Introduction

Abstract Introduction Materials and Methods Results Discussion References

The glutathione S-transferases (GSTs) are a family of multifunctional proteins that are widely distributed and found in diverse species, including plants (1), insects (2), fish (3), birds (4), and mammals (5). These proteins have a role in defending cells against potentially toxic and/or carcinogenic compounds by catalyzing the conjugation of glutathione (GSH) to electrophilic metabolites (6, 7). Conjugation of an electrophile with GSH usually yields a product with decreased reactivity, and therefore generally results in detoxification by reducing or ameliorating covalent binding of reactive intermediates to cellular constituents; this binding is a critical step leading to cellular cytotoxicity and/or carcinogenesis (8). Accordingly, the susceptibility of cells and tissues to cytotoxicities mediated by electrophilic metabolites depends in part on the availability of enzymes, including the GSTs, for conjugation reactions, and hence for detoxification. However, detoxification can also be mediated through a nonenzymatic reaction involving binding of GSTs to electrophiles (11, 12). In this context, it should be noted that in some instances, GST activity produces an intermediate that is more reactive than the parent compound or the primary metabolite, and that conjugation is therefore associated not with detoxification but with toxification (13, 14).

A number of cytosolic GST isozymes have been purified from rat and human tissues, and on the basis of their primary structures are categorized into five separate families, designated as classes alpha, mu, pi, sigma, and theta (5, 15, 16). The GSTs are composed of two subunits and exist as either homo- or heterodimeric proteins. On the basis of decreasing electrophoretic mobilities, three protein bands were resolved and were designated Ya, Yb, and Yc (17). Subsequent studies showed that the Ya and Yc bands represent class alpha GST, whereas the Yb band represents class mu GST (18, 19). These findings were used to devise a class-based nomenclature for GST subunits (20). An additional nomenclature has been proposed by Jakoby and colleagues (21), in which Arabic numerals are used to designate the subunits. In this nomenclature, capital letters are used to represent the alpha (A), mu (M), pi (P), sigma (S), and theta (T) classes, and Arabic numerals are used to represent each of the gene products; for example, class alpha subunits are designated A1, A2, A3, and so forth. The GST heterodimer is signified by hyphenated numerals; for example, the alpha class heterodimer Ya₁Yc₁ is referred to as GSTA1-3.

Expression of GST enzymes in experimental animals exhibits tissue specificities (22), and of importance in this context are the distribution and localization of various GST classes in cell types that reside in particular tissues. This is of special relevance in heterogeneous tissues, such as the lung, in which bioactivating enzymes, including cytochrome P450, are present in a restricted number of cell types. Several studies have examined the cellular localization of various classes of GST in the lung, mainly in the rat. Immunohistochemical studies in rat lung detected class alpha GST (YaYc) throughout the bronchiolar epithelium and alveolar wall (23). However, specific cell types that expressed this GST class were not identified. Other studies, also with rat lung, reported that alpha class GST was localized to the greatest extent in Clara cells of bronchioles and to a lesser extent in alveolar type II cells (24). More recent studies focused on the bronchiolar epithelia of rat and murine lung, and produced data showing that alpha, mu, and pi GST classes were all present in Clara and ciliated cells of rat and murine lung (25). These findings indicated that Clara cells are the predominant sites of GST isozymes; in contrast, there is a paucity of data regarding localization of GST classes in other lung cell types. Data are also unavailable on the potential for inducibility of different GST classes, or for the cell types in which their induction is manifested.

In the present study, we investigated the cellular expression of the GST classes alpha (Ya), mu (Yb₁), and pi (Yp) in murine lung. We also investigated expression of these GST classes under both constitutive and induced conditions, and used as an inducing agent 2(3)-tert-butyl-4-hydroxyanisole (BHA), a phenolic dietary antioxidant that has been reported to induce GST enzymes (25). Because of variability in the findings in reported studies, we used both immunochemical and in situ hybridization techniques to identify more precisely lung cell types that expressed GST protein and messenger RNA (mRNA) for the specific GST classes.

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Animals

Female mice of the SWR strain, weighing 20 to 25 g, were purchased from Charles River Canada (St. Constant, PQ, Canada), and were housed over hardwood bedding (Beta Chip; Northeastern Products Corp., Warrensburg, NY). The mice were maintained on a 12-h light/dark cycle, and were freely provided with food (Purina Rodent Chow; Ralston Purina International, Strathroy, ON, Canada) and drinking water. They were acclimatized to laboratory conditions for at least 7 d before experimental procedures were initiated. The mice were treated intraperitoneally with BHA (Sigma Chemical Co., St. Louis, MO) in corn oil at a dosage of 0.15 mg/g body weight). Mice were given BHA once daily for three consecutive days, and were killed 24 h after the last dose. Control mice received only the vehicle.

