Net Glutathione Secretion across Primary Cultured Rabbit Conjunctival Epithelial Cell Layers

Abstract

purpose. Metabolism and transport of glutathione (GSH), the endogenous thiol antioxidant, in conjunctival tissue to date are poorly understood. The purpose of the present study was to define transport characteristics of GSH in primary cultured rabbit conjunctival epithelial cells (RCECs).

methods. RCECs were grown on membrane filters to exhibit tight barrier properties (transepithelial electrical resistance, TEER, ∼1 kΩ/cm²). Uptake, efflux, and transepithelial transport of GSH were determined in the presence or absence of extracellular Na⁺ under conditions of inhibition of GSH biosynthesis and degradation. Uptake was determined at 15 minutes after instillation of ³H-GSH to the apical or basolateral bathing fluid. GSH efflux was estimated from the time course of release of prebiosynthesized ³⁵S-GSH. Transepithelial transport was assessed by instillation of ³H-GSH in either the apical or basolateral bathing fluid, followed by sampling from respective contralateral sides.

results. Apical uptake and efflux showed Na⁺ dependency up to 65%. GSH uptake in the initial 15 minutes was linear in the presence of 1 mM GSH (labeled and unlabeled) in Na⁺-containing buffer. The uptake rate was higher from the apical fluid than from the basolateral fluid. A Hill analysis of the Na⁺-dependent process yielded a coupling ratio for Na⁺ to GSH of 1.25:1. The efflux rate of GSH into the apical fluid was marginally dependent on the apical presence of Na⁺ and was significantly greater than that in the basolateral fluid. Basolateral efflux of GSH was primarily Na⁺ independent, whereas basolateral uptake almost exclusively was Na⁺ dependent. Depolarizing the RCEC membrane potential decreased GSH efflux into either apical or basolateral fluids (5 pmol/min · 10⁶ cells). Hyperpolarization significantly increased the rate of GSH efflux into the apical fluid (120 pmol/min · 10⁶ cells), whereas the basolateral efflux was not affected. Apparent permeability of GSH across RCEC layers was approximately eight times higher in the basolateral-to-apical (secretion) direction than the opposite (absorption) direction.

conclusions. GSH is transported across RCEC membranes by both Na⁺-dependent and -independent processes. Analysis of the Na⁺-dependent uptake process gave an approximate 1:1 coupling ratio for Na⁺-GSH cotransport. The Na⁺-independent component is highly sensitive to cell membrane potential. Net secretion of GSH into the apical fluid may play a role in the protection of conjunctival tissue and tear film from oxidant insults.

The conjunctiva is a thin, mucus-secreting, vascularized tissue that covers most of the inner surface of the eyelids and is part of the anterior sclera where the cornea begins. ¹ The conjunctiva is thought to function as a passive physical protective barrier and to participate in the maintenance of tear film stability due to the mucus secreted by the resident goblet cells. ¹ In recent years, our laboratory, as well as others, have provided ample evidence for additional functional features of the conjunctiva, including acting as a conduit for drug delivery to the posterior segment of the eye after ocular drug instillation and contributing to the regulation of electrolyte and fluid balance in the microenvironment of its mucosal surface. ² The dynamic nature of conjunctiva has been demonstrated from the identification of several transport mechanisms for Na⁺-absorption: Na⁺-glucose, ^{3 4} Na⁺-amino acid, ⁵ and Na⁺-nucleoside cotransporters, ⁶ in addition to active Cl⁻ secretion. ⁷ Permeability of the conjunctiva to a wide variety of hydrophilic and lipophilic molecules has also been reported. ^{8 9 10}  

Glutathione (GSH) is the most abundant, endogenous thiol-containing tripeptide present in nearly all cell types. Its vital roles in cells include detoxification of electrophiles, maintenance of thiol status of proteins, free radical scavenging, provision of an intracellular store for cysteine, and participation in immunoprotection. ¹¹ No information is available to date on the metabolism, transport, or biosynthesis of GSH in normal conjunctiva, let alone the effect of tissue GSH status and metabolism on development of conjunctival diseases. 

Export of GSH into the extracellular space contributes predominantly to the turnover of cellular GSH. A facilitative GSH efflux system that is Na⁺ independent and functions to mediate GSH transport from high millimolar tissue concentrations to micromolar levels in vascular fluid has been characterized in several cell types. ¹² In recent years, evidence has been presented for a role of the adenosine triphosphate (ATP)-binding cassette (ABC) transporter family as GSH carriers. Some reports suggest that organic anion transporter protein (oatp1), the hepatic basolateral organic solute transporter, and Mrp1 or -2, the ATP-dependent multidrug resistance-associated plasma membrane transport proteins (basolateral or canalicular, respectively), all mediate GSH efflux in hepatocytes. ^{13 14} Evidence for the presence of Na⁺-dependent GSH uptake transporters in tissues such as the lens and liver have also been presented in recent studies, ^{15 16} although the plasma membrane transport mechanisms with respect to asymmetry and regulation have not been studied to date. 

