Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Intragastric instillation of excessive ethanol results in gastric mucosal injury characterized by mucosal edema, subepithelial hemorrhages, cellular exfoliation, and inflammatory cell infiltration (Eastwood and Kirchner, 1974; Laine and Weinstein, 1988). Studies focusing on the pathogenesis of ethanol-induced injury have suggested that several factors are implicated in such processes: products of arachidonate metabolism (e.g., leukotriene) (Peskar et al., 1986), mast cell secretory products (Oates and Hakkinen, 1988), and reactive oxygen species (ROS) (Mizui et al., 1987; Pihan et al., 1987; Szelenyi and Brune, 1988).

With regard to maintaining the integrity of the gastric mucosa, several agents, such as prostaglandins (Guth et al., 1984; Lacy and Ito, 1984; Pihan et al., 1986) or sulfhydryl compounds (Boyd et al., 1981; Szabo et al., 1981), have been shown to protect the stomach from ethanol injury. Although a number of studies on the protective mechanisms by prostaglandins have implicated the maintenance of structural and functional integrity of mucosa (especially mucosal blood flow and the energy-dependent rapid restitution from surviving neck cells) (Lacy and Ito, 1984), the mechanisms by which sulfhydryls protect against ethanol remain unclear. Furthermore, there have been conflicting results on the role of sulfhydryls in ethanol-induced gastric mucosal injury and its protection (Boyd et al., 1981; Szabo et al., 1981; Robert et al., 1984).

Recent studies have suggested that the generation of ROS may be involved in the pathogenesis of ethanol-induced gastric mucosal injury in vivo (Mizui et al., 1987; Pihan et al., 1987; Szelenyi and Brune, 1988). Because the glutathione redox cycle and cellular catalase are potent antioxidant defenses in a variety of tumor cells (Nathan et al., 1980) and in endothelial (Harlan et al., 1984) and epithelial (Hiraishi et al., 1991) cells, these antioxidants may play a role in protecting gastric mucosa against ethanol.

The glutathione redox cycle can be disrupted at several sites in in vitro cultures of these cells (Nathan et al., 1980; Harlan et al., 1984; Hiraishi et al., 1991). The activity of glutathione reductase is selectively inhibited by 1,3-bis(chloroethyl)-1-nitrosourea (BCNU) (Becker and Schirmer, 1995). Glutathione biosynthesis is inhibited by DL-buthionine-(S,R)-sulfoximine (BSO), a selective inhibitor of -glutamylcysteine synthetase (Meister, 1995). Endogenous catalase can be inhibited by 3-amino-1,2,4-triazole (AT) (Sies, 1981).

Because ingested ethanol diffuses into gastric mucosa, targeting both epithelial and endothelial cells (Guth et al., 1984; Lacy and Ito, 1984; Pihan et al., 1986), we have examined and compared the cytotoxic effect of ethanol on cultured epithelial and endothelial cells and evaluated the relative importance of the glutathione redox cycle and endogenous catalase against ethanol injury in these cells.

We found that glutathione redox cycle, but not cellular catalase, maintains the integrity of epithelial cells against ethanol, whereas neither antioxidant seems to play a significant role in protection of endothelial cells against ethanol injury. The distinct significance of these antioxidants in each cell type seems to depend on the capacity of each cell to generate ROS in response to ethanol exposure.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and Reagents. Collagenase (collagenase A from Clostridium histolyticum) was purchased from Boehringer Mannheim Biochemicals (Indianapolis, IN). The 7- to 10-day-old rats of either sex (Sprague-Dawley) were from Charles River (Wilmington, MA) and Dohken (Ibaraki, Japan). Coon's modified Ham's F-12 medium and DMEM were purchased from GIBCO (Grand Island, NY) and adjusted to pH 7.4. FBS was obtained by HyClone Laboratories (Logan, UT). Hyaluronidase (type 1-S), antibiotic/antimycotic, HEPES, BSO, glutathione (GSH), glutathione disulfide, glutathione reductase, 5,5'-dithiobis (2-nitrobenzoic acid), AT, cytochrome c, superoxide dismutase (SOD) (from bovine liver), and Triton X-100 were from Sigma Chemical Co. (St. Louis, MO). BCNU was kindly supplied by Bristol-Meyers (Syracuse, NY) and dissolved in ethanol 100 mg/ml stock just before each experiment. Perchloric acid (HClO₄) was purchased from Aldrich Chemical (Milwaukee, WI). Hanks' balanced salt solution and Earle's balanced salt solution (EBSS) supplemented with 15 mM HEPES were obtained from GIBCO and adjusted to pH 7.4. ⁵¹Cr (sodium chromate, 200-900 Ci/g chromium) was obtained from ICN Biochemicals (Irvine, CA). Tissue culture plates and dishes were from Costar (Cambridge, MA).