Preparation of Cytosol and Nuclear Extracts

Mice were killed by cervical dislocation and their lungs were excised, rinsed, blotted, and weighed. For preparation of cytosol, lung tissues from six mice from either the control or BHA-treated group were pooled and homogenized in ice-cold phosphate-buffered KCl (140 mM KCl, 100 mM K₂HPO₄, 1.5 mM ethylenediamine tetraacetic acid [EDTA], pH 7.4) with a tissue homogenizer. The homogenate was centrifuged at 9,000 × g at 4°C for 20 min, after which the supernatant was removed and centrifuged at 105,000 × g at 4°C for 40 min. Aliquots (150 µl) of the supernatant were dispensed into Eppendorf tubes, rapidly frozen in liquid nitrogen, and stored at 70°C.

Lungs of 20 mice from either the control or BHA-treated group were used for extraction of nuclear proteins. The lungs from each experimental group were pooled, and the nuclear extract was prepared according to a modification of procedures used by Hattori and coworkers (26) and described in our previous studies (27). The lungs were homogenized in 4 volumes of cold buffer A (50 mM Tris-HCl, 0.3 M sucrose, 20 mM KCl, 5 mM MgCl₂, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride [PMSF], 1 mM dithiothreitol [DTT], and 2 µg/ml aprotinin, pH 7.4). The homogenates were filtered and diluted with 2 volumes of cold buffer B (2.2 M sucrose, 50 mM Tris-HCl, 20 mM KCl, 5 mM MgCl, and 1 mM EDTA), yielding a suspension containing a final concentration of 1.56 M sucrose. The homogenate was gently laid over a volume of 3 ml of buffer B and centrifuged at 24,000 rpm for 50 min at 1°C. After centrifugation, the supernatant was removed and the remainder was drained from the nuclear pellet by inverting the centrifuge tubes for 10 min at 4°C. The nuclear pellet was resuspended in 1.0 ml of buffer A containing 0.5 M sucrose, and was centrifuged at 14,000 rpm for 5 min at 4°C. The supernatant was discarded and the pellet was resuspended in 100 µl 25 mM Tris-HCl, pH 7.5, containing 2 mM DTT. An equal volume of 25 mM Tris-HCl containing 800 mM KCl was then added dropwise to this suspension. Extraction of nuclear proteins was done on ice for 60 min. The extracted mixture was then centrifuged at 100,000 × g for 40 min at 4°C. Aliquots of the supernatant, termed the nuclear extract, were stored at 70°C. Protein concentrations were determined by the method of Lowry and associates (28).

Protein Immunoblotting

Protein immunoblotting was performed as described in our previous studies (29). Lung cytosolic proteins (20 to 80 µg) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a 12.5% gel. The proteins were transferred to a 0.45-µm nitrocellulose membrane for 1 h at ~ 35 V in 25 mM Tris-HCl, 192 mM glycine, pH 8.3, and 20% (vol/vol) methanol, using the method of Tobin and colleagues (30). The membrane was immersed for 2 h in a blocking solution consisting of 5% nonfat dried milk in 20 mM Tris-HCl and containing 500 mM NaCl, pH 7.5. It was then incubated overnight at room temperature with a rabbit polyclonal antibody prepared against rat liver cytosolic GST alpha (Ya, 1:1,000), GST mu (Yb₁, 1:500), or mouse liver GST pi (Yp, 1:1,000) (Biotrin International, Inc., Dublin, Ireland). Our preliminary immunohistochemical studies revealed that GST Ya was localized in the cytoplasm as well as in the nuclei of lung cells. As a result, we additionally performed immunoblotting with proteins extracted from the nuclei of cells. Protein bands that bound the antibody were detected by incubation with goat antirabbit IgG conjugated to alkaline phosphatase, with color development in a solution containing p-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt (BRL Life Technologies, Inc., Gaithersburg, MD). Molecular weight standards (Amersham Life Science, Oakville, ON, Canada) were used to estimate the apparent molecular weights of the protein bands.

Tissue Preparation

Mice were anesthetized with sodium pentobarbital (0.12 mg/g body weight, given intraperitoneally) before being killed. The thorax was exposed; a cannula was inserted into the pulmonary artery through the right ventricle, and an incision was made in the right atrium to permit outflow of the perfusate. The lungs were flushed with physiologic saline until the perfusate became clear. Perfusion was initiated with 4% paraformaldehyde in 0.1 M Sorenson's phosphate buffer (69.0 mM Na₂HPO₄, 12.0 mM NaH₂PO₄), pH 7.4. The lungs were then inflated by intratracheal instillation of 0.4 ml of fixative. After ligation of the trachea, the lungs were removed en bloc with the heart and immersed in the same fixative for 2 h at room temperature. The lung lobes were then separated and extraneous tissues were removed. The separated lobes were then placed overnight in fresh fixative at 4°C. Subsequently, the tissues were rinsed in phosphate-buffered saline (PBS), dehydrated, cleared, and embedded in paraffin, using standard procedures. Tissue sections (5 µm) were adhered to acid-cleaned and gelatin-coated slides and used for the immunohistochemical and in situ hybridization experiments.