Our laboratory has developed an air-interface culture of pigmented rabbit conjunctival epithelial cells. ^{17 18} This culture model of conjunctival epithelial cell layers exhibits tight barrier properties ¹⁷ and both active Cl⁻ secretion and Na⁺ absorption. ⁷ In this study, we have examined the metabolism and transport of GSH across primary cultured rabbit conjunctival epithelial cell (RCEC) layers. Evidence for saturable and asymmetrical transport processes for GSH in conjunctival epithelial cells is presented herein. 

Methods

Primary Culture of RCEC Layers

The investigations using rabbits described in this report conformed to the Guiding Principles in the Care and Use of Animals (Department of Health, Education and Welfare Publication, National Institutes of Health 80-23) and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Male, Dutch-belted pigmented rabbits, weighing 2.0 to 2.5 kg, were used for isolation of conjunctival epithelial cells, as described in detail in our previous publications. ^{17 18} Briefly, rabbits were killed with an injection of 85 mg/kg pentobarbital sodium solution into a marginal ear vein. Conjunctival tissues were excised and incubated in 0.2% protease type XIV (Sigma, St. Louis, MO) at 37°C for 30 minutes. Epithelial cells were scraped off using a sterile scalpel blade, suspended in a minimum essential medium (S-MEM) containing 10% fetal bovine serum (FBS) and 0.5 mg/mL DNase I (Sigma, St. Louis, MO), and centrifuged at 200g for 10 minutes. After two consecutive washes of the pelleted cells with S-MEM containing 10% FBS, with filtration of resuspended cells through a 40-μm cell strainer (Becton Dickinson Labware, Franklin Lakes, NJ), the final cell pellet was resuspended in PC-1 medium (BioWhittaker, Walkersville, MD) supplemented with 2 mM l-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 50 μg/mL gentamicin, and 1 μg/mL amphotericin B. These cells at a density of 1.2 × 10⁶ cells/cm² (day 0) were then placed onto polystyrene membrane filters (12 mm diameter, 0.4 μm pore size; Clearwell; Corning-Costar, Cambridge, MA) that were precoated with a mixture of 25 μg/cm² rat tail type I collagen (Collaborative Biomedical Products, Bedford, MA) and cultured in a humidified atmosphere of 5% CO₂ and 95% air at 37°C. The volumes of the apical and basolateral compartments were 0.5 and 1.5 mL, respectively, on days 0 through 4. On day 4 and thereafter, the cells were maintained at an air interface defined by the nominal absence of the apical bathing fluid and 0.8 mL of medium present in the basolateral compartment. These air-interfaced cultures of RCEC layers were used in all our studies. Characterization using light and electron microscopy showed that primary cultures of RCEC layers contained goblet and squamous cells, with the former comprising approximately 4% in the air-interfaced configuration. A sporadic and diffuse periodic acid-Schiff (PAS)-positive staining pattern was seen in all air-interfaced cultures of RCEC layers, suggestive of the presence of mucin-secreting goblet cells. ^{18 19 20} Ion transport characteristics of these cultures are similar to those of the excised conjunctival tissue. 

Measurement of Bioelectric Parameters

Potential difference (PD) and transepithelial electrical resistance (TEER) were monitored from day 2 onward, to assess viability and barrier tightness using an epithelial voltohmmeter (EVOM; World Precision Instruments, Sarasota, FL), and the equivalent short circuit current (I _eq) was calculated using Ohm’s law: I _eq = PD/TEER. ¹⁸ The RCEC layers were used for GSH transport studies after reaching peak electrical parameters from days 5 through 8: peak TEER of approximately 1 kΩ/cm² and PD of approximately 17 mV. ¹⁸ The final cell density of confluent RCEC layers was 0.65 × 10⁶ cells/cm² (day 7). 

GSH Concentration and γ-Glutamyl Transpeptidase and γ-Glutamylcysteine Synthetase Expression in RCECs

The GSH level in RCEC layers was determined using the Tietze recycling assay. ²¹ The presence of γ-glutamyl transpeptidase (GGT) in RCECs and in normal conjunctival tissue was confirmed by Western blot analysis of cell proteins with an antibody kindly provided by Marie Hanigan (University of Virginia Health Sciences Center, Charlottesville, VA). The polyclonal antibody was raised against a peptide sequence consisting of 20 amino acids in the N terminus of the human GGT protein. ²² Anti-rabbit IgG conjugated to horseradish peroxidase was used as the secondary antibody. A human hepatoma cell line (HepG2) was used as a positive control. Procedures for homogenization of conjunctival tissue and RCEC layers and processing for Western blot analysis were similar to that for γ-glutamylcysteine synthetase (GCS), described in the next section. 

