Normalization of Intracellular Ca2� Induces a Glucose-responsiveState in Glucose-unresponsive �-Cells*

Received for publication, April 24, 2002Published, JBC Papers in Press, May 6, 2002, DOI 10.1074/jbc.M203988200

Kohtaro Minami‡§, Masaaki Yokokura‡, Nobuko Ishizuka‡, and Susumu Seino‡¶

From the ‡Department of Cellular and Molecular Medicine, Graduate School of Medicine, Chiba University, 1-8-1,Inohana, Chuo-ku, Chiba 260-8670, Japan and the §Department of Medical Genetics (Novo Nordisk Pharma),School of Medicine, Chiba University, 1-8-1, Inohana, Chuo-ku, Chiba 260-8670, Japan

Although intracellular Ca2� in pancreatic �-cells isthe principal signal for insulin secretion, the effect ofchronic elevation of the intracellular Ca2� concentra-tion ([Ca2�]i) on insulin secretion is poorly understood.We recently established two pancreatic �-cell MIN6 celllines that are glucose-responsive (MIN6-m9) and glu-cose-unresponsive (MIN6-m14). In the present study wehave determined the cause of the glucose unresponsive-ness in MIN6-m14. Initially, elevated [Ca2�]i was ob-served in MIN6-m14, but normalization of the [Ca2�]i bynifedipine, a Ca2� channel blocker, markedly improvedthe intracellular Ca2� response to glucose and the glu-cose-induced insulin secretion. The expression of sub-units of ATP-sensitive K� channels and voltage-depend-ent Ca2� channels were increased at both mRNA andprotein levels in MIN6-m14 treated with nifedipine. As aconsequence, the functional expression of these chan-nels at the cell surface, both of which are decreased inMIN6-m14 without nifedipine treatment, were increasedsignificantly. Contrariwise, Bay K8644, a Ca2� channelagonist, caused severe impairment of glucose-inducedinsulin secretion in glucose-responsive MIN6-m9 due todecreased expression of the channel subunits. Chroni-cally elevated [Ca2�]i, therefore, is responsible for theglucose unresponsiveness of MIN6-m14. The presentstudy also suggests normalization of [Ca2�]i in pancre-atic �-cells as a therapeutic strategy in treatment ofimpaired insulin secretion.

Because intracellular Ca2� is involved in a variety of cellularprocesses such as signal transduction, gene expression, andhormone release (1–6), disturbed intracellular Ca2� homeosta-sis readily induces cell dysfunction (7, 8). In pancreatic �-cells,a rise in the intracellular Ca2� concentration ([Ca2�]i) is thetrigger for insulin secretion. As the extracellular glucose con-centration increases, intracellular ATP is increased and theATP-sensitive K� (KATP) channels are closed, depolarizing theplasma membrane and opening the voltage-dependent Ca2�

channels (VDCCs),1 which allows Ca2� influx. The rise in

[Ca2�]i in the �-cells triggers exocytosis of the insulin granules(9). When the blood glucose level falls, [Ca2�]i returns to basallevel. In this context, persistent hyperglycemia might wellcause sustained elevated [Ca2�]i and abnormalities in glucose-induced insulin secretion. It has been reported that humanpancreatic islets cultured with high glucose show elevated ba-sal [Ca2�]i together with loss of the glucose-induced rise in[Ca2�]i and glucose-induced insulin secretion (10). Normal pan-creatic �-cells exposed to high glucose exhibit an abnormalresponse of intracellular Ca2� and impaired insulin secretion(11), impairments which also are observed in the �-cells ofdiabetic animals (12–14). However, the molecular basis of theeffect of chronic elevation of [Ca2�]i on insulin secretion has notbeen examined in detail, primarily because an appropriate invitro model has not been available.

We recently established two pancreatic �-cell lines with con-trary features, glucose-responsive (MIN6-m9) and glucose-un-responsive (MIN6-m14), and have shown these cell lines to beuseful in �-cell studies (15). MIN6-m9 exhibit glucose metabo-lism and insulin secretion similar to normal pancreatic �-cells,while MIN6-m14 exhibit abnormalities in glucose metabolism,KATP channel activity, VDCC activity, and glucose-induced in-sulin secretion (15).

In the present study we have determined the factors respon-sible for the glucose-unresponsiveness in MIN6-m14. [Ca2�]i inMIN6-m14 is significantly higher than in MIN6-m9. When the[Ca2�]i level was normalized by nifedipine, a Ca2� channelblocker, the glucose-induced insulin secretion was increaseddramatically in MIN6-m14 with a concomitant improvement ofKATP channel and VDCC activities. Accordingly, chronicallyelevated [Ca2�]i is the major factor contributing to the defect inglucose responsiveness of MIN6-m14, and normalization of[Ca2�]i restores the glucose-responsive state of the cells. Be-cause abnormalities of [Ca2�]i in pancreatic �-cells are associ-ated with chronic exposure to high glucose in both normal anddiabetic animals (8, 10–14), the present study suggests nor-malization of [Ca2�]i as a therapeutic strategy for the glucoseunresponsiveness of �-cells in type 2 diabetic patients.

