Introduction

Top Abstract Introduction Materials and methods Results Discussion References

Early studies by Haddow (1938) showed that although polycyclic aromatic hydrocarbons exerted cytotoxic and growth inhibitory effects on structurally normal tissue, the tumors that they induced were resistant to these changes. These findings gave rise to the hypothesis that the "origin of cancer" must involve the emergence of a new phenotype with novel properties that permit growth under conditions that are toxic to normal cells (Haddow, 1938). Subsequent studies have primarily used rodent models of hepatocarcinogenesis to demonstrate expression of a phenomenon initiated by carcinogens and referred to as the "resistance phenotype" (Farber et al., 1976; Farber, 1987; Farber and Sarma, 1987). Hyperplastic hepatocyte nodules induced in rat liver by 2-acetylaminofluorene or ethionine were resistant to the acute hepatotoxic effects of carbon tetrachloride and dimethylnitrosamine, whereas necrosis was manifested in hepatocytes residing in the normal surrounding tissue (Farber et al., 1976). This resistance phenotype is manifested in a small number of hepatocytes, which undergo clonal expansion by cell proliferation to form hepatocyte nodules; in contrast, growth of the vast majority of the nonresistant heptocytes is inhibited (Farber, 1987, 1990).

The concept of resistance in hepatocyte nodules led us to undertake studies to determine whether a similar phenomenon was also manifested in lung tumors. These studies (Forkert et al., 1992) demonstrated that lung tumors induced by the carcinogen urethane are resistant to the cytotoxic effects of three known pneumotoxicants: naphthalene (Mahvi et al., 1977), paraquat (Vijeyaratam and Corrin, 1971) and 1,1-dichloroethylene (DCE)¹ (Forkert and Reynolds, 1982). In contrast to the resistance of the lung tumors, all three chemicals caused necrosis and cell death involving Clara cells of control lung and in those residing in normal tissue of tumor-bearing lung (Forkert et al., 1992). These findings suggested that resistance manifested in tumor foci, irrespective of tumor type, must involve fundamental processes associated with neoplastic transformation.

The basis of the resistance to cytotoxicities in hepatic nodules has been attributed to diminished expression of phase I enzymes, including cytochrome P-450, an enzyme system with a key role in bioactivation of toxicants and carcinogens (Gravela et al., 1975; Farber et al., 1976; Cameron et al., 1976; Feo et al., 1978; Eriksson et al., 1983; Astrom et al., 1983; Farber, 1987). These findings are consistent with the assumption of reduced or deficient metabolic activation in the hepatocyte nodules due to loss of activating enzymes. To determine whether similar alterations occurred in lung tumors, we have initially examined the expression of two P-450 isozymes in lung tumors: the constitutive expression of CYP2B1 was abolished and the inducibility of CYP1A1 was markedly decreased (Forkert et al., 1992, 1996b,d). In contrast, the expression of these P-450s in normal tissue of tumor-bearing lung was conserved (Forkert et al., 1996d). These findings suggested that resistance to chemically induced lung injury is also associated with reduced content of enzymes with capability for metabolic activation, as was found in the hepatocyte nodules. However, these data are indirect and the extent to which metabolites are formed in tumors has not been determined.

This investigation was undertaken to understand the underlying basis for the resistance of lung tumors to cytotoxic chemicals. We hypothesized that the resistance phenomenon manifested in lung tumors is due, in part, to loss of P-450 expression, leading to defective formation of reactive metabolites. To test this hypothesis, we have used a urethane-induced model of lung tumorigenesis (Shimkin, 1955) in conjunction with DCE, a chemical that produces Clara cell injury (Forkert and Reynolds, 1982), as a test xenobiotic. Lung metabolism of DCE is catalyzed mainly by the P-450 isozyme CYP2E1 to an epoxide (Lee and Forkert, 1995; Dowsley et al., 1996, 1995). A major product formed from the epoxide is the glutathione (GSH) conjugate, 2-S-glutathionyl acetate [C] (Dowsley et al., 1996, 1995). Accordingly, identification and localization of conjugate [C] within a tissue represent the sites where the epoxide is generated. Of relevance in this regard is the identification of CYP2E1 as the P-450 isozyme catalyzing epoxide formation. Thus, the location of CYP2E1 should coincide with that for [C], especially since the epoxide is highly reactive and is unlikely to migrate to locations external to the cells in which it is synthesized. In this study, we have performed immunohistochemical studies to localize CYP2E1 and conjugate [C] in control and tumor-bearing lung. The CYP2E1 protein and [C] were identified in adjacent tissue sections to determine whether the cellular sites of CYP2E1 expression coincided with those for [C]. In addition, histochemical staining was performed to confirm the presence of GSH in tumor sites.

