Pancreatic β-cells are susceptible to oxidative stress, which is related closely to the islet dysfunction. In the present study, using the pancreatic cell lines HIT-T15 and RINm5F as known in vitro models of impaired β-cell function as well as primary rat islet β-cells, we observed a relationship between intracellular glutathione levels and oxidative stress-mediated cell dysfunction. Hydrogen peroxide and 4-hydroxy-2-nonenal caused cell death in HIT-T15 and RINm5F cells at lower concentrations compared with non-β-cells, such as HepG2 and NRK-49F cells. The extent of the cytotoxicity caused by the model oxidants was inversely correlated well with intracellular glutathione levels in the cell lines used. Treatment of HIT-T15 and RINm5F cells with l-cysteine or l-cystine significantly augmented the glutathione contents, surpassing the effect of N-acetylcysteine, and abrogated 4-hydroxy-2-nonenal-mediated cytotoxicity almost completely. l-Cysteine increased intracellular glutathione levels in primary β-cells as well. Supplementation of l-cysteine to the RINm5F cell culture inhibited 4-hydroxy-2-nonenal-mediated cytosolic translocation of PDX-1, a key transcription factor for β-cell function. Intrinsic transport activities (Vmax/Km) of the l-cystine/l-glutamate exchanger in HIT-T15 and RINm5F cells were considerably lower than that in NRK-49F cells, although gene expressions of the exchanger were similar in these cells. Results obtained from the present study suggest that the restricted activity of the l-cystine/l-glutamate exchanger controls the levels of intracellular glutathione, thereby making β-cells become susceptible to oxidative stress.
- system xc−
- hydrogen peroxide
- reactive oxygen species
during the progression of Type 2 diabetes, deterioration of impaired insulin secretion by pancreatic β-cell dysfunction results in further progressive hyperglycemia. Such glucose toxicity, associated with irreversible β-cell dysfunction, at least in part involves oxidative stress caused by continuous hyperglycemia (15, 23). Chronic hyperglycemia induces the production of reactive oxygen species (ROS), mainly through the glycation reaction (9, 25), in a great many tissues. ROS induce increases in products of protein oxidation, DNA oxidation, and lipid peroxidation. 4-Hydoroxy-2-alkenals, particularly the most cytotoxic aldehyde 4-hydroxy-2-nonenal (4-HNE), are typical, commonly produced by the lipid peroxidation reaction and induce production of further protein adducts as well as the generation of ROS (6). Diabetic complications have been the focus of attention of the ROS-mediated tissue dysfunction, and evidence indicating pancreatic β-cells as the target of oxidative stress has been mounting. For instance, the levels of 8-hydroxy-2′-deoxyguanosine, a marker of DNA oxidation, and 4-HNE-modified proteins have been shown to increase in the pancreatic β-cells of Goto-Kakizaki rats (10, 16), a model of nonobese Type 2 diabetes, as well as in Type 2 diabetic patients (4). Consequently, oxidative stress originating from hyperglycemia would be one of the causes that eventually provokes impaired islet function at the level of insulin synthesis and secretion.
Glutathione (GSH) is the major low-molecular weight thiol in mammalian cells. The cellular GSH redox buffer is present in cells at millimolar concentrations and forms one major basis of redox homeostasis that maintains an appropriate protein conformation (28). It also plays an important role in the reduction of ROS and electrophiles, including 4-HNE as a co-factor of GSH peroxidases and GSH-S-transferases. The rate-limiting step in de novo synthesis of GSH is a reaction catalized by cytosolic γ-glutamylcysteine synthetase (γ-GCS). On the other hand, l-cysteine availability to the cell is also important factor to maintain GSH levels (36). l-Cysteine is auto-oxidized readily to l-cystine in extracellular circumstances. The concentration of l-cysteine in the plasma (10–20 μM) is 10 times lower than that of l-cystine (100–200 μM) (5). Cells uptake extracellular l-cystine by the l-cystine/l-glutamate exchanger (system xc−), which is a heterodimeric protein complex composed of a light-chain subunit (xCT), which confers substrate specificity and a heavy-chain subunit (4F2hc) common to many amino acid transporters (27). Physiological flux via system xc− involves the entry of l-cystine and exit of l-glutamate. Incorporated l-cystine by system xc− is reduced to form l-cysteine under the intracellular redox state and is utilized as a substrate for GSH synthesis.