Immunohistochemical Procedures

Immunohistochemical localization of the Ya, Yp, and Yb₁ classes of GSTs was performed with the avidin-biotin complex (ABC) technique, and with the same antibodies as used for protein immunoblotting. Tissue sections were deparaffinized by placing the slides in an oven at 60°C for 10 min, and were then cleared in Histoclear and hydrated in a graded ethanol series. The sections were rinsed in PBS and treated with 5% normal goat serum for 20 min to block nonspecific binding of the antibody. The sections were subsequently rinsed in PBS and incubated for 60 min with the primary antibody in PBS containing 2.5% normal goat serum. Reactions with the primary antibody were conducted at dilutions ranging from 1:100 to 1:1,600 to identify the optimal antibody concentration for staining. The sections were then rinsed thoroughly in PBS to remove any unbound primary antibody, and were incubated for 30 min with a biotinylated goat antirabbit secondary antibody (Vector Laboratories, Burlingame, CA). Endogenous peroxidase activity was blocked by immersing the tissue sections for 30 min in a 1% hydrogen peroxide solution in nanopure water. The ABC was prepared according to procedures detailed by the manufacturer (Vector Laboratories), and was allowed to react with the tissue sections for 30 min. Following this, the immunoperoxidase color reaction was developed by incubation for 4 min in a solution containing 0.05% 3,3'-diaminobenzidine tetrahydrochloride and 0.01% hydrogen peroxide in PBS. The reaction was terminated by rinsing the sections in tap water for 5 min. The tissue sections were then incubated in a 0.5% copper sulfate in 0.15 M sodium chloride solution for 5 min, and selected sections were counterstained with hematoxylin. After rinsing in tap water, tap water substitute, and tap water again, the sections were dehydrated, cleared, and coverslipped through the use of Pro-Texx (American Scientific Products, McGraw Park, IL). Controls for specificity of the antibody reaction included incubations performed in the presence of an irrelevant antibody or incubations in which the specific antibody was omitted.

Synthesis of Oligonucleotide Probes and Labeling

In situ hybridization experiments were performed with oligonucleotide probes prepared for the three classes of GST isozymes investigated in this study. For pi-1, the antisense probe consisted of 5'-CGTAGACAGAGGGGTACTCAGAGTGAGG-3', which was complimentary to the selected RNA sequences between nucleotides 1 and 28 (31). The antisense probe for Ya consisted of 5'-CTGTTGCCCACAAGGTAGTCTTG-3', which was complimentary to the selected RNA sequences between nucleotides 266 and 288 (32). For GST Yb₁, the antisense probe consisted of 5'-GGCCCAGCTTGAACTTCTCATTCAG-3', which was complimentary to the selected RNA sequences between nucleotides 151 and 175 (33). Sense probes for the GST isozymes were synthesized and used as controls for the hybridization experiments. The oligonucleotide probes were synthesized at the Oligonucleotide Synthesis Laboratory of the Department of Biochemistry of Queen's University. For the in situ hybridization experiments, 3' end-labeling of the oligonucleotides with [³⁵S]deoxyadenosine triphosphate ([³⁵S]dATP) (Dupont Canada [NEN] Ltd., Mississauga, ON, Canada) was carried out with terminal deoxynucleotidyl transferase (TdT) (GIBCO BRL, Canadian Life Technologies, Burlington, ON, Canada). Approximately 130 ng/µl of oligonucleotide was incubated for 1.5 h at 37°C in a labeling solution containing 5 µl of 5× cobalt reaction buffer (20% vol/vol), 8 µl of diethyl pyrocarbonate (DEPC)-treated water (32% vol/vol), 10 µl of [³⁵S]dATP (~ 125 µCi, 40% vol/vol), and 2 µl of TdT solution (24 U, 8% vol/vol). The labeled probe was purified by passage through a Nensorb 20 column (Dupont Canada [NEN]); 5 µl of 1 M DTT was added to the eluent containing the radiolabeled probe, and the probe was stored at 20°C.