EXPERIMENTAL PROCEDURES

Cell Culture and Measurement of [Ca2�]i—MIN6 cells were culturedin Dulbecco’s modified Eagle’s medium with 25 mM glucose supple-mented with 10% heat-inactivated fetal calf serum under humidifiedcondition of 5% CO2/95% air at 37 °C (15). Cells were loaded with 5 �M

fura-2 acetoxymethyl ester (Fura-2 AM) (Dojindo, Kumamoto, Japan)for 1 h in culture medium without pH indicator or in HEPES-balancedKrebs-Ringer bicarbonate buffer (KRH: 119 mM NaCl, 4.74 mM KCl,2.54 mM CaCl2, 1.19 mM MgCl2, 1.19 mM KH2PO4, 25 mM NaHCO3, and

* This work was supported by Grants-in-Aid for Creative ScientificResearch 10NP0201 and for Scientific Research from the Ministry ofEducation, Culture, Sports, Science and Technology; by a scientificresearch grant from the Ministry of Health, Labour, and Welfare,Japan; and by grants from Novo Nordisk Pharma Ltd., from TakedaChemical Industries Ltd., and from the Yamanouchi Foundation forResearch on Metabolic Disorders. The costs of publication of this articlewere defrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

¶ To whom correspondence should be addressed. Tel.: 81-43-226-2187; Fax: 81-43-221-7803; E-mail: seino@med.m.chiba-u.ac.jp.

1 The abbreviations used are: VDCC, voltage-dependent Ca2� chan-

nel; KRH, Krebs-Ringer HEPES; BSA, bovine serum albumin; LDH,lactate dehydrogenase; MOPS, 4-morpholinepropanesulfonic acid; HK,hexokinase; GK, glucokinase; nt, nucleotide; PM, plasma membrane.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 277, No. 28, Issue of July 12, pp. 25277–25282, 2002© 2002 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 25277

at KK

SA

NW

A K

AG

AK

U K

EN

on July 13, 2008 w

ww

.jbc.orgD

ownloaded from

10 mM HEPES, pH 7.4) containing 0.2% BSA (BSA-KRH) with 0.1 mM

glucose. [Ca2�]i was measured by a dual-excitation wavelength method(340/380 nm) with a fluorometer (Fluoroskan Ascent CF; Labsystems,Helsinki, Finland). [Ca2�]i was calibrated using solutions containingknown Ca2� concentrations (Molecular Probes, Eugene, OR).

Measurement of Insulin Secretion—Cells (1 � 105 cells/well, 48-wellplate) were exposed to 10 �M nifedipine (Sigma), 300 �M diazoxide(Sigma), or 1 �M Bay K8644 (Sigma) for 24 h (preculture). All mediacontained 0.1% Me2SO. The cells were then washed with BSA-KRH andpreincubated for 30 min in the same buffer containing 1 mM glucose.Incubation was performed with various concentrations of glucose withor without other agents as indicated for 1 h at 37 °C. The drugs used inpreculture were omitted throughout the secretion experiments. Re-leased insulin was measured as described previously (15). The amountsof insulin secretion were normalized by the cellular insulin contentsdetermined by acid-ethanol extraction.

Assay for Enzyme Activities—For determination of glucose-phospho-rylating activity, disrupted cells were centrifuged, and supernatantswere incubated in a triethanolamine buffer, pH 7.4, containing 0.5 mM

NADP, 5 mM ATP, 1 unit/ml glucose-6-phosphate dehydrogenase, and0.5 or 50 mM glucose at 30 °C. Velocity of NADPH formation wasmonitored by reading absorbencies at 340 nm. Enzyme activity wasexpressed as nmol NADPH/min/mg of protein (15). Activity of lactatedehydrogenase (LDH) was determined as follows: briefly, cell extractswere incubated in a glycylglycine buffer, pH 10.0, with 1 mM lactate, 5mM NAD, 50 mM glutamate, and 10 unit/ml glutamate-pyruvate trans-aminase at 25 °C. Velocity of NADH formation was monitored by read-ing absorbencies at 340 nm. Enzyme activity is expressed as nmolNADH/min/mg of protein (15).

Determination of ATP Production—For measurement of ATP produc-tion, cells were incubated for 1 h in the presence or absence of 25 mM

glucose. The cells then were washed twice with ice-cold PBS and solu-bilized, and the amount of ATP was measured with an ATP biolumi-nescent assay kit (Roche Molecular Diagnostics, Mannheim, Germany),according to the manufacturer’s instruction.