It has been reported that the expression levels of antioxidant enzymes, including superoxide dismutase (SOD), catalase, and GSH peroxidase (GPx), are low in certain β-cell-derived cell lines as well as in rat islets (32). Overexpression of the antioxidant enzymes could thus confer a beneficial effect on islets exposed to ROS (2, 19, 20, 22, 33). The role of cellular GSH, however, in the protection of β-cell function is still obscure. The present study using HIT-T15 and RINm5F cells as well as pancreatic β-cells presents for the first time evidence suggesting that restricted activity of the system xc− controls the suppressed levels of intracellular GSH, thereby making the β-cells become susceptible to oxidative stress.
MATERIALS AND METHODS
The HepG2 and NRK-49F cells obtained from DNA Bank, Riken BioResource Center (Tsukuba, Japan) with the support of National Bio-Resources Project of Ministry of Education, Culture, Sports, Science and Technology, Japan (MEXT), were cultivated in DMEM supplemented with 10% fetal calf serum, 20 mM HEPES, and antibiotics. The HIT-T15 and RINm5F cells obtained from American Type Culture Collections (Manassas, VA) were cultivated in RPMI 1640 medium supplemented with 10% fetal calf serum, 20 mM HEPES, and antibiotics in poly-l-lysine-coated dishes (Asahi Techno Grass, Chiba, Japan). These cells were maintained in logarithmic growth phase in a humidified 5% CO2 atmosphere at 37°C.
Pancreases from neonatal Wistar rats were digested with 1,000 units/ml dispase for 24 h at 4°C and filtered over 40-μm pore size mesh nylon screen. Islet β-like cells were enriched by subculturing unattached cells three times. More than 95% of cells obtained by this procedure were positive in terms of insulin staining, and these cells secreted insulin in response to glucose. These insulin-producing cells were cultivated in RPMI 1640 medium supplemented with 10% fetal calf serum, 5 mM nicotinamide, and antibiotics.
Cells were seeded in a 96-well multidish (2.5 × 104 cells/well) and maintained for 16 h. Cells were treated with saline, hydrogen peroxide, or 4-HNE (Oxis International, Foster City, CA) for 24 h. Viable cell number was estimated by the 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)carbonyl]-2H-tetrazolium hydroxide (XTT)-formazan assay (29).
Cells were washed twice with ice-cold phosphate-buffered saline and scraped in 1 ml of 2.5% sulfosalicillic acid containing 0.2% Triton X-100. Total GSH (reduced GSH plus oxidized GSSG) contents in the supernatant were determined by 5,5′-dithiobis(2-nitrobenzoic acid)-GSSG reductase recycling assay (1). Protein amounts in the pellet were determined by BCA protein assay (Pierce, Rockford, IL) with bovine serum albumin as the standard.
Total RNA (2 μg) extracted from cells with RNAiso (Takara, Shiga, Japan) was treated with DNase I and was reverse transcribed for 1 h at 42°C with 200 units of M-MLV (Takara) and 300 pmol oligo(dT)18 primer in a 20-μl reaction volume. The PCR reaction (50 μl) consisted of 2 μl of cDNA, 0.5 μM primers, 200 μM dNTP, and 1.25 units of ExTaq (Takara). Amplification was carried out for one cycle at 94°C for 2 min, then 25–40 cycles at 94°C for 30 s, 60°C for 1 min, and 72°C for 1 min, followed by a 72°C step for 5 min. PCR products were separated by electrophoresis on a 2% agarose gel. The primer sets detecting gene products were as follows: xCT, 5′-CCTGGCATTTGGACGTACAT-3′ and 5′-TCAGAATTGCTGTGAGCTTGCA-3′; 4F2hc, 5′-CTCCCAGGAAGATTTTAAAGACCTTCT-3′ and 5′-TTCATTTTGGTGGCTACAATGTCAG-3′; GAPDH, 5′-TGATGACATCAAGAAGGTGGTGAAG-3′ and 5′-TCCTTGGAGGCCATGTAGGCCAT-3′.
