Benecial Effects of Antioxidants in Diabetes Possible Protection of Pancreatic -Cells Against Glucose Toxicity

Hideaki Kaneto, Yoshitaka Kajimoto, Jun-ichiro Miyagawa, Taka-aki Matsuoka, Yoshio Fujitani, Yutaka Umayahara, Toshiaki Hanafusa, Yuji Matsuzawa, Yoshimitsu Yamasaki, and Masatsugu Hori

Oxidative stress is produced under diabetic conditions and possibly causes various forms of tissue damage in patients with diabetes. The aim of this study was to examine the involvement of oxidative stress in the progression of pancreatic -cell dysfunction in type 2 diabetes and to evaluate the potential usefulness of antioxidants in the treatment of type 2 diabetes. We used diabetic C57BL/KsJ-db/db mice, in whom antioxidant treatment (N-acetyl-L-cysteine [NAC], vitamins C plus E, or both) was started at 6 weeks of age; its effects were evaluated at 10 and 16 weeks of age. According to an intraperitoneal glucose tolerance test, the treatment with NAC retained glucose-stimulated insulin secretion and moderately decreased blood glucose levels. Vitamins C and E were not effective when used alone but slightly effective when used in combination with NAC. No effect on insulin secretion was observed when the same set of antioxidants was given to nondiabetic control mice. Histologic analyses of the pancreases revealed that the -cell mass was significantly larger in the diabetic mice treated with the antioxidants than in the untreated mice. As a possible cause, the antioxidant treatment suppressed apoptosis in -cells without changing the rate of -cell proliferation, supporting the hypothesis that in chronic hyperglycemia, apoptosis induced by oxidative stress causes reduction of -cell mass. The antioxidant treatment also preserved the amounts of insulin content and insulin mRNA, making the extent of insulin degranulation less evident. Furthermore, expression of pancreatic and duodenal homeobox factor-1 (PDX-1), a -cell specific transcription factor, was more clearly visible in the nuclei of islet cells after the antioxidant treatment. In conclusion, our observations indicate that antioxidant treatment can exert beneficial effects in diabetes, with preservation of in vivo -cell function. This finding suggests a potential usefulness of antioxidants
From the Department of Internal Medicine and Therapeutics (H.K., Y.K., T.M., Y.F., Y.U., Y.Y., M.H.) and the Department of Internal Medicine and Molecular Science (J.M., T.H., Y.M.), Osaka University Graduate School of Medicine, Suita, Japan. Address correspondence and reprint requests to Dr. Y. Kajimoto, Department of Internal Medicine and Therapeutics, A8, Osaka University Graduate School of Medicine, 2-2 Yamadaoka, Suita City, Osaka 565-0871, Japan. E-mail: kajimoto@medone.med.osaka-u.ac.jp. Received for publication 12 May 1999 and accepted in revised form 16 August 1999. ABC, avidin-biotin complex; AEC, 3-amino-9-ethylcarbozol; BrdU, 5-bromo2-deoxyuridine; BSA, bovine serum albumin; DAB, 3,3 -diaminobenzidine tetrahydrochloride; NAC, N-acetyl- L-cysteine; PDX-1, pancreatic and duodenal homeobox gene-1; ROS, reactive oxygen species; TUNEL, TdTmediated dUTP-biotin nick end labeling. 2398

for treating diabetes and provides further support for the implication of oxidative stress in -cell dysfunction in diabetes. Diabetes 48:23982406, 1999

n general, the development of type 2 diabetes is associated with pancreatic -cell dysfunction occurring together with insulin resistance. Normal -cells can compensate for insulin resistance by increasing insulin secretion, but insufficient compensation leads to the onset of glucose intolerance. Once hyperglycemia becomes apparent, -cell function progressively deteriorates: glucose-induced insulin secretion becomes further impaired and degranulation of -cells becomes evident, often accompanied by a decrease in the number of -cells (14). The signicance of hyperglycemia as a direct cause of these phenomena, i.e., -cell glucose toxicity, has been demonstrated by various studies in vivo (5) and in vitro (610). Chronic hyperglycemia may impair -cell function at the level of insulin synthesis as well as insulin secretion; when -cell derived cell lines are exposed to a high glucose concentration for a long period of time, insulin gene transcription and insulin content are dramatically reduced (610). These changes are often accompanied by a decrease in expression of pancreatic and duodenal homeobox factor-1 (PDX-1) (8,10). PDX-1 (also known as islet duodenum homeobox-1 [IDX-1], somatostatin transactivating factor-1 [STF-1], and insulin promoter factor-1 [IPF1]) (1115), is a -cellspecic transcription factor that plays a major role in maintaining normal -cell function, probably by regulating multiple genes expressed in -cells (14,1618). Under diabetic conditions, reactive oxygen species (ROS) are produced mainly through the glycation reaction (19,20), which occurs in various tissues (21) and may play a role in the development of complications in diabetes (22). Although the induction of the glycation reaction in diabetes was originally found in neural cells and the lens crystalline, which are also known targets of diabetic complications, another target was recently shown to be the -cell (2325). Indeed, advanced glycosylation end products (AGEs) were shown to be detectable in -cells kept under high glucose concentrations (24), and the level of 8-hydroxy-2-deoxyguanosine (8-OHdG), a marker for oxidative stress, is increased in -cells of diabetic Goto-Kakizaki (GK) rats (25). Also, the expression of antioxidant enzymes, such as superoxide dismutase, catalase, and glutathione peroxidase, is known to be very low in islet cells
DIABETES, VOL. 48, DECEMBER 1999

