Hydrogen peroxide alters mitochondrial activation and insulin secretion in pancreatic beta cells

Article

Reference

Hydrogen peroxide alters mitochondrial activation and insulin secretion in pancreatic beta cells

MAECHLER, Pierre, JORNOT, L, WOLLHEIM, Claes

Abstract

The effects of a transient exposure to hydrogen peroxide (10 min at 200 microM H(2)O(2)) on pancreatic beta cell signal transduction and insulin secretion have been evaluated. In rat islets, insulin secretion evoked by glucose (16.7 mM) or by the mitochondrial substrate methyl succinate (5 mM) was markedly blunted following exposure to H(2)O(2). In contrast, the secretory response induced by plasma membrane depolarization (20 mM KCl) was not significantly affected. Similar results were obtained in insulinoma INS-1 cells using glucose (12.8 mM) as secretagogue. After H(2)O(2) treatment, glucose no longer depolarized the membrane potential (DeltaPsi) of INS-1 cells or increased cytosolic Ca(2+). Both DeltaPsi and Ca(2+) responses were still observed with 30 mM KCl despite an elevated baseline of cytosolic Ca(2+) appearing approximately 10 min after exposure to H(2)O(2). The mitochondrial DeltaPsi of INS-1 cells was depolarized by H(2)O(2) abolishing the hyperpolarizing action of glucose. These DeltaPsi changes correlated with altered mitochondrial morphology; the latter was not preserved by the overexpression of the [...]

MAECHLER, Pierre, JORNOT, L, WOLLHEIM, Claes. Hydrogen peroxide alters mitochondrial activation and insulin secretion in pancreatic beta cells. Journal of Biological Chemistry , 1999, vol. 274, no. 39, p. 27905-13

PMID : 10488138

Available at:

http://archive-ouverte.unige.ch/unige:35643

Disclaimer: layout of this document may differ from the published version.

1 / 1

Hydrogen Peroxide Alters Mitochondrial Activation and Insulin Secretion in Pancreatic Beta Cells*

(Received for publication, April 12, 1999, and in revised form, July 8, 1999)

Pierre Maechler‡, Lan Jornot§, and Claes B. Wollheim‡^¶

From the‡Division of Clinical Biochemistry and the§Respiratory Division, Department of Internal Medicine, University Medical Center, CH-1211 Geneva, Switzerland

The effects of a transient exposure to hydrogen per- oxide (10 min at 200 m^M H₂O₂) on pancreatic beta cell signal transduction and insulin secretion have been evaluated. In rat islets, insulin secretion evoked by glu- cose (16.7 m^M) or by the mitochondrial substrate methyl succinate (5 m^M) was markedly blunted following expo- sure to H₂O₂. In contrast, the secretory response in- duced by plasma membrane depolarization (20 m^MKCl) was not significantly affected. Similar results were ob- tained in insulinoma INS-1 cells using glucose (12.8 m^M) as secretagogue. After H₂O₂treatment, glucose no lon- ger depolarized the membrane potential (DC) of INS-1 cells or increased cytosolic Ca²¹. Both DC and Ca²¹ responses were still observed with 30 m^MKCl despite an elevated baseline of cytosolic Ca²¹appearing;10 min after exposure to H₂O₂. The mitochondrialDCof INS-1 cells was depolarized by H₂O₂abolishing the hyperpo- larizing action of glucose. TheseDCchanges correlated with altered mitochondrial morphology; the latter was not preserved by the overexpression of the antiapo- ptotic protein Bcl-2. Mitochondrial Ca²¹was increased following exposure to H₂O₂up to the micromolar range.

No further augmentation occurred after glucose addi- tion, which normally raises this parameter. Neverthe- less, KCl was still efficient in enhancing mitochondrial Ca²¹. Cytosolic ATP was markedly reduced by H₂O₂ treatment, probably explaining the decreased endoplas- mic reticulum Ca²¹. Taken together, these data point to the mitochondria as primary targets for H₂O₂damage, which will eventually interrupt the transduction of sig- nals normally coupling glucose metabolism to insulin secretion.

The control of insulin secretion in the pancreatic beta cell depends on the precise tuning of glucose metabolism leading to signal transduction (1, 2). Indeed, impaired metabolism secre- tion coupling results in inappropriate insulin release poten- tially causing defective blood glucose homeostasis. Dysfunction of the beta cell signal transduction may be of various origin, among which oxidative stress has been proposed to play a critical role.

Type I diabetes, or insulin dependent diabetes mellitus, is an autoimmune disease characterized by altered function and

beta cell death subsequent to exposure to inflammation prod- ucts (3). During insulitis macrophages infiltrate the islets of Langerhans and generate reactive oxygen species such as hy- drogen peroxide (H₂O₂), which exert deleterious actions on the beta cells and on mitochondrial oxidative metabolism. It is noteworthy that activated phagocytes can produce as much as 47 nmol of H₂O₂/10⁶ cells within 30 min corresponding to a concentration of 47m^MH₂O₂ in a diluted volume of 1 ml (4).

Nitric oxide (NO), another free radical precursor produced by macrophages, suppresses mitochondrial activity leading to a defective insulin release in response to nutrient secretagogues (5). Moreover, it has been shown that NO damages islet cell DNA (6) and mitochondrial DNA in beta cells (7). In general, mitochondrial DNA is more sensitive to oxidative stress than nuclear DNA (8, 9). The mitochondria play a key role in the control of nutrient-induced insulin exocytosis by generating 1) ATP to raise cytosolic Ca²¹ concentration ([Ca²¹]_c)¹ through membrane depolarization (1, 2) and 2) additional mitochon- drial factor(s) triggering insulin exocytosis (10, 11).

A defective secretory response to nutrients can be encoun- tered in aged patients (12). In normal rats, insulin secretion is maintained throughout life (13), whereas in perinatal malnour- ished rats aging results in hypoinsulinemia and hyperglycemia (14). Aging has been shown to be associated with the alteration of beta cell function independent of that seen in noninsulin-de- pendent diabetes mellitus (15). A reduced translation of the mitochondrial genome (16) and of the cytochrome c oxidase activity (17) has been reported in elderly subjects, pointing to an age-related mitochondrial dysfunction in this highly oxida- tive organelle. Mitochondrial aconitase, a tricarboxylic acid cycle enzyme, is susceptible to oxidative modification during agingin vivo(18). Moreover, it has recently been demonstrated that the mitochondrial adenine-nucleotide translocase is mod- ified oxidatively during aging together with loss of functional activity (19). In cells the mitochondrion is the main source of oxidants. Indeed, imperfect electron transport generates super- oxide anions, which are spontaneously dismutated to H₂O₂(9, 20). Thus, the mitochondria are pivotal in the control of insulin secretion, whereas at the same time generating reactive oxygen species in the cell mostly in the form of H₂O₂.