In Situ Hybridization

Tissue sections were deparaffinized by placing slides in an oven set at 60°C for 10 min, cleared in Histoclear, and hydrated in a graded ethanol series. The tissue sections were then fixed in 4% paraformaldehyde for 15 min to increase their adherence to the slides. The slides were next immersed in 0.2 N HCl for 20 min and rinsed in nanopure water. The tissue sections were immersed for 15 min in PBS containing 0.3% Triton X-100, after which they were incubated for 15 min in a 1.5 mg/ml proteinase K solution in 5 mM Tris/1 mM CaCl₂, pH 7.5. After three rinses with nanopure water, the slides were acetylated for 10 min in a solution containing 0.1 M triethanolamine, pH 8, and 0.25% (vol/vol) acetic anhydride. Following this step, the slides were dehydrated in 50% ethanol (two immersions of 5 min each) and 95% ethanol (one immersion for 5 min). Hybridization was conducted overnight at 42°C under a siliconized coverslip. The hybridization solution contained the following: 10 mg/ml sheared, boiled, and chilled salmon sperm DNA; 40 µl 5 M DTT; 50% (vol/vol) deionized formamide; 20 ml of 20× standard saline citrate (SSC) (54% vol/vol; 1× SSC = 0.15 M NaCl, 0.015 M sodium citrate, pH 7.4); 2 ml of 50× Denhardt's solution (5% vol/vol); 10 ml 0.2 M sodium phosphate buffer (27% vol/vol), pH 7.0; 10 g dextran sulfate; 5 ml 20% sarkosyl (N-lauroylsarcosine 13.5% vol/vol); and ~ 10⁶ cpm radiolabeled probe for each slide. After hybridization, the slides were rinsed twice at room temperature for 20 min each in 4× SSC containing 2-mercaptoethanol (0.1%); once in 2× SSC containing 2-mercaptoethanol (0.1%) at 50°C; four times at 50°C for 20 min each in 1× SSC containing 2-mercaptoethanol (0.1%); once in 1× SSC at 55°C for 20 min; once in 0.5× SSC at 50°C for 20 min; once in 0.25× SSC at 45°C for 20 min; once in nanopure water at 30°C for 1 min; and four times for 5 min each in nanopure water at room temperature. Tissue sections were then allowed to air-dry for at least 1 h. Controls for the specificity of the oligonucleotide probes consisted of incubation of tissue sections with sense oligonucleotide probes. The slides were subsequently dipped in Kodak NTB2 nuclear track emulsion (Eastman Kodak Co., Rochester, NY), diluted 1:1 with 0.6 M ammonium acetate, and exposed for 10 to 16 d at 4°C. The slides were then developed with D19 developer (Eastman Kodak), fixed with 30% sodium thiosulfate for 5 min, counterstained with hematoxylin, mounted, and coverslipped with Permount.

Quantitative Image Analysis

Data were obtained from measurements made with a Pentium 90 mHz computer-based video densitometry system, using the commercial software package MCID M2 (Microcomputer Imaging Device; Imaging Research Inc., Brock University, St. Catharines, ON, Canada). The computer was interfaced with a Reichert microscope and a video camera that was used to capture images for subsequent data acquisition. After thresholding was performed to enhance the signal-to-noise ratio, the computer counted the number of pixels associated with individual silver grains, and the total number of pixels in a geometric area was determined. The fraction of the selected area occupied by silver grains was identified, and the percentage of the area occupied by silver grains was subsequently calculated. Because alveolar spaces occupy a significant portion of the lung parenchyma, measurements were made to determine the total area occupied by the tissue component, and a correction factor was generated to identify the amount of lung tissue that was present in a circumscribed area. This was determined to be about 22% of the total area measured. The areas occupied by airway epithelium and parenchyma were quantitated by using a cursor to delineate the tissue structures and then measuring their areas. The amounts of silver grains representing GST Yp, Ya, and Yb₁ mRNA transcripts were expressed as the percentages of tissue areas that were occupied by grains. Measurements were obtained from 10 to 12 fields (magnification: ×400) of lung sections mounted on each slide; each slide contained tissue sections from all lung lobes.

Statistical Analysis

Data were analyzed with Kruskal-Wallis one-way analysis of variance (ANOVA) on ranks. All pairwise multiple comparisons were performed according to Dunn's method, and significant differences between experimental groups were determined at P < 0.05.

	Results

Abstract Introduction Materials and Methods Results Discussion References

Protein Immunoblotting

Protein immunoblotting performed with lung cytosolic fractions isolated from untreated and BHA-treated mice showed that the rabbit polyclonal antibodies raised against murine hepatic cytosolic Yp and rat hepatic cytosolic Ya and Yb₁ cross-reacted with lung cytosolic proteins. Single protein bands of approximately 28, 29, and 31 kD were detected for the GST subunits Yp, Ya, and Yb₁, respectively (Figure 1). The reactivities of the protein bands for both the Yp and Ya subunits were greater in protein samples from BHA-treated mice than in those from control mice (Figures 1A and 1B). However, reactivity for the Yb₁ subunit in BHA-treated mice differed to only a small degree from that in the lungs of untreated mice (Figure 1C). Hence, differences in reactivities between lung cytosolic proteins from untreated and BHA-treated mice were more apparent for the Ya and Yp subunits than for the Yb₁ subunit.

