J Mol Cell Cardiol 34, 985ÿ996 (2002)

doi:10.1006/jmcc.2002.2035, available online at http://www.idealibrary.com on1

Depressed PKA Activity Contributes toImpaired SERCA Function and is Linked tothe Pathogenesis of Glucose-inducedCardiomyopathyKaushik Dutta1, Marybeth W. Carmody1, Steven E. Cala2 andAmy J. Davidoff1

1University of New England, College of Osteopathic Medicine, Biddeford, ME, USA;2Wayne State University, School of Medicine, Detroit, MI, USA

(Received 31 January 2002, accepted for publication 26 April 2002)

K. DUTTA, M. W. CARMODY, S. E. CALA AND A. J. DAVIDOFF. Depressed PKA Activity Contributes to Impaired SERCAFunction and is Linked to the Pathogenesis of Glucose-induced Cardiomyopathy. Journal of Molecular and CellularCardiology (2002) 34, 985ÿ996. We have previously described a cardiomyopathy induced by culturing ventricularmyocytes from normal adult rats in a medium containing high concentrations of glucose, which recapitulatescellular changes associated with early onset diabetic cardiomyopathy. This investigation was designed to evaluatecellular mechanisms that could contribute to slowed cytosolic Ca2� removal and myocyte relaxation in glucose-induced cardiomyopathy. Isolated ventricular myocytes were cultured overnight in medium containing normalglucose (n�5.5 mM) or high glucose (HG�25.5 mM). Cytosolic Ca2� removal was monitored with ¯uo-3 andmyocyte mechanics with video-edge detection. Electrically stimulated Ca2� transients were prolonged in HG cells(AT/PK�215�7 ms, n�41) compared to N myocytes (AT/PK�173�5 ms, n�34). By pharmacological and ionicmanipulations, Ca2� removal attributable to SERCA was slower in the HG group (AD/PK�290�17 ms, n�41)compared to N (AD/PK�219�10, n�34), whereas NCX function was similar in both groups of cells. Total PKAactivity was depressed in HG myocytes by 56% compared to N cells. b-adrenergic receptor stimulation with ISO(10ÿ7

M) normalized myocyte relaxation, Ca2� transients and PKA activity in HG myocytes. Furthermore, inhibition ofPKA with H89 (10ÿ5

M) depressed peak fractional shortening (PS) and slowed relengthening (AR/PK) to a greater extentin N (ÿ50% for PS and 92% for AR/PK) than in HG cells (ÿ25% for PS and 48% AR/PK). Depressed cytosolic Ca2� removalwas not, however, associated with changes in basal levels of phosphorylated PLB, nor levels of SERCA, NCX or PLBproteins. We conclude that cellular mechanisms associated with the early onset glucose-induced cardiomyocytedysfunction involves alterations in Ca2� regulation, which may be a common manifestation of other forms ofcardiomyopathies. # 2002 Elsevier Science Ltd. All rights reserved.

KEY WORDS: Hyperglycemia; Calcium transients; EÿC coupling; Diabetes

Introduction

Heart failure is the major cause of death in diabeticpatients.1 A contributing factor to heart failure inthese patients is the development of diabetic cardio-myopathy. Diabetic cardiomyopathy manifests asearly onset diastolic dysfunction and late onset

Please address all correspondence to: Amy J. Davidoff, PhD, UniversiTel: (207) 283-0170, x2824; E-mail: adavidoff@une.edu

0022ÿÿ2828/02/080985�12 $35.00/0

systolic dysfunction.2,3 Abnormal ventricular myo-cyte excitationÿcontraction (EÿC) coupling is thehallmark of diabetic cardiomyopathy, as well as anumber of other diseases including congestive heartfailure and remodeling after myocardial infarction.Although the etiologies may differ among varioustypes of cardiomyopathies, the end results appear to

ty of New England, 11 Hills Beach Rd., Biddeford, ME 04005, USA.

# 2002 Elsevier Science Ltd. All rights reserved.

986 K. Dutta et al.

be similar (e.g. prolonged action potentials (APs),slowed cytosolic Ca2� ¯uxes and slowed myocyteshortening and relengthening).

Diabetic cardiomyopathy has been extensivelycharacterized in animal models (reviewed in Ref. 4).Some of the molecular changes contributing tomyocyte dysfunction have been clearly establishedand include (in rat) depressed outward K� currents(particularly ITO), Na�/Ca2� exchange (NCX) andSR Ca2�uptake,5ÿ10 and changes in myosin isozymedistribution, myo®lament Ca2� sensitivity, andmitochondrial function (e.g. Ca2� ¯uxes and ATPgeneration).11ÿ13 Depressed b-AR signaling (e.g.reduced phosphorylation of phospholamban (PLB))and elevated PKC activity14ÿ17 also contribute toslowed removal of cytosolic Ca2� in cardiac myo-cytes isolated from diabetic animals.5,8,18 Functionalchanges of ion channels, pumps and exchangers mayoccur without changes in expression of these pro-teins, depending on the etiology and/or duration ofthe cardiomyopathy.5,19,20