H. KANETO AND ASSOCIATES

compared with other tissues and cells (26). Therefore, once -cells face oxidative stress, they may be rather sensitive to it, suggesting that glycation and subsequent oxidative stress may in part mediate the toxic effect of hyperglycemia. As direct support for this, we recently showed that glycationmediated ROS production reduces insulin gene transcription (27) and also causes apoptosis of -cells (23). Although animals have their own antioxidant defense systems, the defense can be externally strengthened. This might be especially true for the pancreas, since it has a relatively weak intrinsic defense system against oxidative stress (26). Antioxidants include N-acetyl-L-cysteine (NAC), which scavenges hydrogen peroxide, and vitamins C and E, which are well-known dietary antioxidants. Vitamin E is lipophilic and inhibits lipid peroxidation, scavenging lipid peroxyl radicals to yield lipid hydroperoxides and the tocopheroxyl radical (28). Vitamin C, a water-soluble vitamin, functions cooperatively with vitamin E by regenerating tocopherol from the tocopheroxyl radical. To investigate the implication of oxidative stress in -cell glucose toxicity in vivo and also to nd a potential tool for protecting the -cell function from glucose toxicity, we examined the possible effects of antioxidant treatment (NAC, vitamins C plus E, or both) on the preservation of -cell function in diabetic C57BL/KsJ-db/db mice. In this mouse, a well-known obese model for type 2 diabetes, hyperglycemia is induced because of increasing insulin resistance and the subsequent insufficiency of the -cell compensation (29,30). Our results show that the antioxidant treatment is benecial for treating diabetes and can provide protection to -cells against glucose toxicity in C57BL/KsJ-db/db mice.
RESEARCH DESIGN AND METHODS
Experimental protocol. We used C57BL/KsJ-db/db and nondiabetic C57BL/KsJmisty/misty female mice purchased from Clea Japan (Tokyo). The mice were allowed free access to food and water in a specic pathogenfree environment. At 6 weeks of age, C57BL/KsJ-db/db mice were divided into four groups. Mice in the control group (group 1, n = 40) were kept on regular diet without antioxidants, and mice in the other three groups were given high-antioxidant diets: group 2 (n = 20), 2.5% NAC; group 3 (n = 20), 0.5% vitamin C plus 0.5% vitamin E; group 4 (n = 40), both 2.5% NAC and 0.5% vitamin C plus 0.5% vitamin E. At 10 and 16 weeks of age, certain numbers of mice in each group (see gure legends) were used for physiologic or histologic analyses. The animal studies were conducted in accordance with Principles of Laboratory Animal Care (National Institutes of Health publication no. 8523). Glucose tolerance test. After an overnight fast, mice were injected intraperitoneally with glucose (1.0 g/kg body weight). Blood samples were taken at various time points (0120 min). Blood glucose concentrations were measured by the glucose oxidase method using a glucose analyzer (MS-GR101; Terumo, Tokyo), and serum insulin concentrations were determined using a Lebis radioimmunoassay kit (Shibayagi, Gunma, Japan) with mouse insulin as the standard. Insulin tolerance test. After an overnight fast, mice were injected intraperitoneally with 2 U/kg human regular insulin. Blood samples were taken at various time points (090 min), and blood glucose concentrations were measured as described above. Preparation of pancreas sections and immunohistochemical analyses. The mice were anesthetized using pentobarbital sodium. A midline abdominal incision was made, and pancreases were removed from the mice and xed overnight in a solution of 4% paraformaldehyde. Fixed tissues were processed routinely for paraffin embedding, and ~5-m sections were prepared and mounted on slides. Before each incubation with antibodies, the mounted sections were rinsed three times with phosphate-buffered saline (PBS). For detection of insulin, the avidin-biotin complex (ABC) method was performed using Vectastain ABC Kit (Vector Laboratories, Burlingame, CA). The mounted sections were incubated for 30 min with guinea pig polyclonal anti-insulin antibody (Dako, Glostrup, Denmark) diluted 1:3,000 in PBS containing 1% bovine serum albumin (BSA). They were then incubated for 30 min with biotinylated antiguinea pig IgG (Vector Laboratories), diluted 1:200, as the secondary antibody. The DIABETES, VOL. 48, DECEMBER 1999

FIG. 1. Body weights (A) and nonfasting plasma glucose levels (B) in C57BL/KsJ-db/ db mice (10 and 16 weeks of age) maintained on a diet with and without antioxidants. , group 1; , group 2; , group 3; , group 4. Data are means SE (n = 10). *P < 0.05 vs. group 1. sections were then incubated with ABC reagent for 30 min, and positive reactions were visualized by incubation with the peroxidase substrate solution containing 3,3-diaminobenzidine tetrahydrochloride (DAB) (Zymed Laboratories, San Francisco, CA). Nuclei were counterstained with methyl green. PDX-1 staining followed a similar procedure, except the mounted sections were incubated with Target Retrieval Solution (Dako) at 90C for 5 min before ABC was performed and antiPDX-1 antiserum (16) diluted 1:1,000 in PBS containing 1% BSA was used. Northern blot analyses. Total RNA was isolated from mouse pancreases, and Northern blot analyses were performed. Ten micrograms of total RNA was sizefractionated and transferred onto a Hybond-N+ membrane. After hybridization with a [32P]-labeled mouse insulin cDNA probe at 42C in the presence of 50% formamide, the membrane was washed twice with 1 SSPE (180 mmol/l NaCl, 10 mmol/l sodium phosphate, 1 mmol/l EDTA, pH 7.4) and 0.1% SDS at 50C for 15 min. Kodak XAR lm was exposed with an intensifying screen at 80C. Measurement of insulin content. Four mice in each group (10 and 16 weeks of age) were used to measure the insulin content. After a 6-h fast, the whole pancreas was excised, and insulin content in the pancreatic tissue was determined using a Lebis radioimmunoassay kit (Shibayagi) with mouse insulin as the standard. The 2399