Presence of the regulatory, light subunit (GCS-LS) and catalytic, heavy subunit (GCS-HS) of GCS in conjunctiva was verified by Western blot analysis of cell proteins obtained from RCEC layers and excised conjunctival tissue. Rat liver homogenates were used as a positive control for GCS. Rabbit polyclonal antibodies for GCS-LS and -HS were a gift from Terrence Cavanagh (University of Washington, Seattle, WA). 

RCEC layers or excised conjunctival tissues were homogenized with the use of a tissue grinder (Pyrex brand 16 × 150 mm; Corning, Inc., Corning, NY) in a homogenizing buffer (250 mM sucrose, 10 mM phosphate [pH 7.4], with antiprotease cocktails [Sigma]). Homogenates were centrifuged at 13,000g for 10 minutes at 4°C. The supernatants were used for Western blot analysis, as described previously. ²³ Because the synthetic peptides used to generate GCS antibodies were linked to ovalbumin (OVA), we used an OVA-containing blocking agent in Western blot analysis. The secondary antibody was horseradish peroxidase-conjugated goat anti-rabbit IgG (Roche Molecular Biochemicals, Indianapolis, IN). 

GSH Uptake

RCECs grown on 12-mm internal diameter, 0.4-μm pore size membrane filters (∼0.7 × 10⁶ cells/well) were pretreated for 30 minutes with 1 mM acivicin and 10 mM d,l-buthionine sulfoximine (BSO) in NaCl buffer composed of 135 mM NaCl, 1.2 mM MgCl₂, 0.81 mM MgSO₄, and 25 mM HEPES (pH adjusted to 7.4 using Tris-base) before all uptake studies. ²⁴ To dissociate GGT-mediated breakdown and resynthesis from intact GSH uptake, we performed all studies in the presence of acivicin and BSO, inhibitors of GSH breakdown and synthesis, respectively. Acivicin is an irreversible inhibitor of GGT, and BSO is an inhibitor of the rate-limiting enzyme of GSH biosynthesis GCS. ¹¹ Evidence for the expression of these two enzymes in normal rabbit conjunctival tissue and in primary RCECs was established by Western blot analysis. In initial experiments, the optimal incubation time (for assessing the linear uptake from the apical and basolateral fluids) was determined by incubating the BSO-acivicin-treated cells for 2, 5, 10, 15, 30, and 60 minutes with ³H-GSH ([glycine-2-³H]; 50 Ci/mmol, Perkin Elmer Life Sciences Inc., Boston, MA) at 2.5 μCi/mL (>98% purity estimated by HPLC) containing 1 mM unlabeled GSH in NaCl buffer. To assess temperature dependency, cellular uptake was performed in three wells per time point, individually for each fluid domain (e.g., apical or basolateral) at 37°C and 4°C. The latter data represent nonspecific binding of GSH to RCECs. Cellular uptake was terminated by suctioning off the dosing solution at indicated time points, and the accumulated cellular radioactivity was determined after three consecutive washes in ice-cold NaCl buffer, 100 mL each. RCEC layers were then cut out, and cells were lysed with a 0.5-mL mixture of 0.1% Triton X-100 and 0.1 N NaOH for 1 hour at room temperature. All transport data for labeled GSH were corrected for nonspecific adsorption at 4°C. Some RCEC layers were incubated with trypsin-EDTA (0.05% and 0.02%, respectively) for 10 minutes at 37°C to liberate the cells from the polystyrene filters, and the number of cells was then counted for normalization of GSH transport data. 

Concentration Dependency and Na⁺ Requirement of GSH Uptake

To determine whether GSH entry into RCEC layers exhibits sodium dependency at varying GSH concentrations, apical or basolateral uptake studies were performed for 15 minutes. For this purpose, 0.01, 0.1, and 1 mM unlabeled GSH was dissolved in an NaCl or choline chloride buffer containing both 2 mM dithiothreitol (DTT) and 2.5 μCi ³H-GSH. DTT was used to maintain GSH in reduced form for the duration of incubations in our transport studies. The choline chloride buffer had the same composition as the NaCl buffer, except that 135 mM choline chloride was substituted for 135 mM NaCl. 

Effect of varying the sodium concentration on GSH uptake was evaluated using the data obtained from 0.1 mM GSH uptake studies, where 2 mM DTT and 2.5 μCi ³H-GSH were present in incubation buffers containing 0, 5, 10, 20, 35, 85, or 135 mM NaCl. These buffers were prepared by mixing the NaCl and choline chloride buffers in different proportions. Uptake of GSH from apical and basolateral fluids were measured as described earlier. GSH uptake data were plotted against Na⁺ concentration and analyzed using the logarithmic form of the Hill equation: log [v/(V _max − v)] = n log (S)− log K′) as described earlier, ²⁵ where v, V _max, n, and K′ represent the initial velocity, maximal velocity, number of substrate binding sites, and a constant comprising multiple interaction factors and the intrinsic dissociation constant, respectively. The Hill analysis was performed by computer (Origin, ver.6.0; Microcal Software, Inc., Northampton, MA). 