	Materials and Methods

Top Abstract Introduction Materials and methods Results Discussion References

Chemicals and Reagents. The following chemicals and reagents were used in this study: DCE (>99% purity) (Aldrich Chemical Co., Montreal, Quebec, Canada); glycine (ICN Biomedicals, Montreal, Quebec, Canada); urethane (ethyl carbamate), bovine serum albumin (BSA), hydrogen peroxide (30% v/v), glutaraldehyde (70% aqueous), paraformaldehyde, 3',3'-diaminobenzidine tetrahydrochloride mercury orange (Sigma Chemical Co., St. Louis, MO); sodium pentobarbital (Somnotol; MTC Pharmaceuticals, Hamilton, Ontario, Canada); biotinylated goat anti-rabbit IgG, biotinylated horse anti-goat IgG, avidin-biotin blocking reagent (Vector Laboratories, Inc., Burlingame, CA); streptavidin conjugated to horseradish peroxidase (Zymed Laboratories, San Francisco, CA); and isopentane (Fisher Scientific Co., Toronto, Ontario, Canada). An IgG purification kit [ImmunoPure (A/G)] was purchased from Pierce Chemical Co. (Rockford, IL). The polyclonal antibody for CYP2E1 was obtained from Oxford Biomedical Research, Inc. (Oxford, MI). All other chemicals were of reagent grade and were purchased from standard commercial suppliers.

Animal Treatment. Strain A/J mice (The Jackson Laboratory, Bar Harbor, ME) of 6 weeks of age were housed on hardwood bedding and maintained on a 12-h light/dark cycle in a temperature-controlled (25 ± 1°C) room. Mice were provided free access to food and water. After being allowed to equilibrate to laboratory conditions for at least 1 week, the mice were randomly assigned to either the control or urethane-treated group. Mice were treated with a single dose of urethane (1 mg/g body wt., i.p.) dissolved in saline; control mice were given only the vehicle. At 2, 4, or 12 months after urethane treatment, saline- and urethane-treated mice were administered DCE (125 mg/g body wt., i.p.) in corn oil. All mice were sacrificed 1 h after DCE treatment. Selection of this time point was based on results from preliminary experiments showing that [C] was formed rapidly in the lungs of DCE-treated mice. For the GSH histochemical staining experiments, mice were sacrificed 4 months after urethane treatment.

Tissue Preparation. Lung tissue was processed for histopathologic evaluation and immunohistochemical staining experiments by using methods described previously (Forkert, 1995). Mice were anesthetized with sodium pentobarbital (0.12 mg/g body wt., i.p.), the thorax was exposed, a cannula was inserted into the pulmonary artery via the right ventricle, and the lungs were flushed with saline to clear the tissues of blood. When the perfusate became clear, perfusion was continued with 4% paraformaldehyde in 0.1 M Sörensen's phosphate buffer (12.0 mM NaH₂PO₄, 69.0 mM Na₂HPO₄), pH 7.4. The lungs were then inflated with 0.3 ml of this fixative by tracheal instillation. After ligation of the trachea, the lungs and heart were excised en bloc and immersion-fixed for an additional 4 h at room temperature. The lung lobes were separated and fixed overnight at 4°C, rinsed thoroughly in phosphate-buffered saline (PBS), dehydrated, cleared, and embedded in paraffin. Tissue sections (5 µm) were placed on acid-cleaned, gelatin-coated slides. Contiguous sections were used for histopathologic evaluation and the immunolabeling of the CYP2E1 and [C] proteins. Some tissue sections subjected to immunohistochemical labeling were counterstained with hematoxylin, whereas other sections were stained with H&E for histopathologic observation of lung tumors and identification of tumor types.