ANTIOXIDANT TREATMENT PROTECTS -CELLS

data were normalized with respect to protein concentration in the extract, which was measured using a protein assay (BioRad Laboratories, Richmond, CA). Detection of apoptosis. Apoptotic cells were detected by the TdT-mediated dUTP-biotin nick end labeling (TUNEL) method using an in situ apoptosis detection kit (Takara Biochemicals, Kyoto, Japan). Sections were treated with proteinase K (20 g/ml) for 15 min at room temperature and incubated with TdT enzyme for 60 min at 37C. After washing in PBS, the sections were further incubated with antiuorescein isothiocyanate HRP conjugate for 30 min at 37C and visualized with DAB. Using the same sections, insulin staining was also performed as described above except 3-amino-9-ethylcarbozol (AEC) (Dako) was used for visualization and nuclei were counterstained with hematoxylin. Evaluation of cell proliferation. 5-Bromo-2-deoxyuridine (BrdU) (Sigma, St. Louis, MO), freshly dissolved in saline, was injected intraperitoneally into C57BL/KsJ-db/db mice at a dose of 100 mg/kg 6 h before excision of the pancreas. Sections were incubated for 60 min at 37C with a mouse monoclonal anti-BrdU antibody diluted 1:10. After washing in PBS, sections were incubated for 30 min at 37C with sheep anti-mouse IgG antibody (diluted 1:10) conjugated with alkaline phosphate (Cell Proliferation Kit; Amersham Japan, Tokyo) followed by visualization with DAB. To detect insulin-producing cells, the same sections were stained for insulin as described above using AEC for visualization, and nuclei were counterstained with hematoxylin. Morphometry. The total area of the pancreas was measured using a television monitor connected to a light microscope and microvideoscope system with a tablet measure unit (Krypton-40; Flovel, Tokyo). -Cell numbers were calculated by counting the nuclei surrounded by cytoplasm positive for insulin staining. The measurement and calculation were done with a total of 12 sections of the pancreas from 4 mice (3 sections per mouse) in each group. The total number of islets included in the 12 sections from each group was approximately 500700. Statistical analyses. Data are means SE. Statistical comparisons of means among individual groups used analysis of variance (ANOVA) followed by post-hoc testing with Fishers least signicant difference test.

RESULTS

Glucose tolerance. No differences in food intake or body weight were observed among the four groups (group 1, untreated; group 2, NAC alone; group 3, vitamins C plus E;

group 4, NAC and vitamins C plus E) at 10 or 16 weeks of age (Fig. 1A). As shown in Fig. 1B, the nonfasting blood glucose levels in mice in groups 2 and 4 were lower than those in mice in group 1 at both 10 and 16 weeks of age. No significant difference was observed between groups 1 and 3 (Fig. 1B). Intraperitoneal glucose tolerance tests performed at 10 and 16 weeks of age revealed that glucose tolerance in mice in groups 2 and 4 was ameliorated (Fig. 2A and B). Also, insulin secretion during the glucose tolerance test was improved in mice in groups 2 and 4 compared with those in group 1 (Fig. 2C and D), indicating that NAC is benecial for preservation of glucose tolerance. In contrast, the effect of antioxidative vitamins is not evident when used alone: there was no difference in blood glucose or insulin levels between groups 1 and 3 at either 10 or 16 weeks of age (Fig. 2AD). When used in combination with NAC, however, the antioxidative vitamins exerted some beneficial effect: there were signicant differences in blood glucose levels (at 30 and 60 min after the glucose injection) and insulin levels (at 30 and 120 min) between groups 2 and 4 at 16 weeks of age (Fig. 2B and D). In contrast to the positive effects of the antioxidants in diabetic C57BL/KsJ- db/db mice, the same set of antioxidants (NAC and vitamins C plus E) did not alter the glucose tolerance in nondiabetic C57BL/KsJ-misty/misty mice (data not shown). To investigate the possible effects of antioxidant administration on insulin sensitivity/resistance, an intraperitoneal insulin tolerance test was performed. As shown in Fig. 3, reduction of blood glucose levels in response to the injected insulin (2 U/kg) was similar in the four groups at both 10 and 16 weeks of age.