Culture of Epithelial and Endothelial Cells. Primary culture of the gastric fundic mucosa from rats was prepared as described previously (Hiraishi et al., 1993). More than 90% of the cells have been previously identified as mucus-producing epithelial cells and shown to possess the capability to synthesize DNA as well as cyclic nucleotides, to produce and secrete mucous glycoprotein, and to generate prostaglandins (Hiraishi et al., 1993). Confluent epithelial monolayers were studied 3 days after seeding. Endothelial cells from bovine aorta were isolated, identified, and maintained as described previously (Schwarz, 1978; Hiraishi et al., 1992). First-passage endothelial cells at the confluent state 2 days after passage were used for experiments (Hiraishi et al., 1992).

⁵¹Cr Release Assay. Cytotoxicity was quantified by measuring ⁵¹Cr release from prelabeled cells (Hiraishi et al., 1991, 1992). Confluent monolayers of epithelial or endothelial cells in 24-well culture plates were labeled overnight with 2 µCi ⁵¹Cr/well/0.5 ml of each culture medium. In experiments with BSO, cells were incubated with or without BSO during the overnight labeling for 16 h. In a preliminary study, we found that GSH biosynthesis was dose dependently inhibited by incubation with BSO at 0.33 to 10 µM in epithelial cells (Hiraishi et al., 1991) and at 10 to 333 µM in endothelial cells. After prelabeling, cells were washed three times with Hanks' balanced salt solution and incubated with increasing concentrations (4-12%) of ethanol in 1 ml EBSS for up to 3 h. In some experiments, labeled cells were preincubated with either BCNU or vehicle control in EBSS for 10 min or with either AT or EBSS for 60 min before ethanol exposure. After the cytotoxicity assay, aspirated buffer was centrifuged at 1000g for 2 min, and 100 µl of cell-free supernatant buffer was removed for the determination of specific ⁵¹Cr release (A  B/C  B) × 100%, where A represents the mean test ⁵¹Cr cpm released, B represents the mean spontaneous ⁵¹Cr cpm released, and C represents the mean maximum ⁵¹Cr cpm released. Maximum ⁵¹Cr release was determined by incubating cells in 0.2% Triton X-100. Spontaneous ⁵¹Cr release was determined in control monolayers incubated in only EBSS and was 8 to 12% of maximum ⁵¹Cr release after 3 h of incubation. ⁵¹Cr radioactivity was counted with a Gamma 7000 Counting System (Beckman Instruments, Fullerton, CA).

Assay for Enzyme Activities and GSH Content. Quadruplicate 60-mm dishes of epithelial or endothelial cells incubated with 14 ml EBSS containing BCNU or ethanol (0.1%) vehicle were solubilized by incubation with 0.2% Triton X-100 at room temperature for 1 h and frozen at 20°C, and the supernatant buffer was assayed within 24 h for glutathione reductase and catalase (Carlberg and Mannervik, 1985; Hiraishi et al., 1991). The activities of glutathione reductase and catalase were expressed as (m)IU/mg protein, as determined according to the method of Bradford (1976). GSH content was measured in BSO-treated and control cells (Hiraishi et al., 1994) and expressed in "GSH equivalents" [GSH + glutathione disulfide (nmol)/10⁶ cells].