GSH Efflux

For GSH efflux studies, RCEC layers were prelabeled on day 6 with 10 μCi/mL ³⁵S-cysteine (Cysteine, l-[³⁵S]; >1000 Ci/mmol, Perkin Elmer Life Sciences Inc., Boston, MA) from both bathing fluids for 12 hours in the presence of 1 mM DTT and 1 mM acivicin. Cell layers were washed three times with respective buffers (i.e., NaCl or choline chloride buffers) before efflux measurements to remove extracellular radioactivity. Next, efflux studies were performed in fresh NaCl and choline chloride buffers. Efflux of radiolabeled compounds in RCECs to apical or basolateral bathing fluids was determined by measuring the released radioactivity in 100-μL aliquots of extracellular fluid at 1, 2, 4, 5, 15, and 30 minutes. Depolarization of an inside negative cell membrane potential was achieved by incubating RCEC layers in an apical high-K⁺ buffer (i.e., 135 mM KCl and 5 mM NaCl, with other ionic compositions being the same as those in NaCl buffer) for 10 to 15 minutes at 37°C, with or without the presence of 1 μM of the K⁺ ionophore valinomycin in both bathing fluids. Such a buffer composition creates conditions in which the buffer’s K⁺ concentration closely resembles or (by the inclusion of valinomycin) approaches that in the cytoplasm of cells, making the cell membrane potential less negative. Hyperpolarization of the cell membrane was achieved by the addition of 1 μM valinomycin to the regular NaCl incubation buffer for 10 to 15 minutes at 37°C, from either the apical or basolateral fluids, depending on the extracellular fluid compartment in which the efflux of radiolabeled compounds was being determined. Valinomycin added to the apical compartment allows K⁺ to escape from the high (>100 mM) cytoplasmic concentration of RCEC layers, into the NaCl incubation buffer, in which there is a low K⁺ concentration (5 mM), making the potential across the apical membrane more negative. The same rationale applies to basolateral membrane hyperpolarization. 

Transepithelial GSH Flux

Transepithelial flux of GSH was determined by spiking the ipsilateral bathing fluid with 2.5 μCi ³H-GSH and sampling aliquots at 5, 15, 30, and 60 minutes from the contralateral bathing fluid. ^{8 18 26} These studies were conducted in the presence of 0.01 mM unlabeled GSH in NaCl buffer at 37°C, under conditions of inhibition of GSH synthesis and degradation (pretreatment with 10 mM BSO and 1 mM acivicin, respectively). The integrity of the epithelial barrier was monitored by measuring the PD and TEER of RCEC layers, before and after flux experiments. The apparent permeability coefficient (P _app) was estimated from the steady state slope of a plot of the cumulative amount of the radiolabeled tracer appearing in the contralateral fluid versus time. ¹⁸ Apparent permeability is expressed as P _app = J/(A · C ₀), where the steady state flux (J) of GSH is normalized by both the effective surface area (A) and initial dose concentration (C ₀). 

Statistical Analysis

The results are expressed as the mean ± SEM for four to six determinations per data group. An unpaired, two-tailed Student’s t-test was used to determine the statistical difference between two group means. When comparing three or more group means, one-way analysis of variance (ANOVA) was used. Statistical significance among the group means was determined by the modified Fisher least-squares difference approach. Differences were considered statistically significant when P ≤ 0.05. 

Results

GSH Content, GCS and GGT Expression, and GSH Turnover in RCECs

Intracellular GSH content of RCEC layers averaged 9.1 ± 0.2 nmol/10⁶ cells. Assuming that RCECs are spherically shaped with radii ranging from 5 to 10 μm, they have an estimated cell volume of, at most, 9 to 10 pL. The intracellular concentration of GSH in RCECs would then be at least 1 mM. Western blot analysis revealed the presence of both the regulatory (30 kDa) and catalytic (78 kDa) subunits (GCS-LS and GCS-HS, respectively) of the rate-limiting enzyme in GSH biosynthesis (Fig. 1A) . In addition, expression of GGT protein, the enzyme involved in GSH catabolism, in excised conjunctiva and cultured RCECs was confirmed by a band at 66 kDa in Western blot analysis (Fig. 1B) . Additional information on the turnover of GSH was also obtained in experiments on kinetics of the disappearance of GSH in the presence of BSO, an inhibitor of GSH biosynthesis. When cells were treated with BSO, a time-dependent decrease in the cellular GSH level was observed. An estimate of the half-life for BSO depletion of cellular GSH content was approximately 4 hours (Fig. 1C) . 

