were used to establish the diagnosis of FL: (1) a clinical history of acute episodes several hours after exposure to moldy hay; (2) lung function and/or radiographic criteria (interstitial infiltrates) consistent with hypersensitivity pneumonitis (acute infiltrates were absent on chest radiographs in both the symptomatic patients and asymptomatic controls described subsequently); and (3) the demonstration of specific IgG antibodies against Saccharopolyspora rectivirgula, Thermoactinomyces vulgaris, or Aspergillus fumigatus.

A group of 10 asymptomatic farmers (AF) (two males and eight females, 43 ± 1 yr old) served as a control group. Eight of the AF tested positive for IgG antibodies against the antigens mentioned earlier. All of the patients with FL reported shortness of breath on exertion, whereas AF complained of no specific respiratory symptoms. All individuals included in the study were nonsmokers and were instructed to avoid antigen contact for 1 wk before the exposure testing done as part of the study.

Exposure Tests

Standardized hay exposures of all subjects were conducted in a designated chamber in our department, as described elsewhere (8). In brief, a mixture of moldy hay specimens obtained from patients with FL was used as the antigen source; the hay was tossed by the study participants for 1 h. Previous experiments done with personal sampling devices had shown that the hay dust exposure achieved with this method is comparable to the dust load on a farmer's workday (10). The observation period after exposure was 6 h. During this time the onset of general symptoms (i.e., fever, chills, myalgias, headache, nausea) was documented. A set of lung function measurements, including VC, DL_CO, and Pa_O2 was obtained before exposure (baseline). Each subject's pulmonary reaction was evaluated by measurement of VC (every hour after exposure), DL_CO (1, 3, 5, and 6 h after exposure), and Pa_O2 (3 and 6 h after exposure). To ensure that these parameters reflected a process in the interstitium of the lung, but not airway obstruction, we also measured FEV₁ (%) and specific airway resistance (SRaw) (before exposure and every hour after exposure). Data for VC, DL_CO, and Pa_O2 were analyzed, provided that FEV₁ (%) and SRaw measured at the same time point were within the normal range. After 6 h of observation, fiberoptic bronchoscopy and bronchoalveolar lavage (BAL) were performed.

All lung function tests were done with standard equipment (Jaeger & Thoennies, Würzburg, Germany). Lung volumes were referenced to standard values as published by the European Community for Steel and Coal (11). Blood gas analysis was performed with arterialized capillary blood from the ear lobe (double values).

The study protocol was approved by the local ethics committee, and written informed consent was obtained from all participants in the study.

Biologic Samples

Serum and bronchoalveolar lavage fluid (BALF) were obtained through standard techniques. Aliquots of BALF were taken for GSH assays (as described subsequently), total cell counts with an automated counter (Coulter, Inc., Hialeah, FL), and cytocentrifuge preparations for differential counts. The cells were pelleted and the supernatants were used to assess the volume of epithelial lining fluid (ELF).

Estimation of Respiratory ELF

To determine the volume of ELF, we applied the urea dilution method as described by Rennard and colleagues (12). Concentrations of urea nitrogen in serum and BALF supernatants (following centrifugation at 3,000 × g for 10 min) were measured with the urea nitrogen 65-UV Kit (Sigma Chemical Co., St. Louis, MO). To avoid uncontrolled urea diffusion during the lavage procedure, a standardized lavage protocol was applied, in which instillation of a 20-ml aliquot of BALF through the bronchoscope was followed by a 20-s suction period and the procedure was repeated five times, until a total volume of 100 ml of fluid per lung segment had been administered. A total of three segments from the middle lobe, the lingula, and the left upper lobe were lavaged in every study participant.

Cell Counts

Total cell count was measured with a Coulter counter. Differential cell counts were assessed with cytocentrifuge preparations stained with May-Grünwald-Giemsa. Cell counts were expressed as percent of BALF cells and per milliliter of BALF recovered.

GSH Concentration and Form

GSH concentrations were measured in freshly obtained BALF, using standard techniques as previously reported (13). All determinations were made in triplicate, and the average value was calculated. For measurement of total glutathione (GSH_T = GSH + 2 × glutathione disulfide [GSSG]), a 100-µl aliquot of BALF supernatant (3,000 × g for 10 min) was mixed immediately after BAL with 1.1 ml of 0.1 M sodium phosphate buffer, pH 7.0, containing 1 mM ethylenediaminetetraacetic acid (EDTA), 0.2 mM nicotinamide adenine dinucleotide phosphate (NADPH), 63.5 µM 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB), and 4 U/ml glutathione reductase (all from Sigma or Serva, Heidelberg, Germany). The rate of reduction of DTNB was recorded spectrophotometrically at a wavelength of 412 nm. The GSH_T concentration in the BALF sample was calculated with an internal GSH standard having a concentration of 0.84 µM (14).