Electrophysiological Analyses—Whole cell recordings of ATP-sensi-tive K� current were performed as described previously (15). The ex-tracellular solution contained 135 mM NaCl, 5 mM KCl, 5 mM CaCl2, 2mM MgSO4, 5 mM HEPES, and 3 mM glucose, pH 7.4. The pipettesolution contained 107 mM KCl, 11 mM EGTA, 2 mM MgSO4, 1 mM

CaCl2, and 11 mM HEPES, pH 7.2. Whole cell Ba2� currents throughthe VDCCs were recorded as described (15). Briefly, Ba2� was used asa charged carrier for measurement of VDCC currents. The extracellularsolution contained 40 mM Ba(OH)2, 20 mM 4-aminopyridine, 90 mM

tetraethylammonium hydroxide, 10 mM tetraethylammonium chloride,140 mM methanesulfonate, and 10 mM MOPS, pH 7.4. The pipettesolution contained 10 mM CsCl, 130 mM cesium aspartate, 10 mM

EGTA, 5 mM Mg-ATP, and 10 mM MOPS (pH 7.2). Cells were main-tained at a holding potential of �60 mV, and square pulses of 400-msecduration at potentials between �40 and �70 mV in steps of 10-mV were

applied every 4 s. Recordings were made using the EPC-7 amplifier(List Electronics, Darmstadt, Germany).

RNA Blotting—Total RNA (10 �g) from cells was subjected to form-aldehyde-agarose gel electrophoresis. RNA was transferred to a nylonmembrane and UV-cross-linked. Membrane was hybridized with�-[32P]dCTP-labeled probes corresponding to the cDNA of mouse hex-okinase-I (HK-I) (GenBankTM accession no. J05277, nt 1464–1903),mouse glucokianse (GK) (L38990, nt 98–652), mouse lactate dehydro-genase-A (LDH-A) (NM_010699, nt 291–980), mouse Kir6.2 (U73626, nt753–1427), hamster sulfonylurea receptor 1 (SUR1) (L40623, nt 3126–4049), mouse �1-subunit of VDCC (NM_009781, nt 4889–5446), ormouse �3-subunit of VDCC (NM_007581, nt 801–1600). Rediprime ran-dom primer labeling kit (Amersham Biosciences) was used to label theprobes. Blots were exposed to Kodak X-OMAT AR film (Eastman KodakCo., Rochester, NY) at �80 °C.

Subcellular Fractionation and Immunoblotting—Cells were scrapedinto a lysis buffer termed buffer A, containing 150 mM NaCl, 50 mM

Tris, pH 7.4, 1 mM EDTA, 1 mM EGTA, and 1 mM dithiothreitol withprotease inhibitors and sonicated (TOMY Sonicator UD201; Tomy,Tokyo, Japan; 1 � 10-s burst; output power � 1) and then centrifugedat 700 � g for 15 min. The pellet, which contained nuclei and undis-rupted cells, was discarded. The supernatant was centrifuged at 8000 �g for 15 min. The 8,000 � g pellet was referred to as the plasmamembrane-enriched fraction. The supernatant was then ultracentri-fuged at 100,000 � g for 1 h. The pellet was enriched in internalmembranes. The final supernatant obtained was the cytosolic fraction.All fractions were resuspended in the same volume of buffer A with 1%Triton X-100. Aliquots from each subcellular fraction (volumes equal toPM fractions containing 5 �g of protein) were subjected to SDS-PAGE.Resolved proteins were transferred to polyvinylidene fluoride mem-branes (Millipore, Bedford, MA) and probed with antibodies againstKir6.2 (16), SUR1 (17), the �1-subunit of VDCCs (Alomone, Jerusalem,Israel), or the �3-subunit of VDCCs (Alomone). Secondary antibodieswere conjugated to horseradish peroxidase and visualized by enhancedchemiluminescence reagent (Amersham Biosciences). Kir6.2 was de-tected as a multimeric complex with SUR1 because it was difficult todenature completely. Mobility of the complex on the gel was determined

FIG. 1. Intracellular Ca2� concentration ([Ca2�]i) in MIN6cells. A, [Ca2�]i in standard culture condition. Cells (1 � 105 cells/well,96-well plate) were treated with or without 10 �M nifedipine. The cellswere loaded with 5 �M Fura-2 AM, and [Ca2�]i was measured in theculture medium without a pH indicator. n � 8, **, p � 0.01. m9,MIN6-m9, m14, MIN6-m14, m14-Nif, nifedipine-treated MIN6-m14. B,[Ca2�]i change in response to glucose. Cells were treated with or with-out 10 �M nifedipine and loaded with 5 �M Fura-2 AM. [Ca2�]i wasmeasured in BSA-KRH with 0.1 mM glucose at 30-s intervals, and 25mM glucose (final concentration) was added to each well at the indicatedtime points. m14, MIN6-m14, m14-Nif, nifedipine-treated MIN6-m14.n � 6.

FIG. 2. Insulin secretory response in MIN6-m14 before andafter normalization of [Ca2�]i by nifedipine. Cells (1 � 105 cells/well, 48-well plate) were treated with or without 10 �M nifedipine for24 h. The cells were washed and pre-incubated for 30 min at 37 °C inBSA-KRH containing 1 mM glucose without nifedipine. Incubation wasperformed in BSA-KRH with indicated concentrations of glucose (A),glibenclamide (B), or Bay K8644 (C) for 1 h at 37 °C. Note that nifedip-ine was absent throughout the secretion experiment. Values of secretedinsulin were normalized by cellular insulin contents. m14, MIN6-m14;m14-Nif, nifedipine-treated MIN6-m14. n � 4, *, p � 0.05, and **, p �0.01.