FIG. 2. Glucose tolerance in C57BL/KsJ-db /db mice treated with antioxidants. Intraperitoneal glucose tolerance tests were performed in C57BL/KsJ-db / db mice in each group at 10 (A and C) and 16 (B and D) weeks of age. After an overnight fast, glucose was injected intraperitoneally at a dose of 1 g/kg, and blood glucose ( A and B ) and insulin ( C and D ) levels were measured. , Group 1 (untreated); , group 2 (diet with NAC supplementation); , group 3 (diet with vitamin C plus E supplementation); , group 4 (diet with NAC and vitamin C plus E supplementation). Data are means SE (A , n = 10; B, n = 12; C and D, n = 6 each). *P < 0.05 vs. group 1; **P < 0.01 vs. group 1; # P < 0.05 vs. group 2. 2400 DIABETES, VOL. 48, DECEMBER 1999

H. KANETO AND ASSOCIATES

FIG. 3. Insulin resistance in C57BL/KsJ-db /db mice treated with antioxidants. Intraperitoneal insulin tolerance tests were performed in C57BL/KsJ-db/db mice (10 and 16 weeks of age) given diet with and without antioxidants. After an overnight fast, insulin was injected at a dose of 2 U/kg. , Group 1; , group 2; , group 3; , group 4. Data are means SE (n = 8).

Islet morphology. To elucidate how glucose tolerance is preserved in antioxidant-treated mice, insulin immunostaining was performed with the pancreases from C57BL/KsJ-db/db mice in each group. The results revealed that islets of the 16-week-old mice in group 4 (Fig. 4B) were larger and more numerous than those in group 1 mice (Fig. 4A and B). Mice

in group 2 showed immunostaining results similar to group 4, while mice in group 3 were similar to group 1 (data not shown), suggesting that the antioxidative vitamins did not exert signicant effects on islet morphology. To further elucidate the potency of antioxidants in preserving -cell function in diabetes, we tried to examine the

FIG. 4. Effects of antioxidants on -cell mass. A representative result is shown for the insulin immunostaining performed with pancreatic tissue sections derived from C57BL/KsJ-db /db mice (16 weeks of age) in group 1 (A ) and group 4 (B ). In group 1 mice, islets were irregular in shape and islet cells showed marked insulin degranulation. In contrast, islets in group 4 mice were plump and revealed intense insulin immunostaining. Bar, 200 m. C: -Cell mass was quantitatively evaluated by counting -cell numbers per square millimeter of the pancreatic section. The counting was performed in the mice (10 and 16 weeks of age) in group 1 ( ) and group 4 (). Data are means SE ( n = 4). *P < 0.05. DIABETES, VOL. 48, DECEMBER 1999 2401

ANTIOXIDANT TREATMENT PROTECTS -CELLS

FIG. 5. Insulin content in C57BL/KsJ-db / db mice treated with antioxidants. Insulin content in pancreases was measured at 10 and 16 weeks of age in C57BL/KsJ-db/ db mice in group 1 ( ) and group 4 ( ). Data are means SE (n = 4). * P < 0.05; **P < 0.01.

mechanism underlying the preservation of insulin secretion and islet morphology. For this purpose, we focused on the mice in group 4, in which we assumed the maximal potency of antioxidants might be revealed. -Cell mass. Using pancreas sections from mice in groups 1 and 4, we quantitatively evaluated the -cell mass. For each group, sections containing ~500700 islets were prepared, and the -cell number per square millimeter was calculated. As shown in Fig. 4C, the number of -cells in the 10-week-old mice in group 4 was higher than that of the mice in group 1, and this difference became more evident at 16 weeks of age. Insulin synthesis. The results of insulin staining also revealed that insulin degranulation was less evident in the islets of mice in group 4 (Fig. 4A and B). Data for the insulin content (Fig. 5), which showed a signicant difference between the mice in groups 1 and 4, supported this observation. Although the decrease in insulin content in the whole pancreas was in part due to the decrease in -cell mass (Fig. 4C), the insulin content per cell, estimated by normalizing the data (Fig. 5) with respect to -cell number (Fig. 4C), was also higher in the mice in group 4 (data not shown). Consistent with this observation, the results of Northern blot analyses revealed that the insulin mRNA level in the mice in group 4 is signicantly higher than that in the mice in group 1 (Fig. 6). Again, the preservation in the insulin mRNA in part depends on the preservation in -cell mass (Fig. 4C), but the insulin mRNA per cell, normalized with respect to -cell number, was also higher in the mice in group 4 (data not shown). Immunostaining for PDX-1. To understand the background for the preservation in insulin biosynthesis, we examined the possible effects of the antioxidants on the expression of a transcription factor, PDX-1 (1118). Immunostaining for PDX-1 was performed in the pancreases of C57BL/KsJ-db/db mice in groups 1 and 4. Consistent with a previous observation obtained with -cells after partial pancreatectomy (5), we found that PDX-1 expression was reduced in the mice in group 1. In group 4 mice, however,
2402

FIG. 6. Insulin mRNA levels in C57BL/KsJ-db /db mice treated with antioxidants. A representative result is shown for Northern blot analyses performed using mouse insulin cDNA as the probe. Total RNA was isolated from pancreases of C57BL/KsJ-db/db mice in group 1 and group 4 at 10 (A) and 16 (B ) weeks of age. C: Relative insulin mRNA levels were evaluated by densitometry, with those of 10-weekold mice in group 1 arbitrarily set at 1. Data are means SE (n = 4). *P < 0.05; **P < 0.01.