Of particular importance is the high sensitivity of pancreatic beta cells to oxidative stress. Moreover, the diabetic state is associated with increased oxidative stress and free radical damage (20). In fact, the expression of the H₂O₂-inactivating enzymes catalase and glutathione peroxidase in rat pancreatic islets is twenty times lower than in the liver (21). As a conse-

* This work was supported by the Swiss National Science Foundation Grants 32-49755.96 (to C. B. W.) and 31.37799.93 (to L. J.) and by a grant from the Silva-Casa Foundation attributed through the AETAS Foundation for Research on Aging (Geneva) (to C. B. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement”

in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

¶To whom correspondence should be addressed: Division de Bio- chimie Clinique, Centre Me´dical Universitaire, 1 rue Michel-Servet, CH-1211 Geneva 4, Switzerland; Tel.: 41-22-702-55-48; Fax: 41-22-702- 55-43; E-mail: claes.wollheim@medecine.unige.ch.

1The abbreviations used are: [Ca²¹]c, cytosolic calcium concentra- tion; [Ca²¹]m, mitochondrial calcium concentration; KRBH, Krebs- Ringer bicarbonate HEPES buffer;DCc, cell membrane potential;DCm, mitochondrial membrane potential; NO, nitric oxide; ER, endoplasmic reticulum; FCCP, carbonyl cyanide p-trifluoromethoxyphenyl- hydrazone.

THEJOURNAL OFBIOLOGICALCHEMISTRY Vol. 274, No. 39, Issue of September 24, pp. 27905–27913, 1999

© 1999 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org

27905

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

quence the basal catalase activity in islets is very low (1.3 units/mg protein in the rat) and was reported to result in high susceptibility to cytotoxicity during a 16-h exposure to H₂O₂ (10 –500m^M) (22). The rat beta cell line INS-1, which was used in the present study, exhibits similar low baseline catalase levels (1.0 units/mg protein) as primary rat islets (22). Taken together, these bindings call for a better understanding of the cellular mechanisms linking oxidative stress to impaired insu- lin secretion.

In mouse pancreatic beta cells H₂O₂hyperpolarizes the cell membrane coupled with an increase of cell membrane conduct- ance (23). Moreover, it has recently been shown that H₂O₂ increases intracellular Ca²¹, decreases the ATP/ADP ratio, and inhibits glucose-stimulated insulin secretion from isolated mouse islets (24). The present work was designed to dissect in more detail oxidative stress-induced cell damage and to paral- lel mitochondrial parameters with the secretory response in insulin-secreting cells stimulated with glucose after exposure to H₂O₂. We used H₂O₂ as a well established model of a bio- logically active oxygen-derived intermediate (20) at the concen- tration of 200m^M. Rat islets and the beta cell line INS-1 have been shown to be sensitive to this H₂O₂concentration without exhibiting major toxicity (22). Our results show that the defec- tive glucose-induced insulin secretion observed after H₂O₂ treatment correlates with altered mitochondrial activation seen as a loss of mitochondrial membrane potential (DCm), decreased ATP generation (measured on-line in living cells), and impaired responses of mitochondrial Ca²¹ concentration ([Ca²¹]_m). In addition, the mitochondrial morphology, which depends mainly on DCm, was examined. In PC12 cells, the antiapoptotic protein Bcl-2 (for a review see Ref. 25) was re- ported to prevent the H₂O₂-inducedDCmloss and the subse- quent apoptosis (26). Here we also examined whether overex- pression of Bcl-2 could prevent the alteration of mitochondrial morphology during H₂O₂treatment.

EXPERIMENTAL PROCEDURES

Materials—Coelenterazine, Mitotracker, rhodamine-123, and bisoxo- nol were obtained from Molecular Probes (Eugene, OR); luciferin was

from Promega (Madison, WI); bovine serum albumin, doxycycline, methyl succinate, firefly lantern extract, and FCCP were from Sigma;

anti-insulin antibody was from Linco (St. Charles, MO); Bcl-2 mouse monoclonal antibody was from Santa Cruz Biotechnology (Santa Cruz, CA); G418 and hygromycin were from Calbiochem.

Cell Culture—INS-1 cells were cultured in RPMI 1640 medium as described previously (27–29). Stable clones of INS-1 cells expressing the Ca²¹-sensitive photoprotein aequorin in the cytosol (INS-1/C-29) (28) or targeted either to the mitochondria (INS-1/EK3) (28) or to the endo- plasmic reticulum (ER) (INS-1/ER#18) (30) were cultured in the pres- ence of 250mg/ml G418 for continuous selection of cells expressing the plasmid with the associated neomycin resistance. Clonal INS-1 lines expressing cytosolic luciferase under the control of doxycycline-depend- ent transcriptional transactivator (INS-r3-LUC7) were used for cytoso- lic ATP monitoring in living cells (29). Pancreatic islet cells were iso- lated by collagenase digestion from male Wistar rats weighing;200 g (31) and were cultured free floating in RPMI 1640 medium (11.1 mM

glucose) for 2– 4 days.

Insulin Secretion—For perifusion protocols, islets or trypsinized cells were kept in spinner culture for 2 h in glucose-free RPMI 1640 supple- mented with 25 mMHEPES and 1% new born calf serum prior to perifusion in modified Krebs-Ringer bicarbonate HEPES buffer (KRBH) composed of 135 mMNaCl, 3.6 mMKCl, 10 mMHEPES, pH 7.4, 5 mMNaHCO3, 0.5 mMNaH2PO4, 0.5 mMMgCl2, 1.5 mMCaCl2, and 2.8 mMglucose. Cells were placed in a thermostatted chamber (35 rat islets or 10⁶clonal cells/chamber) and perifused at a flow rate of 1 ml/min. For static incubations, INS-1 cells (2 3 10⁵cells/well in polyornithine- treated 24-well plates) were seeded and cultured for 3–5 days in com- plete RPMI 1640 medium. Prior to the experiments, cells were main- FIG. 1.Effect of H2O2on insulin secretion in rat islets stimulated by glucose, methyl succinate, or KCl. Rat pancreatic islets were maintained in culture 2– 4 days prior to perifusion experiments in KRBH buffer containing basal 2.8 mMglucose. Insulin secretion was stimulated for 10 min with 16.7 mMglucose (A), 5 mMmethyl succinate (C), or 20 mMKCl (E). The oxidative stress was imposed for 10 min by perifusing 200 m^MH2O2followed by a 5-min recovery period in basal buffer supplemented with 100 units/ml catalase. Secretagogues were then applied for 10 min:

16.7 mMglucose (B), 5 mMmethyl succinate (D), or 20 mMKCl (F). Traces are the mean1S.E. of 4 – 6 preparations (see Table I for statistics). The thin linesinAandBare representative control traces of three independent experiments done in triplicate without glucose stimulation.

