Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Glutathione (-glutamylcysteinylglycine) (GSH) appears to play an important role in a number of cellular functions (1). GSH is critical to cellular defense either as a cofactor for conjugation to electrophilic species or by acting as a scavenger of reactive oxygen species (2). One of the key pathways of GSH depletion is the glutathione S-transferase (GST)-mediated conjugation of electrophiles produced by cytochrome P450 monooxygenases (3). Diethylmaleate (DEM), a substrate requiring GST for conjugation, depletes most cellular GSH with the exception of the mitochondrial pool in hepatocytes (4, 5). A similar response occurs in lung where naphthalene (NA) exposure depletes GSH to 20% of control (6). Therefore, regulation of the GSH pool may be critical to the survival of specialized epithelial target cells, especially nonciliated bronchiolar epithelial (Clara) and alveolar epithelial (type II) cells.

A clear understanding of the role of GSH in protecting lung cells from injury depends on understanding the capability of individual cell populations to maintain and regulate their GSH pools. The strategies used by a cell population to combat oxidative stress and maintain the availability of reduced GSH stored in intracellular pools may elucidate the mechanism involved in cellular resistance to cytotoxic agents. It is apparent that in an organ such as the lung, with over 40 different cell phenotypes, not all of the cells are equally susceptible to injury, and the mechanisms involved in cellular resistance to cytotoxic agents may vary between cells. There is a significant body of literature regarding the ability of GSH to protect alveolar type II cells from a variety of oxidant stressors (7). However, at the present time, how the GSH pool is regulated in other target cell populations within the lung is unclear. Lung-slice studies and the use of isolated perfused lungs (10, 11) indicate that elevating the total cellular GSH level provides significant protection from xenobiotic cytotoxicants. It is not known whether this elevation is associated with protection of Clara cells, a primary target cell type for oxidant gases (such as ozone) and bioactivated xenobiotic compounds such as NA (12) or dichloroethylene (13). Currently, the most frequently used methods for detection and quantification of GSH are assays that detect average GSH content from a population of cells and mask differences between individual cells. Although these assays may be quite sensitive (14), they are not suited for specific analysis of subpopulations of cells within a heterogenous tissue, such as the lung. This has been shown to have particular importance in evaluating GSH heterogeneity of tumor cells (15). Whereas analysis of whole tumors and control tissue populations showed significant differences, analysis on a cell-by-cell basis revealed that the variance could be attributed to specific subpopulations that contained higher levels of GSH (15).

In the lung, the injury response of cells to bioactivated toxicants varies both regionally and within specific microenvironments (16). For example, low doses of NA administered intraperitoneally injure Clara cells in the most distal areas of the lung, whereas higher doses cause the injury pattern to extend proximally (12). However, even at low doses, Clara cells in the distal airways do not respond uniformly. These differences in susceptibility may represent differences in activity of phase I and/or phase II metabolic enzymes or cofactors/cosubstrates associated with these enzymes. Accordingly, because GSH has been demonstrated to be an important nucleophilic detoxifying agent for NA, the current studies address the question of whether differences in Clara cell susceptibility are due to variations in GSH levels within individual Clara cells.

The current study was designed to address three issues: (1) define GSH pools within the Clara cell, a primary target cell population for pulmonary toxicants; (2) determine whether differences exist in steady-state GSH levels on a cell-by-cell basis; and (3) evaluate whether heterogeneity contributes to the differences in susceptibility to NA injury.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

All chemicals used were reagent grade or better. High-performance liquid chromatography (HPLC) solvents were obtained from Fisher Scientific, Pittsburgh, PA. Lyophilized porcine elastase was obtained from Worthington Biochemical, Malvern, PA. Monobromobimane (MBrB) and monochlorobimane (MClB) were purchased from Calbiochem, San Diego, CA. The following fluorochromes were purchased from Molecular Probes, Eugene, OR: propridium iodide, ethidium homodimer-1 (EthD-1), 6-diamidino-2-phenylindole (DAPI), dihydrochloride fluorescein isothiocyanate (FITC), cell tracker green (BODIPY), and tetramethylrosamine (TMR) chloride. DEM was purchased from Aldrich, Milwaukee, WI. GSH-monoethylester (GSH-e) was purchased from Bachem Bioscience, Philadelphia, PA. Dulbecco's modified Eagle's medium (DMEM)/F12 medium was obtained from GIBCO Labs, Grand Island, NY. Teflon culture dish inserts coated with fibrillar collagen and fibronectin were purchased from Becton Dickinson Labware, Bedford, MA. Insulin, hydrocortisone, transferrin, penicillin/streptomycin, gentamicin, and retinol were purchased from Sigma, St. Louis, MO. Epidermal growth factor was purchased from Upstate Biotechnical, Lake Placid, NY. Bovine hypothalamus extract was a generous gift from Dr. Reen Wu, University of California, Davis.