l-Cystine uptake was determined essentially according to the reported procedure (8). Briefly, cells were seeded on a 24-well multidish (2 × 105 cells/well for NRK-49F cells and 4 × 105 cells/well for HIT-T15 and RINm5F cells) 16 h before the experiment. Cells were washed with 1 ml of extracellular fluid (ECF) buffer consisting of 122 mM NaCl, 25 mM NaHCO3, 3 mM KCl, 1.4 mM CaCl2, 1.2 mM MgSO4, 0.4 mM K2HPO4, 10 mM d-glucose, and 10 mM HEPES (pH 7.4) at 37°C. Uptake was started by adding 200 μl of ECF buffer containing various concentrations of cold cystine and 3.7 × 103 Bq l-[14C(U)]-cystine (9.25 × 109 Bq/mmol, PerkinElmer Life and Analytical Sciences, Waltham, MA) at 37°C for 30 min unless otherwise stated. The incubation was terminated by washing cells twice with ice-cold ECF buffer, and cells were solubilized in 750 μl of 1% Triton X-100. An aliquot (25 μl) was taken for BCA protein assay with bovine serum albumin as a standard. Another portion (500 μl) was subjected to a measurement of radioactivity using a liquid scintillation counter (LSC 1000, Aloka, Tokyo, Japan). The Michaelis-Menten constant (Km), maximum uptake rate (Vmax), and nonsaturable uptake rate constant (Pnon) of l-cystine uptake were calculated from the equation below for kinetic studies.
Nuclear extracts were prepared using NE-PER nuclear extraction reagents (Pierce Biothechnology, Rockford, IL) essentially according to the manufacturer's instructions. Immunoblot analysis was carried out using 2 μg of nuclear proteins. The blot was sequentially probed with anti-pancreas duodenum homeobox-1 (PDX-1) antibody (TransGenic, Kumamoto, Japan) and anti-TATA binding protein antibody (Abcam, Cambridge, UK).
Cells grown on a poly-l-lysine coated multi-culture slide (BD Biosciences) were washed twice with ice-cold PBS and fixed with 3% formaldehyde for 15 min at room temperature. After being washed with PBS, cells were blocked with PBS containing 0.1% bovine serum albumin and 0.05% Tween 20 for 1 h at room temperature and then incubated with anti-PDX-1 antibody diluted with the blocking solution (×125) at 4°C for 16 h. Cells were washed and probed with Alexa Fluor 546-conjugated secondary antibody (Invitrogen) for 2 h at room temperature. Nuclei were stained with 4′,6-diamino-2-phenylindole (DAPI). Subcellular localization of the PDX-1 was observed using an inverted fluorescent microscopy (Keyence BZ-8000).
All data are given as the means ± SD. Significance of the difference between the control and treated group(s) was assessed by Welch's t-test or one-way ANOVA followed by Dunnett's test for data from two or multiple groups, respectively. Bonferroni test was performed to compare data from multiple groups with each other. Values of P < 0.05 were taken to be significant.
The cytotoxic effects of oxidative stress on pancreatic islet-derived HIT-T15 and RINm5F cells were compared with those on HepG2 hepatoma and NRK-49F fibroblastic cells (Fig. 1A). Oxidative stresses were generated by the addition of hydrogen peroxide as a direct oxidant and 4-HNE as a pathophysiological stressor. Treatment with hydrogen peroxide proved to be significantly cytotoxic to the β-cells, as demonstrated previously, and also to NRK-49F cells, although to a lesser extent. HepG2 cells were shown to be resistant to hydrogen peroxide-mediated cytotoxicity. IC50 values of hydrogen peroxide for NRK-49F, HIT-T15, and RINm5F cells were calculated to be 578, 121, and 42.2 μM, respectively. 4-HNE was also cytotoxic to HIT-T15 and RINm5F cells but did not produce cytotoxic effects on HepG2 or NRF49F cells at all in the observed range of concentrations. The IC50 values of 4-HNE for HIT-T15 and RINm5F cells were 41.2 and 26.3 μM, respectively.