PDX-1 expression was markedly retained, with clear immunostaining in the nuclei (Fig. 7). -Cell apoptosis. To examine the background for the preservation of -cell mass in the antioxidant-treated mice, -cell apoptosis was evaluated. Using double-immunostaining for insulin and apoptotic cells, ~500700 islets were examined for each group of mice. In the islets of mice in group 1, 4.8 TUNEL-positive -cells per 100 islets were recognized. On the other hand, only a few TUNEL-positive -cells (0.6 per 100 islets) could be identied in the islets of the mice in group 4 (Fig. 8). -Cell proliferation rate. To further understand the background for the preservation of -cell mass in antioxidanttreated mice, we examined the possible effects of the antioxDIABETES, VOL. 48, DECEMBER 1999

H. KANETO AND ASSOCIATES

FIG. 7. Immunostaining for PDX-1 in C57BL/KsJ-db/ db mice treated with antioxidants. Pancreases were immunostained for PDX-1 in mice (16 weeks of age) in group 1 (A) and group 4 (B ). In group 1 mice, weak immunostaining for PDX-1 was observed in islet cells. This contrasts with the clear immunostaining in nuclei of the islet cells in the mice in group 4. Bar, 50 m. Similar results were obtained with 12 pairs of sections from 4 mice each of groups 1 and 4.

idants on the cell proliferation rate of the -cells. For this purpose, we measured the percentage of S-phase cells by monitoring BrdU incorporation (31). In mice from groups 1 and 4, ~500700 islets were examined, and the ratio of the BrdU-positive -cells was calculated. As shown in Fig. 9A and B, some BrdU-positive cells were detectable in the islets of both group 1 and group 4. Double-immunostaining for BrdU and insulin revealed that most of the BrdU-positive cells in the islets were insulin-containing -cells. There was no difference in the ratio of BrdU-positive -cells between groups 1 and 4 at either 10 or 16 weeks of age (Fig. 9C).
DISCUSSION

In this study, we have shown that antioxidant treatment using NAC can improve glycemic control with preservation

of pancreatic -cell function in diabetic C57BL/KsJ-db/db mice. Since the same set of antioxidants did not alter the glucose tolerance in nondiabetic C57BL/KsJ-misty/misty mice, the antioxidant treatment probably exerts its effect in association with the presence of hyperglycemia; i.e., by protecting -cells from the toxic effects of ROS produced under hyperglycemic conditions. Indeed, the antioxidant treatment increased the -cell mass (Fig. 4) and preserved insulin content (Fig. 5) and mRNA amount (Fig. 6). Since decreases in -cell mass and insulin biosynthesis are considered to be associated with the development of diabetes in various animal models with type 2 diabetes (32,33), it is likely that those effects of the antioxidants contributed to partial preservation of the glucose tolerance in the antioxidant-treated mice (Figs. 1B and 2).

FIG. 8. Effects of antioxidants on -cell apoptosis. A representative result for double-immunostaining for insulin (red) and apoptotic cells (brown) (visualized by TUNEL method) performed on pancreases of C57BL/KsJ-db /db mice (16 weeks of age) in group 1 (A) and group 4 ( B). The arrow indicates the -cell identified as being apoptotic. Bar, 50 m. In each group, 12 pairs of sections from 4 mice each of groups 1 and 4 were examined, and similar results were obtained. DIABETES, VOL. 48, DECEMBER 1999 2403

ANTIOXIDANT TREATMENT PROTECTS -CELLS

FIG. 9. Effects of antioxidants on -cell proliferation. A representative result for double-immunostaining for insulin (red) and BrdU (dark brown) performed in pancreases of C57BL/KsJ- db/db mice (16 weeks of age) in group 1 (A) and group 4 (B). Several BrdU-positive -cells per islet (arrows) were recognized in both groups. Bar, 50 m. In each group, 12 pairs of sections from 4 mice each of groups 1 and 4 were examined, and similar results were obtained. C : Percentage of BrdU-positive cells in -cells calculated in mice (10 and 16 weeks of age) in group 1 ( ) or group 4 (). Data are means SE (n = 4).

There was no difference in insulin sensitivity between groups 1, 2, 3, and 4 in the insulin tolerance test (Fig. 3), suggesting that improvement of glycemic control was mainly initiated by a benecial effect of antioxidant on -cells. However, we cannot totally deny the possibility that the antioxidant treatment could have exerted an inuence on target tissues other than the -cells such as muscle and fat; it was recently reported that oxidative stress impaired insulininduced GLUT4 translocation in 3T3-L1 adipocytes (34). Thus, it would be safe to conclude that antioxidant treatment has benecial effects on preservation of -cell function in diabetes, although the effects may not be exerted totally through its direct action on -cells. Also, regardless of the inuence on insulin sensitivity, because the antioxidant treatment indeed reduced blood glucose levels, it must have
2404

reduced glucose toxicity and contributed in part to the prevention of a decrease of -cell mass and insulin content. Once the vicious circle of the glucose toxicity is prevented by somehow achieving good glycemic control, the circle then starts moving in the opposite direction: the toxic effects of high glucose will be reduced, resulting in further improvement of glycemic control. Nonetheless, we assume that the antioxidants used in this study probably kept providing protection of -cells from the toxic effects of glucose. While the average nonfasting glucose levels in mice treated with antioxidants (NAC and vitamins C plus E) were >20 mmol/l at 10 and 16 weeks of age (Fig. 1B), Robertson and colleagues (610) have shown that chronic exposure of the -cellderived HITT15 or TC6 cells to a glucose level of 11.1 mmol/l caused dramatic reductions in insulin gene transcription and insulin
DIABETES, VOL. 48, DECEMBER 1999