View larger version (33K): [in this window] [in a new window]   Figure 1.   Protein immunoblotting of lung cytosol for GST Yp, Ya, and Yb1. Cytosolic proteins from control or BHA-treated mice were electrophoretically separated on a 12.5% SDS polyacrylamide gel and were transferred to nitrocellulose membranes. The membranes were reacted with one of the following rabbit polyclonal antibodies: anti-mouse-liver cytosolic Yp (A), and anti-rat-liver cytosolic Ya (B) and Yb1 (C). The membranes were subsequently incubated with goat antirabbit IgG conjugated to alkaline phosphatase. (A and B) Lane 9: prestained molecular weight standards. Lanes 1, 3, 5, and 7, and lanes 2, 4, 6, and 8 were loaded with cytosolic proteins from lungs of BHA-treated and untreated mice, respectively. Protein contents of the lanes were as follows: lanes 1 and 2, 80 µg; lanes 3 and 4, 60 µg; lanes 5 and 6, 40 µg; lanes 7 and 8, 20 µg. (C) Lane 1 contained prestained molecular standards. Lanes 2, 4, and 6, and lanes 3, 5, and 7 contained cytosolic proteins from BHA-treated and untreated mice, respectively. Lanes were loaded as follows: lanes 2 and 3, 20 µg; lanes 4 and 5, 40 µg; lanes 6 and 7, 60 µg.

Immunoblots prepared from nuclear proteins extracted from lungs of untreated and BHA-treated mice reacted positively with the anti-Ya antibody. A protein species of about 29 kD was detected, whose molecular weight was similar to the apparent molecular weight of the cytosolic proteins (Figure 2). Immunoreactivity was slightly greater in protein bands containing nuclear extracts from BHA-treated mice than in those from untreated mice. However, the discrepancy in reactivity between cytosolic proteins from untreated and BHA-treated mice was greater than that obtained with the nuclear proteins.

View larger version (28K): [in this window] [in a new window]   Figure 2.   Protein immunoblotting of cytosolic and extracted nuclear proteins from lungs of untreated and BHA-treated mice. Cytosolic and nuclear proteins were electrophoretically separated on a 12.5% SDS polyacrylamide gel and were transferred to a nitrocellulose membrane. The membrane was reacted with a rabbit anti-rat-liver Ya polyclonal antibody, and was subsequently incubated with a goat antirabbit IgG conjugated to alkaline phosphatase. Lanes were loaded as follows: lane 1, molecular weight standards; lane 2, nuclear proteins from untreated mice; lane 3, nuclear proteins from BHA-treated mice; lane 4, cytosolic proteins from untreated mice; lane 5, cytosolic proteins from BHA-treated mice. All of the lanes were loaded with 65 µg of protein.

Immunohistochemical Distribution of GST Yp, Ya, and Yb₁

The cell distribution of GST Yp, Ya, and Yb₁ proteins was determined immunohistochemically in lung tissue sections from untreated and BHA-treated mice. Optimal staining was obtained at antibody dilutions of 1:800, 1:600, and 1:200 for the Ya, Yp, and Yb₁ subunits, respectively. Immunoreactivity was observed in lungs of both control and BHA-treated mice; the pattern of distribution of lung cells that were labeled was similar in tissues from the two experimental groups (Figure 3). All staining for Ya (Figure 3a), Yp (Figure 3c), and Yb₁ (Figure 3e) was localized in bronchiolar epithelial cells, and was highly concentrated in Clara cells. Relative staining levels for all three GST subunits in lung tissue from untreated mice were not markedly different from one another, but varied in tissue sections from BHA-treated mice. Based on staining properties and the amounts of antibodies required to produce optimal immunohistochemical staining, increased reactivity induced by BHA was most apparent for the Ya (not shown) and Yp subunits (Figure 3c), but was low for the Yb₁ subunit (Figure 3e). The increased labeling was manifested predominantly in the bronchiolar epithelium and in Clara cells. Immunoreactivity was also found in the parenchyma, and staining for Ya (Figure 3b), Yp (Figure 3d), and Yb₁ (Figure 3f) was observed to the greatest extent in alveolar type II cells in the lungs of both untreated and BHA-treated mice. In addition, Ya staining was seen in alveolar type I and capillary endothelial cells (Figure 3b). However, magnitudes of staining for all three GST subunits were greater at all times in Clara than in type II cells. In the control incubations, staining was negligible in lung sections incubated in the presence of a nonspecific antibody or in reactions in which the specific antibody was omitted (not shown).