Synthesis of MClB and MBrB Conjugate Standards

MClB and MBrB differ in that the MClB requires GST to conjugate GSH. MBrB does not require a transferase-mediated conjugation and binds all reduced sulfhydryls optimally at basic pH. The end products of these two reactions yield the same GSH bimane conjugate (GS-bimane), with the exception that in biologic mixtures chlorobimane yields only one conjugate, GS-bimane, whereas bromobimane yields several additional sulfhydryl-conjugated bimane adducts. To produce the GS-bimane standard, pure GSH was incubated with MBrB. The bimane-GSH standard was synthesized by the addition of excess GSH to 3.7 µmol of MBrB dissolved in 50 mM NaOH. The reaction mixture was stirred for 4 h at room temperature under N₂. The pH of the mixture was adjusted to 3.5 by the addition of 5% acetic acid. The mixture was applied to an SM-16 solid-phase extraction column pre-equilibrated with 1% acetic acid. The purified product was eluted from the cartridge with a 1:1 solution of acetonitrile/water, evaporated under an N₂ stream, and resuspended in methanol. The identity of the final product was confirmed by electrospray mass spectrometric analysis using a VG Quattro BQ triple quadropole mass spectrometer in positive-ion mode. The sample was dissolved in methanol/water (1:1) and was delivered at a flow rate of 5 µl/min. Capillary voltage was 3.6 kV, cone voltage was 50 V, and source temperature was 70°C. Parent ion was observed at 497.7 amu.

GSH Determination by HPLC

GSH was determined according to the method of Fahey and Newton (17). Briefly, 10⁶ isolated cells were homogenized in a glass homogenizer on ice in 100 µl of 200 mM methanesulfonic acid (MSA) containing 5 mM diethylenetriaminepentacetic acid (DTPA). The pH of the sample was adjusted to 8.0, and samples were derivatized by the addition of 2 mM MBrB. The GS-bimane adduct was separated and eluted by HPLC on a Novapak C₁₈ column (5 × 100 mm) using a mobile phase of 7.5% acetonitrile, 15 mM tetrabutylammonium phosphate, and 0.25% acetic acid, pH 3.4. The fluorescent derivative was measured with a Shimadzu RF-535 fluorescence detector at excitation (ex) and emission (em) wavelengths of 360 and 460 nm, respectively. Standard curves were linear to 100 pmol/sample.

Clara Cell Isolation

Clara cells were isolated as described previously (18). Briefly, animals were killed with an overdose of pentobarbital. The trachea was cannulated and lungs were perfused via the pulmonary artery with a balanced salt solution. Perfused lungs were removed from the chest, lavaged with a solution containing 2 mM ethyleneglycol-bis-(-aminoethyl ether)-N,N'-tetraacetic acid, and instilled with an elastase solution (4.3 U/ml) for 20 min. To select Clara cells, elastase-digested lungs were minced, filtered through gauze, and centrifuged through a layer of calf serum. Macrophages were removed by molecular panning and viable Clara cells measured by exclusion of erythrosine B (19). Before the imaging experiments, nonviable cells were identified by incubation with EthD-1 (3 µg/ml, ex 520 nm). Only those cells that excluded the dye were analyzed in the imaging experiments. For every image, 16 frames were averaged. For each group in each experiment, intensity from 100 cells was collected. Each experiment was repeated a minimum of three times with cells from three separate isolations.

Clara Cell Culture

Cells were isolated as previously described (18) and plated on a Teflon insert coated with fibrillar collagen and fibronectin. Cells were seeded at a density of 10⁶ cells/dish. DMEM/F12 medium was supplemented with insulin (2.5 mg/ml), epidermal growth factor (12.5 µg/ml), bovine hypothalamus extract (1 µl/ml), hydrocortisone (1 mM), transferrin (2.5 mg/ml), penicillin (10,000 U/ml), streptomycin (10 mg/ml), gentamicin (50 mg/ml), amphotericin (250 µg/ml), and retinol (1 mM) as previously described (20). Viability was measured by exclusion of erythrosine B dye. For these experiments cultured cells remained viable as assessed by erythrosine B (19) for up to 7 d.

Semiquantitative Epifluorescence

Images of isolated Clara cells were produced using a Photon Technology International (PTI-DELTASCAN 4000) system mounted on a Nikon Diaphot inverted microscope equipped for epifluorescence using a CF Fluor 100× oil objective. Cell suspensions were illuminated with a 75-W xenon arc lamp, and fluorescence was detected with a Hamamatsu Intensified CCD camera. The signal was analyzed using the PTI RF-D4022 workstation. The ex wavelength that gave the peak emission for each fluorochrome was determined using the -scan mode. For MClB, MBrB, and DAPI the cells were excited at 385, 395, and 340 nm, respectively, using a Chroma DM-400 dichroic mirror with a 460/50 bandpass emission filter. Image shift between ex wavelengths was not observed on the basis of the ex/em pattern of multispectral fluorescence microscopy standards (Multispec; Molecular Probes). To determine the concentration of GSH-MClB conjugate detected by image analysis, 10⁶ cells previously evaluated by PTI analysis were pelleted, the media removed, and the cells resuspended in 100 µl of 200 mM MSA containing 5 mM DTPA. Average imaging values were plotted against average HPLC measurements of the same cells under different experimental conditions (depleted and augmented GSH) to establish the concentration of GSH represented by different imaging values. Assuming GSH binds MClB 1:1, a semiquanititative value was calculated by regression analysis.