GSH Uptake in RCEC Layers: Time Course, Sodium Dependency, and Asymmetry

Labeled GSH uptake in the presence of 1 mM unlabeled GSH from the apical and basolateral fluids of RCEC layers was determined at 37°C and 4°C at 2, 5, 15, 30, and 60 minutes (data not shown). These measurements were recorded in three wells per time point, individually for each fluid domain (e.g., apical or basolateral). Uptake at 4°C, an estimate of nonspecific binding, did not increase significantly with time. Corrected GSH uptake, which is defined as the difference between uptake at 37°C and adsorption at 4°C for each time point, was linear for 15 minutes. 

Requirement of Na⁺ for GSH uptake was examined next with the duration of these studies set at 15 minutes. Uptake data of ³H-GSH shown in Figure 2 were obtained after a 30-minute pretreatment of RCECs with both acivicin and BSO at 37°C in NaCl or choline chloride buffer. Labeled GSH uptake from either apical or basolateral fluids was competitively inhibited by unlabeled GSH and was partially Na⁺ dependent at all unlabeled GSH concentrations (Fig. 2) . Apical ³H-GSH uptake was greater than basolateral uptake in the presence of Na⁺, except for 0.10 mM GSH, as determined by ANOVA (Fig. 2) . It should be noted that TEER of RCEC layers (1000 ± 100 Ω/cm², n = 18) did not change significantly during the uptake period when incubated in Na⁺-containing or Na⁺-free buffers compared with that in the regular culture medium. 

Hill Analysis of GSH Uptake

We performed GSH uptake studies in which the uptake of GSH at a fixed concentration of 100 μM was evaluated as a function of varying concentrations of Na⁺ in the incubation buffer. Hill coefficients for apical (1.2:1) and basolateral (1.3:1) GSH uptakes as a function of Na⁺ were slightly greater than 1, suggesting at least a 1:1 coupling ratio for Na⁺-GSH for the Na⁺-dependent uptake process in both apical and basolateral cell membranes (Fig. 3) . The calculated K _m in both apical and basolateral uptake studies was approximately 14 ± 1 mM Na⁺. V _max for Na⁺-dependent GSH uptake, however, was 29 ± 2 pmol/10⁶ cells in apical studies and 19 ± 2 pmol/10⁶ cells in basolateral studies. 

GSH Efflux in RCEC Layers: Time Course, Sodium Requirement, and Asymmetry

RCEC layers were metabolically prelabeled for 12 hours with 10 μCi of ³⁵S-l-cysteine in the presence of 1 mM acivicin starting on day 6 of culture. In preliminary experiments, we verified that, after 12 hours of incubation, the cellular radioactivity (∼0.4% of the original dose) was predominantly (>96%) in the form of GSH under these experimental conditions. Time-course experiments conducted up to 30 minutes in NaCl or choline chloride buffers showed that GSH efflux (picomoles per million cells) was linear for the first 5 minutes in the presence or absence of Na⁺ in the incubation buffer (Fig. 4) . The apical efflux was slower with Na⁺-containing buffer than without Na⁺ in the buffer by ANOVA (Fig. 4A) , whereas there was no significant difference between those in the presence or absence of Na⁺ in the basolateral fluid (Fig. 4B) . The initial rates of efflux in the apical fluid in the presence and absence of Na⁺ were 50 and 90 pmol/min per 10⁶ cells, respectively. The basolateral efflux with or without Na⁺ was approximately the same at 20 pmol/min per 10⁶ cells. 

Membrane Potential Dependency of GSH Efflux

Depolarization of the cells by incubating RCEC layers in apical high-K⁺ buffer, regardless of copresence in apical and basolateral fluids of 1 μM of the K⁺ ionophore valinomycin, resulted in approximately 90% reduction of GSH efflux into the apical and basolateral fluids (Fig. 4B) . Hyperpolarizing cell membranes by the addition of 1 μM valinomycin to the NaCl incubation buffer from apical or basolateral fluids stimulated the rate of efflux more than twofold into apical fluid, but had no detectable effect on efflux into the basolateral fluid by ANOVA (Figs. 4A 4B) . TEER (900 ± 100 Ω/cm², n = 18) of the cell layers estimated before and after the experiments did not change significantly. 