Although a signi®cant amount is known about thecellular consequences in heart muscle of long-termdiabetes, little is known about the pathogenesis ofdiabetic cardiomyopathy. Recently it was shownthat there is a progressive change in the expression ofEÿC coupling proteins. For example, increases in PLBexpression precede decreases in sarco(endo)plasmicreticulum Ca2� ATPase (SERCA) and ryanodinereceptor (RYR) expression by several weeks.5 Cel-lular changes such as prolonged AP duration andimpaired myocyte mechanics appear to occur asearly as a few days after induction of diabetes.9,20,21

Thus, abnormal cardiomyocyte EÿC coupling isseen very early in the development of the disease.

Although the pathogenic factor(s) contributing tothis cardiomyopathy have not been clearly estab-lished, hyperglycemia appears to play a signi®cantrole.22 We have recently found that a diabetic-likecardiomyopathy (i.e. impaired EÿC coupling) isrecapitulated by culturing normal adult rat ventri-cular myocytes in a high glucose (HG) medium.22,23

After one day of culture in HG the most prominenteffects are slowed cytosolic Ca2� removal and pro-longed myocyte relengthening, which are not due toacute effects of glucose (,12 h) or increased osmo-larity.22 These results are consistent with the earlyonset of diastolic dysfunction in whole hearts. Ourprevious work revealed prolonged AP duration inthese HG-cultured myocytes. Prolonged AP alonewould be suf®cient to impair relaxation24ÿ26 inour HG cells, but other cellular changes includingdepressed SERCA and NCX, and/or depressed phos-phorylation of PLB may also be components of thephenotype.

In the present study we evaluated cytosolicCa2� removal by SERCA and NCX in ventricularmyocytes that were cultured overnight in HG(25.5 mM). We also evaluated the in¯uence ofcAMP-PKA signaling on intracellular Ca2� (Ca2�

i )transients and twitches, and measured PKA activity.Our results support the view that SERCA function isdepressed in HG myocytes because of lower basalPKA activity rather than changes in SERCA content.This study provides insight into cellular processesthat may precede overt changes in gene expression ofmacromolecules (e.g. SERCA, PLB and NCX) char-acteristic of abnormal EÿC coupling in diabeticcardiomyopathy.

Methods

Cell isolation and culture

Single ventricular myocytes were isolated enzy-matically from the hearts of adult male SpragueÿDawley rats (200ÿ250 g) using the methoddescribed previously.22 In brief isolated myocyteswere plated under sterile conditions on glass cover-slips pre-coated with laminin (10 mg/mL, Collabora-tive Biochemical Products, Bedford, MA, USA) andmaintained overnight in a de®ned medium consist-ing of Medium 199 (Sigma) with Earle's saltscontaining 25 mM HEPES and NaHCO3, supple-mented with albumin (2 mg/mL), L-carnitine(2 mM), creatine (5 mM), taurine (5 mM), glucose(5 mM), insulin (0.1 mM), penicillin (100 U/mL),streptomycin (100 mg/mL), and gentamicin(100 mg/mL). The medium also contained eithernormal concentrations of glucose (n�5.5 mM) orhigh glucose (HG�25.5 mM).

Measurement of Ca2�i transients

Myocytes attached to coverslips were placed in aLyden chamber (Harvard Apparatus, Holliston, MA,USA) and were loaded with membrane-permeant¯uo 3/AM (1 mM; Molecular Probes Inc. Eugene, OR,USA) in HEPES buffer (see below) containing 20%pluoronic acid for 30 min in the dark, at roomtemperature. The composition of the HEPES buffer(unless otherwise indicated) was (in mM): 131 NaCl,4 KCl, 1 CaCl2, 1 MgCl2, 10 glucose, 10 HEPES, atpH�7.4. The chamber was constantly perfused withHEPES buffer (�2 mL/min) and the cells were ®eldstimulated to contract at a frequency of 0.5 Hz(unless otherwise indicated). Cells were placed onan inverted microscope (Olympus IX 70) equipped

Figure 1 Representative traces of twitches (a) andintracellular Ca2� transients (b) recorded from adult ratventricular myocytes cultured�24 h in medium contain-ing either normal glucose (n�5.5 mM) or high glucose(HG�25.5 mM) concentrations. Relaxation (AR/PK) isdescribed by the area under relengthening (AR) normalizedby peak twitch amplitude (PK). The time course of cyto-solic Ca2� transients were recorded in ¯uo-3 loaded cellsand were described by the area under the entire transient(AT) normalized to PK. Steady-state twitches and Ca2�

transients were recorded in control buffer, while cells werestimulated at 0.5 Hz. Traces result from averaging �10steady-state signals. Peaks of traces were normalized toeach other to better illustrate differences in time courses ofdecay.