Confocal Microscopy

To compare the distribution of cytoplasmic GSH between mitochondria and other cytoplasmic compartments, cells were imaged with a Bio-Rad MRC-600 Scanning Confocal Imaging System using COMOS version 6.01 software, mounted on an Olympus BH2 RFL (reflected fluorescence light) microscope with standard filter blocks for green (545 nm) and blue (490 nm) ex. The confocal imaging system was equipped with an argon laser emitting wavelengths of 488 and 514 nm. Filter blocks A1 and A2 are designed to capture the em signal from one corresponding fluorochrome. These filter blocks were selected to capture two-thirds of the maximal signal present for each fluorochrome. For comparison of the green and red images, the gain and black level settings on the image system were optimized for each channel as each image was collected.

Determination of Intracellular GSH Pools

To analyze the distribution of GSH pools between nuclear and cytoplasmic compartments, cells were imaged by semiquantitative epifluorescence as described earlier, using a modification of the protocol described by Bellomo and colleagues (21). Briefly, cells were incubated with MClB as described earlier, and imaged. Cells were then fixed and bleached with ethanol/acetic acid (75%/25%), incubated with MBrB (2 mM; pH 8.0), and imaged. Next, the cells were incubated with DAPI (2 µg/ml) and imaged. Finally, the cells were incubated with FITC (50 µM) and imaged.

To determine the mitochondrial distribution of sulfhydryls, freshly isolated cells were labeled with the 5 µM TMR chloride for 45 min and 1 µM BODIPY for 20 min at room temperature, and analyzed by cofocal microscopy. Labeled cells were recovered by centrifugation, and resuspended in fresh DMEM/F12 medium, and imaged. To eliminate red fluorescence of TMR chloride (574 nm em) in the green channel, a bandpass filter (520 nm ± 25) was used. After the collection of the image on both the green and red channels, the 520-barrier filter was inserted and an additional image was collected using the previous gain and black level settings. This allowed the estimation of the amount of green signal without the contribution of the red signal. After the simultaneous collection of both the green and red images, the two images were merged using 256 optimized colors. From this image the colocalization of both fluorochromes was evaluated.

Measurement of Cellular GSH Levels

Three approaches were used to determine the heterogeneity of Clara cell GSH by incubation with MClB: (1) analysis of isolated Clara cells; (2) analysis of cultured Clara cells, and (3) analysis of Clara cells in microdissected airway explants. Isolated Clara cells were incubated with MClB for 20 min, rinsed, and then imaged. To image GSH in cells in culture, cells were incubated in MBCl for 20 min, then rinsed with fresh media three times and imaged. To determine Clara cell GSH levels in situ, lungs were filled with 1% agarose as previously described (22) and airways down the axial pathway of the right cardiac lobe were then exposed by microdissection within 45 min of the animal's death. After microdissection, explants were incubated with Waymouth's medium containing MClB for 20 min. The intensity of GSH-MClB conjugate both in situ and in vitro was determined by the same imaging system used for isolated cells, except the intensified CCD camera was mounted on an upright microscope (Nikon Optiphot-2) using a water immersion objective (Olympus WPlan FL40X UV). Detection limits were set on the basis of the area of highest signal. For in situ measurements, this was in cells of the lobar bronchus. Images were collected from six airway levels: lobar bronchus, midlevel bronchus (six to eight generations), proximal bronchioles (four and eight generations from the terminal bronchiole), and the most distal terminal bronchiole.

GSH and Susceptibility

To determine whether cells with lower GSH are more susceptible to injury, isolated cells were treated with agents that are known to modulate intracellular GSH levels. Clara cells were incubated with diethyl maleate (a GSH-depleting agent) or GSH monomethyl ester (an agent that augments GSH levels) and imaged to analyze the differences in intracellular GSH levels. Specific concentrations for specific experiments are included in the caption for Figure 6. In a separate experiment, isolated cells were treated with 0.5 mM NA for 3 h. Permeability to EthD-1(ex = 530, em = 645) was used as an indicator of viability. At 1.5 and 3 h, cells were incubated with MBCl and EthD-1 and imaged. Using semiquanitative epifluorescence, the percentage of cells with high or low signal was determined by ranking the fluorescent intensity.