View larger version (144K): [in this window] [in a new window]   Figure 3.   Distribution of GST Ya (a and b), Yp (c and d), and Yb1 (e and f ) subunits in lung tissue from untreated and BHA-treated mice. (a) Localization of Ya staining in the bronchiolar epithelium of untreated mice. (b) Alveolar type II cells were reactive, with staining that was seen in both the nucleus and cytoplasm (arrows). Staining was also localized in alveolar type I and endothelial cells (arrowheads). (c) Labeling for Yp was localized in the bronchiolar epithelium and concentrated in the Clara cells (arrow) of BHA-treated mice. (d) In lung parenchyma, alveolar type II cells contained staining in both the cytoplasm (arrowheads) and nucleus (arrows). (e) Staining for the Yb1 subunit in the bronchiolar epithelium of BHA-treated mice. (f ) In the parenchyma, staining was observed mainly in the cytoplasm (arrowheads) of type II cells. Bars = 20 µm.

Subcellular localization of the GST subunits in the bronchiolar epithelium was heterogeneous, and was found in both the nuclear and cytoplasmic compartments. No unique pattern of distribution for individual subunits was apparent in the bronchioles. The subcellular distribution of the subunits in alveolar type II cells was more distinct than in bronchiolar epithelial cells. Staining for the Ya and Yp subunits occurred in both the nuclei and cytoplasm of type II cells (Figures 3b and 3d). However, the nuclear staining for Ya was more prominent than that for the Yp subunit. On the other hand, staining for Yb₁ was predominantly cytoplasmic (Figure 3f). Treatment of mice with BHA did not alter patterns of subcellular staining in lung cells as compared with those seen in untreated mice.

In Situ Hybridization: Distribution of GST Yp, Ya, and Yb₁ mRNA

In situ hybridization experiments with lung sections done with the antisense ³⁵S-labeled oligonucleotide probes for GST Yp, Ya, and Yb₁ revealed specific localization of silver grains representing individual GST mRNAs. Grains specific for all three GST subunits were highly concentrated over bronchiolar epithelial cells (Figures 4a, 4c, and 4e). Grains were also localized in alveolar septa, and were numerous over alveolar type II cells (Figures 4b and 4d). In the controls, hybridization of lung sections with the sense probe yielded low background labeling, with no specific distribution of silver grains over either the bronchiolar epithelium or alveolar septal cells (Figure 4f).

View larger version (146K): [in this window] [in a new window]   Figure 4.   Distribution of GST Ya (a and b), Yp (c and d), and Yb1 (e and f ) mRNA transcripts in lung sections from untreated and BHA-treated mice. Silver grains representing Ya mRNA transcripts were concentrated in the bronchiolar epithelium (a) and alveolar septa in lung parenchyma of BHA-treated mice (b). Numerous grains were localized over alveolar type II cells (arrowheads). (c) Transcripts for the Yp subunit were also localized in the bronchiolar epithelium of untreated mice. (d) In lung parenchyma of BHA-treated mice, grains associated with Yp mRNA transcripts were located in type II cells (arrowheads). (e) Grains representing Yb1 transcripts were found in the bronchiolar epithelium of BHA-treated mice. (f ) Grains were sparse in bronchiolar epithelium of control lung section from BHA-treated mice. Bar = 20 µm.

Quantitative Image Analysis

Quantitative image analysis was performed on lung sections hybridized with ³⁵S-labeled oligonucleotide probes to determine the relative distribution of silver grains representing mRNA transcripts specific for GST Ya, Yp, and Yb₁. The analysis was performed on two major lung tissue areas, which were demarcated into bronchiolar epithelial and alveolar or parenchymal regions. The data are summarized in Figure 5. Silver grains were most abundant in the bronchiolar epithelium, and in the lungs of untreated mice, levels were comparable for all three GST subunits examined. In BHA-treated mice, quantities of silver grains were highest in the bronchiolar epithelium in lung sections hybridized with the Ya oligonucleotide. Significantly increased numbers of grains were also detected for Yp in the bronchiolar epithelium. However, the BHA-induced effect in bronchioles was weaker for the Yb₁ subunit than for either the Ya or Yp subunit. In the parenchymal portions of lung tissue from untreated mice, the relative abundance of grains was similar for all GST subunits. Treatment with BHA produced increases in areas occupied by grains in the parenchyma for all subunits. Thus, the relative amounts of grains corresponding to Ya and Yp induced by BHA were maximal in the bronchioles, and were most pronounced for Ya. The effects of BHA treatment on mRNA transcripts in the parenchyma were less striking.