Importance of [Ca2�]i in Glucose Responsiveness of �-Cells25278

at KK

SA

NW

A K

AG

AK

U K

EN

on July 13, 2008 w

ww

.jbc.orgD

ownloaded from

by Kir6.2/SUR1-co-transfected Ltk cell extracts. Blots were quantifiedby scanning densitometry (Amersham Biosciences). To verify subcellu-lar fractionation, Na�/K�-ATPase (Upstate Biotechnology, Lake Placid,NY) and sarco/endoplasmic reticulum Ca2�-ATPase (Santa Cruz Bio-

technology, Santa Cruz, CA), markers for plasma membrane and inter-nal membrane, respectively, were detected by immunoblotting.

Statistical Analysis—Values are expressed as means � S.E. Thesignificance of differences between test groups was evaluated by un-paired Student’s t test or one-way analysis of variance followed byTukey’s test. p � 0.05 was considered significant.

RESULTS

[Ca2�]i of MIN6-m9 and MIN6-m14—In the standard cul-ture condition (Dulbecco’s modified Eagle’s medium with 10%fetal calf serum and 25 mM glucose), [Ca2�]i in MIN6-m14(317 � 7 nM, n � 8) was significantly higher than in MIN6-m9(227 � 4 nM, n � 8) (Fig. 1A). Application of the Ca2� channelblocker nifedipine (10 �M for 24 h) lowered [Ca2�]i of MIN6-m14 to levels similar to those in MIN6-m9 (Fig. 1A). Intracel-lular Ca2� was poorly responsive to glucose in control MIN6-m14 (without nifedipine treatment), but the response wasmarkedly improved in nifedipine-treated MIN6-m14 (Fig. 1B).

Insulin Secretion before and after Normalization of [Ca2�]i—Although normalization of [Ca2�]i by nifedipine did not alterbasal levels of insulin secretion in MIN6-m14, glucose-inducedinsulin secretion was dramatically increased after normaliza-tion of [Ca2�]i (Fig. 2A). Because the cellular insulin contentalso was increased by nifedipine (226.2 � 4.2 versus 505.2 � 12pmol/mg of protein, n � 4; control MIN6-m14 versus nifedipinetreated-MIN6-m14), insulin secretion was normalized by thecontents. The insulin secretion at 25 mM glucose was 20.8 � 1.1and 121.4 � 10.9 pmol/h/mg of protein in control MIN6-m14and nifedipine-treated MIN6-m14, respectively. The nifedipinewas washed out before the secretion experiments and wasomitted thereafter throughout the experiments. These resultsindicate that the normalization of [Ca2�]i restored the glucoseresponsiveness of MIN6-m14 both in intracellular Ca2� andinsulin secretion.

Glibenclamide- and Bay K8644-stimulated Insulin Secre-tion—We then examined the effects of normalization of [Ca2�]ion glibenclamide- or Bay K8644-stimulated insulin secretion.Glibenclamide, a sulfonylurea, stimulates insulin secretion byinhibiting the KATP channels (18), while Bay K8644 stimulatesinsulin secretion by activating the VDCCs of pancreatic �-cells(19). Neither glibenclamide (1 �M) nor Bay K8644 (1 �M) had asignificant stimulatory effect on insulin secretion in controlMIN6-m14, but effects were observed when [Ca2�]i was nor-

FIG. 3. Enzyme activities and ATP production. A, glucose phos-phorylating activity. Cells were grown on a culture dish (10-cm diam-eter) with or without 10 �M nifedipine for 24 h. The cells were lysed andincubated in a buffer containing 0.5 mM NADP, 5 mM ATP, 1 unit/mlglucose-6-phosphate dehydrogenase, and 0.5 or 50 mM glucose at 30 °C.Enzyme activity is expressed as nmol of NADPH/min/mg of protein.Glucose phosphorylating activity at 0.5 mM glucose represents HKactivity, and GK activity is the activity at 50 mM glucose minus that at0.5 mM glucose. B, LDH activity. Cells (1 � 105 cells/well, 48-well plate)were treated with or without 10 �M nifedipine for 24 h. Cell extractswere incubated in buffer with 1 mM lactate, 5 mM NAD, 50 mM gluta-mate, and 10 unit/ml glutamate-pyruvate transaminase at 25 °C. En-zyme activity was expressed as nmol of NADH/min/mg of protein. C,cellular ATP contents. Cells treated with or without 10 �M nifedipinewere incubated for 1 h in the presence or absence of 25 mM glucose. TheATP concentration of the cell lysate was measured with an ATP biolu-minescent assay kit. m14, MIN6-m14; m14-Nif, nifedipine-treatedMIN6-m14. n � 4. There was no significant difference between treat-ments in all experiments.

FIG. 4. Electrophysiology of KATP channels and VDCCs. A, nor-malized peak KATP channel conductance. Cells cultured on collagen-coated slide glass were treated with or without 10 �M nifedipine for 24 hand then used for recording. Because the membrane area of each cellvaries, the ATP-sensitive conductance was normalized by dividing bythe membrane capacitance for each cell. m14, MIN6-m14; m14-Nif,nifedipine-treated MIN6-m14. n � 7–9, *, p � 0.05. B, current-voltagerelationships of VDCCs. Whole-cell Ba2� currents through VDCCs wererecorded. m14, MIN6-m14; m14-Nif, nifedipine-treated MIN6-m14. n �19–22. **, p � 0.01.