View larger version (80K): [in this window] [in a new window]   Figure 6.   Confocal microscopic comparison of GSH distribution in isolated Clara cells based on GSH-BODIPY fluorescence in response to DEM and GSH-e. (A) Left panel: Range of intensity of GSH-BODIPY fluorescence in untreated Clara cells. Right panel: Representative high-signal Clara cell after incubation with 1 mM GSH-e for 1 h from the same isolated cell population. Bar = 5 µm. (B) Left panel: Clara cell depleted of GSH by DEM (2.125 mM for 1 h). Middle panel: Same Clara cell, repleted by incubation with GSH-e (1 mM) for 0.5 h. Right panel: Same Clara cell, repleted by incubation with GSH-e (1 mM) for 1 h. (C) Left panel: Clara cells depleted of GSH with DEM (4.25 mM) for 1 h. Right panel: Partial repletion of same cells by incubation with GSH-e for 1 h.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

GS-Bimane Conjugate Specificity and Signal Stability

HPLC analysis of Clara cell extracts incubated with MClB showed that the chlorobimane specifically bound cytosolic GSH (Figure 1A). The peak corresponding to the GS- bimane conjugate contained 98% of the total area counts. When cells were incubated as long as 60 min, the conjugation of GSH to MClB (1.0 mM), as measured by fluorescent image analysis, reached a plateau by 10 min (Figure 1B). Under the conditions used in the imaging experiments, a loss of GS-bimane conjugate from cells could not be detected by image analysis over a 1-h incubation (Figure 2A). This was true for cells with low and high levels of conjugate (Figure 2A). In addition, there was no change in the background fluorescence in the medium surrounding the cells. This was confirmed by HPLC measurement showing little GS-bimane conjugate in the media of isolated cells. Clara cell viability, assessed by erythrosine B exclusion, did not change significantly during the incubation period (Figure 2B).

View larger version (13K): [in this window] [in a new window]   Figure 1.   (A) HPLC fluorescence analysis of GS-bimane conjugate after incubation of isolated Clara cells with 1 mM MClB for 20 min. A peak observed at 31 min comprising 98% of the total area counts in the cytoplasmic fraction elutes with the peak that corresponded with GS-bimane conjugate standard. (B) Time course of GSH conjugation by MClB in isolated mouse Clara cells. Clara cells were incubated with 1 mM MClB for 5, 10, 20, 30, 40, 50, or 60 min; centrifuged for 10 min to pellet cells; placed in fresh medium; and analyzed by image analysis. Values represent means ± standard deviation (SD) for three separate experiments.

View larger version (20K): [in this window] [in a new window]   Figure 2.   (A) Image analysis of MClB signal (ex 385 nm) in four isolated Clara cells recorded at 30-s intervals. The background (BKG) signal for a space between the four cells (lines 1, 2, 3, and 4) was also recorded. The signal was stable in both high- and low-intensity cells for at least the first 60 min after the end of incubation with MClB. (B) Viability and GSH concentration in isolated Clara cells over 4 h. Viability (squares) of control Clara cells was measured by erythrosine B. The concentration of GSH-MClB (circles) and GSH-MBrB (diamonds) in Clara cells was measured by image and HPLC analysis, respectively. Values represent means ± SD for three separate experiments.

Intracellular GSH levels, measured by HPLC fluorescence of the MBrB-GSH derivative, were 1.29 ± 0.16 nmol/10⁶ cells and remained at this level over a 240-min period (Figure 2B). Standard deviations for GS-bimane intensity varied by as much as 50% of the mean even when data for 300 cells/time point were collected. HPLC analysis of GSH-MBrB closely matched concentrations estimated by image analysis of GS-bimane produced with MClB. Previous work in isolated hepatocytes suggested that GS-bimane conjugate translocates to the nucleus (23). To determine whether GS-bimane conjugate localizes in the nucleus after conjugation in the cytoplasm, Clara cells were incubated with MClB and/or permeablized with Triton X-100. The nucleus was identified by EthD-1 (Figures 3A and 3B) staining. After permeablization, the epifluorescence from the EthD-1 intercalation into DNA in the nucleus did not overlap with the epifluorescent signal produced by excitation of the GS-bimane conjugate in the cytoplasm.

View larger version (8K): [in this window] [in a new window]   Figure 3.   To establish that the GS-bimane conjugate was excluded from the nucleus, isolated Clara cells were incubated with 1 mM MClB for 20 min and the fluorescence was captured with a digital camera. Cells were permeabilized with Triton X-100 in the presence of the DNA-intercalating fluorochrome EthD-1 for 20 min. (A) Ex at 520 nm delineated positive spheroidal zones of EthD-1 emission within the cells, indicating the position of the nucleus. (B) The nuclear zones did not emit an MClB signal when the same cells were excited at 385 nm. A crescent-shaped profile was apparent that did not include the EthD-1-positive area.

Modulation of Intracellular GSH

DEM (5.0 mM) effectively decreased the GSH pool to 40% of control, as measured by HPLC (Table 1). Addition of DEM to the incubation decreased the total GSH pool as assessed with MBrB (decreased to 24% of control).