987Glucose-mediated Cardiomyopathy

with a heated (27�1�C) and light tight chamber,and imaged through a 20�UAPO water immersionobjective. Fluo 3-loaded myocytes were excited bymercury-arc lamp system at a 480 nm wavelengththrough an epi¯uorescence attachment (505 nmdichroic mirror; Chroma Tech. Corp., Brattleboro,VT, USA). Fluorescence (535 nm) was detected withphotomultiplier tube and recorded at a samplingrate of 1000 Hz (IonOptix Corporation, Milton, MA,USA). The intensity of the ¯uorescence at 535 nmincreases with an increase in Ca2�

i . The time coursesof the Ca2� transients were of interest (rather thanabsolute values of [Ca2�]), and ¯uorescent signalswere not calibrated. However, relative magnitudes ofCa2�

i were estimated by expressing the changes inpeak ¯uorescence amplitude normalized to baseline¯uorescence (F/F0).

Measurement of myocyte shortening

Mechanical properties of cultured ventricular myo-cytes were assessed by a video-based edge-detectionsystem (IonOptix Corp., Milton, MA, USA) as pre-viously reported.22,27 Shortening of rod-shapedmyocytes was detected at both longitudinal edgesat a video-sweep speed of 240 Hz, while beingcontinually superfused with HEPES buffer at 27�C.In certain experiments, cells were exposed to HEPESbuffer supplemented with either a b-AR agonist,isoproterenol (ISO�10ÿ7

M), a PKA inhibitor H89(10ÿ5

M), or H89� ISO. For experiments using H89,cells were pre-incubated in H89 for 30 min prior torecording.

The indices used to describe isotonic shorteninghave been previously reported22 [and illustrated inFig. 1(a)] include; peak fractional shortening (PS;peak shortening amplitude normalized to resting celllength) and area under the shortening phase normal-ized to peak shortening amplitude (AC/PK). The indexused to describe isotonic relengthening is area underthe relengthening phase normalized to peak ampli-tude (AR/PK). All indices of shortening/relengtheningwere determined after averaging �10 steady-statetwitches for each myocyte, and analyzed off-lineusing Clamp®t (Axon Instruments, Union City,CA, USA).

Rapid switching of superfusate

Pharmacological interventions were used to evalu-ate cytosolic Ca2� removal attributable to SERCA,NCX or slow processes (e.g. mitochondrial Ca2�

uptake), using a rapid switching device. This system

was comprised of a multi-barreled pipette that waspositioned close to the myocyte being recorded(within �300 mm). The pipette was made of 4 PEtubes converging to a single ef¯uent tube (dead spacewas estimated to be less than 10 mL). Electronicallycontrolled valves allowed for rapid switching of thegravity feed ef¯uent, with a single stream superfusingthe cell. All baseline recordings were made whilesuperfusing myocytes with control buffer (see recipeabove). The superfusate was then switched to isolateeither SR Ca2� uptake or Na�/Ca2� exchange (seeresults section for details).

Assessment of SERCA function

We have previously reported22,23 that electricallystimulated (ES) Ca2� transients decline slower inmyocytes cultured in HG compared to N [Fig. 1(b)].To evaluate SERCA mediated cytosolic Ca2� removal

Figure 2 Functional assessment of SERCA. (a) Electricallystimulated (ES) steady-state Ca2� transients were recordedin control buffer, then the extracellular solution wasrapidly changed to a Ca2�-free buffer (0 Ca2�) for 10 s.Once ES transients stopped, the stimulus was shut off. After10 s in 0 Ca2�, cells were superfused with a Na�- and Ca2�-free buffer (0 Na/0 Ca2�). After another 10 s, cells wererapidly superfused with a brief (100 ms) caffeine (10 mM)puff in 0 Na�/0 Ca2�. Under these conditions Na�/Ca2�

exchange is inhibited, caffeine releases SR Ca2� andcytosolic Ca2� removal is primarily due to SERCA. Hashmarks re¯ect a change in time scale during which solutionsare changed. (b) Representative traces of Ca2� transientsfrom N and HG cells after the short-caffeine exposure. Theintegration of the rising phase (As) and declining phase(AD) were calculated. (c) AS and AD were normalized topeak transient amplitude (F* ms/Fpeak). Summary data forAS/PK and AD/PK after the short-caffeine protocol. Open barsrepresent N and solid bars represent HG myocytes. Datarepresent mean� SE (n�34ÿ41 cells/group). * Indicatessigni®cantly different from n (P , 0.05).