Assay of Cytochrome c Reduction. Detection of superoxide anion (O₂) was based on its ability to reduce cytochrome c (O'Brien, 1984; Mayo and Curnutte, 1990). Monolayers of epithelial or endothelial cells in 35-mm culture dishes were incubated at 37°C for 3 h with 4.8 ml EBSS (without phenol red) containing graded concentrations of ethanol (0-10%) and cytochrome c (10 nmol) so that cells at the confluent state were exposed to the same doses of ethanol as in the cytotoxicity assay. A portion of the buffer was incubated at 37°C without the cultured cells for use as a blank. After incubation, the buffer was cleared by centrifugation at 3000g for 5 min. In some experiments, to examine SOD inhibition of cytochrome c reduction, monolayers in 35-mm dishes were incubated with ethanol (10%), cytochrome c (10 nmol), and SOD (500 U/ml) for 3 h. Cytochrome c reduction in an aliquot of the supernatant buffer was determined by measuring absorbance at 550 nm on a Gilford 260 Spectrophotometer (Oberlin, OH), using the blank as reference. E_mM (ferrocytochrome c  ferricytochrome c) at 550 nm was taken as 18.5 (Margoliash and Frohwirt, 1959).

Statistical Analysis. Data were expressed as mean ± S.E.M. ANOVA was performed when more than two groups were compared, and when significant (P < .05), Tukey's multiple comparison test was applied to test for the differences between individual groups. The significance of the differences between two groups was determined with an unpaired Student's t test; P  .05 values were considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Ethanol-Induced Cytotoxicity to Epithelial and Endothelial Cells. The effects of increasing concentrations of ethanol (4-12%) on epithelial and endothelial cell ⁵¹Cr release were examined in preliminary studies. Ethanol-induced cell injury was dependent on both the concentration of ethanol applied and the duration of exposure (Fig. 1, A and B). The earliest signs of injury were seen after 3 h of exposure to 8% ethanol in both epithelial and endothelial cells. In the following experiments to assess the role of antioxidants, 6 to 10% of ethanol was used as a cytotoxic agent.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Dose-dependence of ethanol-induced51Cr release from cells. Epithelial (A) or endothelial (B) cells were labeled overnight with 51Cr, washed, and incubated with increasing concentrations of ethanol (4-12%) for up to 3 h. Aliquots of supernatant buffer were sampled for determination of specific 51Cr release. Values represent mean ± S.E.M. of quadruplicate determinations in a single experiment.

Effects of Disruption of GSH Redox Cycle on Ethanol-Induced ⁵¹Cr Release in Epithelial Cells. The pretreatment of epithelial cells with BCNU (50 µg/ml) caused a left shift of the dose-response curve for ethanol (6-10%) (Fig. 2A). Pretreatment with BCNU alone, up to 50 µg/ml for 10 min, did not significantly increase ⁵¹Cr release during a 3-h incubation, compared with ethanol control. Figure 3A demonstrated that ethanol-induced lysis directly correlated with the concentrations of BCNU. Gastric epithelial susceptibility to lysis by ethanol was inversely related to endogenous glutathione reductase activity.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effect of BCNU on ethanol-induced 51Cr release. 51Cr-labeled epithelial (A) or endothelial (B) cells were preincubated for 10 min with control buffer (0.05% ethanol) or BCNU (50 µg/ml) and then incubated with increasing concentrations of ethanol (6-10%) for 3 h. Values represent mean ± S.E.M. of the triplicate determinations on three separate preparation. *P < .05 and ***P < .001, significant differences compared with control values (BCNU).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Effects of BCNU concentration on ethanol-induced 51Cr release and glutathione reductase activity of epithelial or endothelial cells. Glutathione reductase activity and ethanol-induced 51Cr release were determined in epithelial (A) or endothelial (B) cells following a 10-min incubation with varying concentrations of BCNU (0-50 µg/ml). Specific 51Cr release was determined after incubation with ethanol (8%) for 3 h. Glutathione reductase activity was determined in Triton X-100-solubilized cells. Values for 51Cr release represent mean ± S.E.M. of the quadruplicates on three separate preparation. Values for glutathione reductase activity represent mean ± S.E.M. of the triplicate determinations in a single experiment. *P < .05, **P < .01, and ***P < .001, significant differences compared with control values.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Effect of BSO on ethanol-induced 51Cr release. Epithelial (A) or endothelial (B) cells were preincubated with 10 or 100 µM BSO, respectively, during a 16-h labeling period with 51Cr and then incubated with increasing concentrations of ethanol (6-10%) for 3 h. Values represent mean ± S.E.M. of the quadruplicate determinations on three separate preparation. ***P < .001, significant differences compared with control values (BSO).