Oxidized glutathione (glutathione disulfide, GSSG) was measured according to the method described by Adams and associates (15). After centrifugation (3,000 × g for 10 min), BALF supernatant was mixed with an equal volume of 10 mM N-ethylmaleimide in 0.1 M potassium phosphate buffer, pH 6.5, containing 17.5 mM EDTA. A total of 250 µl of the mixture was passed through a SEP-PAK C₁₈ cartridge (Waters Associates, Milford, MA) that had been prewashed with 3 ml of methanol followed by 3 ml of distilled water. GSSG was eluted from the column with 1 ml of 0.1 M potassium phosphate buffer, pH 7.5, with 5 mM EDTA, 800 µM DTNB, 2 U/ml glutathione reductase, and 1 mM NADPH. The rate of reduction of DTNB was recorded spectrophotometrically at 412 nm. Standards of GSSG (Boehringer, Mannheim, Germany) of known concentration were processed exactly as the BALF supernatants and were used to generate standard curves.

The concentration of reduced glutathione (GSH) was calculated from the following equation: GSH = GSH_T  2 × GSSG.

Statistics

Statistics were calculated using SPSS software version 8.0 for Windows (SPSS Inc., Chicago, IL). Data are expressed as mean ± SEM. For group comparisons, the nonparametric Mann-Whitney test was used. Values of p < 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline Lung Function

The results of the baseline lung function measurements in patients with FL and in AF are summarized in Table 1. There were no severe lung function impairments in either group. Significant reductions of DL_CO and of Pa_O2 during steady state exercise were found in FL patients as compared with AF (Table 1). There were no significant differences in FL patients and AF with respect to lung volumes. The results for FEV₁ and SRaw were similar in the two groups, and did not indicate airway obstruction.

Exposure Test

Following standardized hay exposure, the group of patients with FL exhibited a remarkable decline in all evaluated lung function parameters: VC (31 ± 4% of baseline), DL_CO (17 ± 3% of baseline), and Pa_O2 (14 ± 2% of baseline) (Figure 1). These changes fulfilled the recommendations of the Deutsche Gesellschaft für Pneumologie for a significant pulmonary reaction in antigen exposure tests for hypersensitivity pneumonitis in every case of FL, thus confirming the clinical diagnosis (16). In contrast, AF showed only minor changes in lung function: VC (4 ± 5% of baseline), DL_CO (9 ± 3% of baseline), and Pa_O2 (5 ± 2% of baseline) (Figure 1). These changes did not fulfill the previously mentioned criteria (16).

View larger version (19K): [in this window] [in a new window]   Figure 1.   Changes in lung function parameters after acute hay exposure in AF and in patients with FL. The bars for VC, DLCO, and PaO2 following exposure ("post") represent the mean ± SEM of the minimum value of the respective lung function parameter observed 3 to 6 h after exposure, expressed as a percent of the baseline value (baseline = 100%). See text for interpretation.

BALF Cell Counts

Six hours after hay exposure, BAL was performed in all participants in the study. The percentages and absolute numbers of neutrophils in the recovered BALF were strongly increased above the reported normal range in both the FL and AF groups (17), and were significantly greater in the FL patients than in the AF (Table 2). The percentages and absolute numbers of lymphocytes in the recovered BALF were also higher in patients with FL than in AF, but the difference did not reach statistical significance (Table 2). As a consequence, the percentages but not the absolute numbers of alveolar macrophages were lower in patients with FL than in AF (Table 2). The T-helper/T-suppressor cell ratio was not significantly different in the two groups (AF: 2.3 ± 1.3; FL: 1.7 ± 0.4; p = NS).

ELF Volumes

The BALF volumes, recovered from three segments (middle lobe, lingula, and left upper lobe), each lavaged with 100 ml saline, did not differ significantly in the FL patients and AF (AF: 85.2 ± 11.0 ml, FL: 78.2 ± 10.9 ml; p = NS). The obtained ELF volume was higher in patients with FL than in AF (AF: 0.49 ± 0.08 ml, FL: 1.19 ± 0.09 ml; p < 0.05).