TABLE I

Effect of H2O2on insulin secretion (% of cell content) in rat islets stimulated with different secretagogues

The values are the mean6S.E. (n54 – 6) of areas under the curve calculated for individual traces presented in Fig. 1 during 10 min of stimulation with the mentioned secretagogues without (control) or with (oxidative stress) a 10-min pretreatment with 200m^MH2O2. NS, not significant.

Groups Glucose Methyl succinate KCl

16.7 mM 5 mM 20 mM

Control 0.23160.028 0.41960.126 0.30860.114 Oxidative stress 0.06960.003 0.12460.010 0.28960.097 Difference 270%,p,0.005 274%,p,0.05 26%, NS

Oxidative Stress and Pancreatic Beta Cell Function 27906

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

tained for 2 h in the glucose-free culture medium (unless otherwise stated), washed, and preincubated in glucose-free KRBH for 30 min.

After further washing, cells were incubated with or without 200m^M H2O2for 10 min, washed again in the presence of 100 units/ml catalase, and finally stimulated for 30 min. Bovine serum albumin (0.1%) was added to buffers as carrier, and insulin secretion and cellular insulin content extracted with acid ethanol were determined by radioimmuno- assay using rat insulin as a standard (27).

Measurements of Luminescence—Luciferase or aequorin-expressing cells were seeded on 13-mm diameter coverslips 3–5 days prior to analysis and maintained in the same medium as above except for the addition of G418 and hygromycin. Prior to luminescence measure- ments, cells were maintained in glucose- and glutamine-free RPMI 1640 containing 10 mMHEPES for 2–5 h at 37 °C. This period also served to load aequorin-expressing cells with 2.5m^Mcoelenterazine, the prosthetic group of aequorin (28). For ER#18 cells, the latter procedure was undertaken in Ca²¹-free buffers to conserve ER aequorin before the measurements as detailed elsewhere (30). Luminescence was measured by placing the coverslip in a 0.5-ml thermostatted chamber at 37 °C approximately 5 mm from the photon detector. We used a home-built photomultiplier apparatus (EMI 9789, Thorn-EMI, United Kingdom), and data were collected each second on a computer photon counting board (EMI C660) prior to calibration for [Ca²¹]cand [Ca²¹]mas described previously (28). Suspensions of islet cells were perifused with the same buffers as INS-1 cells using a perifusion apparatus (28). Intact cells were perifused with KRBH, and beetle luciferin (10m^M) was added to the buffer for luciferase-expressing cells (29). Luciferase luminescence was used for the monitoring of ATP in living cells as described previously (29).

ATP Measurements in Cell Extracts—According to the protocol of Stanley and Williams (32), cells were stimulated for 10 min at 37 °C and scraped into 1 ml of 0.4 N HClO4to terminate the reaction. Following neutralization with 2 N K2CO3, cell extracts were incubated with a luciferin-luciferase mixture in arsenate buffer (0.1MNa2HAsO4z7H2O, pH 7.4, 20 mM MgSO4z7H2O), and the resultant luminescence was measured.

Membrane Potential—After a 3–5 day culture period, cells were trypsinized (0.025% trypsin, 0.27 mMEDTA), and the cell suspension was maintained for 2 h in a spinner culture with glucose-free RPMI 1640 plus 1% newborn calf serum at 37 °C. For cell membrane potential (DCc) measurements, 2310⁶cells were pelleted, resuspended in 2 ml of KRBH with 100 nMbisoxonol, and transferred to a thermostatted cuvette. Cells were then excited at 540 nm, and emission was recorded at 580 nm.DCmwas measured as described (10, 33). Briefly, after the spinner culture period, cells were loaded with 10mg/ml rhodamine-123 for 10 min at 37 °C. After centrifugation, the cells were resuspended and transferred to the fluorimeter cuvette, and fluorescence, excited at 490 nm, was measured at 530 nm. All membrane potential measure- ments were performed at 37 °C with gentle stirring in an LS-50B fluorimeter (Perkin-Elmer).

Transient Transfection with Human Bcl-2—Cells were seeded on A-431 extracellular matrix-coated glass coverslips at 2310⁵cells/2 ml of RPMI 1640 medium in 35-mm dishes. Two to three days later the cells were washed with phosphate-buffered saline. For each well, 20ml of the polycationic lipid LipofectAMINE (Life Technologies, Inc.) was diluted in 200ml of RPMI 1640 medium (10 m^MHEPES), and 4mg of

plasmid with human Bcl-2 (kindly donated by Dr. Jean-Claude Marti- nou, Ares Serono, Geneva, Switzerland) was diluted in another 200ml of the same medium. The liposome and plasmid solutions were then mixed and the volume was adjusted to 500ml with RPMI 1640-HEPES medium. After a 30-min incubation at room temperature, this transfec- tion mixture was added to the well prior to a 5-h incubation at 37 °C in air/5% CO2. The cells were then washed twice with RPMI 1640-fetal calf serum (10%) medium and further cultured in the same medium for 48 h before the experiment. This transfection procedure resulted in;25% of transfected cells as judged by immunofluorescence, using an anti-hu- man Bcl-2 antibody that did not react with the endogenous rat Bcl-2.

Mitochondrial Staining and Immunofluorescence—Rat islet cells and INS-1 cells were cultured for 2–3 days on A-431 matrix-coated glass coverslips prior to the experiments (11). The cells were washed twice with KRBH, loaded with 100 nMMitotracker for 25 min at 37 °C, and washed again with KRBH. Then the cells were exposed to KRBH only, with 200m^MH2O2or 1m^MFCCP for 10 min. Treatment was stopped by adding catalase (100 units/ml) and giving fresh KRBH for another 20-min period. The cells were then washed, fixed in 4% paraformalde- hyde, and permeabilized with 0.1% Triton X-100 in phosphate-buffered saline, 1% bovine serum albumin. The preparation was then blocked with phosphate-buffered saline bovine serum albumin before staining using mouse monoclonal anti-human Bcl-2 immunoglobulin followed by fluorescein-labeled goat anti-mouse IgG as second antibody. Cells were viewed using a Zeiss laserscan confocal 410 microscope.

Statistical Analysis—Where applicable, values are expressed as mean6S.E., and significance of difference was calculated by a Stu- dent’sttest for unpaired data. Traces without S.E. values are repre- sentative of at least three independent experiments.