Antibody Production and Purification. The hapten used for immunization was chemically synthesized [C] conjugated to BSA as the carrier protein with glutaraldehyde as the chemical cross-linker and was prepared as described (Forkert et al., 1997). Polyclonal antibodies to [C] were raised in rabbits. The properties of these antibodies have been characterized and demonstrated to be effective in recognizing [C] but not the carrier protein or the cross-linker (Forkert et al., 1997). The antiserum was affinity-purified using a protein A/G column; the IgG fraction obtained was used for immunolabeling. Protein concentrations were determined by the Bradford protein assay (Bradford, 1976).

Immunohistochemical Detection of CYP2E1 and [C]. Immunohistochemical labeling was performed as described in our previous studies (Forkert, 1995), with modifications. Tissue sections were deparaffinized, cleared, hydrated in a graded ethanol series, and rinsed in PBS. To immunolabel CYP2E1, the sections were incubated in 5% normal horse serum in PBS for 20 min to block nonspecific binding of antibody. To immunolabel [C], the sections were incubated in PBS containing 5% normal goat serum and 3% BSA. After rinsing in PBS, sections were incubated for 60 min with an antibody for either CYP2E1 or an antibody for [C]. The CYP2E1 antibody was diluted (1:400) in PBS containing 2.5% normal horse serum. The antibody for [C] was diluted in PBS containing 1% normal goat serum and 0.1% of the conjugate, glycine-glutaraldehyde-BSA, which was synthesized as described previously (Forkert et al., 1997). Inclusion of the conjugate in antibody reactions effectively blocked nonspecific antibody binding. All tissue sections were then incubated with the avidin-biotin reagent for 15 min to block reactions with endogenous biotin. The sections were rinsed thoroughly to remove unbound primary antibody, and reacted for 10 min with a biotinylated horse anti-goat IgG for CYP2E1 or a biotinylated goat anti-rabbit IgG to immunostain for [C]. Endogenous peroxidase activity was blocked by reacting tissue sections for 30 min with 1% hydrogen peroxide in nanopure water. Sections were then incubated with streptavidin conjugated to horseradish peroxidase for 10 min and the immunoperoxidase color reaction was developed by incubation in PBS containing 0.05 3',3'-diaminobenzidine tetrahydrochloride and 0.01% hydrogen peroxide; this was terminated by rinsing the sections in tap water for 5 min. After incubating the sections for 5 min in 0.15 M sodium chloride containing 0.5% copper sulfate, they were dehydrated, cleared, and mounted. Controls for the specificity of the reactions included incubations in which the specific antibodies were omitted or incubations performed in the presence of a nonspecific antibody.

Histochemical Localization of GSH. Histochemical staining for GSH was performed on frozen lung sections by following the procedures detailed in our previous studies (Forkert and Moussa, 1989). The azo dye mercury orange reacts with tissue sulfhydryls. However, specificity for GSH staining is based on the considerably faster rate of reaction with GSH than with proteinSH groups (Bennet, 1951; Ashgar et al., 1975). Substantial GSH staining is achieved within 5 min of the reaction, whereas staining for proteinSH groups requires an incubation time of several hours (Ashgar et al., 1975). Frozen lung sections (5 µm) were prepared with a cryostat and incubated in 50 µM mercury orange in toluene for 5 min. In controls, tissue sections were incubated in toluene alone. All of the sections were then mounted.

	Results

Top Abstract Introduction Materials and methods Results Discussion References

Histopathology. Lung tissues from untreated mice exhibited normal alveolar and bronchiolar structure. In particular, the bronchiolar epithelium was lined by ciliated and nonciliated Clara cells that had characteristic features. Hyperplastic foci, adenomas, and carcinomas were prevalent in the lungs of urethane-treated mice that had been maintained for 2, 4, and 12 months, respectively, and were identified by previously described criteria (Mostofi and Larsen, 1951; Shimkin, 1955; Kaufman et al., 1979). Areas of hypercellularity were designated as hyperplastic foci, whereas spherical structures with compact masses of cuboidal cells were classified as solid adenomas. Papillary adenomas were structures with an open tubular pattern lined by nonciliated columnar epithelial cells resembling Clara cells, and carcinomas were large tumors with cellular anaplasia, hyperchromatic nuclei, and invasion that compressed the surrounding lung parenchyma. Hyperplastic foci, adenomas, and carcinomas were not observed in the lungs of untreated control mice. At 1 h after DCE treatment, morphologic alterations were not apparent in lung cells residing in either neoplastic or normal tissue of tumor and nontumor-bearing lung.