H. KANETO AND ASSOCIATES

content. Also, according to a result of uorescence-activated cell sorting (FACS) analysis, there was a signicant increase in ROS amount in the HIT-T15 cells kept under 20 mmol/l glucose compared with cells kept under 5 mmol/l glucose (H.K., Y.K., unpublished observations). Thus, we assume that even after the moderate decrease of blood glucose levels was achieved in the antioxidant-treated mice, the blood glucose levels in those mice remained high enough to cause glucose toxicity, and that the antioxidants kept protecting the -cells against it by neutralizing the toxic effects of oxidative stress provoked due to hyperglycemia. As possible mechanisms underlying the preservation of the -cell mass, we found that apoptosis was suppressed in the mice in group 4 (Fig. 8). -Cell apoptosis has been suggested as being involved in physiologic and pathologic decrease of -cell mass (33,35,36). According to a report by Donath et al. (35), hyperglycemia induces -cell apoptosis in Psammomys obesus islets. Also, Pick et al. (33) showed that apoptosis plays a role in the failure of -cell mass compensation in Zucker diabetic fatty rats, providing support for the possible involvement of apoptosis in -cell glucose toxicity. Although not very many TUNEL-positive cells were identied even in the untreated mice (group 1), it should be noted that the frequency of apoptosis tends to be underestimated, since apoptotic cells are usually eliminated rather quickly. While glycation-mediated ROS production induces apoptosis in -cells in vitro (17), our present results support the hypothesis that ROS, provoked by hyperglycemia in vivo, play a signicant role in decreasing the -cell mass in type 2 diabetes through induction of apoptosis. Failure to detect a difference in rate of -cell proliferation between the groups with and without antioxidant treatment also supports this possibility (Fig. 9). The homeodomain containing transcription factor PDX-1 is involved in pancreas development and, possibly, also in regeneration after birth (18,3740). It is expressed in mature -cells and functions as an important transcription factor common to multiple genes essential for the -cell function (1418). When the -cellderived cell line HIT-T15 was kept under a high glucose concentration for a long period of time, a reduction of PDX-1 activity was identied in association with a reduction of insulin gene transcription (8,10). Also, Zangen et al. (5) have shown that the expression level of PDX-1 is reduced in islets from partially pancreatectomized rats, which were exposed to chronic hyperglycemia in vivo. Although the limitations of mouse islet cell preparation prevented reliable quantication concerning PDX-1 expression levels and DNA-binding activity, we observed clear immunostaining in nuclei of the islet cells in group 4, which contrasts with weak immunostaining for PDX-1 in islet cells in group 1 (Fig. 7). These results provide further support for the vulnerability of PDX-1 to chronic high glucose conditions and also suggest the involvement of oxidative stress in hyperglycemiadependent reduction of PDX-1 activity, agreeing with the observation that PDX-1 activity in HIT cells was reduced by induction of oxidative stress in vitro (27). Clinically, it was shown that heterozygous mutations of PDX-1 cause diabetes, typically maturity-onset diabetes of the young type 4 (MODY 4) (41,42). Also, a phenotype of the heterozygous PDX-1 knockout mouse, which displays a decreased -/-cell ratio and subsequent impaired glucose tolerance, indicates that the PDX-1 gene has a dose-dependent effect (43). Whereas
DIABETES, VOL. 48, DECEMBER 1999

these observations suggest that the amount of PDX-1 expression is important for maintaining normal -cell function, it is possible that the suppression of PDX-1 function by ROS in diabetes is implicated in further deterioration of -cell function and thus mediates glucose toxicity. Although both NAC and vitamins C and E have antioxidative activity, only NAC ameliorates glucose tolerance when used alone (Figs. 1B and 2). The main function of vitamins C and E is to suppress lipid peroxidation, which occurs in the plasma membrane and damages membrane structure and permeability. Therefore, the antioxidative vitamins C and E may not have been able to exert an effect on intracellular events, such as those involved in apoptosis and gene transcription. In other words, inhibition of lipid peroxidation by antioxidative vitamins (C and E) was not enough to improve -cell function. On the other hand, NAC scavenges hydrogen peroxide, which is produced in the cytoplasm and probably the nuclei of cells as a direct consequence of the glycation reaction (19,20). Once produced, hydrogen peroxide has a relatively long lifespan and can transfer anywhere by penetrating nuclear and plasma membranes. Indeed, hydrogen peroxide is known to mediate the glycation-dependent degradation of several proteins and is widely involved in the damage of various tissues in diabetes (20,44). Therefore, it seems reasonable that NAC, which is known to decrease the intracellular hydrogen peroxide level in -cells (23), was effective for preventing -cell damage. Although the antioxidative vitamins did not reveal signicant effects when used alone, they exerted some benecial effects when used in combination with NAC (Fig. 2B and D). While many of the molecules constituting the machinery for the glucose-responsive insulin secretion are membrane proteins, it seems possible that vitamins C and E improved the outcome of NAC treatment through their effects on such membrane proteins. In conclusion, our present results show for the rst time that antioxidants can exert benecial effects on pancreatic -cell function in diabetes. Thus a sufficient supply of antioxidants may prevent or delay -cell dysfunction in diabetes by providing protection against glucose toxicity.
ACKNOWLEDGMENTS