FIG. 5. RNA blotting of molecules involved in glucose metabo-lism (A) and KATP channels and VDCCs (B). Total RNA was isolatedfrom MIN6-m14 treated with or without 10 �M nifedipine for 24 h, and10 �g of aliquot was electrophoresed and transferred onto a nylonmembrane. The membrane was probed with 32P-labeled DNA frag-ments corresponding to cDNA indicated in the figure. Photographs of 18S RNA are shown to confirm equal loading. m14, MIN6-m14; m14-Nif,nifedipine-treated MIN6-m14.

Importance of [Ca2�]i in Glucose Responsiveness of �-Cells 25279

at KK

SA

NW

A K

AG

AK

U K

EN

on July 13, 2008 w

ww

.jbc.orgD

ownloaded from

malized by nifedipine (Fig. 2, B and C). These results suggestthat normalization of [Ca2�]i by nifedipine treatment improvesthe function of the KATP channels and/or the VDCCs, both ofwhich are impaired in control MIN6-m14 (15).

Effects of Normalization of [Ca2�]i on Glucose Metabolism inMIN6-m14—To evaluate the effect of normalization of [Ca2�]iby nifedipine on glucose metabolism in MIN6-m14, we meas-ured activity of enzymes involved in glucose metabolism andATP production. Differently than in normal �-cells, glucose-phosphorylating activity in MIN6-m14 is due mostly to HK, alow Km isoform of the glucose-phosphorylating enzyme (15).Normalization of [Ca2�]i by nifedipine did not alter either HKor GK activity (Fig. 3A). Lactate dehydrogenase activity, whichis increased in control MIN6-m14 as compared with MIN6-m9(14), was not altered by nifedipine (Fig. 3B). Furthermore,normalization of [Ca2�]i had no influence on ATP production inMIN6-m14 (Fig. 3C). These results demonstrate that normal-ization of [Ca2�]i by nifedipine does not change glucose metab-olism in MIN6-m14.

Electrophysiological Analyses of KATP Channels and VD-CCs—We then performed functional analyses of the KATP chan-nels and the VDCCs by patch clamp technique. Normalizationof [Ca2�]i by nifedipine significantly increased KATP channelconductance in MIN6-m14 (Fig. 4A). In addition, VDCC cur-rents at the whole cell level also were restored by normalizationof [Ca2�]i by nifedipine. Peak current was significantly greaterin nifedipine-treated MIN6-m14 than in control MIN6-m14

(Fig. 4B). These data show that both KATP channel and VDCCactivities are significantly improved in MIN6-m14 after nor-malization of [Ca2�]i.

Expressions of Various Genes Important in Glucose-inducedInsulin Secretion—Normalization of [Ca2�]i by nifedipine didnot alter mRNA expression of either GK or HK (Fig. 5A).mRNA levels of LDH were decreased after normalization of[Ca2�]i (Fig. 5A). mRNA expression of Kir6.2, the pore-formingsubunit of the �-cell KATP channel, was markedly increased bytreatment with nifedipine (Fig. 5B). SUR1, the regulatory sub-unit of KATP channels, showed a slight increase in mRNA levels(Fig. 5B). Expression of the �1-subunit of the VDCCs wasincreased strongly in MIN6-m14 after normalization of [Ca2�]i,but there was no significant difference in mRNA expression ofthe �3-subunit of VDCCs (Fig. 5B).

Subcellular Localization of the Subunits of KATP Channelsand VDCCs—We investigated the subcellular localization ofsubunits of KATP channels and VDCCs before and after nor-malization of [Ca2�]i. Subcellular fractionation was verified byimmunoblot analysis of Na�/K�-ATPase and sarco/endoplas-mic reticulum Ca2�-ATPase, markers for plasma membraneand internal membrane fractions, respectively. Normalizationof [Ca2�]i by nifedipine increased the levels of Kir6.2 in bothplasma membrane and internal membrane fractions signifi-cantly to a similar degree (Fig. 6B). SUR1 was somewhatincreased by treatment with nifedipine, but the differenceswere small (Fig. 6C). Expression level of the �1-subunit ofVDCCs was increased significantly both in plasma membraneand internal membrane fractions of MIN6-m14 after normal-ization of [Ca2�]i (Fig. 6D), while the expression level of the�3-subunit was unaffected by nifedipine (Fig. 6E). These re-sults indicate that the Kir6.2 subunit of KATP channels and the�1-subunit of VDCCs are increased at the protein level, while

FIG. 6. Immunoblotting of the subunits of KATP channels andVDCCs in subcellular fractions. Cells treated with or without 10 �M

nifedipine were fractionated into plasma membrane-enriched fraction(PM), internal membrane-enriched fraction (IM), and cytosol fraction(Cyt). Aliquots of each fraction (5 �g of protein equivalent to PM frac-tion) were subjected to SDS-PAGE, and blots were probed with anti-bodies indicated in the figures. A, verification of specificity for subcel-lular components. Antibodies to Na�/K�-ATPase to sarco/endoplasmicreticulum Ca2�-ATPase were used as markers for plasma and internalmembranes, respectively. B–E, quantification of KATP channel (B,Kir6.2; and C, SUR1) and VDCC (D, �1- and E, �3-) subunit proteins.Data form four separate experiments are illustrated, normalized to theintensity of the signals of PM fractions (arbitrary units; means � S.E.).Photographs are representative results. m14, MIN6-m14; m14-Nif,nifedipine-treated MIN6-m14. *, p � 0.05, **, p � 0.01.