View larger version (30K): [in this window] [in a new window]   Figure 5.   Relative content and distribution of GST mRNA in lung parenchyma (Par) and bronchioles (Bron) of control and BHA-treated mice. Quantities of silver grains corresponding to GST Ya, Yp, and Yb1 mRNA transcripts were measured through computerized image analysis. Data are expressed as mean ± SD of the percentage of areas containing mRNA transcripts. All pairwise multiple comparisons were performed according to Dunn's method, and significant differences between groups were set at P < 0.05. a: Significantly different from lung parenchyma of control mice for each GST subunit. b: Significantly different from bronchioles of control mice and lung parenchyma of BHA-treated mice. c: Significantly different from Yp and Yb1 in bronchioles of BHA-treated mice. d: Significantly different from Yb1 in bronchioles of BHA-treated mice. e: Significantly different from Yb1 in bronchioles of untreated mice.

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

The importance of adopting a cellular approach in metabolic studies of the respiratory system is well established, primarily because of the heterogeneity of cell types that reside in its tissue and the cell-specific nature of cytotoxic events that are manifested in it. In accord with this premise, the cell localization and distribution of the GST family of enzymes in lung tissue have been examined. However, because of previous lack of uniformity in GST nomenclature, ambiguities exist with regard to identification of individual GST classes and subunits within tissues; such difficulties may also be ascribed to the heterodimeric composition of many GST isozymes, and to the lack of monospecificity of antibody reagents prepared against GST proteins. Nevertheless, data are increasingly available for comparative purposes and for understanding the underlying basis of the cell-selective damage caused by chemical xenobiotics, particularly in tissues such as the lung.

Previous studies have detected proteins of the alpha class of GST in bronchial and bronchiolar epithelial and alveolar-wall cells in untreated rats (24). The Ya₁ protein was found to the greatest extent in Clara cells (24). In other studies, a polyclonal antibody prepared against liver isolates, and which recognizes the alpha and mu classes of hepatic GST isozymes, identified homologous GST proteins in rat lung that were localized in Clara and alveolar type II cells (34). A Clara-cell location has also been reported for GST-P, a homodimer (YpYp) of the pi class in rat lung. This is a form of GST that is highly localized in human lung and is found in the bronchial and bronchiolar epithelium (35). The Yb₁ subunit is also detected, but in low amounts, in the bronchiolar epithelium, including Clara and ciliated cells (25). These findings suggested that the application of monospecific antibodies that recognize individual GST subunits in immunochemical and immunohistochemical studies of lung tissue is desirable.

In the present study, protein immunoblotting confirmed that each of the antibody reagents used detected a single protein species of appropriate molecular mass; the proteins were of apparent molecular weights of 28, 29, and 31 kD for the GST Yp, Ya, and Yb₁ subunits, respectively, in control and BHA-treated mice (Figure 1). These results supported the specificity of our immunohistochemical experiments showing localization of Ya, Yb₁, and Yp subunits in both bronchiolar epithelial and parenchymal regions of the lung. Immunoreactivities for all subunits were markedly higher in the bronchiolar epithelium than in the parenchyma. In the bronchioles, Clara cells were labeled to the greatest extent in both untreated and BHA-treated mice (Figures 3a, 3c, and 3e). The magnitudes of bronchiolar labeling for Ya, Yb₁, and Yp subunits were variable, with amounts of staining in the rank order Ya > Yp > Yb₁. In lung parenchyma, alveolar type II cells were highly labeled for all GST subunits (Figures 3b, 3d, and 3f). Staining was also found for the Ya subunit in alveolar type I and capillary endothelial cells, but at lower levels (Figure 3b). These immunohistochemical findings were corroborated by mRNA data from our in situ hybridization experiments done with synthesized oligonucleotides having sequences unique for the Ya, Yp, and Yb₁ subunits. The localization of mRNA transcripts specific for the GST subunits coincided with the sites at which the respective proteins were found, and silver grains were most abundant in the bronchiolar epithelium in both untreated and BHA-treated mice (Figures 4a, 4c, and 4e). Grains were also localized in the alveolar septa, and were most prominent in type II cells (Figures 4b and 4d). Measurement of mRNA transcripts by computerized image analysis confirmed that relative quantities of mRNA transcripts in the bronchioles were consistent with the rank order defined for the subunit proteins: Ya > Yp > Yb₁ (Figure 5). Levels of mRNA transcripts in the parenchyma were significantly lower than in bronchioles, and differences between the subunits were not observed. These results showed colocalization of subunit proteins and mRNA species at the same cellular sites. Furthermore, there was good agreement in the relative extents to which the proteins and mRNA transcripts were expressed in the different lung regions. These findings strongly supported the specificity of the identities of the lung cells containing the specific GST subunits.