Intracellular GSH levels in HIT-T15 and RINm5F cells were compared with those in HepG2 and NRK-49F cells (Fig. 1B). The highest GSH levels were observed in HepG2 cells, and relatively lower GSH contents were observed in the β-cells. Compared with HepG2 cells, GSH levels in NRK-49F, HIT-T15, and RINm5F cells were 60, 24, and 10%, respectively. These results, therefore, indicate that the intracellular GSH levels are inversely correlated with susceptibility to oxidative stress.
Inhibition of GSH synthesis by pretreatment with buthionine sulfoximine (BSO), an inhibitor of γ-GCS, rendered NRK-49F and RINm5F cells more susceptible to hydrogen peroxide. The BSO-mediated increase in the susceptibility was also observed in RINm5F cells treated by 4-HNE (Fig. 2).
To supply a GSH precursor, cells were treated with a membrane permeable l-cysteine derivative, N-acetyl-l-cysteine (NAC), and GSH levels were determined (Fig. 3A). Treatment of NRK-49F cells with 10 mM NAC for 12 h did not increase GSH levels, suggesting that l-cysteine is adequately supplied to maintain intracellular GSH levels at least at this time point in these cells. On the other hand, NAC treatment increased GSH levels in HIT-T15 and RINm5F cells by 2.4- and 5.2-fold, respectively, indicating that the intracellular GSH levels in these β-cell lines are restricted by l-cysteine availability. Treatment with 1 mM l-cyteine did not change GSH levels in NRK-49F cells but increased the levels significantly in HIT-T15 and RINm5F cells by 4.5- and 11-fold, respectively (Fig. 3A). Intracellular GSH levels increased in a concentration-dependent manner; as low as 0.05 mM l-cysteine showed significant effects, and 0.5 mM showed similar effect to 1 or 3 mM (Fig. 3B). Treatment of HIT-T15 and RINm5F cells with l-cystine induced an increase in GSH levels as similar extents to l-cysteine (Fig. 3C). Steady-state GSH levels in islet β-cells obtained from neonatal rats was in the range of the β-cell lines (Fig. 3D). Treatment with l-cysteine significantly increased GSH levels in the primary culture of islet β-cells (Fig. 3D) as well. These results strongly suggest that the l-cysteine availability, namely the function of system xc−, is the rate-limiting step in a GSH synthesis of islet β-cells.
Next, experiments were conducted to determine the relative cytotoxicity of the oxidative stressor in the β-cells supplemented with l-cysteine in culture. Cells were pretreated with 1 mM l-cysteine for 12 h and then treated with various concentrations of 4-HNE for 24 h. The l-cysteine supplementation almost completely ameliorated the 4-HNE-induced cytotoxicity in HIT-T15 cells as well as in RINm5F cells (Fig. 3E).
To address a role of GSH in oxidative stress-mediated β-cell dysfunction, intracellular distribution of PDX-1, a major transcription factor for insulin gene, was investigated. 4-HNE showed limited effect on total PDX-1 levels in RINm5F cells up to 8 h (data not shown) but induced a decrease in nuclear PDX-1 levels in a time-dependent manner (Fig. 4A). Cells exposed to 4-HNE showed less nuclear signals stained by PDX-1 antibody compared with the control cells (Fig. 4C). Pretreatment of the cells with l-cysteine overcame the 4-HNE-mediated decrease in nuclear PDX-1 levels (Fig. 4B) and the nuclear PDX-1 signals (Fig. 4C). These results indicate that an increase in the intracellular GSH content in the β-cells reverses susceptibility to oxidative stress.
It is possible that the aberrant function of system xc− is caused by a low level of gene expression. To investigate this possibility, genetic expression of the subunits of system xc−, xCT, and 4F2hc was analyzed by RT-PCR. Under the present experimental conditions, both rat and mouse xCT mRNAs gave 182-bp PCR products and 4F2hc mRNAs gave 141-bp fragments, as shown in Fig. 5, as a typical event. Results from various numbers of PCR cycles (Fig. 5 and data not shown) revealed no significant difference in the genetic expression of these subunits in NRK-49F and RINm5F cells.