FIG. 7. Effects of diazoxide in MIN6-m14cells (A–C) and BayK8644 in MIN6-m9 cells (D–F). MIN6-m14 cells (A–C) and MIN6-m9cells (D–F) were treated with and without 300 �M diazoxide (A–C) or 1�M Bay K8644 (D–F) for 24 h. A and D, glucose-induced insulin secre-tion. n � 4, *, p � 0.05, **, p � 0.01. B and E, [Ca2�]i change in responseto glucose. C, F, immunoblotting of Kir6.2 and �1-subunits in differentsubcellular fractions. m14, MIN6-m14; m14-Dzx, diazoxide-treatedMIN6-m14; and m9. MIN6-m9, m9-Bay K; Bay K8644-treatedMIN6-m9.

Importance of [Ca2�]i in Glucose Responsiveness of �-Cells25280

at KK

SA

NW

A K

AG

AK

U K

EN

on July 13, 2008 w

ww

.jbc.orgD

ownloaded from

membrane trafficking of these channel subunits is not affectedby [Ca2�]i.

Effects of Other Ca2� Modulating Agents on MIN6-m14 andMIN6-m9 Cells—To confirm that the improved glucose respon-siveness in MIN6-m14 after nifedipine treatment was due tothe normalization of [Ca2�]i, we measured the effect of diazox-ide, which activates the KATP channel and hyperpolarizes theplasma membrane, thereby inhibiting Ca2� influx through theVDCCs (20). Pre-exposure of MIN6-m14 to diazoxide (300 �M

for 24 h) normalized [Ca2�]i and restored the Ca2� response toglucose (Fig. 7B). At the same time, the glucose-induced insulinsecretion also was significantly improved (Fig. 7A). Expres-sions of the Kir6.2 and the �1-subunit were increased signifi-cantly by diazoxide in both plasma membrane and internalmembrane fractions in MIN6-m14 (Fig. 7C). We also examinedthe effects of Bay K8644 (1 �M), a Ca2� channel agonist, inglucose-responsive MIN6-m9. Bay K8644 increased [Ca2�]i inMIN6-m9 (227 � 4 and 339 � 2 nM, n � 6, before and afterapplication of Bay K8644, respectively). Glucose-induced insu-lin secretion, as well as the Ca2� response to glucose, wasseverely impaired after treatment with Bay K8644 in MIN6-m9(Fig. 7, D and E). Expressions of both the Kir6.2 and the�1-subunits were decreased by Bay K 8644 (Fig. 7F). Thesedata suggest that chronically high [Ca2�]i impairs glucose re-sponsiveness in MIN6 cells by decreasing the expression ofboth KATP channels and VDCCs.

DISCUSSION

We used two clonal pancreatic �-cell lines, MIN6-m9 andMIN6-m14, in the present study as models of glucose-respon-sive and glucose-unresponsive �-cells, respectively. We previ-ously demonstrated decreased activities of KATP channels andVDCCs in glucose-unresponsive MIN6-m14 (15). KATP chan-nels and VDCCs serve a crucial role in coupling glucose me-tabolism to exocytosis of the insulin granules in pancreatic�-cells (18). Here we show that MIN6-m14 exhibit elevated[Ca2�]i compared with glucose-responsive MIN6-m9, and thatnormalization of [Ca2�]i by the Ca2� channel blocker nifedipinerestores the activities of both KATP channels and VDCCs torepair glucose-induced insulin secretion in MIN6-m14. Itshould be noted that normalization of [Ca2�]i also was achievedby the KATP channel opener diazoxide with effects similar tothose of nifedipine. Moreover, the Ca2� channel agonist BayK8644 exerted effects contrary to the channel blocker in glu-cose-responsive MIN6-m9. These findings confirm that chron-ically elevated [Ca2�]i is the major cause of the glucose unre-sponsiveness in MIN6-m14.