Immunohistochemical Detection of CYP2E1 and Conjugate [C]. Nontumor-bearing lung revealed positive reactivity for CYP2E1 in the bronchiolar epithelium, with preferential staining in Clara cells (Fig. 1a). Alveolar type II cells were also stained but the amount of labeling was minimal compared with that in Clara cells. This staining distribution was also observed in uninvolved tissue of tumor-bearing mice. In contrast, CYP2E1 staining was present at only a low level in hyperplastic foci and was negligible in adenomas or carcinomas (Fig. 1d). The lack of CYP2E1 labeling in the tumors was observed regardless of whether they were solid or papillary types. An exception to this deficient CYP2E1 reactivity was at the periphery of papillary tumors, where epithelial cells lining the papillae or finger-like projections stained positively (Fig. 1g). Staining was most pronounced in the apical portions of columnar cells. CYP2E1 staining patterns were similar in untreated and DCE-treated mice and in nontumor- and tumor-bearing lungs.

View larger version (129K): [in this window] [in a new window]   Fig. 1.   Immunohistochemical staining for CYP2E1 and conjugate [C] in lung tissue. Immunohistochemistry was performed with the avidin-biotin peroxidase complex technique. a, bronchiole in lung section from a DCE-treated nontumor-bearing mouse. Staining for CYP2E1 is highly localized in the nonciliated Clara cells (arrowhead). b, bronchiole in lung section adjacent to that depicted in a exhibits staining for [C] in the Clara cells (arrowheads). c, bronchiolar epithelium (arrowheads) in tissue section from nontumor-bearing control untreated mouse that was reacted with the antibody for [C]. d, minimal CYP2E1 staining in a tumor from a mouse treated with urethane 4 months previously. e, bronchiole lying adjacent to a tumor; the epithelium is labeled for CYP2E1 (arrowhead). f, lung section adjacent to that depicted in e, in which staining for [C] is seen in the bronchiolar epithelium (arrowhead). g, periphery of a papillary tumor (12 months after urethane). Tissue processes are lined with epithelium that stains for CYP2E1 (arrowhead). h, lung section adjacent to that depicted in g, in which staining for [C] is found in the epithelial lining cells (arrowhead). Bar, 50 µm. Original magnification, 260 × (a, b, g, and h); 200 × (c and d); 130 × (e and f).

Histochemical Localization of GSH. Histochemical staining for GSH in normal lung tissue confirmed our earlier observations in which staining was observed in the bronchiolar epithelium and alveolar septa, and has been described in our previous studies (Moussa and Forkert, 1992; Forkert and Moussa, 1993). Staining for GSH was most prominent in the bronchiolar epithelium and was highly concentrated in the Clara cells. The pattern of GSH staining was similar in lung tissue from nontumor-bearing mice and structurally normal lung tissue from tumor-bearing mice. Staining was present in tumor foci and was of comparable intensity as seen in normal tissue (Fig. 2).

View larger version (61K): [in this window] [in a new window]   Fig. 2.   Histochemical staining for GSH in tissue section from tumor-bearing lung. Histochemical staining for GSH was performed with mercury orange in frozen tissue sections. GSH-mercury orange complexes viewed under epifluorescence were localized in the bronchiolar epithelium of normal lung tissue (a, arrow) and tumor tissue (b, arrows). Bar, 20 µm. Original magnification, 250 × (a); 320 × (b).