The MClB detectable GSH pool, measured by HPLC, was elevated by compounds that are known to increase cellular GSH. N-acetyl-L-cysteine (NAC) produced a greater increase in GSH-MBrB conjugate (187% of control; Table 1) than in GSH-MClB conjugate (119% of control) when both conjugates were measured by HPLC. At identical concentrations (0.5 mM), NAC produced a greater increase in GSH-MBrB conjugate than did GSH-e. GSH added to the outside of cells did not increase the pool as measured by MClB (Table 1), although a large increase in the total GSH pool was detected by MBrB (450% of control; data not shown). To determine whether incubation with extracellular GSH resulted in GSH binding to the outside of the cell, aliquots of cells were repeatedly centrifuged and resuspended in fresh media. With each wash the amount of GSH detectable by MBrB decreased (in nmol/10⁶ cells): first wash, 5.65 ± 2.45; second wash, 1.30 ± 0.55; third wash, 0.70 ± 0.27; control, 0.98 ± 0.23.

Intracellular GSH Pools

GSH localization was determined using dyes that identified nonprotein cellular sulfhydryls, the nucleus, and mitochondria. In most cases the highest intensity for GS-bimane conjugate was found toward the center of a relatively circular profile (Figure 4A). This was true for both high- and low-intensity cells.

View larger version (24K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   Figure 4.   Sequential incubation with a series of fluorochromes was used to indicate differences in GSH and protein sulfhydryls in cytoplasmic and nuclear compartments. (A) The intracellular GSH accessible to GSTs was identified by incubation of isolated Clara cells with MClB excited at 385 nm. (B) The distribution of protein sulfhydryls within the same cells was determined by bleaching and extraction with ethanol/acetic acid followed by incubation with MBrB excited at 397 nm. (C) The position of the nucleus within the same cells was identified by binding of DAPI to nuclear DNA excited at 345 nm. (D) The distribution of cellular protein was determined by binding of FITC excited at 520 nm. (E) Comparison of profiles of intensity for the fluorochromes illustrated in A-D in a cell with high levels of MClB-GSH (marked by *) and a cell with low levels of MClB-GSH (marked by **) for MClB, MBrB, DAPI, and FITC. The position of the profile is marked by the white line in A.

To determine the distribution of the GSH within the entire thiol pool, the signals created by both MClB and MBrB were compared. The fluorescent intensity of MClB-GSH conjugate was equal to or greater than the fluorescent signal produced by the MBrB (Figure 4B). When the distribution of specific GSH labeling with MClB was compared to nonspecific sulfhydryl MBrB labeling by linear profiling across the cell, the intensity peaks were similar in low-intensity cells (Figure 4E). In high-intensity cells the MClB intensity was greater than that for MBrB.

When the position of the nucleus was identified with DAPI, it varied within the fluorescent profile created by MClB (Figure 4C). MClB and DAPI labeling distribution were also compared by linear profiling across the cell. The intensity peak created by DAPI in high-intensity cells was not centered under the peak created by the MClB (Figure 4E). In low-intensity cells the DAPI and GS-bimane intensity peaks were more similar. However, in these cells the intensity of GS-bimane conjugate was uniform across the cell and was the same in areas with a nucleus (where DAPI signal was high) as in those without a nucleus (where DAPI signal matched background).

To determine whether cell size contributes to increased GSH level, cellular protein was labeled with FITC to estimate cell size. The GS-bimane signal strength varied by a factor of 2 or less between high- and low-intensity cells. The heterogenous signal strength of GS-bimane conjugate did not correlate with differences in cell size. Larger cells, as measured by FITC, did not necessarily have a higher GS-bimane signal.

Confocal microscopy, using TMR chloride to label mitochondria and BODIPY cell tracker green to label cellular thiols, indicated that these two dyes colocalize (Figure 5). Images were collected simultaneously on both the green (Figure 5A) and red (Figure 5B) channels. When the 520-nm barrier filter was inserted to remove all red emissions (Figures 5C and 5D), the red channel was completely obscured (Figure 5D) but the green signal remained (Figure 5C). The TMR chloride (red) signal appeared in large spherical bodies and smaller punctate spherical bodies. The BODIPY thiol conjugate (green signal) had a similar distribution as well as a diffuse background fluorescence throughout the cytoplasm. The highest and most homogeneous level of signal was detected in the large spherical bodies that were distributed throughout the peripheral regions of the cells. The smaller punctate spherical compartments displayed a more heterogeneous level of signal that was usually lower as compared with the larger bodies. To colocalize the images, the gain level on the microscope was optimized to allow visualization of both distinctive cellular pools (gain = 6.8) (Figure 5A). Then the green and red images were merged; the colocalization of the green and red fluorescence appeared yellow (Figure 5E). The previously red punctate profiles noted earlier were now yellow and the other portions of the cell appeared green. Nonprotein sulfhydryl derivative of BODIPY dye was detected in a diffusely distributed pattern throughout the cytoplasm. Scanning through the entire cell at gain settings that optimized signal detection in the spheroidal inclusions (gain = 6.8; Figure 5A) indicated that in most cases there was an irregularly shaped spheroidal zone that contained much less fluorescent signal than did the surrounding cytoplasm. This zone had the conformation associated with the nucleus.