Transepithelial Transport of GSH in Cultured RCECs

The flux of GSH, estimated from the steady state slope of a plot of the cumulative amount of the radiolabeled tracer appearing in the contralateral fluid versus time, ¹⁸ was several times higher than the paracellular passive diffusion marker mannitol and exhibited directionality unlike mannitol. Unidirectional flux in the apical-to-basolateral direction (J _ab) and that in the basolateral-to-apical direction (J _ba), were 37.7 pmol/h · cm² and 308 pmol/h · cm², respectively. Thus, net flux (J _ba − J _ab) of GSH was 270 pmol/h · cm² in the basolateral-to-apical direction, suggesting net carrier-mediated GSH secretion into apical fluid. The transepithelial permeabilities (P _app = J/[A · C ₀]) for GSH measured in the presence and absence of Na⁺ ions are summarized in Table 1 . In the absorptive (i.e., apical-to-basolateral) direction, the presence or absence of Na⁺ in the incubation buffer of both apical and basolateral compartments made no significant difference in P _app for GSH. However, in the secretory (i.e., basolateral-to-apical) direction, the P _app for GSH was two times higher in the presence of Na⁺ than in its absence in the choline chloride incubation medium (Table 1) . Additionally, basolateral-to-apical permeability for GSH was eight and four times greater than that of the apical-to-basolateral direction in the presence and absence of Na⁺, respectively. GSH P _app was two orders of magnitude greater than that for mannitol. 

Discussion

The major finding of the present study relates to transepithelial flux of GSH across the conjunctiva. Our data show that a net secretion of GSH occurred in the basolateral-to-apical direction across conjunctiva under the experimental conditions used (Table 1) . GSH transport across cultured RCEC membranes was by both Na⁺-dependent and -independent processes, and the Na⁺-independent component was driven by membrane potential. 

At the outset we established the presence of biosynthetic and degradative machinery for GSH in primary cultured RCECs and in the excised conjunctival tissue. Western blot analysis showed the expression of both subunits of the rate-limiting enzyme of GSH synthesis, GCS, and of the degradative enzyme, GGT. Although there is no report on enzymatic activity or expression of GCS in conjunctiva, GGT activity was demonstrated in human tear fluid in one study. ²⁷ The 1 mM steady state concentration of GSH in primary cultured RCEC, calculated from the data in Figure 1 (assuming 5–10-μm radii for spherical epithelial cells) is similar to several epithelial cell types but is significantly lower than that of lens epithelium or the hepatocyte. ²⁸ The turnover rate of approximately 4 hours for GSH in conjunctiva also lies between that in liver (∼2 hours) and the lens (∼24 hours). ^{11 16 28}  

To dissociate GGT-mediated breakdown and resynthesis from intact GSH uptake, we performed all studies in the presence of acivicin and BSO, inhibitors of GSH breakdown and synthesis, respectively. Both Na⁺-dependent and -independent uptake of GSH were found in RCECs. Sodium-dependent GSH uptake in RCECs was greater from apical fluid and almost exclusively mediated the GSH transport across the basolateral membrane (Fig. 2) . The Na⁺-dependent GSH transporter has been reported to mediate GSH uptake against a concentration gradient. ^{29 30} Sodium-GSH cotransport has been described only in some tissues that include the eye, brain endothelium, and epithelial cells from kidney and intestine. ^{24 29 30 31 32 33 34 35} Furthermore, in addition to sodium, other cations such as K⁺, Ca²⁺, and Mg²⁺ increase GSH uptake in intestine. ^{34 35} Whereas uptake is electroneutral with monovalent cations, electrogenic transport was demonstrated with divalent cations in brush border membrane vesicle preparations. ³⁵ In our present studies, Na⁺-dependent GSH transport in RCECs was electroneutral. Hill analysis of Na⁺-coupled GSH uptake in RCEC confirmed the requirement of at least 1 mole of sodium for each 1 mole of GSH taken up (Fig. 3) . However, the exact mechanisms underlying Na⁺-coupled GSH transport are unknown to date. 

The Na⁺-independent transport of GSH has been shown in a number of hepatic and nonhepatic tissues. ¹⁶ The facilitative GSH transporter mediates GSH efflux under physiological conditions from high-millimolar concentrations in the cells ²⁸ to micromolar levels in extracellular fluids. The human tear film GSH concentration is approximately 110 μM. ^{36 37} Accordingly, GSH efflux was dominant compared with uptake in RCECs. The rates of GSH efflux into either apical or basolateral fluid were significantly greater than those observed for uptake. Moreover, apical efflux rate of GSH was greater than the basolateral rate. The significantly lower apical efflux in the presence of sodium than that in its absence may be due to Na⁺-dependent apical reuptake of the GSH secreted into the apical fluid. 

Glutathione is negatively charged at physiological pH. Hence, the resting electrochemical potential acting across cell membranes favors and essentially regulates the efflux of GSH. ^{16 34 35} Cell membrane potential is primarily determined by the relative impermeability of cells to potassium ions. At steady state, there is a high cytoplasmic concentration of potassium compared with that of extracellular fluid levels. The presence of the K⁺ ionophore valinomycin caused potassium to leak from the RCECs, hyperpolarizing the existing inside-negative cell membrane potential. A more negative membrane potential resulted in an increasing GSH efflux rate into apical fluid when valinomycin was included in the apical bathing solution. Basolateral efflux did not change drastically under the same conditions. Moreover, when a buffer containing high potassium concentrations was introduced to the apical compartment in the presence or absence of valinomycin at both sides of RCEC layers, the GSH efflux into the apical and basolateral fluids was virtually abolished. High extracellular potassium (135 mM KCl and 5 mM NaCl) depolarizes the cell membrane potential, making it less negative. The presence of the ionophore valinomycin probably accelerates the equilibration of cytoplasmic and extracellular concentrations of potassium across cell membranes, but apparently was not needed to achieve the depolarizing effect on RCEC apical and/or basolateral membranes. 