	Discussion

Top Abstract Introduction Materials and methods Results Discussion References

Previous studies regarding hepatocarcinogenesis in rodents found a correlation between diminished expression of phase I enzymes, including cytochrome P-450, and formation of reactive metabolites in hepatocyte nodules (Gravela et al., 1975; Astrom et al., 1983; Farber, 1984; Rinaudo and Farber, 1986; Farber, 1987). The amounts of [9-¹⁴C]2-acetylaminofluorene covalently bound to DNA, RNA, and proteins in hepatocyte nodules induced by diethylnitrosamine were markedly reduced (Rinaudo and Farber, 1986). Formation of an acetylaminofluorene-derived DNA adduct was inhibited in hepatic tumors or enzyme-altered foci (Huitfeldt et al., 1988). Moreover, activation of different arylamines to mutagenic species, as assessed by the Salmonella mutagenesis assay, was depressed in livers with hepatocyte nodules, compared with normal livers (Ozawa et al., 1990). These data are consistent with the assumption that the diminished presence of activating enzymes in hepatocyte nodules is responsible for decreased metabolic activation and formation of reactive intermediates in hepatocyte nodules; this mechanism has been proposed to be responsible for the resistance phenomenon (Farber, 1987; Farber and Sarma, 1987).

Similar studies linking xenobiotic resistance with altered formation of reactive metabolites in lung tumors have not been reported, despite the availability of well established animal models of tumorigenesis. In this context, our previous studies have demonstrated that urethane-induced lung tumors are resistant to the cytotoxic effects of chemical xenobiotics, including DCE (Forkert et al., 1992). In contrast, bronchiolar Clara cells in structurally normal tissue of tumor-bearing lung are damaged and appear to be highly vacuolated. An identical Clara cell lesion is also elicited in nontumor-bearing lung tissue from mice treated with DCE (Forkert and Reynolds, 1982; Forkert et al., 1992). Some of the metabolic mechanisms of DCE-induced lung injury have been elucidated. The Clara cell cytotoxicity is mediated by metabolic activation of DCE by CYP2E1, a P-450 isozyme that is highly localized in this cell type (Forkert, 1995; Lee and Forkert, 1995; Dowsley et al., 1996). The major metabolite formed from CYP2E1-dependent oxidation of DCE is the highly reactive DCE-epoxide, which is strongly implicated as the ultimate toxic species (Dowsley et al., 1995, 1996; Forkert et al., 1996a,c). Initial reaction of GSH with the epoxide yields the intermediate, 2-S-glutathionylacetyl chloride, which either undergoes hydrolysis to form [C] or reacts with a second GSH molecule to produce 2-(S-glutathionyl) acetyl glutathione (Dowsley et al., 1995, 1996). Under aqueous conditions, 2-(S-glutathionyl) acetyl glutathione hydrolyzes to produce [C]. Thus, [C] is a major product formed and this GSH conjugate may be regarded as an index of formation of the epoxide.

An objective of this study was to investigate resistance of lung tumors to cytotoxicities induced by chemical xenobiotics. Here, we have used a murine model in which lung tumors of the papillary and solid types were induced in mice treated with the carcinogen urethane (Forkert et al., 1992). An immunohistochemical approach was used to examine the cellular distribution of CYP2E1 in nontumor- and tumor-bearing lung. As was found in nontumor-bearing lung, CYP2E1 was highly expressed in Clara cells residing in uninvolved tissue of tumor-bearing lung (Fig. 1, a and e); the presence of tumors within the same lung did not affect CYP2E1 expression. Slightly higher CYP2E1 immunoreactivity was detected in hyperplasias, compared with the amounts in adenomas and carcinomas; this indicates increased P-450 loss at more advanced stages of tumor growth. The extent of CYP2E1 decrease was not different in solid versus papillary tumors, suggesting that differences in tumor structure were not associated with P-450 modification. The lack of CYP2E1 in lung tumors should impair activation of substrates that are selectively metabolized by this P-450 isoform. Of pertinence to P-450-mediated oxidation is the presence of NADPH-cytochrome P-450 reductase, which is also reduced in lung tumors, suggesting that the coordinate regulation of the reductase and P-450 is conserved (Forkert et al., 1996d). Thus, the enzyme components essential for P-450-dependent oxidation of DCE are available for metabolic activation in the target Clara cells of normal tissue, but are diminished or abolished in neoplastic tissues. This depressed metabolic state occurred early in neoplastic transformation because it was manifested in hyperplastic foci. These findings supported our hypothesis that resistance of lung tumors to cytotoxic chemicals is associated with reduced or lack of P-450-dependent formation of reactive metabolites. In the case of DCE, the loss of CYP2E1 in the tumors coincided with diminished formation of the epoxide, as assessed by production of conjugate [C].