988 K. Dutta et al.

we followed a protocol previously described by Yaoet al.28 For each myocyte, we recorded steady-stateCa2� transients during ES in control buffer, followedby a rapid change to a Ca2�-free solution for 10 s[see Fig. 2(a) for protocol]. The composition of theCa2�-free (0 Ca2�) solution was the HEPES bufferwithout added Ca2� plus EGTA (5 mM). The timerequired to remove all extracellular Ca2� was as-certained by cessation of ES transients. The ®eldstimulationwasthenshutoffandthe superfusatewaschanged to a Na�- and Ca2�-free (0 Na�/0 Ca2�)solution. Equimolar LiCl replaced NaCl in the 0 Na�/0 Ca2� buffer (no osmolarity compensation wasmade for CaCl2 or EGTA). If the superfusate waschanged simultaneously from control buffer to a

0 Na�/0 Ca2� solution, myocytes developed Ca2�

oscillations, presumably because extracellular Ca2�

was not immediately removed. Yao et al.29 haveshown that in rat myocytes (unlike other species)extracellular Ca2�may be trapped within the extra-cellular matrix, and is therefore more dif®cult toremove by rapid superfusion. We also observed thatwith our rapid switching device, it took �2 s toremove extracellular Ca2� from our myocytes (deter-mined when ES contractions were inhibited). Aftersuperfusing with 0 Na�/0 Ca2� solution for 10 s, themyocytes were rapidly exposed to 10 mM caffeine(dissolved in 0 Na�/0 Ca2� buffer), for 100 ms,followed by caffeine-free 0 Na�/0 Ca2� solution[Fig. 2(a)]. Using this protocol the peak caffeine-inducedSRCa2�transientalwaysoccurred>200 msfrom the onset of the short-caffeine exposure,providing suf®cient time to wash out the caffeineprior to the decline of the Ca2� transient [Fig. 2(b)].Thus, the decline of the Ca2� transient under thisshort-caffeine protocol re¯ects SERCA function (pre-dominantly), since the NCX was disabled in the0 Na�/0 Ca2� solution. Indices to evaluate caffeine-induced Ca2� release and uptake are described inthe results section.

Assessment of NCX function

To assess NCX function, we exposed myocytes to along caffeine protocol.28 In this protocol [Fig. 3(a)],each myocyte was paced at 0.5 Hz in control bufferuntil steady-state Ca2� transients were established.ES was then shut off for 10 s, followed by a rapid andsustained (>4 s) exposure to control buffer contain-ing 10 mM caffeine. Caffeine was then washed out bysuperfusing the myocyte with control buffer and ESwas restarted. Once steady-state Ca2� transientswere re-established, the protocol was repeatedusing the caffeine buffer supplemented with nickel(5 mM). Only myocytes with fully recovered EStransients were used in the analyses. Under theseconditions the decline of the caffeine-induced Ca2�

transient in control buffer was primarily dependenton the NCX and slow processes (i.e. mitochondrialCa2� uptake and SL Ca2�-ATPase), since the con-tinuous presence of caffeine prevents net SR Ca2�

accumulation.30 Exposure to caffeine plus Ni2�

blocks NCX and therefore, the decline of theCa2� transient is dependent on Ca2� removal bythe slow processes. As depicted by the representativetransients in Figure 3(a), the rate of Ca2�

i decline isfaster with caffeine alone than with caffeine�Ni2�.To quantify the time course for cytosolic Ca2�

removal during this long caffeine protocol,

Figure 3 Functional assessment of NCX. (a) ES steadystate Ca2� transients were recorded in control buffer,followed by a 10 s period without ES, cells were rapidlysuperfused with caffeine (10 mM) dissolved in control bufferfor >5 s. Under these conditions, Ca2� is constantly beingreleased by the SR, therefore only NCX and slow Ca2�

uptake processes (e.g. mitochondrial and SL pumps)contribute to cytosolic Ca2� removal (and not SERCA).The same myocyte was re-exposed to control bufferwithout caffeine, and ES steady-state Ca2� transients re-established. The long-caffeine protocol was repeated inpresence of nickel (Ni2�; 5 mM) in order to block NCX.Under these conditions both SERCA and NCX are blocked,therefore cytosolic Ca2� decline is due to only the slowprocesses. The difference between areas under the Ca2�

transient decline (4 s chosen arbitrarily) before and afternickel (�AT) represents NCX function. (b) Summary dataof �AT/PK (�AT normalized to peak ¯uorescence intensity)for N (open bar) and HG myocytes (solid bar). (c) Summarydata of peak Ca2� release expressed as F/F0. Data representmean� SE for 42ÿ46 cells/group.

989Glucose-mediated Cardiomyopathy

we measured the area under the Ca2� transient (AT)for 4 s starting at the rising phase. The differencebetween AT before and after application of Ni2�

represents NCX-dependent Ca2� ef¯ux (�AT).