RESULTS

Effect of H₂O₂on Insulin Secretion in Rat Islets and INS-1 Cells—Rat pancreatic islets were maintained in culture (11.1 mM glucose) for 2– 4 days prior to the experiments. Insulin secretion was stimulated with 16.7 mMglucose for 10 min after basal perifusion at 2.8 mM glucose (Fig. 1A). When 200 m^M H₂O₂was added for 10 min a transient release of insulin was observed. The cells were then washed for 5 min in the presence of catalase to block any remaining extracellular H₂O₂. The subsequent stimulation with 16.7 mM glucose was largely blunted (Fig. 1B), because only 30% of the secretory response was preserved in terms of area under the curve (see Table I).

Control representative traces of islets perifused without glu- cose stimulation are shown in Fig. 1,AandB(thin lines). The tricarboxylic acid cycle intermediate succinate, rendered cell permeant by the ester binding of a methyl group (34), evoked an insulin secretory response (Fig. 1C), which was significantly inhibited by 74% after H₂O₂treatment (Fig. 1Dand Table I). A nonnutrient stimulation of insulin release was induced by KCl, which triggers insulin exocytosis by a simple rise of [Ca²¹]_c consequent to membrane depolarization (Fig. 1E). In this case TABLE II

Effect of H2O2on insulin secretion (% of cell content) in INS-1 cells stimulated with glucose following different preincubation conditions Cells were preincubated in glucose-free medium (groups a and b) or in the presence of 11.1 mMglucose (groups c-f) before exposure to 200m^M H2O2(groups e and f) or KRBH only (groups a-d) and then incubated for 30 min with glucose at 2.8 mM(basal) or 12.8 mM(stimulated). The values are the mean6S.E. of a representative independent experiment (out of three) done in quadruplicate. NS, not significant.

Preincubation

conditions H2O2exposure Insulin secretion

during H₂O₂exposure Glucose incubation

Insulin secretion during glucose

incubation

2 h 10 min 30 min

a. Glucose-free Nil 1.3060.12 2.8 mM 1.9360.20

(basal)

b. Glucose-free Nil 1.3460.15 12.8 mM 16.7461.41

(stimulated) p,0.0001versusa.

c. Glucose 11.1 mM Nil 1.5760.06 2.8 mM 5.1460.12

(basal) p,0.0001versusa.

d. Glucose 11.1 mM Nil 1.7060.15 12.8 mM 16.8160.75

(stimulated) p,0.0001versusc.

e. Glucose 11.1 mM 200m^MH2O2 3.8360.35 2.8 mM 4.4360.49

p,0.001versusc. (basal) NSversusc.

f. Glucose 11.1 mM 200m^MH2O2 2.8960.19 12.8 mM 6.4661.12

p,0.005versuse. (stimulated) NSversuse.

Oxidative Stress and Pancreatic Beta Cell Function 27907

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

a similar secretory response to KCl was observed after H₂O₂ exposure compared with the control stimulation (Fig. 1F). In static incubations at basal 2.8 mMglucose, the insulinoma cells INS-1 released insulin during a 10-min exposure to H₂O₂in a dose-dependent manner (Fig. 2A). The threshold for the action of H₂O₂on insulin secretion was 200m^M (152%,p,0.02), a concentration used in all following experiments. This 10-min treatment with 200 m^M H₂O₂ did not result in apparent cell morphology alteration when observed by phase-contrast mi- croscopy (not shown). Following the 10-min treatment with H₂O₂, the basal secretion (at 2.8 mMglucose) measured during a 30-min incubation was elevated (Fig. 2B). The clonal cells were more sensitive to H₂O₂than primary cells, because stim- ulation with 12.8 mMglucose after the oxidative stress period was unable to evoke any secretory response (Fig. 2B). More- over, because the basal release was elevated, no further effect of 30 mMKCl was observed above the secretory level reached in control preparations. It should be noted that these experiments were performed following a 2-h preincubation period in glu- cose-free medium, which slightly decreases basal insulin re- lease (Table II). This could possibly sensitize the cells to oxi- dative stress because of fuel depletion. Therefore, secretion experiments were repeated without a starvation period and

showed a similar pattern as in Fig. 2,AandB,in terms of H₂O₂ sensitivity and blunted glucose (12.8 mM) response (Table II).

Effect of H₂O₂onDCcin INS-1 Cells—Bisoxonol fluorescence was used to monitor theDCcin a suspension of INS-1 cells.

After a baseline at 2.8 mM glucose, 10 mM sugar was added (12.8 mMfinal), which produced a depolarization ofDCc, and it was further depolarized by the subsequent addition of 30 m^M KCl (Fig. 3A). When 200 m^M H₂O₂ was first added to the cuvette, a transient hyperpolarization ofDCcwas followed by a slight, gradual depolarization (Fig. 3B). After a 10-min inter- FIG. 2.Effect of H2O2on insulin secretion in INS-1 cells stim-

ulated by glucose or KCl. INS-1 cells were cultured for 3–5 days and then maintained for 2 h in a glucose-free medium prior to preincubation in glucose free KRBH for 30 min. After further washing, cells were incubated with increasing concentrations of H2O2for 10 min and insu- lin release was determined (A). InB, cells were incubated without (Control) or with (Oxidative Stress) 200m^MH2O2for 10 min, washed again in the presence of 100 units/ml catalase, and finally stimulated with 12.8 mMglucose or 30 mM KCl for 30 min. Basal condition refers to KRBH buffer containing 2.8 mMglucose. Values are the mean6S.E. of quadruplicates from one of three (A) or four (B) experiments. *, p, 0.02; **,p,0.002; ***,p,0.0005versus A, control value without H2O2

(1.7560.11% content);B, basal conditions in control groups.

FIG. 3.Effect of H2O2onDCcin INS-1 cells. INS-1 cells were kept in suspension for 2 h in a spinner culture with glucose-free medium at 37 °C. ForDCcmeasurements, 2310⁶cells were resuspended in 2 ml of KRBH containing the fluorescent dye bisoxonol and transferred to a thermostatted cuvette. InA, the depolarizing effect of 10 mMglucose (12.8 mMfinal) was tested followed by a control depolarization using 30 mMKCl. InB, 200m^MH2O2was added for 10 min followed by 100 units/ml catalase before testing the effect of 10 mMglucose (12.8 mM

final) and finally KCl (30 mM). InC, 200m^MH2O2was applied during glucose-inducedDCcdepolarization. Each trace is representative of 3–5 independent experiments.