The activities of KATP channels and VDCCs can be regulatedby various factors, including gene expression, phosphorylation/dephosphorylation, and trafficking to the plasma membrane(21–36). Our data show that normalization of [Ca2�]i increasesthe expression of both KATP channels and VDCCs. In particu-lar, Kir6.2, the pore-forming subunit of the KATP channels ofpancreatic �-cells (23), and the �1-subunit of the VDCCs, alsothe pore-forming subunit of the channel (24), were stronglyup-regulated at both the mRNA (Fig. 5) and protein levels (Fig.6). Contrariwise, Bay K8644, a Ca2� channel agonist, de-creased the expression of the Kir6.2 subunit of KATP channelsand the �1-subunit of VDCCs (Fig. 7F). Little is known of theregulation of these channel expressions, but it has been re-ported that high glucose leads to marked decreases in bothKir6.2 and SUR1 expression in isolated rat pancreatic islets aswell as in the INS-1 �-cell line (25). A decrease in Kir6.2expression also has been noted in pancreatic islets of Zuckerdiabetic fatty rats (26). Furthermore, mRNA expression of the�- and �-subunits of VDCCs are down-regulated in high glu-cose-infused rats, but diazoxide restores expression (27, 28).

mRNA levels of the �1-subunit of VDCCs also are reduced inpancreatic �-cells of Zucker diabetic fatty rats (29). Becausepersistent hyperglycemia causes sustained elevation of [Ca2�]iin the �-cells (8, 10), these observations may well be the resultof alteration in [Ca2�]i. Considering these findings together,[Ca2�]i most likely regulates the expression of KATP channelsand VDCCs at the transcriptional level. It recently has beenshown that intracellular membrane trafficking of ion channelsubunits is important in the functional expression of thesechannels at the cell surface (30–37). To evaluate the effect ofintracellular Ca2� on membrane trafficking of the channels, wemeasured the protein levels of the channel subunits in differentsubcellular fractions. However, the relative abundance of theKATP channel subunits and the VDCC subunits in plasmamembrane fraction and internal membrane fraction was simi-lar in MIN6-m14 before and after nifedipine treatment, indi-cating that membrane trafficking of these channel subunits isnot affected by [Ca2�]i.

MIN6-m14 exhibit abnormalities not only in glucose respon-siveness but also in glucose sensitivity. As in normal pancreaticislets, half-maximal insulin secretion occurs at 15 mM glucosein MIN6-m9, while the value is below 1 mM in MIN6-m14 cells(15). Although glucose responsiveness (intracellular Ca2� re-sponse to glucose and amount of secreted insulin) was dramat-ically improved by normalization of [Ca2�]i, glucose sensitivityremained unchanged (Figs. 2A and 7A). GK is a rate-limitingenzyme in glycolysis in �-cells and is thought to be a glucosesensor for glucose-induced insulin secretion (38, 39) because ithas a Km value higher than the physiological concentration ofglucose. However, MIN6-m14 predominantly expresses HK-I, alow Km isoform of the glucose-phosphorylating enzyme, and �90% of the activity occurs at 0.5 mM of glucose (15), whichmight well lead to abnormal glucose sensitivity. Normalizationof [Ca2�]i did not alter the expression of two isoforms of theenzyme (Fig. 5) or their activities (Fig. 3A), which might ac-count for the unchanged glucose sensitivity of MIN6-m14 be-fore and after normalization of [Ca2�]i by nifedipine. Takentogether with the findings on LDH activity and ATP productionin MIN6-m14 cells, glucose metabolism apparently is not influ-enced by changes in [Ca2�]i and is regulated independently ofKATP channel and VDCC activities.

In conclusion, chronically elevated [Ca2�]i induces glucoseunresponsiveness in MIN6-m14 by decreasing the expressionof both KATP channels and VDCCs at the protein level. Nor-malization of [Ca2�]i by nifedipine or diazoxide induces a glu-cose-responsive state from a glucose-unresponsive state by re-storing the functional expressions of these channels. Sinceabrogation of intracellular Ca2� homeostasis in pancreatic�-cells has been shown to be associated with impaired insulinsecretion (8, 10–13), the present study suggests normalizationof [Ca2�]i in pancreatic �-cells as a therapeutic strategy in thetreatment of the impaired glucose-induced insulin secretionseen in type 2 diabetes.

REFERENCES

1. Ghosh, A., and Greenberg, M. E. (1995) Science 268, 239–2472. Berridge, M. J. (1994) Mol. Cell. Endocrinol. 98, 119–1243. Bito, H., Deisseroth, K., and Tsien, R. W. (1997) Curr. Opin. Neurobiol. 7,

419–4294. Hardingham, G. E., and Bading, H. (1999) Microsc. Res. Tech. 46, 348–3555. Wollheim, C. B., and Sharp, G. W. (1981) Physiol. Rev. 61, 914–9736. Artalejo, C. R., Adams, M. E., and Fox, A. P. (1994) Nature 367, 72–767. Verkhratsky, A., and Toescu, E. C. (1998) Trends. Neurosci. 21, 2–78. Levy, J. (1999) Endocrine 10, 1–69. Prentki, M. (1996) Eur. J. Endocrinol. 134, 272–286

10. Bjorklund, A., Lansner, A., and Grill, V. E. (2000) Diabetes 49, 1840–184811. Okamoto Y., Ishida, H., Taminato T., Tsuji, K., Kurose, T., Tsuura, Y., Kato, S.,

Imura, H., and Seino, Y. (1992) Diabetes 41, 1555–156112. Tsuji, K., Taminato, T., Ishida, H., Okamoto, Y., Tsuura, Y., Kato, S., Kurose,

T., Okada, Y., Imura, H., and Seino, Y. (1993) Metabolism 42, 1424–142813. Kato, S., Ishida, H., Tsuura, Y., Tsuji, K., Nishimura, M., Horie, M., Taminato,

T., Ikehara, S., Odaka, H., Ikeda, I., Okada, Y., and Seino, Y. (1996) J. Clin.