Immunoblot assays and antibodies

Myocytes cultured in either N or HG medium werescraped and pelleted, then resuspended in 1 ml of a1% sodium dodecyl sulfate (SDS) solution and heatedfor 3 min at 70ÿ80�C. Cell homogenates in SDSwere collected and assayed for protein using the

method of Lowry. SDS-polyacrylamide gel electro-phoresis and immunoblotting were performed aspreviously described by Cala and Miles.31 Each gelcontained molecular weight markers (14ÿ200 kDa)(BioRad, Hercules, CA, USA). Proteins (80 mg/lane)separated by SDS-PAGE were transferred electro-phoretically onto nitrocellulose membranes(0.45 mm; BioRad) in 50 mM sodium phosphate(pH 7.2). The antibodies for the NCX were pur-chased from Swant (Bellinzona, Switzerland), anti-[16Ser-P]-PLB from Upstate Biotechnology (LakePlacid, NY, USA) and anti-[17Thr-P]-PLB were agift from Dr John Colyer (University of Leeds, UK).Monoclonal antibodies to SERCA2a (A7R5) and PLB(2D12) were provided by Dr Larry Jones (IndianaUniversity School of Medicine). Antigen-antibodycomplexes were detected using [125I] protein A (NENLife Science Products, Boston, MA) and autoradio-graphy using Kodak BioMax MS ®lm. To avoidcomplications associated with non-linearity ofX-ray ®lm, quanti®cation of radioactivity was doneby excising bands from nitrocellulose and gammacounting. Some ®lms were overexposed simply toillustrate low signals in Figure 4.

Protein kinase A activity

PKA activity was measured using a non-radioactive,ELISA based, Protein Kinase Assay Kit (Calbiochem,San Diego, CA, USA). Brie¯y, isolated myocytescultured �24 h in N or HG media were washed inice-cold phosphate buffered saline. The cells werepelleted, re-suspended and lysed in cold samplepreparation buffer (in mM); 50 Tris-HCl, 50b-mercaptoethanol, 10 EGTA, 5 EDTA, 1 PMSF,and 10 Benzamidine (at pH 7.5). The samples weresonicated, then 1.2 ml of the sample was centrifugedat 100 000� g for 1 h at 4�C. Protein concentrationof the supernatant was determined by the Lowrymethod. Approximately 2ÿ10 mg of protein was runin the PKA assay. The O.D. at 490 nm was read inmicroplate reader (Bio-Tek Instruments, Inc.Winooski, VT, USA). PKA activity was expressed asa ratio of PKA to protein content in each sample.

Statistical analyses

Data are presented as mean� SE. Statistical signi®-cance was determined for myocyte mechanics andCa2� transients by analysis of variance (SYSTAT).Follow-uptestsformultiplecomparisons(Bonferroni)were chosen depending on whether signi®cance(P , 0.05) was identi®ed in main effects and/orinteraction terms.

Figure 4 Representative immunoblots assessing protein content for SERCA, NCX and PLB in myocytes cultured for�24 h in N and HG media. (a) Cardiac myocytes cultured under N or HG conditions were scraped and pelleted at 4�C, thendissolved in 1% SDS-containing buffer for protein determination. Protein homogenates (80 mg/lane) were electrophoresedon gradient (5ÿ18% acrylamide) or 6.5% gels, transferred to nitrocellulose (0.45 mm), and visualized with Amido blackstain (left panel). Immunoblotting was done with antibodies to SERCA, NCX and PLB, and antigen-antibody complexesvisulaized [125I] protein A binding and autoradiography (right panel). Protein content for SERCA, NCX and PLB wasquanti®ed in ®ve separate pairs of cultures (three shown). No signi®cant differences in steady-state levels of these proteinswere found between N and HG cells. (b) Cell homogenates from N and HG cells were analyzed by SDS-PAGE in duplicatehalves, and the nitrocellulose transfer was cut into two pieces and immunoblotted with either anti-[16Ser-P]- or anti-[17Thr-P]-PLB antibodies. Visualization of antigen-antibody complexes was with [125I] protein A and autoradiography.(c) Representative sample of the extent of ISO (10ÿ7

M) stimulation of PLB phosphorylation (16Ser) compared to basal(Con) phosphorylation in HG myocytes. There were no signi®cant differences in either basal or ISO-stimulated PLBphosphorylation (on 16Ser or 17Thr) between N and HG myocytes (summary data not shown).

990 K. Dutta et al.

Results

Electrically stimulated twitches and Ca2� transients inN and HG myocytes

We have previously reported that culturing adult ratventricular myocytes for�24 h in a de®ned medium

containing a high concentration of glucose, prolongsrelengthening and slows cytosolic Ca2� removal,similar to that seen in myocytes isolated from diabeticanimals.20,22,23,27 The indices used to evaluatemyocyte mechanics (twitches) in this investigationare illustrated in Figure 1(a), and the summary dataare presented in Figures 5(a) and (b). All twitches

Figure 5 Selected indices of myocyte twitches and Ca2�

transients from myocytes cultured�24 h in N (open bars)and HG (solid bars) media. Steady-state (a) peak fractionalshortening (dPS), (b) myocyte relengthening (AR/PK) and(c) Ca2� transients (AT/PK) evoked by ES (at 0.5 Hz) andrecorded in control buffer (c), or buffer supplemented witheither isoproterenol (ISO; 10ÿ7

M), H89 (10ÿ5M) or

H89� ISO. Ca2� transients were recorded in ¯uo-3 loadedcells and myocyte twitches were recorded in ¯uo-3unloaded cells. Data represent mean� SE for 8ÿ29 cells/group/condition. * Indicates signi®cantly different from Nin the same buffer (P , 0.05).