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Effects of BSO concentration on ethanol-induced 51Cr release and glutathione content of epithelial or endothelial cells. Glutathione content and ethanol-induced 51Cr release were determined in epithelial (A) or endothelial (B) cells after a 16-h incubation with 0.33 to 10 µM or 10 to 333 µM BSO, respectively. Specific 51Cr release was determined after incubation with ethanol (8%) for 3 h. Glutathione content was extracted and assayed, as described in the text. Values for 51Cr release represent mean ± S.E.M. of the quadruplicates on three separate preparation. Values for glutathione content represent mean ± S.E.M. of the quadruplicate determinations in a single experiment. **P < .01 and ***P < .001, significant differences compared with control values.

Effects of Disruption of GSH Redox Cycle on Ethanol-Induced ⁵¹Cr Release in Endothelial Cells. Pretreatment of endothelial cells with BCNU (50 µg/ml) did not affect the dose-response curve for ethanol (6-10%) (Fig. 2B). Endothelial cell lysis brought about by ethanol was not influenced by inhibition of glutathione reductase activity with BCNU (Fig. 3B).

Effect of Inhibition of Endogenous Catalase by AT on Ethanol-Induced Injury. Treatment with 50 mM AT for 60 min inhibited endogenous catalase activity by 87%, from 3.2 ± 0.3 to 0.4 ± 0.1 IU/mg protein in epithelial cells, and by 90%, from 15.0 ± 1.2 to 1.5 ± 0.2 IU/mg protein in endothelial cells (mean ± S.E.M. of four determinations), without any alteration in cellular GSH levels or the activity of glutathione reductase in both cells (data not shown). Inhibition of cellular catalase by AT in epithelial and endothelial cells did not affect their susceptibility to ethanol (6-10%) (Table 1).

View this table: [in this window] [in a new window]   TABLE 1 Effect of aminotriazole on ethanol-induced 51Cr release from epithelial and endothelial cells 51Cr-labeled cells were preincubated for 60 min with buffer (AT) or 50 mM 3-amino-1,2,4-triazole (+AT) and then incubated with increasing concentrations of ethanol. Aliquots of supernatant buffer were sampled after a 3-h incubation for determination of specific 51Cr release. Values represent mean ± S.E.M. of triplicates on three separate preparations (n = 9).

Generation of O₂ by Epithelial and Endothelial Cells in Response to Ethanol Exposure. A 3-h incubation of epithelial cells with ethanol caused an increase in cytochrome c reduction as the concentrations of ethanol (6-10%) increased (Table 2). Exposure of epithelial cells to 10% ethanol for 3 h increased the reduction in cytochrome c from undetectable level (in control) to 2.96 ± 0.03 nmol/10⁶ cells, and the presence of SOD (500 U/ml) in the reaction mixture nearly completely prevented the increase. In contrast, 3-h exposure of endothelial cells to ethanol (6-10%) did not increase cytochrome c reduction compared with control incubation without ethanol.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The current study demonstrated that the glutathione redox cycle, but not endogenous catalase, plays a critical role in maintaining the integrity of gastric epithelial cells against ethanol-induced damage, whereas neither antioxidant defense seems to play a significant role in protection of endothelial cells against such injury. The distinct difference of the role of antioxidants in protection against ethanol appears to depend on the capability of each cell to produce cytotoxic ROS.

The pathogenesis of ethanol-induced gastric mucosal damage has been suggested to include several factors, such as arachidonate metabolites (Peskar et al., 1986), mast cell secretory products (Oates and Hakkinen, 1988), and further ROS (Mizui et al., 1987; Pihan et al., 1987; Szelenyi and Brune, 1988). Although the glutathione redox cycle and endogenous catalase may contribute to the detoxification of ROS (Nathan et al., 1980; Harlan et al., 1984; Hiraishi et al., 1991), there has been debate as to the role of glutathione in maintaining the gastric mucosal integrity against ethanol (Boyd et al., 1981; Szabo et al., 1981; Robert et al., 1984). Earlier studies focusing on the role of glutathione in vivo have used an electrophilic agent [i.e., diethyl maleate (DEM)] (Boyd et al., 1981; Robert et al., 1984) that depletes stores of GSH through the formation of a thioether conjugate in a reaction catalyzed by endogenous glutathione-S-transferase (Chasseaud, 1979), and a significant decrease in mucosal GSH levels by DEM was found (Boyd et al., 1981; Robert et al., 1984). Importantly, however, the obtained results are conflicting; although sulfhydryls offer protection against ethanol (Szabo et al., 1981), GSH depletion by DEM may induce mucosal ulceration for itself (Boyd et al., 1981) or may ameliorate ethanol-induced ulceration in the rat stomach (Robert et al., 1984). Therefore, the use of DEM as a depletor of mucosal GSH levels has been criticized in examination of the role of the glutathione redox cycle in gastric mucosal protection and injury.