The subcellular location of cytosolic GST subunits has been examined in previous studies in rat and murine lung (25, 34). In an ultrastructural study of the bronchiolar epithelium, labeling for Ya, Yp, and Yb₁ was observed in the nuclear and cytoplasmic compartments of Clara and ciliated cells (25). Comparable levels for the three subunits were detected in the nuclei as well as in the cytoplasm of these cells, although the content of Yp was slightly higher in murine than in rat lung. Other studies reported that GST-P was present in the nuclei and cytoplasm of Clara and ciliated cells (34). Our immunohistochemical data are generally in agreement with these findings, and showed that Ya and Yp are localized in the nucleus and cytoplasm (Figures 3b and 3d), whereas Yb₁ resides mainly in the cytoplasm (Figure 4f). Because Ya appeared to be strongly expressed in nuclei, we were interested in ascertaining the identity of the nuclear Ya protein and its potential inducibility. Protein immunoblotting of nuclear proteins extracted from lung cells confirmed that Ya is a protein of about 29 kD, and is similar in apparent molecular weight to the cytoplasmic species (Figure 2). These results indicated that the antibody that detected cytosolic Ya also recognized the same epitopes in the nucleus, suggesting that the same subunit resides in both nucleus and cytoplasm. More significantly, the nuclear Ya was inducible by BHA treatment, as was the cytoplasmic Ya subunit. To the best of our knowledge, this represents the first report of inducibility of the Ya subunit within the nuclei of lung cells. The localization of GST in the cell nucleus is not unexpected, since previous studies have identified a nonhistone-DNA-binding protein from the nucleus of the rat as GST YbYb; this GST was localized in discrete interchromatic nuclear domains (38). Interestingly, when this GST is isolated from cell nuclei and injected into the cytoplasm, it was found to translocate rapidly into the nucleus (39). It is not known whether YbYb is a unique form of nuclear GST or whether it is a cytosolic isozyme. These findings lend credence to the hypothesis that an additional role of GST isozymes is to function as carrier proteins that transport molecules such as steroids from the cytoplasm to the nucleus (40).

Induction of GST activities in rodents, including rats and mice, has been reported with at least 100 different chemicals in numerous tissues, including liver, lung, stomach, small and large intestine, pancreas, and esophagus (22). The dietary antioxidant BHA has been used as an inducing agent for lung GST in rats and mice (41). Dietary consumption of BHA increased levels of cytosolic GST in these rodents; however, the increase in enzyme activity was more pronounced in mice. Studies with mice showed that BHA induces lung GST II (pI 8.7) and III (pI 7.9), and although it is not clear to which classes these belong, they may be assigned to the pi and mu classes on the basis of their pI values (42). Our results showed increased amounts of GST protein after treatment of mice with BHA (Figures 1 and 3). Induced protein content was detected for all three GST subunits, with increases for Ya and Yp being more marked than for Yb₁ (Figures 1 and 3). Significantly increased amounts of mRNA transcripts were expressed for all three subunits after BHA treatment (Figures 4 and 5). The inductive effects were manifested regionally and were found in both parenchyma and bronchioles. In the parenchyma, increased mRNA levels were comparable for all three subunits. However, the increases in the bronchioles were not uniform, and were greater for Ya than for Yp. Increased amounts of mRNA transcripts were also detected for Yb₁ in the bronchioles, but the inductive effect for this subunit was the lowest achieved for any of the three subunits. In other studies, treatment of mice with BHA markedly increased the hepatic expression of the alpha and mu GST classes, and this was associated with increased levels of the respective mRNAs (42). However, our data and those for murine liver are not directly comparable, because the latter were derived from Northern blotting and cloning and our data in lung were obtained from in situ hybridization coupled with quantitative image analysis. Nevertheless, these findings implicate transcriptional activation as a mechanism responsible for BHA-mediated GST isozyme alterations in both lung and liver.

The GST enzymes have a major role in facilitating conjugation of reactive intermediates to GSH, thereby enhancing the efficiency of detoxification, although this outcome is by no means always the case. In support of the protective role of the GST enzymes, previous studies have shown that administration of dietary antioxidants including BHA decreased the incidence of carcinogenesis induced by benzo[a]pyrene in murine lung (42). This has been ascribed to induction of GST isozymes implicated in the detoxification of reactive metabolites of benzo[a]pyrene, and these have been proposed to be GST II and III. These findings underscored the importance of identifying the types of lung cells that express specific GST isozymes, and implied that it would also be of importance to identify cells in which activating enzymes, including cytochrome P450, reside. The rationale for this assertion is that highly reactive metabolites may require GST-mediated conjugation at the site of their formation. Numerous P450 isozymes, including 1A1, 2B1, 3A2, 4B1, and 2E1 are localized predominantly in Clara and alveolar type II cells (29, 43). These data correlated with our findings that the GST alpha, pi, and mu subclasses were expressed to the greatest extent in Clara and type II cells. Interestingly, our previous studies also found the highest levels of GSH in these two cell types (44). Taken together, localization of the cytochrome P450 and GST enzyme systems, as well as GSH, within the same lung-cell populations probably provides optimal conditions for the detoxification of reactive intermediates formed from potential pneumotoxicants.