991Glucose-mediated Cardiomyopathy

were recorded in cells that were not loaded with the¯uorescent probe, ¯uo 3/AM. Fluo 3 was used toevaluate the time course of intracellular Ca2�

transients in myocytes cultured in either N or HG

medium. The time course of the ¯uorescence signaldecay was described by the area under the transient,normalized to the peak ¯uorescent intensity (AT/PK).We chose this method of analysis because it describeschanges in cytosolic [Ca2�] during the entire con-tractile cycle without making the assumption thatcytosolic Ca2� removal follows a speci®c time course(e.g. a single exponential decay27). A larger AT/PK

indicates a slower decline in cytosolic [Ca2�]. Repre-sentative transients are shown in Figure 1(b) andsummary data in Figure 5(c). Ca2� transient decay,recorded during electrical stimulation (ES) in control(C) buffer, was signi®cantly slower in myocytescultured in HG medium compared to N medium.The following protocols were designed to determinewhether slowed cytosolic Ca2� removal in HG cells isattributable to slowed Ca2�uptake by SERCA and/orCa2� extrusion by NCX.

SERCA function in N and HG myocytes

As described in the methods section, Figure 2(a)illustrates the protocol used to evaluate SERCAfunction in the absence of Ca2�-induced Ca2� release(via electrical stimulation) or Ca2� ef¯ux throughNCX. Figure 2(b) depicts representative caffeine-induced Ca2� transients from N and HG myocytesevoked by the short-caffeine protocol. We dividedeach transient into a rising (or systolic) phase (AS),which corresponds to Ca2� release from the SR, and adeclining phase (AD), which corresponds to Ca2�

uptake by SERCA. It should be noted that SERCAfunctions throughout the release and uptake phases,such that SR Ca2� release and uptake cannot becompletely separated. This protocol was, however,designed to block Ca2� ef¯ux through NCX.28

Summary data for areas under the rising anddeclining phases were normalized to the peak amp-litude for each cell AS/PK and AD/PK, respectively[Fig. 2(c)]. Rate constants (� ) were not used to assessrise and fall of cytosolic Ca2�because the signals werenot well ®t with exponential functions. Integratingthe signal and normalizing it to peak ¯uorescentamplitude22 avoids the complicating factor thatincreased SR Ca2� release increases the rate ofcytosolic Ca2� removal, as previously described.32

AS/PK was similar between N and HG myocytes,suggesting that caffeine inducible SR Ca2� releasewas normal in HG cells. AD/PK was signi®cantly larger(i.e. slower) in HG myocytes than in N cells [Fig. 2(c)].Therefore, the slowed decline in cytosolic [Ca2�]during electrically stimulated transients is consistentwith slowed SERCA-medicated Ca2� removal.

992 K. Dutta et al.

NCX in N and HG myocytes

To determine whether slowed cytosolic Ca2�

removal in HG myocytes during ES is also attribut-able to impaired NCX function, we exposed myocytesto a long caffeine protocol as described in the methodssection. The integrated signal (AT) was used as ameasure of SR Ca2� release and cytosolic Ca2�

removal attributable to NCX and slow processes(e.g. mitochondrial Ca2� uptake and SL Ca2� extru-sion). SR Ca2� load and caffeine-releasable Ca2�

were assessed by comparing peak Ca2� signals(F/F0). NCX-dependent Ca2� ef¯ux was estimatedby comparing the difference between AT before andafter application of Ni2� (�AT). Ca2� removal attri-butable to NCX was the same in N and HG myocytes[Fig. 3(b)]. In addition, SR Ca2� load was similar inboth groups [Fig. 3(c)].

To further validate this protocol as a means toquantify NCX function, we determined the reprodu-cibility (and recovery) of the ES- and caffeine-inducedtransients. We found that after the long-caffeineexposure, ES-induced Ca2� transients recoveredafter �15ÿ20 s in control buffer. Although theshape of the Ca2� transient during the long-caffeineexposure differed amongcells, it was extremelyrepro-ducible within each cell. For example, upon re-establishing ES-induced Ca2� transients, subsequentlong-caffeine exposures produced the same timecourse and shape, multiple times (data not shown).

Measurements of Ca2� regulatory protein levels inN and HG myocytes

To determine whether reductions in steady-statelevels of critical Ca2� handling proteins account forabnormal mechanical and Ca2� transient data, wecarried out immunoblotting of SDS homogenatesfrom ®ve pairs of cultures (N vs HG) using antibodiesto SERCA2a, NCX, and PLB [Fig. 4(a)]. Steady-statelevels of SERCA2a, NCX and PLB were unchangedbetween N and HG myocytes (summary data notshown). Of note, we found two immunoreactivebands for NCX. While proteolysis cannot be com-pletely ruled out, a doublet was consistently seen inevery sample and no other immunoreactive productswere detected, suggesting a high degree of selectivityfor the antibody (data not shown). The upper band isclose to the predicted molecular weight from thecDNA sequence of NCX1 (106 kDa found in rat).A lower bis-acrylamide concentration (we used0.8%) could account for this shift in mobility,which differs from other reports of 120 kDa.