Metabolite formation has commonly been determined by measuring covalent binding of a radiolabeled substrate to tissue constituents, including proteins and nucleic acids. Here, we have used an immunohistochemical approach to evaluate metabolite formation in lung tissue, including tumor foci. This strategy produces highly sensitive signals, which permit detection of metabolites in lung cells and in structures such as hyperplasias and small tumors. It also allows comparisons of results in different tumor foci as well as in adjacent uninvolved tissues. This is of importance because of the cell-specific nature of xenobiotic metabolism within the lung. Immunohistochemical experiments were performed to detect conjugate [C] in lung cells, and hence the sites of DCE-epoxide formation and GSH conjugation. The DCE-epoxide-derived [C] was detected in the bronchiolar epithelium and localized primarily in the Clara cells of nontumor-bearing- and normal tissue of tumor-bearing lung (Fig. 1f). The [C] antigen was not detected in lung tissue from mice that were not administered DCE, confirming that [C] is a product associated with DCE metabolism. In contrast, tumor foci, including hyperplasias, adenomas, and carcinomas, were devoid of staining for [C]. These findings demonstrated that neither CYP2E1 nor [C] were present in the tumor foci; hyperplasias, adenomas, and carcinomas were all deficient in these proteins. A different cell distribution of these proteins was seen in normal lung tissue where both the CYP2E1 and [C] proteins were present at appreciable levels in the Clara cells. Taken together, these data are consistent with the view that lung tumor cells lack properties allowing them to bioactivate DCE to reactive metabolites, and strongly suggest that the basis for the resistance phenomenon in lung tumors is similar to that reported in hepatocarcinogenesis. These results indicate a certain degree of commonality in the mechanisms mediating resistance to xenobiotics.

Notwithstanding our data strongly supporting the premise that lung tumor cells are defective in activating DCE to reactive metabolites, a question arises regarding whether the absence of [C] in tumor foci can also be attributed to the lack of availability of GSH for conjugation with the DCE-epoxide. The results of our histochemical experiments showed that GSH staining was present at appreciable amounts in tumor foci (Fig. 2) and confirmed the existence of GSH in tumor cells for conjugation reactions. There is also a possibility that the deficient localization of [C] in the tumors is due to efflux from the cells as a result of the action of the GS-X pump, a GSH conjugate export carrier (Müller et al., 1994). An additional possibility is that DCE is not distributed to tumor cells due to alterations in the architecture of lung tissue in neoplastic states. However, these are unlikely scenarios in view of our findings showing the lack of both CYP2E1 and conjugate [C] at the same tumor sites.

It should be emphasized that the results of these studies in mice and with DCE as the test chemical cannot easily be extrapolated to other species or xenobiotics. Other chemicals that elicit Clara cell damage are 4-ipomeanol and naphthalene, which are metabolized mainly by CYP4B1 and Cyp2f2, respectively (Czerwinski et al., 1991; Buckpitt et al., 1995). Moreover, there are marked species differences in their metabolisms. Rabbit CYP4B1 was the most efficient in metabolic activation of 4-ipomeanol (Czerwinski et al., 1991). On the other hand, naphthalene metabolism was substantially higher in the lungs of mice than in those of hamsters and rats and correlated closely with the levels of CYP2F2 present (Buckpitt et al., 1995). It is not known whether there is diminished expression of all P-450s in tumor tissue or whether the loss is selective to isozymes including CYP2E1. Interestingly, lung tumors from mice treated with naphthalene were also resistant to cytotoxicity, whereas Clara cells in normal surrounding tissue were damaged (Forkert et al., 1992). These findings suggested that activation of naphthalene may be diminished as in the case of DCE. However, other factors are likely to be involved.

In summary, our results demonstrated that [C] only accumulated in structurally normal tissue and were observed in cells in which CYP2E1 was expressed. The negligible levels of [C] and CYP2E1 in lung tumors are consistent with diminished or lack of metabolic activation of DCE and was not associated with a deficit of GSH for conjugation. These findings supported our hypothesis that resistance of tumor foci to cytotoxicities induced by xenobiotics is mediated in part by diminished formation of reactive intermediates at these sites. To the best of our knowledge, these data represented the most direct evidence to date showing a deficiency of lung tumor cells to generate reactive metabolites from potential cytotoxic chemicals.