A composite, hypothetical scheme built from extrapolation of the rates of conjunctival GSH transport we observed to those that might be operating in vivo is presented in Figure 5 . The transport rates (uptake and efflux at each cell membrane, as well as paracellular diffusion) shown in Figure 5 were determined by making some assumptions. We assumed GSH concentrations of 0.107 and 0.015 mM for the tear side and serosal side respectively based on literature data, ^{28 36} whereas the intracellular GSH concentration measured in this study was approximately 1 mM. The calculated unidirectional flux rates of GSH transport across each cell membrane and net movement are expressed as picomoles of GSH per hour per square centimeter. The efflux rates of GSH were significantly higher than those for uptake for both apical and basolateral membranes and thus are the determining factors for direction of net GSH flux across RCEC layers. Net GSH secretion of 1080 pmol/h · cm² was estimated, which is the result of the vectorial sum of all flux rates. It should be noted that the permeability coefficients include a passive diffusional component of transepithelial GSH transport (which should be the same in either apical-to-basolateral or opposite direction, with a similar GSH concentration gradient across the epithelium). The net secretion of GSH in the basolateral-to-apical direction is thought to play an important role in protection of the conjunctiva from injury by exogenous toxic chemicals and oxidants. However, to what extent conjunctival GSH secretion contributes to overall levels in tear film in health and disease remains to be studied further. 

In summary, in our study GSH was transported across the conjunctival epithelial cell membranes by both Na⁺-dependent and -independent processes in a predominantly saturable manner. GSH efflux represented a major component of GSH transport in RCEC layers. The observed net secretion into the apical fluid may play an important role in the protection of conjunctival tissue from oxidant insult. 

 Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2001.

 Supported by National Eye Institute Grants EY11135 (RK) and EY12356 (VHLL); National Heart, Lung, and Blood Institute Grants HL38658 (K-JK) and HL64365 (K-JK, VHLL); and American Heart Association Grant AHA-GIA 990542N. HJG is the recipient of a Predoctoral Fellowship from the American Foundation for Pharmaceutical Education.

 Submitted for publication September 12, 2001; revised December 17, 2001; accepted January 2, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Western blot analysis of GCS-HS and GCS-LS (A) and GGT (B) expressed in excised conjunctival tissues and RCEC cultures. Each lane was equally loaded with 150 μg protein. The molecular size of each protein is shown next to the band of interest. The bands shown at 78 and 30 kDa in (A) were found to be the major bands for GCS-HS and -LS, respectively, whereas minor faint bands were seen near the origin (not shown). The 66-kDa band in (B) was the only band seen in Western blot analysis of GGT. (C) Total intracellular GSH levels as a function of incubation time with 0.1 mM BSO, an inhibitor of GCS, the rate-limiting enzyme of GSH biosynthesis. BSO was added to both bathing fluids of the RCEC layers grown on membrane filters at 37°C for the time indicated, and GSH levels at each time point were estimated. Cells remained viable throughout the 24-hour experimental period, because BSO exposure did not alter the TEER significantly (range, 800–1000 Ω/cm²). Data are mean ± SEM (n = 4).

Figure 2.

 

³H-GSH uptake at three extracellular GSH concentrations and effect of Na⁺ replacement with choline on uptake from either apical or basolateral fluids of cultured RCEC layers pretreated for 30 minutes with 1 mM acivicin and 10 mM BSO, inhibitors of enzymes of GSH degradation and synthesis. Uptake was conducted in NaCl and choline chloride buffers. Apical GSH uptake in the presence of Na⁺ was significantly greater than basolateral, as determined by ANOVA for 0.01 mM and 1.00 mM GSH, and did not reach statistical significance (modified Fisher least-squares difference for 0.10 mM GSH). *Significant difference at P < 0.05 compared with Na⁺-containing group by one-way ANOVA. ‡Significant difference at P < 0.02 compared with apical uptake in the presence of Na⁺. Data are the mean ± SEM (n = 6).

Figure 3.