Accordingly, the present study addressed the role of the glutathione cycle in protection against ethanol by disrupting the cycle at more than one site in cultured epithelial and endothelial cells to define its significance more clearly; these sites included selective inhibition of glutathione reductase by 10-min incubation with BCNU (Becker and Schirmer, 1995) and selective inhibition of -glutamylcysteine synthetase, a rate-limiting enzyme of GSH biosynthesis, by 16-h incubation with BSO (Meister, 1995).

Earlier studies on isolated or cultured gastric epithelial cells from rats (Mutoh et al., 1990; Nagy et al., 1994) and cultured bovine endothelial cells (Kviety et al., 1990) have shown that ethanol induces dose-dependent injury. The present study has demonstrated that impairment of the glutathione redox cycle of gastric epithelial cells at two independent sites rendered the cells more susceptible to ethanol. First, inhibition of glutathione reductase with BCNU (1-50 µg/ml) potentiated ethanol-induced lytic injury dose dependently, corresponding with the degree of inhibition of the enzyme (Figs. 2A and 3A). Second, selective inhibition of -glutamylcysteine synthetase with BSO (0.33-10 µM), resulting in diminished GSH biosynthesis, significantly increased epithelial cell susceptibility to ethanol-induced lysis (Figs. 4A and 5A). These observations, together with our previous finding on DEM (Mutoh et al., 1990), provide strong evidence that the glutathione redox cycle is specifically involved in protection of gastric epithelial cells against ethanol.

The technique to culture endothelial cells from rat gastric microvessels is not yet available; therefore, we used endothelial cells isolated from bovine aorta because of availability and the ability to easily produce pure endothelial cell culture. There are, however, phenotypic differences between the endothelium of large vessels and capillary microvessels and between species (Schor and Schor, 1986; Zetter, 1988). Also, extracellular matrix components (fibronectin, collagen IV, and laminin) provide structural support for gastric mucosal cells and influence cell migration, attachment, differentiation, and proliferation (Irwin et al., 1995). The response of each cell type to ethanol is therefore likely to be modified by the substratum on which the cells are grown. Because both cell types were cultured on plates and dishes from the same vendor without the addition of extracellular matrix components in the present study, the absence of extracellular matrix in the epithelial and endothelial cell cultures may result in a suppression of the characteristics exhibited by these cells in vivo.

With these limitations in mind, the present study has shown that impairment of the glutathione redox cycle did not result in increased susceptibility to ethanol injury in endothelial cells. Pretreatment of the cells with graded concentrations of BCNU (1-50 µg/ml), while inducing dose-dependent inhibition of glutathione reductase activity, failed to affect ethanol-induced lysis (Figs. 2B and 3B). In addition, inhibition of -glutamylcysteine synthetase with BSO (10-333 µM), while decreasing glutathione biosynthesis, failed to affect the sensitivity to ethanol (Figs. 4B and 5B). These findings indicate that the cycle plays a lesser role in maintaining the integrity of endothelial cells against ethanol, which is in clear contrast to the results on epithelial cells. The current study also suggested that endogenous catalase, which also metabolizes H₂O₂, is not involved in protecting either cell against ethanol. Inhibition of intracellular catalase did not diminish the resistance of either cell to ethanol (Table 1), indicating that endogenous catalase does not play a major role in maintaining the integrity of both cell types against ethanol.