Oxidative Stress and Pancreatic Beta Cell Function 27908

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

val, 100 units/ml catalase was added to neutralize the remain- ing extracellular H₂O₂. Subsequent exposure to 10 mMglucose caused no significant depolarization of the DCc. In contrast, KCl (30 mM) evoked further depolarization (Fig. 3B). When H₂O₂ treatment was applied after glucose stimulation, the glucose-induced cell depolarization was reversed without inhi- bition of the effect of KCl added subsequently (Fig. 3C).

Effect of H₂O₂ on [Ca²¹]_c in INS-1 Cells—Attached INS-1 cells stably expressing cytosolic aequorin, a Ca²¹-sensitive photoprotein, were used to measure [Ca²¹]_c in a perifusion setup. Glucose (12.8 mM) and KCl (30 mM) raised [Ca²¹]_c(Fig.

4A). The perifusion of 200 m^M H₂O₂ for 10 min resulted in a retarded and sustained increase of [Ca²¹]_c to ;400 n^M (Fig.

4B). When glucose (12.8 mM) was added after H₂O₂treatment, no further augmentation of [Ca²¹]_cwas observed (Fig. 4C). In contrast, KCl (30 mM) caused a normal rise in [Ca²¹]_c even after the oxidative stress (Fig. 4D).

Effect of H₂O₂onDCmin INS-1 Cells—TheDCmwas meas- ured in a suspension of INS-1 cells by monitoring rhodamine- 123 fluorescence. The addition of 10 mMglucose (12.8 mMfinal) potently hyperpolarized the DCm, and 1 m^M protonophore FCCP depolarized it (Fig. 5A). H₂O₂(200m^M) induced an initial moderate and rapid depolarization ofDCmfollowed by a slow, progressive further depolarization (Fig. 5B). After catalase

treatment, 10 mMglucose (12.8 mMfinal) failed to hyperpolar- izeDCmand even accelerated the depolarization. The subse- quent addition of 1m^MFCCP completed theDCmdepolariza- tion first initiated by H₂O₂. As a control of enzymatic efficacy, catalase was added before H₂O₂treatment. In these conditions, H₂O₂ failed to alter the DCm, and glucose added thereafter hyperpolarized the DCm in a normal manner (Fig. 5B, thin line). When H₂O₂ was applied after glucose stimulation, the hyperpolarization of DCmresulting from glucose metabolism FIG. 4.Effect of H2O2on cytosolic [Ca²¹] in INS-1 cells. Adher-

ent INS-1 cells expressing cytosolic aequorin were cultured for 3–5 days before loading with coelenterazine and monitoring of photon emission in a photon counting chamber perifused with KRBH containing basal 2.8 mMglucose.Ashows a control trace using cytosolic Ca²¹raising agents,i.e.12.8 mMglucose and 30 mMKCl.B,200m^MH2O2was added for 10 min followed by 100 units/ml catalase (Cat.) (thin line,control trace without H2O2). Effects of 12.8 mMglucose (C) and 30 mMKCl (D) following H2O2treatment are shown. Each trace is representative of 3–5 independent experiments.

FIG. 5. Effect of H2O2 on DCm in INS-1 cells. The DCm was measured in a suspension of 2310⁶INS-1 cells/2 ml of KRBH using rhodamine-123 fluorescence after a spinner culture period. InA, the glucose-induced (12.8 mMfinal) hyperpolarization ofDCmwas tested followed by the complete depolarization of DCm using 1 m^M of the uncoupler FCCP. InB, 200m^MH2O2was added for 10 min followed by 100 units/ml catalase before testing the effect of 10 mMglucose (12.8 mM

final) and finally FCCP (1m^M). As a control, catalase (Cat.) was also added before H2O2treatment (B,thin line). InC, 200m^MH2O2was applied during glucose-induced (12.8 mMfinal)DCmhyperpolarization.

Each trace is representative of 3–5 independent experiments.

Oxidative Stress and Pancreatic Beta Cell Function 27909

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

was counteracted leading to depolarization ofDCm(Fig. 5C).

H₂O₂ almost completely depolarized the DCmbecause FCCP had only minor effects when added after the oxidative stress.

Effect of H₂O₂on [Ca²¹]_min INS-1 Cells—[Ca²¹]_mwas mon- itored in perifused INS-1 cells stably expressing mitochondri- ally targeted aequorin. In control recordings glucose (12.8 m^M) raised [Ca²¹]_mup to 1.3m^M(Fig. 6A). The addition of 200m^M H₂O₂for 10 min rapidly increased the [Ca²¹]_mfrom a baseline of 200 nM to the micromolar range followed by a continuous augmentation up to 1.2m^M(Fig. 6B). After the removal of H₂O₂ and its extracellular depletion by catalase (100 units/ml for 2 min), [Ca²¹]_mstabilized at a plateau of approximately 1.1m^M. This oxidative stress completely inhibited the glucose response (Fig. 6C). In contrast, H₂O₂ treatment did not affect the [Ca²¹]_mrise evoked by KCl-induced cell depolarization (Fig.

6D), which also augments [Ca²¹]_c(Fig. 4D).

Effect of H₂O₂on ATP Generation in INS-1 Cells—Cellular ATP was measured after appropriate incubation in cell extracts (static) and monitored in living INS-1 cells expressing cytosolic luciferase (perifusion). When measured from static incuba- tions, glucose (12.8 mM) increased the ATP content by 34%

compared with nonstimulated cells. A 10-min treatment with 200m^MH₂O₂reduced the cellular ATP by 57%. When the cells were stimulated with 12.8 mM glucose after the oxidative stress, the ATP content was still 40% lower compared with

control cells despite a significant increase relative to cells ex- posed to H₂O₂(Table III). When monitored in perifusion, glu- cose (12.8 mM) increased ATP production approximately 40%

above baseline (Fig. 7A). When oxidative stress was applied during glucose stimulation (12.8 m^M), the glucose-induced in- crease in cytosolic ATP was returned to basal levels (Fig. 7B).

At 2.8 mM glucose, the addition of 200m^M H₂O₂ provoked a continuous decrease of the cytosolic ATP concentration until the removal of H₂O₂at the end of the 10-min period of oxidative stress (Fig. 7C). After a recovery phase corresponding to the catalase treatment, the cells were exposed to 12.8 mMglucose, which slightly increased cytosolic ATP. Because the inhibition of ATP generation has been shown to result in a marked de- crease of ER Ca²¹ (24), we have also monitored the latter parameter in INS-1 cells expressing aequorin in the ER com- partment. The levels of ER Ca²¹were lowered by exposure of the cells to 200m^MH₂O₂(Fig. 7D). The lag time for the decrease of ER Ca²¹was 46.863.9 s and the amplitude was 58.3610.8 m^Mat 3 min after the addition of H₂O₂(n54).