 

Hill plot of GSH uptake as a function of Na⁺ concentration in primary cultured RCEC layers grown on membrane filters. Cells were pretreated for 30 minutes with 1 mM acivicin and 10 mM BSO, inhibitors of GSH breakdown and synthesis. Apical or basolateral uptake was determined individually for each domain using ³H-GSH at 2.5 μCi/mL in the presence of 100 μM unlabeled GSH. Incubation buffers containing 0, 5, 10, 20, 35, 85, and 135 mM NaCl were used. At concentrations less than 135 mM NaCl, the buffers were made by mixing NaCl buffer with choline chloride buffer at different proportions. The slopes of 1.2 (R ² = 1.00) and 1.3 (R ² = 0.99) for uptake from apical and basolateral fluids, respectively, were estimated, suggesting a coupling ratio of at least one Na⁺ for each GSH molecule taken up.

Figure 4.

 

Time course of GSH efflux from primary cultured RCEC layers studied on day 7. Efflux studies were performed at 37°C by measuring radioactivity in aliquots taken from apical (A) and basolateral (B) fluids of RCEC layers prelabeled for 12 hours with 10 μCi ³⁵S-cysteine in the presence of 1 mM acivicin (to block degradation of GSH), starting on day 6 of culture. The apical GSH efflux was significantly slower with Na⁺-containing buffer than without Na⁺ in the buffer (A, ▴ versus ○), whereas there was no significant difference between those in the presence or absence of Na⁺ in the basolateral fluid (B, ▴ versus ○). Depolarization of RCEC layers resulted in significant reduction of GSH efflux into both apical (A, ▵) and basolateral (B, ▵) fluids. Hyperpolarization of RCEC layers significantly increased efflux into apical fluid (A, •), but did not significantly change GSH efflux into basolateral fluid (B, •). val., valinomycin. Statistical comparisons using ANOVA are as follows: in (A) ▴ vs. ○ P < 0.02; in (A) and (B) ▵ vs. ▴ or ○ P < 0.05; in (A) • vs. ▴ or ○ P < 0.05. Data are mean ± SEM (n = 6).

Table 1.

 

Transepithelial Permeability of GSH in RCECs

	Na+	Apical-to-Basolateral	Basolateral-to-Apical
GSH	+	560 ± 310	4540 ± 230* , † , ‡
GSH	−	520 ± 330	2170 ± 110* , ‡
Mannitol	±	2.10 ± 0.15	2.49 ± 0.11

 Transport of 10 μM GSH (labeled and unlabeled combined) was studied during day 7. RCEC (TEER 1000 ± 100 Ω/cm²). Data are mean P _app ± SEM (n = 6) expressed as centimeters per second × 10⁻⁷. +, measured in the NaCl buffer; −, measured in the choline chloride buffer, ±, measured in both the NaCl and choline chloride buffers.

*  Significantly higher than apical-to-basolateral permeability.

†  Significantly higher than basolateral-to-apical permeability under Na⁺-free conditions.

‡  Significantly higher than mannitol permeability.

Figure 5.

 

Schematic diagram illustrating the rationale behind net basolateral-to-apical secretion of GSH in rabbit conjunctival epithelial cells. Solid arrows: rates of GSH uptake and efflux at each membrane. Their magnitudes were extrapolated from the slopes of initial linear regions of time-course plots observed for GSH uptake or efflux. All transport rates are expressed in whole numbers in picomoles per hour per square centimeter, in contrast to Figures 2 3 and 4 , for ease of comparison. GSH levels in tear and plasma were taken from the literature. ^{28 36} The cytoplasmic GSH level was calculated from our current studies, assuming cultured RCECs are spherically shaped with radii ranging from 5 to 10 μm with an estimated cell volume of, at most, 9 to 10 pL. The calculations of rates of GSH transport take into account the number of cells per well on day 7 (0.73 ± 0.02 × 10⁶) and surface area (1.13 cm²) and are standardized per hour for uniformity. For example, for uptake at approximately 0.10 mM GSH, the calculations are equal to (29 pmol/10⁶ cells per 10 min) × (60 min/1 h) × (0.73 × 10⁶ cells) × (1/1.13 cm²) and represent 100 pmol/h · cm² for apical side in the figure. Dashed arrow: net directional transport, describing either the net directional flux at each individual membrane (e.g., apical or basolateral) or the overall transepithelial flux, which is the sum of the net directional fluxes at each membrane, minus the paracellular contribution. The paracellular contribution to GSH transport is extrapolated from the apparent permeability coefficient of mannitol reported in Table 1 , assuming that mannitol (182 Da) and GSH (307 Da) have approximately the same paracellular passive diffusive permeability across cultured RCECs and adjusting the concentration-gradient-dependent magnitude of this arrow to the reported tear-to-plasma (0.107:0.015 mM) GSH levels. The net GSH flux is obtained by subtracting 750 minus 70 from 1900 (1900 − 750 − 70 = 1080 pmol/h · cm²). Parallel lines: tight junctions. Question mark: possible unknown carriers.

The authors thank Bin Ouyang for outstanding technical assistance in these studies.