Importance of [Ca2�]i in Glucose Responsiveness of �-Cells 25281

at KK

SA

NW

A K

AG

AK

U K

EN

on July 13, 2008 w

ww

.jbc.orgD

ownloaded from

Invest. 97, 2417–242514. Roe, M. W., Philipson, L. H., Frangakis, C. J., Kuznetsov, A., Mertz, R. J.,

Lancaster, M. E., Spencer, B., Worley, J. F., 3rd, and Dukes, I. D. (1994)J. Biol. Chem. 269, 18279–18282

15. Minami K., Yano, H., Miki, T., Nagashima, K., Wang, C. Z., Tanaka, H.,Miyazaki, J. I., and Seino, S. (2000) Am. J. Physiol. 279, E773–E781

16. Suzuki, M., Fujikura, K., Inagaki, N., Seino, S., and Takata, K. (1997) Diabetes46, 1440–1444

17. Kawaki, J., Nagashima, K., Tanaka, J., Miki, T., Miyazaki, M., Gonoi, T.,Mitsuhashi, N., Nakajima, N., Iwanaga, T., Yano, H., and Seino, S. (1999)Diabetes 48, 2001–2006

18. Seino, S. (1999) Annu. Rev. Physiol. 61, 337–36219. Panten, U., Zielmann, S., Schrader, M. T., and Lenzen, S. (1985) Naunyn

Schmiedebergs Arch. Pharmacol. 328, 351–35320. Malaisse, W. J., Pipeleers, D. G., and Mahy, M. (1973) Diabetologia 9, 1–521. Inagaki, N., Gonoi, T., Clement, J. P., 4th, Namba, N., Inazawa, J., Gonzalez,

G., Aguilar-Bryan, L., Seino, S., and Bryan, J. (1995) Science 270,1166–1170

22. Levitan, I. B. (1994) Annu. Rev. Physiol. 56, 193–21223. Beguin, P., Nagashima, K., Nishimura, M., Gonoi, T., and Seino, S. (1999)

EMBO J. 18, 4722–473224. Ertel, E. A., Campbell, K. P., Harpold, M. M., Hofmann, F., Mori Y., Perez-

Reyes, E., Schwartz, A., Snutch, T. P., Tanabe, T., Birnbaumer, L., Tsien,R. W., and Catterall, W. A. (2000) Neuron 25, 533–535

25. Moritz, W., Leech, C. A., Ferrer, J., and Habener, J. F. (2001) Endocrinology

142, 129–13826. Tokuyama, Y., Fan, Z., Furuta, H., Makielski, J. C., Polonsky, K. S., Bell, G. I.,

and Yano, H. (1996) Biochem. Biophys. Res. Commun. 220, 532–53827. Iwashima, Y., Pugh, W., Depaoli, A. M., Takeda, J., Seino, S., Bell, G. I., and

Polonsky, K. S. (1993) Diabetes 42, 948–95528. Iwashima, Y., Abiko, A., Ushikubi, F., Hata, A., Kaku, K., Sano, H., and Eto,

M. (2001) Biochem. Biophys. Res. Commun. 280, 923–93229. Tokuyama, Y., Sturis, J., DePaoli, A. M., Takeda, J., Stoffel, M., Tang, J., Sun,

X., Polonsky, K. S., and Bell, G. I. (1995) Diabetes 44, 1447–145730. Zerangue, N., Schwappach, B., Jan, Y. N., and Jan, L. Y. (1999) Neuron 22,

537–54831. Griffith, L. C. (2001) Curr. Biol. 11, R226–R22832. Gao, T., Chien, A. J., and Hosey, M. M. (1999) J. Biol. Chem. 274, 2137–214433. Bichet, D., Cornet, V., Geib, S., Carlier, E., Volsen, S., Hoshi, T., Mori, Y., and

De Waard, M. (2000) Neuron 25, 177–179034. Beguin, P., Nagashima, K., Gonoi, T., Shibasaki, T., Takahashi, K., Kashima,

Y., Ozaki, N., Geering, K., Iwanaga, T., and Seino, S. (2001) Nature 411,701–706

35. Prince, L. S., Workman, R. B., Jr., and Marchase, R. B. (1994) Proc. Natl. Acad.Sci. U. S. A. 91, 5192–5196

36. Ko, Y. H., Delannoy, M., and Pedersen, P. L. (1997) Biochemistry 36,5053–5064

37. Bradbury, N. A. (1999) Physiol. Rev. 79, S175–S19138. Newgard, C. B., and McGarry, J. D. (1995) Annu. Rev. Biochem. 64, 689–71939. Matschinsky, F. M. (1996) Diabetes 45, 223–241

Importance of [Ca2�]i in Glucose Responsiveness of �-Cells25282

at KK

SA

NW

A K

AG

AK

U K

EN

on July 13, 2008 w

ww

.jbc.orgD

ownloaded from