Effect of H₂O₂on Mitochondrial Morphology in Cells Over- expressing Bcl-2—INS-1 cells were cultured on glass coverslips transiently transfected with human Bcl-2, and mitochondria were stained with Mitotracker prior to treatment with 200m^M H₂O₂or 1m^MFCCP for 10 min. Control cells exhibited normal filament-like mitochondria (Fig. 8A). Exposure to H₂O₂modi- fied the mitochondrial morphology into a more condensed, glob- ular pattern (Fig. 8B). Treatment with the uncoupling protono- phore FCCP produced similar effects on the mitochondria (Fig.

8C). Cells overexpressing the antiapoptotic protein Bcl-2 were identified by immunostaining using anti-human Bcl-2 antibody (Fig. 8, E, G, and I). This antibody did not react with the endogenous rat Bcl-2, revealing only the transgene. In control cells, human Bcl-2 co-localized with mitochondria. In addition, there was a diffuse staining, because Bcl-2 is located predom- inantly, albeit not exclusively, in the outer mitochondrial mem- brane but also on the endoplasmic reticulum and in the nuclear membrane (Fig. 8E). Overexpression of Bcl-2 did not protect against the alteration of mitochondrial morphology induced by H₂O₂or by the depolarization evoked by FCCP (Fig. 8,FandH, respectively).

DISCUSSION

Mitochondria play a key role in the control of nutrient- induced insulin secretion by coupling glycolysis to distal events leading to exocytosis (1, 2, 35). However, this highly aerobic organelle also produces H₂O₂from dismutated superoxide an- ions (9) generating an intracellular oxidative stress that may damage neighboring molecules. Among these, mitochondrial enzymes such as aconitase (18) and adenine-nucleotide trans- FIG. 6.Effect of H2O2on mitochondrial [Ca²¹] in INS-1 cells.

Adherent INS-1 cells expressing mitochondrial aequorin were cultured for 3–5 days before loading with coelenterazine and monitoring photon emission in a photon counting chamber perifused with KRBH contain- ing basal 2.8 mMglucose.Ashows the effect of 10 mMglucose (12.8 mM

final) on mitochondrial Ca²¹.Bshows the effect of a 10-min treatment with 200m^MH2O2followed by the addition of 100 units/ml catalase.C andDshow the effects of 12.8 mMglucose (C) and 30 mMKCl (D) following H2O2treatment. Each trace is representative of 3–5 inde- pendent experiments.

TABLE III

Effect of H2O2on ATP levels in INS-1 cells stimulated with glucose The values are the mean6S.E. of ATP levels in cell extracts from five independent experiments. INS-1 cells were incubated for 10 min in the presence of 2.8 mMglucose (control) with 200m^MH2O2for the groups undergoing oxidative stress before stimulation with 12.8 mMglucose for 10 min.

Groups ATP

Differences versus control

Differences versusH2O2

pmol/mg protein

Control 34.961.6

(2.8 mMglucose)

Glucose 46.761.5 134%

(12.8 mM) p,0.0005

H2O2-control 15.061.1 257%

(Oxidative stress) p,0.0001

Glucose (12.8 mM) 20.962.6 240% 139%

After oxidative stress p,0.001 p,0.05

Oxidative Stress and Pancreatic Beta Cell Function 27910

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

locase (19) are susceptible to oxidative modification. Mitochon- drial DNA is also highly sensitive to oxidative stress (8, 9). In the present study, we used H₂O₂as a recognized biologically important oxidant (20). The cells were exposed to 200m^MH₂O₂, a level seen during activation of phagocytesin vitro(4).

Exposure of rat islets to H₂O₂ resulted in a transient in- crease in insulin release at basal nonstimulatory glucose con- centration and impaired the glucose-induced secretory re- sponse tested subsequent to the oxidative stress. A similar pattern was obtained using the mitochondrial substrate methyl succinate as a secretagogue, suggesting that pathways down- stream of glycolysis are susceptible to oxidative alterations. On the contrary, the KCl-evoked cell depolarization was still able to promote insulin exocytosis following exposure to H₂O₂in the islet perifusion experiments. This points to the mitochondrion rather than to the exocytotic process as the sensitive site. These results are in accordance with a recent report from our labora- tory (36) using the diabetogenic compound alloxan known to generate H₂O₂and free radicals (20). Exposure of INS-1 cells to alloxan for 5 min resulted in elevated basal insulin release;

thereafter glucose failed to stimulate oxidative metabolism and insulin secretion, whereas the secretory response to KCl was largely preserved (36). Taken together, these data suggest that oxidative stress primarily alters nutrient-evoked insulin secre- tion and that the elevated basal insulin release is the conse- quence of an increase in [Ca²¹]_c. Indeed, treatment of INS-1 cells with either H₂O₂(present study) or alloxan (36) results in the elevation of [Ca²¹]_c. The steady state [Ca²¹]_c attained under these conditions was similar to that measured with KCl on the plateau phase, which could explain why the effects of

H₂O₂and KCl on insulin secretion in static incubation were not additive.

In the mitochondria, exposure to H₂O₂caused a rapid eleva- tion of Ca²¹, even at early time points when [Ca²¹]_cwas still at basal level, excluding a simple transfer between these two compartments. Rather, the first phase of [Ca²¹]_mrise might be explained by a direct inhibition by H₂O₂of the Na¹/Ca²¹an- tiporter that controls the efflux of Ca²¹from the mitochondrial matrix (37). Transfer of reducing equivalents to the electron transport chain increases DCm, which enhances the driving force for mitochondrial Ca²¹uptake mediated by a low affinity uniporter (38). The failure of glucose to hyperpolarize DCm

following oxidative stress undoubtedly explains the lack of the normal [Ca²¹]_m increase during glucose stimulation (28).

These mitochondrial alterations led to the diminished ATP synthesis and probably to an insufficient generation of addi- tional mitochondrial factors required for nutrient-evoked insu- lin exocytosis (10). The depolarized DCmexplains the altered mitochondrial morphology, which was not prevented by over- expression of Bcl-2 (Fig. 8). Both alterations are also observed in mitochondrial DNA-deficient INS-1 cells associated with impaired nutrient-induced insulin secretion (39). It is notewor- thy that a mutation in the human mitochondrial tRNA^Leu(UUR) gene, which is associated with mitochondrial diabetes, also yields a similar mitochondrial morphology (40).