Because regulation of SERCA by PLB depends uponthe state of phosphorylation on 16Ser and 17Thr,33

changes in levels of phospho-PLB due to HG treat-ment might also account for depressed SERCA activ-ity. To investigate the possibility that basal levels ofphosphorylated PLB were altered in HG cells, phos-pho(16Ser)- and phospho(17Thr)-PLB levels weredetermined by immunoblotting [Fig. 4(b)]. Althoughthere was variability among cultures [Fig. 4(b)],there were no statistical differences in levels of PLBphosphorylation between N and HG treated cells(summary data not shown). Levels of basal phospho-PLB in quiescent myocytes appeared low comparedto total PLB based upon relative immunoreactivitiesusing the three different antibodies.

To explore whether the variability observedbetween N and HG phospho-PLB [Fig. 4(b)] wasphysiologically relevant or perhaps a consequence ofvery low signal levels, we exposed myocytes to ISO(10ÿ7

M) for 15 min, a concentration and timesuf®cient to stimulate maximal inotropic and lusi-tropic effects. Figure 4(c) shows a representative blotwhich illustrates the difference between basal (con)and ISO stimulated phospho-PLB. The extent ofISO-stimulated phosphorylation of PLB was substan-tially (�7-fold) greater than that occurring in basalstate. Thus, it is unlikely that the variability seen inphosphorylation of either the serine or threonineresidues with HG treatment re¯ected a physiologic-ally relevant change in PLB phosphorylation.

Effects of ISO and H89 on Ca2� removal andmyocyte mechanics

Steady-state myocyte mechanics and Ca2� trans-ients were recorded in control buffer (C) during ES at0.5 Hz. As previously reported22,34 and repeatedherewithin, HG myocytes show normal peak frac-tional peak shortening (PS), prolonged relaxation(AR/PK), and slowed Ca2�

i removal (AT/PK; Fig. 5).b-adrenergic receptor stimulation by isoproterenol(ISO; 10ÿ7

M) produced positive inotropic [i.e.increased PS, Fig. 5(a)] and lusitropic effects [i.e.decreased AR/PK, Fig. 5(b)], and faster Ca2�

i decline[i.e. decreased AT/PK, Fig. 5(c)] in both N and HG cells.Furthermore, ISO stimulation abolished the differ-ences between N and HG cells. Inhibiting basal PKAactivity by pre-incubating cells for 15 min with H89(10ÿ5

M) slowed both cytosolic Ca2� removal andmyocyte relengthening, and blunted peak shorten-ing in both N and HG myocytes. Differences betweenN and HG cells in control buffer were abolished withH89 [Figs 5(b) and (c)]. PS, which did not differ incontrol buffer, was signi®cantly lower in N cells thanHG myocytes with H89 [Fig. 5(a)]. Consistent amongall of these indices is that the inhibitory effect of H89

Table 1 Relative PKA activity in myocytes cultured�24 h in either N or HG media

Groups Total PKA activity(activity/mg

protein)

ISO stimulatedPKA activity

(activity/mg protein)

N 0.99�0.29 0.68�0.17HG 0.53�0.12 0.75�0.19HG/N 0.56�0.06 1.10�0.06

Data represent mean� SE from 3 or 4 cultures. HG/N ratios werecalculated for each culture before averaging.

993Glucose-mediated Cardiomyopathy

was more pronounced in N cells than in HG myo-cytes, suggesting that basal levels of PKA were higherin N cells. In other words, twitches and Ca2� trans-ients in HG myocytes are less dependent on PKA thanin N cells. H89 prevented ISO-induced effects on allindices (in both groups) when compared to ISO alone(i.e. prevented the inotropic, lusitropic, and Ca2�

removal effects). However, when compared to H89alone, there were some residual effects of ISO, evidentwith the H89� ISO treatment. For example, relax-ation and Ca2� decline were faster in both N and HGcells H89� ISO than with H89 alone, whereasPS was greater in N cells but unchanged in HGmyocytes.

PKA activity in N and HG myocytes

In order to determine whether cAMP-PKA signalingis altered in glucose-induced cardiomyopathy, PKAactivity was evaluated in N and HG cells (Table 1).The assay was qualitative and values were normal-ized to total protein content in each culture dish inorder to account for potential differences in cellnumber per dish. In four separate cultures, totalPKA activity was depressed by an average of56.4�6.1% in HG cells compared to N. The effectof ISO stimulation on PKA activity was measured inthree of these cultures. This was accomplished byincubating N and HG cells in ISO (10ÿ7

M) for 15 minprior to the assay. ISO normalized PKA activity inHG myocytes, such that it was 110.3�5.5% of Ncells incubated in ISO. PKA activity after incubationwith ISO represents submaximal stimulation of theenzyme (Table 1). These data are consistent with ourphysiological measurements presented in Figure 5,which indirectly indicate that basal PKA was greaterin N myocytes than in HG cells.