The next approach was conducted to investigate the transporter functionality of system xc− in the β-cells by analyzing the transport kinetics of radiolabeled l-cystine. The uptake of [14C]-l-cystine, as revealed by the cell-to-medium ratio, was increased linearly for at least 60 min of incubation in both NRK-49F and RINm5F cells (data not shown). Kinetic analyses of the l-cystine uptake were, therefore, carried out by a 30-min incubation procedure in the latter experiments. In an analysis using Eadie-Scatchard plots, the l-cystine uptake was found to consist of saturable and nonsaturable processes in NRK-49F, RINm5F, and HIT-T15 cells. Results from three independent experiments are summarized in Table 1. Km values tended to be higher and Vmax values significantly lower in RINm5F and HIT-T15 cells than in NRK-49F cells. Intrinsic transport activities (Vmax/Km) of the system xc− in HIT-T15 and RINm5F cells were 1/10 and 1/6.4 of those observed in NRK-49F cells. The values showing nonspecific uptake as revealed by Pnon in the β-cell models were also significantly lower than those in NRK-49F cells.
Oxidative stress induces aberrant macromolecule function and a pathophysiological imbalance that eventually leads to numerous kinds of disease. Although it has well been known that oxidative stress is involved in the development and progression of diabetic complications, recently obtained evidence indicates that progressive deterioration of glucose homeostasis, a fundamental characteristic of Type 2 diabetes, is also associated with exacerbation of oxidative stress (23). Namely, the continuous hyperglycemia often associated with hyperlipidemia causes a decrease in the level of cellular antioxidants and increases in the production of pro-oxidants, including ROS as well as electrophilic substances, resulting in continued β-cell dysfunction and further deterioration. Such glucose toxicity may be relevant to the susceptibility of β-cells to oxidative stress, since both the functional activities and expression levels of antioxidant enzymes, such as SOD, catalase, and GPx, are limited in these cells (17), and forced expression of these enzymes has been shown to overcome the susceptibility (31). On the other hand, regulation of the levels of the major nonprotein antioxidant GSH in β-cells is still obscure. Therefore, the purpose of the present study was to explore the machinery that governs the regulation of GSH levels in β-cells.
HIT-T15 and RINm5F cells, which retain the capacity for insulin secretion, were used as the β-cell model in the present study. It has been reported that secretion and gene expression of insulin are decreased when HIT-T15 cells are cultured for a long period in a high glucose medium, characteristics resembling the β-cell dysfunction observed in Type 2 diabetes (18). On the other hand, HIT-T15 cells are a hamster-derived cell line; therefore, rat-derived RINm5F cells were used to detect gene expression and to observe common responses in β-cell lines. To produce oxidative stress on cells with hydrogen peroxide, the most often employed experimental oxidant, and 4-HNE, an end product of lipid peroxidation that rises in Type 2 diabetic model animals (10, 30) as well as patients (3, 34), were used throughout the study. The β-cell models employed were significantly more sensitive to the oxidant exposure than the other tissue-derived cells. 4-HNE exhibited only marginal effects on the β-cell viability at pathologically workable concentrations; in contrast, hydrogen peroxide displayed a much more striking effect. Hydrogen peroxide has been shown to induce calcium influx via the TRPM2 channel and concomitant cell death in RINm5F cells (12), suggesting that this oxidant-mediated cytotoxicity was provoked by specific cellular signaling rather than by generalized oxidative stress.
Although GSH levels in red blood cells have been reported to decrease in Type 2 diabetes patients (21), information of the levels in the islets are thus far only limitedly available (26, 35). The present study demonstrated that steady-state GSH levels in the β-cell models as well as the primary culture of rat islet β-cells are considerably lower than cells derived from other tissues. These observations are consistent with a previous study (26) in which there were reportedly lower GSH levels in mouse β-cell derived βTC3 cells than in the pancreatic AR42J acinar cells. In addition, steady-state GSH levels inversely coincided with the susceptibility to oxidative stress in the cells used, and the inhibition of GSH synthesis by BSO augmented the sensitivity to the oxidants. Moreover, supplementation of intracellular GSH reversed the oxidant-mediated cytosolic translocation of PDX-1 as well as cytotoxicity in the β-cell models. Similar results showing l-cysteine-mediated protection of RINm5F cells from hydrogen peroxide have been reported (24). Consequently, these observations reveal that low intracellular GSH levels sustained in the β-cells are directly involved in the susceptibility to oxidative stress.