This work was supported in part by grants from Suzuken Memorial Foundation (to Y.K.) and a grant-in-aid for Scientic Research from the Ministry of Education of Japan (to Y.K. and Y.Y.). We thank Noriko Fujita and Katsumi Yamamori for the excellent technical assistance. We also thank Dr. M. Alan Permutt of the University of Washington School of Medicine and Dr. John M. Chirgwin of the University of Texas Health Science Center for kindly providing the mouse insulin II cDNA plasmid.
REFERENCES
1. Porte D Jr: Banting lecture 1990: -cells in type II diabetes mellitus. Diabetes 40:166180, 1991 2. DeFronzo RA, Bonadonna RC, Ferrannini E: Pathogenesis of NIDDM: a balanced overview. Diabetes Care 15:318368, 1992 3. Yki-Jarvinen H: Glucose toxicity. Endocrine Rev 13:415431, 1992 4. Vinik A, Pittenger G, Rafaeloff R, Rosenberg L, Duguid W: Determinants of pancreatic islet cell mass: a balance between neogenesis and senescence/apoptosis. Diabetes Rev 4:235263, 1996 5. Zangen DH, Bonner-Weir S, Lee CH, Latimer JB, Miller CP, Habener JF, Weir GC: Reduced insulin, GLUT2, and IDX-1 in -cells after partial pancreatectomy. Diabetes 46:258264, 1997 2405

ANTIOXIDANT TREATMENT PROTECTS -CELLS

6. Robertson RP, Zhang H-J, Pyzdrowski KL, Walseth TF: Preservation of insulin mRNA levels and insulin secretion in HIT cells by avoidance of chronic exposure to high glucose concentrations. J Clin Invest 90:320325, 1992 7. Olson LK, Redmon JB, Towle HC, Robertson RP: Chronic exposure of HIT cells to high glucose concentrations paradoxically decreases insulin gene transcription and alters binding of insulin gene regulatory protein. J Clin Invest 92:514519, 1993 8. Sharma A, Olson LK, Robertson RP, Stein R: The reduction of insulin gene transcription in HIT-T15 cells chronically exposed to high glucose concentration is associated with loss of RIPE3b1 and STF-1 transcription factor expression. Mol Endocrinol 9:11271134, 1995 9. Poitout V, Olson LK, Robertson RP: Chronic exposure of TC-6 cells to supraphysiologic concentrations of glucose decreases binding of the RIPE3b1 insulin gene transcription activator. J Clin Invest 97:10411046, 1996 10. Moran A, Zhang H-J, Olson LK, Harmon JS, Poitout V, Robertson RP: Differentiation of glucose toxicity from beta cell exhaustion during the evolution of defective insulin gene expression in the pancreatic islet cell line, HIT-T15. J Clin Invest 99:534539, 1997 11. Ohlsson H, Karlsson K, Edlund T: IPF1, a homeodomain-containing-transactivator of the insulin gene. EMBO J 12:42514259, 1993 12. Leonard J, Peers B, Johnson T, Ferreri K, Lee S, Montminy MR: Characterization of somatostatin transactivating factor-1, a novel homeobox factor that stimulates somatostatin expression in pancreatic islet cells. Mol Endocrinol 7:12751283, 1993 13. Miller CP, McGehee RE, Habener JF: IDX-1: a new homeodomain transcription factor expressed in rat pancreatic islets and duodenum that transactivates the somatostatin gene. EMBO J 13:11451156, 1994 14. Stoffers DA, Thomas MK, Habener JF: Homeodomain protein IDX-1: a master regulator of pancreas development and insulin gene expression. Trends Endocrinol Metab 8:145151, 1997 15. Weir GC, Sharma A, Zangen DH, Bonner-Weir S: Transcription factor abnormalities as a cause of beta cell dysfunction in diabetes: a hypothesis. Acta Dia betol 34:177184, 1997 16. Watada H, Kajimoto Y, Umayahara Y, Matsuoka T, Kaneto H, Fujitani Y, Kamada Y, Kawamori R, Yamasaki Y: The human glucokinase gene -cell-type promoter: an essential role of insulin promoter factor 1 (IPF1)/PDX-1 in its activation in HIT-T15 cells. Diabetes 45:14781488, 1996 17. Waeber G, Thompson N, Nicod P, Bonny C: Transcriptional activation of the GLUT2 gene by the IPF-1/STF-1/IDX-1 homeobox factor. Mol Endocrinol 10:13271334, 1996 18. Kaneto H, Miyagawa J, Kajimoto Y, Yamamoto K, Watada H, Umayahara Y, Hanafusa T, Matsuzawa Y, Yamasaki Y, Higashiyama S, Taniguchi N: Expression of heparin-binding epidermal growth factor-like growth factor during pancreas development: a potential role of PDX-1 in transcriptional activation. J Biol Chem 272:2913729143, 1997 19. Sakurai T, Tsuchiya S: Superoxide production from nonenzymatically glycation protein. FEBS Lett 236:406410, 1988 20. Hunt JV, Smith CC, Wolff SP: Autoxidative glycosylation and possible involvement of peroxides and free radicals in LDL modication by glucose. Diabetes 39:14201424, 1991 21. Myint T, Hoshi S, Ookawara T, Miyazawa N, Suzuki K, Taniguchi N: Immunological detection of glycated proteins in normal and streptozotocin-induced diabetic rats using anti hexitol-lysine IgG. Biochem Biophys Acta 1272:7379, 1995 22. Baynes JW: Role of oxidative stress in development of complications in diabetes. Diabetes 40:405412, 1991 23. Kaneto H, Fujii J, Myint T, Miyazawa N, Islam KN, Kawasaki Y, Suzuki K, Nakamura M, Tatsumi H, Yamasaki Y, Taniguchi N: Reducing sugars trigger oxidative modication and apoptosis in pancreatic -cells by provoking oxidative stress through the glycation reaction. Biochem J 320:855863, 1996 24. Tajiri Y, Moller C, Grill V: Long term effects of aminoguanidine on insulin