View larger version (19K): [in this window] [in a new window]   Figure 5.   Colocalization of the green cellular thiol dye BODIPY and the red mitochondrial dye TMR chloride, by confocal microscopy. The green (A) and red (B) images were collected simultaneously. Insertion of a 520-nm barrier filter removed all red emissions from the green channel (C) and totally obscured the red channel (D). The confocal images from TMR chloride and BODIPY were merged using 256 optimized colors, so that colocalization of the green and red fluorescence appears yellow (E).

This experiment was repeated on cells treated with two different concentrations of DEM to deplete GSH (Figures 6B and 6C). The wide variation in fluorescence within the population observed with the MClB experiments was also detectable by confocal microscopy using BODIPY (Figure 6A). DEM-depleted cells were similar to untreated cells and contained three detectable zones of fluorescence (Figure 6B). The least-fluorescent zone was the irregular spheroidal area roughly in the center of the cell. The next highest level of intensity was found in the surrounding cytoplasmic area, and the most intense fluorescence was observed in spheroidal or punctate areas, which were also labeled with TMR chloride. In the cells with the least signal, no areas emitted fluorescence for BODIPY. In cells containing slightly more signal, the central area and the majority of the cytoplasm contained no signal but there was a faint positive emission from the TMR-positive punctate areas. As with cells containing detectable fluorescence in the cytoplasm, the TMR-positive zone had even more intense fluorescence. Although detectable, these remaining signals were decreased from the control cells (Figure 6A), as evidenced by the need to increase the gain to allow imaging of these compartments (gain = 8.0). The nuclear compartment was unchanged (Figure 6B).

Incubation with GSH-e augmented the levels of BODIPY detectable signal. In control cells incubated with 1 mM GSH-e most fluorescent compartments appeared as large spherical bodies (Figure 6A). After depletion with 2.125 mM DEM, repletion with 1 mM GSH-e resulted in a significant level of labeling detectable in the large spherical compartment, and cytoplasmic levels appeared similar to control (gain = 6.0) (Figure 6B). Depletion with a higher concentration of DEM (4.25 mM) decreased the signal to a very low level (gain = 10) (Figure 6C), mostly apparent in the cytoplasm. Incubation with 1 mM GSH-e after depletion with 4.25 mM DEM elevated the signal only in the cytoplasmic compartment but not the spherical bodies (Figure 6C).

Heterogeneity

Imaging of the GS-bimane conjugate in isolated cells indicated that the GSH concentration in untreated cell populations was heterogenous (Figure 7A). A majority of the population (40%) contained between 0.01 and 0.49 fmol GS-bimane per cell, and approximately 78% of the cells contained less than 1.5 fmol GS-bimane per cell. A small percentage of the population (6%) contained 4.5 or more fmol GS-bimane per cell. This heterogeneity is also observed in Clara cells evaluated in situ in microdissected airway preparations. In proximal bronchi (Figure 8A), the signal had a wider range of intensities than in the terminal bronchioles in the same airway path (Figure 8B). Numerous cells in the epithelial lining of both airway generations had little or no detectable signal, whereas others contained amounts of conjugate in the range of intensity observed in the isolated cells. This variablity appeared to be greater in isolated Clara cells maintained in culture. In general, the GS-bimane signal was higher in cultured Clara cells than in explants or isolated Clara cells. On a cell-by-cell basis, the GS-bimane conjugate signal was heterogeneous after 4 (Figure 8C) and 7 d (Figure 8D) in culture.

View larger version (30K): [in this window] [in a new window]   Figure 7.   Percentage of isolated cells that contain 0 to 4.5 fmol or greater GSH-MClB/cell as determined by image analysis in control (A), DEM-depleted (B), or GSH-e-augmented (C) Clara cells. A total of 300 cells was analyzed from three isolation experiments for each treatment. (A) Clara cells incubated with 1 mM MClB for 1 h were analyzed by image analysis and placed into categories according to the level of GSH-MClB. Mean = 1.3 ± 0.5 fmol/cell. (B) Cells were incubated with 5 mM DEM for 1 h, then treated similarly to control cells. Mean = 0.34 ± 0.30 fmol/ cell. (C) Cells were incubated with 1 mM GSH-e for 1 h, then treated similarly to controls. Mean = 3.0 ± 1.9 fmol/cell.