GSH Concentrations

The GSH_T concentration in native BALF was decreased in patients with FL as compared with AF (AF: 6.4 ± 0.8 µM; FL: 4.8 ± 0.8 µM; p < 0.05). This difference was even more pronounced when the respective GSH_T concentrations in ELF were calculated (AF: 1,185.0 ± 189.9 µM; FL: 292.5 ± 27.5 µM; p < 0.001) (Figure 2). The concentration of the reduced form of glutathione (GSH) was also decreased in patients with FL, both in native BALF (AF: 5.7 ± 0.7 µM; FL: 4.2 ± 0.7 µM; p < 0.05) and in ELF (AF: 1,054.5 ± 172.9 µM; FL: 256.8 ± 22.1 µM; p < 0.001) (Figure 2). Accordingly, the concentration of the oxidized (GSSG) form of GSH was increased in the ELF of AF (AF: 65.2 ± 10.0 µM; FL: 17.8 ± 4.3 µM; p < 0.001) (Figure 2). However, the ratio of the GSSG to the GSH_T concentration (expressed as a percent of GSH_T) did not differ significantly between the FL and AF groups (AF: 5.7 ± 0.5%; FL: 5.5 ± 0.7%; p = NS). As compared with a group of healthy, nonsmoking controls reported earlier (GSH_T: 568.0 ± 44.5 µM; GSH: 505.6 ± 52.5 µM) (14), GSH_T and GSH concentrations were considerably decreased in patients with FL and approximately doubled in AF in our study.

View larger version (14K): [in this window] [in a new window]   Figure 2.   Concentrations of total, reduced, and oxidized GSH in ELF obtained from AF and patients with FL by BAL at 6 h after acute hay exposure. See text for interpretation. Open circles represent AF; closed circles represent patients with FL. Bars = mean ± SEM. *p < 0.001.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In our study, we could confirm previous findings of an acute neutrophilic alveolitis after antigen exposure (5), which occurred in patients with FL and to a relatively minor degree also in AF. Moreover, we observed significantly decreased concentrations of GSH_T and GSSG in native BALF and in ELF in symptomatic patients with FL at 6 h after hay exposure, as compared with healthy nonsmoking controls (14), whereas AF exhibited increased GSH levels. In this respect, patients with FL are strikingly similar to patients with idiopathic pulmonary fibrosis, who also exhibit low GSH levels in ELF (14, 17), which have been linked to the underlying inflammatory and fibroproliferative process in this disease (18, 19).

In contrast, increased GSH concentrations in ELF, similar to those seen in the AF in our study, have previously been reported only in cigarette smokers (20). Interestingly, hypersensitivity pneumonitis rarely occurs in smokers (1, 2, 21). The high levels of GSH observed in cigarette smokers have been interpreted as a defense mechanism of the lungs of smokers. Recently, it has been shown that high GSH levels in cigarette smokers decrease epithelial permeability and protect alveolar epithelial cells from injury (22, 23). Since all of the participants in our study were never-smokers, our data suggest that hay-dust exposure may upregulate GSH concentrations in the lungs of AF. As a consequence, decreased epithelial permeability, as indicated by a lower ELF volume in the AF than in the FL patients in our study, may decrease antigen contact with immunocompetent cells and may therefore preclude disease manifestation.

Since we did not perform BAL at baseline, it remains unknown whether the differences in GSH levels in patients with FL and in AF are chronic or induced acutely by hay exposure. Irrespective of this specific detail, AF may, either constitutionally or by virtue of a special inflammatory response, be able to upregulate their pulmonary GSH levels, which in turn may protect them from disease manifestation. In contrast, individuals who cannot upregulate their GSH concentrations may be prone to developing FL. Genetic predisposition may therefore be the underlying cause of the different glutathione levels in AF and FL patients, and may be one of the reasons for the epidemiologic finding that only about 10% of exposed individuals develop manifest FL (24).

We are aware that this interpretation of our results includes speculation. At this point, we also cannot exclude the possibility that differences in the pulmonary GSH concentration of patients with FL and AF are epiphenomena. In particular, the decreased GSH concentrations of patients with FL could be caused by the inflammatory disease process itself, since inflammatory cytokines such as transforming growth factor-1 have been shown to downregulate gene expression for enzymes critical to GSH biosynthesis (25). Although we cannot provide a clear-cut cause-and-effect relationship between GSH levels and disease manifestation, our results provide circumstantial evidence supporting the hypothesis that the individual's ability to upregulate GSH concentrations in ELF, in response to hay exposure or cigarette smoke, may be a host susceptibility factor that determines the manifestation of hypersensitivity pneumonitis. Prospective clinical trials are needed to prove this hypothesis.