release and biosynthesis: evidence that the formation of advanced glycosylation end products inhibits B cell function. Endocrinology 138:273280, 1997 25. Ihara Y, Toyokuni S, Uchida K, Odaka H, Tanaka T, Ikeda H, Hiai H, Seino Y, Yamada Y: Hyperglycemia causes oxidative stress in pancreatic -cells of GK rats, a model of type 2 diabetes. Diabetes 48:927932, 1999 26. Tiedge M, Lortz S, Drinkgern J, Lenzen S: Relation between antioxidant enzyme gene expression and antioxidative defense status of insulin-producing cells. Diabetes 46:17331742, 1997 27. Matsuoka T, Kajimoto Y, Watada H, Kaneto H, Kishimoto M, Umayahara Y, Fujitani Y, Kamada T, Kawamori R, Yamasaki Y: Glycation-dependent, reactive oxygen species-mediated suppression of the insulin gene promoter activity in HIT cells. J Clin Invest 99:144150, 1997 28. Stahl W, Sies H: Antioxidant defense: vitamins E and C and carotenoids. Dia betes 46 (Suppl. 2):1418, 1997 29. Hummel KP, Dick MM, Coleman DL: Diabetes, a new mutation in the mouse. Science 153:11271128, 1966 30. Coleman DL, Hummel KP: Studies with the mutation, diabetes, in the mouse. Diabetologia 3:238248, 1966 31. deFrazio A, Leary JA, Hedley DW, Tattersall NHN: Immunohistochemical detection of proliferating cells in vivo. J Histochem Cytochem 35:571577, 1987 32. Tokuyama Y, Sturis J, DePaoli AM, Takeda J, Stoffel M, Tang J, Sun X, Polonsky KS, Bell GI: Evolution of -cell dysfunction in the male Zucker diabetic fatty rat. Diabetes 44:14471457, 1995 33. Pick A, Clark J, Kubstrup C, Levisetti M, Pugh W, Bonner-Weir S, Polonsky KS: Role of apoptosis in failure of -cell mass compensation for insulin resistance and -cell defects in the male Zucker diabetic fatty rat. Diabetes 47:358364, 1998 34. Rudich A, Tirosh A, Potashnik R, Hemi R, Kanety H, Bashan N: Prolonged oxidative stress impairs insulin-induced GLUT4 translocation in 3T3-L1 adipocytes. Diabetes 47:15621569, 1998 35. Donath MY, Gross DJ, Cerasi E, Kaiser N: Hyperglycemia-induced -cell apoptosis in pancreatic islets of Psammomys obesus during development of diabetes. Diabetes 48:783744, 1999 36. Scaglia L, Smith FE, Bonner-Weir S: Apoptosis contributes to the involution of cell mass in the post partum rat pancreas. Endocrinology 136:54615468, 1995 37. Jonsson J, Carlsson L, Edlund T, Edlund H: Insulin-promoter-factor 1 is required for pancreas development in mice. Nature 37:606609, 1994 38. Offield MF, Jetton TL, Labosky PA, Ray M, Stein RW, Magnuson MA, Hogan BLM, Wright CVE: PDX-1 is required for pancreas outgrowth and differentiation of the rostral duodenum. Development 122:983995, 1996 39. Waguri M, Yamamoto K, Miyagawa J, Tochino Y, Yamamori K, Kajimoto Y, Nakajima H, Watada H, Yoshiuchi I, Itoh N, Imagawa A, Namba M, Kuwajima M, Yamasaki Y, Hanafusa T, Matsuzawa Y: Demonstration of two different processes of -cell regeneration in a new diabetic mouse model induced by a selective perfusion of alloxan. Diabetes 46:12811290, 1997 40. Sharma A, Zangen DH, Reitz P, Taneja M, Lissauer ME, Miller CP, Weir GC, Habener JF, Bonner-Weir S: The homeodomain protein IDX-1 increases after an early burst of proliferation during pancreatic regeneration. Diabetes 48:507513, 1999 41. Stoffers DA, Zinkin NT, Stanojevic V, Clarke WL, Habener JF: Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence. Nat Genet 15:106110, 1997 42. Stoffers DA, Ferrer J, Clarke WL, Habener JF: Early-onset type-II diabetes mellitus (MODY4) linked to IPF1. Nat Genet 17:138139, 1997 43. Dutta S, Bonner-Weir S, Montminy M, Wright C: Regulatory factor linked to late-onset diabetes? Nature 392:560, 1998 44. Sagara M, Satoh J, Wada R, Yagihashi S, Takahashi K, Fukuzawa M, Muto G, Toyota T: Inhibition with N-acetylcysteine of development of peripheral neuropathy in streptozotocin-induced diabetic rats. Diabetologia 39:263269, 1996

2406

DIABETES, VOL. 48, DECEMBER 1999