[Ca²¹]_cwas elevated following H₂O₂treatment, and glucose added thereafter failed to elicit a [Ca²¹]_c response. This was not the case for the KCl-evoked depolarization ofDCc, which resulted in normal [Ca²¹]_c rise, suggesting that the voltage sensitive L-type Ca²¹ channels were still functional. Similar FIG. 7.Effect of H2O2on cytosolic ATP and ER Ca²¹in INS-1 cells. Adherent INS-1 cells expressing cytosolic luciferase or ER aequorin were cultured for 3–5 days before recording in a photon counting chamber perifused with KRBH containing basal 2.8 mMglucose.Ashows the changes of cytosolic ATP generated by glucose (12.8 mMfinal).Bshows the effect of 200m^MH2O2during glucose stimulation (12.8 mM) while monitoring cytosolic ATP.Cshows the effect of 200m^MH2O2at basal glucose (2.8 mM) followed by the addition of 100 units/ml of catalase for 5 min before testing the effect of glucose (12.8 mMfinal).Dshows ER Ca²¹changes in control cells (thin line) and cells exposed to 200m^MH2O2(thick line) at basal glucose (2.8 mM). Each trace is representative of 3–5 independent experiments.

Oxidative Stress and Pancreatic Beta Cell Function 27911

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

effects of H₂O₂ have been reported recently in the CRI-G1 insulin secreting cell line (41) and in mouse islets (24). Krip- peit-Drewset al.(24) postulated that the first phase of intra- cellular Ca²¹ rise is because of Ca²¹ mobilization from the mitochondria, whereas the second phase would reflect Ca²¹ influx through pathways distinct fromL-type Ca²¹channels.

Hersonet al.(41) suggested that the first phase is caused by mobilization of intracellular Ca²¹stores insensitive to thapsi- gargin and therefore proposed the mitochondria as the source of the initial [Ca²¹]_crise. Regarding the second phase, these authors (41) postulated that H₂O₂opens a nonselective cation channel in the plasma membrane, which was described in a previous study as a novel Ca²¹influx pathway activated by oxidative stress (42). We did not observe the second phase [Ca²¹]_crise probably because of the 50 times lower concentra- tion of H₂O₂used in the present study. The measurements of Ca²¹changes both in the mitochondrial and ER compartments permitted us to conclude that the first phase of [Ca²¹]_celeva- tion is probably because of diminished pumping of Ca²¹into the ER as a consequence of a decrease in cytosolic ATP (30, 43).

Indeed, ER Ca²¹was lowered by H₂O₂treatment secondary to the marked decrease in cytosolic ATP and before the elevation of [Ca²¹]_c. These results substantiate the critical role of cyto- solic ATP in determining the filling state of the ER Ca²¹stores.

The following sequence of events for the actions of H₂O₂can be proposed. We suggest that hydrogen peroxide or derivative products such as free radicals directly inhibits the mitochon- drial Na¹/Ca²¹antiporter, as previously observed (37), block- ing the efflux of Ca²¹from the mitochondria. This inhibition would lead to the rapid rise in [Ca²¹]_mindependent of any changes in the cytosolic compartment, an effect recently re- ported in endothelial cells (37). Meanwhile, oxidative stress inhibits tricarboxylic acid cycle enzymes such as aconitase (18), attenuating the generation of reducing equivalents, in partic-

ular NADH. This would lead to inactivation of the electron transport chain with the concomitant loss of DCm and the decline in ATP generation. Moreover, oxidative modification of the adenine-nucleotide translocase would inhibit the transloca- tion of ATP to the cytosol (19). The depolarization of DCm is unlikely to be because of the opening of the transition pore, at least on this short time scale, because overexpression of Bcl-2 did not prevent the mitochondrial effects evoked by H₂O₂treatment.

The retarded elevation in [Ca²¹]_cfollowing exposure to H₂O₂ may well be explained by the observed decrease in cytosolic ATP. Indeed, lowering of cytosolic ATP by the mitochondrial uncoupler FCCP results in a marked decrease of ER Ca²¹as monitored in INS-1 cells expressing aequorin in this cellular compartment (30). Reduction of cytosolic ATP results in im- paired pumping of Ca²¹into the ER, an effect mimicked by the sarco/endoplasmic reticulum Ca²¹-ATPase inhibitors. The lat- ter effect has been demonstrated in insulin-secreting cell lines (30, 41) and has also been shown to lead to apoptosis (44, 45).

Thus, the lowering of ER Ca²¹by H₂O₂treatment is another demonstration of the close functional coupling between mito- chondria and ER Ca²¹stores (45, 46). This effect on ER Ca²¹ certainly accounts for the rise in [Ca²¹]_cobserved shortly after H₂O₂treatment. Ca²¹influx from the extracellular space was also reported to occur following exposure to H₂O₂in mouse islets (24) or to alloxan (generating free radicals) in INS-1 cells (36). It has been postulated that a nonselective cation channel mediates this retarded Ca²¹influx, which is promoted by oxidative stress (41). Whatever the mechanism, it can be surmised that both the inhibition of Ca²¹pumping by the sarco/endoplasmic reticulum Ca²¹-ATPase into the ER and the Ca²¹influx from the extra- cellular space contribute to the elevation of [Ca²¹]_c.

Dysfunction of the beta cell by oxidative stress is implicated in type I diabetes and in aging. The present study provides evidence for the involvement of the mitochondria in impaired FIG. 8.Effect of H2O2on mitochondrial morphology in INS-1 cells. Cells were cultured on glass coverslips and (D–I) transfected with Bcl-2.

Mitochondria were stained with Mitotracker prior to the appropriate treatment for 10 min followed by fixation and visualization under a laserscan confocal microscope.A,control cells;B,cells treated with 200m^MH2O2;C,cells treated with 1m^Mof the uncoupler FCCP. InD–I, cells were transiently transfected with human Bcl-2. Cells expressing the transgene were identified by immunostaining using anti-human Bcl-2 antibody (E, G,andI).D,F,andHshow the mitochondrial staining of untransfected cells and Bcl-2-transfected cells (arrow) for control, H2O2, and FCCP conditions, respectively.

Oxidative Stress and Pancreatic Beta Cell Function 27912

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from

signal transduction, coupling glucose metabolism to insulin secretion following an oxidative stress. Moreover, a model is suggested for the sequence of events leading to the elevation of [Ca²¹]_c. Further work will address the identification of factors normally required for nutrient-induced insulin secretion but are missing in the damaged beta cell.

Acknowledgment—We are indebted to C. Bartley, G. Chaffard, and H. Petersen for expert technical assistance.