View larger version (139K): [in this window] [in a new window]   Figure 8.   Comparison of GS-bimane intensity in epithelial cells lining intrapulmonary conducting airways of the mouse and Clara cells maintained in vitro. (A) Second-generation intrapulmonary bronchus. Compared with the terminal bronchiole (B), there was a wide range of signal in proximal airways. The cells with the highest signal (*) had more GS-bimane signal than did the bronchiole. There were also fewer low-signal cells (**). (B) Terminal bronchiole. Compared with proximal bronchi (A), the cells with the highest signal (*) were less intense and there was a greater abundance of cells with minimal signal (**) in distal bronchioles. (C and D) Clara cell cultures. A wide range of GS-bimane signal is apparent after both 4 (C) and 7 (D) d in culture. Clara cells with the highest signal (*) were also less abundant than cells with moderate to low (**) GS-bimane intensity.

GSH and Susceptibility

DEM treatment caused a decrease in the intracellular concentration of the GS-bimane conjugate throughout the entire population of cells (Figure 7B). In this group the majority of Clara cells (52%) contained the lowest concentration of the GS-bimane derivative (0.01 to 0.49 fmol/cell) compared with controls (38%) (Figure 7). Only 5% of the DEM-treated Clara cells contained concentrations of the GS-bimane conjugate of 3 fmol/cell or greater, compared with > 10% of control. Conversely, after treatment with 1 mM GSH-e, the proportion of the cell population containing the lowest concentration of conjugate (0.01 to 0.49 fmol/cell) decreased 40% from control, and the population of cells in the three highest concentration ranges (3.5 fmol/ cell and above) increased by > 300% (Figure 7C).

GSH was depleted by incubation of the cells with the aromatic hydrocarbon NA (Figure 9). More than 65% of the isolated cell population was characterized as having the lowest GS-bimane intensity value after a 90-min incubation with NA. At 3 h this low GS-bimane category contained > 75% of the isolated cell population. At 90 min of incubation, all the cells permeable for EthD-1 were in the low GS-bimane- intensity group; approximately 20% of the cells with this level of intensity were permeable. At 3 h, over 80% of the cells with the lowest level of GS-bimane intensity were EthD-1-permeable. Clara cells that retained higher intensity levels of GS-bimane were not EthD-1-permeable.

View larger version (26K): [in this window] [in a new window]   Figure 9.   To determine the relative GSH levels of susceptible Clara cells, isolated cells were treated with 0 and 0.5 mM NA for 1.5 and 3 h. Cells were then incubated with MClB and EthD-1 and imaged. The distribution of GS-bimane conjugate intensity was then plotted as a population histogram. After 1.5 h, > 65% of the cells had the lowest GS-bimane levels measured and 20% of these cells were EthD-1-permeable (inset). After 3 h of NA exposure, approximately 75% of the cells had lost their GSH and 80% of these cells were EthD-1-permeable (inset). No EthD-1- permeable cells having detectable GS-bimane conjugate were observed.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We conclude from this study that: (1) there is a wide variation in the steady-state intracellular GSH content between individual Clara cells; (2) intracellular pools of GSH within Clara cells are compartmentalized, and differences in these pools contribute to the heterogeneity between individual Clara cells; (3) heterogeneity of cellular GSH, in turn, appears to be a significant factor in determining the susceptibility of individual cells within the Clara-cell population to NA injury.

We used two strategies to define intracellular compartmentalization of GSH. First, using semiquantitative epifluorescence, the location of the nuclear compartment was compared with the high-intensity areas of GS-bimane signal in the cytoplasm. Second, colocalization of cellular thiols and mitochondria were defined in Clara cells by confocal microscopy. Recent work in hepatocytes has demonstrated a nuclear compartmentalization of transferase-accessible GSH as detected by MClB (21). A follow-up study using cultured hepatocytes concluded that identification of GSH in the nucleus might be an artifact of GS-bimane transfer after cytoplasmic conjugation (23). We have been unable to reproduce either finding in freshly isolated Clara cells. The majority of GS-bimane fluorescence is cytoplasmic (Figure 4E), and the most intense signal does not overlap with the area identified by nuclear DNA dyes. Our studies indicate BODIPY conjugates do not appear to label the nucleus but do present a diffuse labeling throughout the cytoplasm of all cells.