There may be several possibilities as to the greater role of the glutathione redox cycle than endogenous catalase in the protection of epithelial cells against ethanol. First, catalase is concentrated in peroxisomes, whereas the glutathione redox cycle is localized in the cytosol and mitochondria (Jones et al., 1981); thus, the subcellular organelle targeted by ethanol may be either cytosol or mitochondria. Second, catalase displays a substantially higher K_m value for H₂O₂, whereas the glutathione redox cycle is operating at moderate intracellular H₂O₂ concentrations (Mannervik, 1985); thus, the accumulation of H₂O₂ within epithelial cells (possibly from disputation of O₂ generated in response to ethanol exposure, as described in the following paragraph), appears to be negligible or it may be an unimportant part of ethanol-mediated cell injury. Third, the catalase activity of epithelial cells (~3 IU/mg protein) is low compared with those of endothelial cells (~15 IU/mg protein) and cultured hepatocytes (~80 IU/mg protein; unpublished observation), so the enzyme may not efficiently detoxify accumulated H₂O₂. Finally, a chain of oxidative events leading to the formation of toxic lipid peroxides may be exclusively metabolized by the glutathione redox cycle (Lieber, 1988); indeed, ethanol has been shown to induce lipid peroxidation in rat gastric mucosa in vivo (Mizui et al., 1987).

With regard to the specificity of the reduction of exogenously applied extracellular cytochrome c as a marker for O₂ production, anything else released in the process of cell death (e.g., cytochrome c reductase) may produce a positive signal (O'Brien, 1984). Thus, the extent to which O₂ is responsible for the cytochrome c reduction is determined by measuring the amount of the reduction that is sensitive to SOD (Mayo and Curnutte, 1990); the difference in cytochrome c reduction in the presence and absence of SOD was used as a measure of O₂ production. We have found that gastric epithelial cells are capable of producing O₂ in response to exposure to ethanol, without the involvement of blood flow or polymorphonuclear neutrophils (Table 2). In contrast, however, O₂ production from endothelial cells did not increase when cells were exposed to ethanol. These results indicate that epithelial cells, when exposed to ethanol, generated O₂ as a function of ethanol concentration, whereas endothelial cells did not, and this is the first report to demonstrate that the capability to produce ROS in response to ethanol exposure depends on cell type.

With respect to the source of O₂ generation by gastric epithelial cells, one possible source may be the oxidative metabolism of ethanol by microsomes involving cytochrome P-450 (Lieber, 1988). However, this may be less likely because P450IIE1, a microsomal P450 enzyme inducible by ethanol, is not detected in the rat gastric mucosa either constitutively or after ethanol treatment (Shimizu et al., 1990). Another possibility may be cellular xanthine oxidase system. Gastric mucosae of the rat and human are shown to possess alcohol dehydrogenase activity (Seitz et al., 1989), which can convert ethanol to acetaldehyde; acetaldehyde can be a substrate of xanthine oxidase to produce ROS including O₂ (Steinbeck et al., 1993). Indeed, two inhibitors of xanthine oxidase, allopurinol and oxypurinol, have been shown to provide protection against ethanol-induced injury in the rat stomach (Mizui et al., 1987). These issues on the possible source of ROS generation are currently under investigation in our laboratory.

In summary, impairment of the glutathione redox cycle of cultured gastric epithelial cells at two independent sites rendered these cells more susceptible to ethanol injury, and inhibition of endogenous catalase did not affect the resistance of these cells to ethanol. In contrast, neither disruption of the glutathione cycle nor inhibition of catalase influenced the susceptibility of endothelial cells to ethanol. These results suggest that the glutathione redox cycle plays a critical role in maintaining the integrity of gastric epithelial cells against ethanol-induced damage, whereas endogenous catalase does not. Neither of these antioxidants seems to play a significant role in protection of endothelial cells against ethanol damage. The distinctly different significance of antioxidants seems to depend on the capacity of each cell type to generate cytotoxic ROS in response to ethanol exposure.

Because ingested ethanol diffuses into the gastric mucosa, targeting both epithelium and endothelium (Guth et al., 1984; Lacy and Ito, 1984; Pihan et al., 1986), and because extracellular GSH can be taken up by gastric epithelial cells through mediation of -glutamyl transpeptidase and -glutamylcysteine synthetase (Hiraishi et al., 1994), dietary GSH may offer protection of gastric epithelium against intragastric ethanol through accumulation of mucosal GSH in vivo as well.