REFERENCES 1. Matschinsky, F. M. (1996)Diabetes45,223–241

2. Wollheim, C. B., Lang, J., and Regazzi, R. (1996)Diabetes Rev.4,276 –297 3. Rabinovitch, A., Suarez-Pinzon, W. L., Strynadka, K., Lakey, J. R., and

Rajotte, R. V. (1996)J. Clin. Endocrinol. Metab.81,3197–3202 4. Anderson, R. (1992)J. Immunol. Methods155,49 –55

5. Welsh, N., Eizirik, D. L., Bendtzen, K., and Sandler, S. (1991)Endocrinology 129,3167–3173

6. Fehsel, K., Jalowy, A., Qi, S., Burkart, V., Hartmann, B., and Kolb, H. (1993) Diabetes42,496 –500

7. Wilson, G. L., Patton, N. J., and LeDoux, S. P. (1997)Diabetes46,1291–1295 8. Yakes, F. M., and Van Houten, B. (1997)Proc. Natl. Acad. Sci. U. S. A.94,

514 –519

9. Beckman, K. B., and Ames, B. N. (1998)Physiol. Rev.78,547–581 10. Maechler, P., Kennedy, E. D., Pozzan, T., and Wollheim, C. B. (1997)EMBO J.

16,3833–3841

11. Maechler, P., Kennedy, E. D., Wang, H., and Wollheim, C. B. (1998)J. Biol.

Chem.273,20770 –20778

12. Coordt, M. C., Ruhe, R. C., and McDonald, R. B. (1995)Proc. Soc. Exp. Biol.

Med.209,213–222

13. Starnes, J. W., Cheong, E., and Matschinsky, F. M. (1991)Am. J. Physiol.260, E59 –E66

14. Garofano, A., Czernichow, P., and Bre´ant, B. (1999)Diabetologia42,711–718 15. Shimizu, M., Kawazu, S., Tomono, S., Ohno, T., Utsugi, T., Kato, N., Ishi, C.,

Ito, Y., and Murata, K. (1996)Diabetes Care19,8 –11

16. Hayashi, J., Ohta, S., Kagawa, Y., Kondo, H., Kaneda, H., Yonekawa, H., Takai, D., and Miyabayashi, S. (1994)J. Biol. Chem.269,6878 – 6883 17. Isobe, K., Ito, S., Hosaka, H., Iwamura, Y., Kondo, H., Kagawa, Y., and

Hayashi, J. I. (1998)J. Biol. Chem.273,4601– 4606

18. Yan, L. J., Levine, R. L., and Sohal, R. S. (1997)Proc. Natl. Acad. Sci. U. S. A.

94,11168 –11172

19. Yan, L. J., and Sohal, R. S. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 12896 –12901

20. Yu, B. P. (1994)Physiol. Rev.74,139 –162

21. Tiedge, M., Lortz, S., Drinkgern, J., and Lenzen, S. (1997)Diabetes 46, 1733–1742

22. Benhamou, P. Y., Moriscot, C., Richard, M. J., Beatrix, O., Badet, L., Pattou, F., Kerr-Conte, J., Chroboczek, J., Lemarchand, P., and Halimi, S. (1998) Diabetologia41,1093–1100

23. Krippeit-Drews, P., Lang, F., Haussinger, D., and Drews, G. (1994)Pflugers Arch. Eur. J. Physiol.426,552–554

24. Krippeit-Drews, P., Kramer, C., Welker, S., Lang, F., Ammon, H. P., and Drews, G. (1999)J. Physiol.(Lond.)514,471– 481

25. Kroemer, G. (1997)Nat. Med.3,614 – 620

26. Satoh, T., Enokido, Y., Aoshima, H., Uchiyama, Y., and Hatanaka, H. (1997) J. Neurosci. Res.50,413– 420

27. Asfari, M., Janjic, D., Meda, P., Li, G., Halban, P. A., and Wollheim, C. B.

(1992)Endocrinology130,167–178

28. Kennedy, E. D., Rizzuto, R., Theler, J. M., Pralong, W. F., Bastianutto, C., Pozzan, T., and Wollheim, C. B. (1996)J. Clin. Invest.98,2524 –2538 29. Maechler, P., Wang, H., and Wollheim, C. B. (1998)FEBS Lett.422,328 –332 30. Maechler, P., Kennedy, E. D., Sebo, E., Valeva, A., Pozzan, T., and Wollheim,

C. B. (1999)J. Biol. Chem.274,12583–12592

31. Pralong, W. F., Spat, A., and Wollheim, C. B. (1994)J. Biol. Chem.269, 27310 –27314

32. Stanley, P. E., and Williams, S. G. (1969)Anal. Biochem.29,381–392 33. Duchen, M. R., Smith, P. A., and Ashcroft, F. M. (1993)Biochem. J.294,35– 42 34. MacDonald, M. J., and Fahien, L. A. (1988)Diabetes37,997–999

35. Maechler, P., and Wollheim, C. B. (1998)Biofactors8,255–262

36. Janjic, D., Maechler, P., Sekine, N., Bartley, C., Annen, A. S., and Wollheim, C. B. (1999)Biochem. Pharmacol.57,639 – 648

37. Jornot, L., Maechler, P., Wollheim, C. B., and Junod, A. F. (1999)J. Cell Sci.

112,1013–1022

38. Gunter, T. E., Gunter, K. K., Sheu, S. S., and Gavin, C. E. (1994)Am. J.

Physiol.267,C313–C339

39. Kennedy, E. D., Maechler, P., and Wollheim, C. B. (1998)Diabetes47,374 –380 40. van den Ouweland, J. M. W., Maechler, P., Wollheim, C. B., Attardi, G., and

Maassen, J. A. (1999)Diabetologia42,485– 492

41. Herson, P. S., Lee, K., Pinnock, R. D., Hughes, J., and Ashford, M. L. J. (1999) J. Biol. Chem.274,833– 841

42. Herson, P. S., and Ashford, M. L. (1997)J. Physiol.(Lond.)501,59 – 66 43. Landolfi, B., Curci, S., Debellis, L., Pozzan, T., and Hofer, A. M. (1998)J. Cell

Biol.142,1235–1243

44. Zhou, Y. P., Teng, D., Dralyuk, F., Ostrega, D., Roe, M. W., Philipson, L., and Polonsky, K. S. (1998)J. Clin. Invest.101,1623–1632

45. Berridge, M. J., Bootman, M. D., and Lipp, P. (1998)Nature395,645– 648 46. Rizzuto, R., Pinton, P., Carrington, W., Fay, F. S., Fogarty, K. E., Lifshitz,

L. M., Tuft, R. A., and Pozzan, T. (1998)Science280,1763–1766

Oxidative Stress and Pancreatic Beta Cell Function 27913

at Bibliotheque Faculte Medecine Geneve, on October 2, 2012www.jbc.orgDownloaded from