Cell fractionation studies have indicated that nuclear and mitochondrial pools of GSH may exist (24). The present study demonstrates that this compartmentalization of GSH may occur within individual Clara cells. We delineated three distinct sulfhydryl-containing regions. BODIPY-labeled cellular thiols were present as inclusions that filled the cytoplasm of most cells. These inclusions colocalized with TMR, a fluorescent dye previously shown to reflect mitochondrial oxidative function (25). Although fluorescent dyes such as TMR may well label lysosomes in addition to mitochondria, nonciliated (Clara) cells possess few such organelles (26, 27). Mitochondria, especially in mice, represent an overwhelming percent (> 30%) of the cytoplasm (26). Further, the TMR fluorescent signal from these colocalization studies appeared morphologically identical to mouse mitochondria (29, 30). This, in combination with the finding that we were able to modulate the BODIPY signal with compounds known to modulate GSH, supports the conclusions of our present work. When isolated Clara cells were treated with DEM, the mitochondrial thiol signal was nearly eliminated (Figure 6). This is in contrast to what occurs in mitochondria isolated from fractionated hepatocytes, where Meredith and Reed (4) previously reported that the hepatic mitochondrial GSH pool is long-lived (t_1/2 = 30 h) compared with cytosolic GSH (t_1/2 = 2 h). Cytosolic GSH levels decreased to 40% of control levels in hepatocytes incubated with DEM for 1 h, whereas the mitochondrial GSH level remained unchanged. This biphasic decline was also noted in hepatocytes under oxidative stress (31, 32). In contrast, the mitochondrial as well as cytosolic GSH pools in intact Clara cells were responsive to GSH modulators. DEM decreased both cytoplasmic and mitochondrial sulfhydryls. This difference may represent a difference in mitochondrial GSH transport rates. Given the close relationship between intracellular GSH and mitochondrial sulfhydryl pools in modulating intracellular oxidative stress (33), the apparent susceptibility of mitochondrial GSH in Clara cells may be a factor in the sensitivity to pulmonary cytotoxicants. By our measurements, cellular GSH levels appear to be highest in the mitochondria, next highest in the cytoplasm, and lowest in the nucleus. In Clara cells, differences in these pools also appear to be reflected by the wide heterogeneity in intracellular GSH.

To determine the individual variation on a cell-by-cell basis, isolated Clara cells and cells in microdissected airways and in culture were compared by imaging. We determined the actual concentrations of GSH associated with relative amounts of fluorescence on a per-cell basis by evaluating aliquots of the same cells by two methods. Single-cell GSH content was determined by cellular conjugation of GSH with MClB, followed by concurrent image analysis; HPLC analysis of the GS-bimane conjugate yielded a quantitative measurement of average cellular GSH concentration. Analysis of the isolated Clara cell population by epifluorescent imaging showed much greater diversity in GSH levels between cells than has been found in tumor cell lines (34). By applying the fluorescent imaging approach to intact bronchioles and Clara cells in culture, we found this to be true for Clara cells under a variety of conditions and also confirmed previous histochemical studies, which showed that this wide range of variability is present in Clara cells (35). Although the average steady-state GSH pool is approximately 1 fmol GSH per Clara cell, there is more than a 4-fold variation between individual cells. Most of the heterogeneity in the isolated cells appears to result from the existence of two subpopulations of cells. One subpopulation contains low steady-state levels of GSH, which could be elevated by GSH-e, and another subpopulation contains high steady-state levels of GSH. GSH depletion by DEM decreased GSH content by at least 60% when averaged over the whole population by HPLC analysis, but imaging experiments indicated that all cells within the population did not respond uniformly. Previous studies using GS-bimane conjugates to label tumor cells for flow cytometry (34, 36, 37) and dye-transfer studies (38, 39) have also described differences between cells on the basis of changes in relative fluorescence. Our studies now establish the wide variability between Clara cells by determining GSH concentrations of individual cells.

The heterogeneity in GSH content in Clara cells appears to modulate the susceptibility to NA injury. Earlier work showing that DEM-pretreated mice were substantially more susceptible to NA-induced Clara cell injury supported the importance of GSH in the detoxification of reactive NA metabolites (6). Treatment of isolated cells with NA caused GSH levels to decrease, as measured by semiquantitative epifluorescence. Likewise, we have found that decreases in intracellular GSH precede a decrease in Clara cell viability (Figure 9). Subpopulations of Clara cells with low levels of intracellular GSH are the only cells susceptible to injury by NA, as judged by loss of membrane integrity.

In summary, the heterogeneity in GSH content of the Clara cell subpopulations within the respiratory tract may be related to the differing cell susceptibility in response to pulmonary cytotoxicants (12). At low doses, compounds such as NA injure a small population of Clara cells at a selective site (terminal bronchioles) within the lung. Higher doses injure a greater proportion of the Clara cells within the terminal bronchiole and extend the injury to Clara cells in more proximal airways. Even at LD₅₀ doses of NA, not all Clara cells lining the airway tree are injured. When Clara cells are isolated from throughout the airway tree, as performed for this study, they maintain the same range of susceptibility to injury by NA and its metabolites (40). The current study shows that Clara cell populations maintain a heterogeneous distribution of GSH concentration and a wide variation in response to GSH depletion and augmentation. This emphasizes the relationship between the heterogeneity of the GSH pool and the susceptibility to cytotoxicants. Of particular significance is the sensitivity of the fluorescence assays used in the present study to evaluate the GSH concentration on a single-cell basis to gain an understanding of the differences that exist between individual cells within the same population. The demonstration of mitochondrial GSH and its modulation strengthens the hypothesis that compartmentalization of intracellular GSH pools may be involved to a certain extent in the mechanism of pulmonary cellular toxicity by bioactivated compounds.