The precise physiological role and molecular basis of the subtle antioxidant system in islets has not been adequately elucidated. The present study demonstrated how the total GSH levels are regulated in the β-cells. The de novo GSH synthesis is controlled by the rate-limiting enzyme γ-GCS as well as intracellular l-cysteine levels resulting from system xc− function. Intermediate expression levels in the catalytic subunit of γ-GCS, compared with other metabolic tissues, in rat islets have been reported (35). Treatment of the β-cell models as well as islet β-cells from neonatal rats with l-cysteine or l-cystine at concentrations much lower than NAC forcefully increased intracellular GSH levels, although no effect was observed in NRK-49F cells. l-Cystine, the added concentration of which (0.05–0.1 mM) was several times lower than that in the cell culture medium (0.21 mM as l-cystine), may increase intracellular l-cysteine in parallel with extracellular circumstances by system xc− and thereby augment the intracellular GSH levels. These results suggest that the transporter system xc− is a critical factor to maintain GSH levels, and the enzymes required for its synthesis adequately function in the β-cells. l-Cysteine added exogenously effectively increased GSH levels as well and, therefore, could be substituted for l-cystine. The genetic expressions of xCT and 4F2hc in RINm5F cells were similar to those in NRK-49F cells; however, the intrinsic transport activities as expressed by Vmax/Km of system xc− in the β-cell models were significantly lower than those in fibroblastic cells. Consequently, it is suggested that the limited intrinsic transport activity of system xc− plays a role in the restricted GSH levels and, thereby, makes β-cells become susceptible to oxidative stress.
Growing evidences suggest the efficacy of antioxidants in the protection and/or preservation of the β-cell function from glucose toxicity. Forced expression of catalase and GPx or in combination with SOD has been shown to provide protection of insulin-secreting cultured cells (20, 31, 33) as well as islets (2, 22, 35) from oxidative insults. It has been reported (14) that ingestion of NAC in combination with vitamin C and E in db/db mice, a well known obese model for Type 2 diabetes, reversed mass loss in islets and the nuclear levels of PDX-1. Similar preservative effects of antioxidants on β-cell function have been observed in other models of Type 2 diabetes, such as Zucker diabetic fatty rats (30) and Goto-Kakizaki rats (11). Therefore, it is expected that deterioration of β-cell function during continued hyperglycemia could be protected by antioxidant therapy, which would be likely to be particularly effective in the early phase of Type 2 diabetic patients. On the other hand, the targets of the anti-“glucose toxicity” therapy have been focused on anti-oxidant enzymes such as catalase, SOD, and GSH peroxidase, based on only a small number of research papers that specifically demonstrated reduced activity or content of these enzymes (17, 32). The present study suggests system xc− to be another valuable target of antioxidant therapy in diabetes and possibly acute pancreatitis, which is particularly attractive because simple supplementation of l-cysteine would be as effective as gene therapies as well as the combination of several antioxidants.
It has been reported (13) recently that treatment of islet cells with l-cysteine at concentrations higher than those used in the present study suppresses insulin secretion. The authors concluded that the hydrogen sulfide produced from l-cysteine in the cells inhibits the insulin secretory process, although they did not observe the GSH level. We have also observed lower concentrations of l-cysteine, which sufficiently increase GSH levels, induced a significant decrease in insulin secretion in HIT-T15 and RINm5F cells (unpublished results). In conjunction with the present results, it might be suggested that treatment of islets with l-cysteine may result in an increased intracellular GSH in addition to the production of hydrogen sulfide, which would cause a decrease in glucose-dependent insulin secretion in a short period of exposure. This idea is consistent with the hypothesis proposed by Fridlyand and Philipson (7), who suggested adaptive inhibition of ROS production leads to a decrease in insulin secretion. Therefore, an increase in the antioxidant in β-cells may not be the sole therapy required for the glucose toxicity that occurs in Type 2 diabetes. In any event, the results from the present study do suggest that an increase in GSH protects against a reduction of the islet cell mass that results from long-term exposure to the oxidative stress.
This work was supported by a Grant-in-Aid for Scientific Research C (17590108) as well as by “High-Tech Research Center” Project for Private Universities: matching fund subsidy from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (2004–2007).
Pacific Edit reviewed the manuscript before submission.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society