Discussion

The signi®cant ®ndings of this study are that glucose-induced cardiomyopathy, which includes slowed

cytosolic Ca2� removal and impaired myocyterelengthening, involves slowed SR Ca2� uptake,with normal NCX. It is likely that slowed SR Ca2�

uptake results from depressed PKA regulation ofSERCA rather than through depressed SERCAexpression. In addition, both expression and functionof NCX appear to be normal. These conclusions aresupported by the data showing that protein contentof SERCA, PLB and NCX are not affected by culturingmyocytes in HG (Fig. 4), that PKA activity is lower inHG myocytes (Table 1), and that manipulating PKA(i.e. either activating or inhibiting this pathway)normalizes myocyte mechanics and Ca2� transients(Fig. 5). For example, stimulating the b-AR systemwith ISO abbreviates Ca2� transients and relaxationtimes for both N and HG cells, and normalizes all EÿCcoupling parameters measured in HG cells (e.g.AT/PK, PS and AR/PK). Conversely, blocking PKAactivity with H89, blunts peak shortening and slowsCa2� transients and relaxation in both N and HGcells, however the effect is much greater in N cellsthan in HG cells (Fig. 5). Thus depressed basal PKAactivity with subsequent decreased SERCA activity,could account for slowed Ca2� removal and myocyterelengthening in HG cells. Taken together these datasupport the view that exposure to HG for a shortperiod (�24 h) produces changes in regulation ofEÿC coupling that precede overt changes in geneexpression observed in later stages of diabeticcardiomyopathy.

Depressed SERCA activity is common to variousmodels of diabetes,5,35,36 and could contribute toimpaired relaxation by slowing cytosolic Ca2�

removal. Regardless of the etiology or duration oftype 1 diabetes, SR Ca2� uptake appears to bedepressed in isolated SR vesicles and cytosolic Ca2�

decline is slower in intact myocytes.18,20,35 How-ever, the effects on gene expression and proteincontent of SERCA are quite variable and dependenton duration of disease.5,19,35 Thus, a functionalchange in SERCA may occur with or without anychanges in mRNA or protein levels. It should be notedthat these ®ndings are not consistent among allmodels of type 2 diabetes, including animals that areglucose intolerant, obese and insulin resistant.36ÿ38

Our data provide some evidence that SR Ca2�

release is not impaired in HG myocytes. For example,fractional Ca2� release can be estimated by calculat-ing the ratio of peak Ca2�amplitude during ES to thatduring maximal caffeine exposure (e.g. during theNCX protocol). This ratio was unaffected by culturingmyocytes in HG [Fig. 3(c)]. Furthermore, the indexAS/PK, measured using the protocol to assess SERCA,was not altered by HG [Fig. 2(c)]. One limitation ofthis interpretation is that in intact myocytes we

994 K. Dutta et al.

cannot be certain that SR Ca2� load is the same in Nand HG cells. Under our experimental conditions, thebest that can be said is that each cell was allowed todevelop steady-state conditions (i.e. at 0.5 Hz stimu-lus frequency) prior to each measurement.

b-adrenergic signaling and subsequent contrac-tile responses are depressed in late stages of dia-betic cardiomyopathy.18,39,40 In chronic diabetes,b-AR subtypes appear to be differentially down-regulated,41 and b-AR signaling is depressed.18 Inour acute model, b-AR signaling appears to be intactsince stimulation by ISO results in both positiveinotropic and lusitropic responses in HG cells, whichare comparable to (or exceed) responses of N myo-cytes. For example, impaired relaxation, prolongedCa2� transients and depressed PKA activity in HGcells are normalized with ISO stimulation (Fig. 5 andTable 1). This is consistent with observations that inearly stages of diabetic cardiomyopathy, hearts mayactually be hypersensitive to conditions that elevateCa2�

i .19 ISO stimulation has also been shown tonormalize EÿC coupling mechanisms in severalmodels of cardiomyopathy,42ÿ44 but not in allmodels (e.g. end stage heart failure44).

In the absence of changes in protein content ofSERCA or PLB in HG mediated cardiomyopathy, wefocused on regulation through cAMP-PKA depend-ent pathways. PLB is a key regulator for SERCAactivity. In its dephosphorylated state it serves as aninhibitor for SERCA activity; when phosphorylatedits inhibitory effect is removed, thus enhancing SRCa2�uptake. In a recent study it has been shown thatin the absence of b-AR stimulation, inhibition ofPKA with H89 depresses contraction in isolatedmyocytes.45 In the present study, the PKA inhibitorH89 decreased peak fractional shortening [Fig. 5(a)]and prolonged the duration of the twitches and Ca2�

transients [Figs 5(b) and (c)] to a greater extent inN cells than in HG myocytes. H89 also blunted (butdid not completely prevent) the response of ISOstimulation. We cannot exclude the possibilitiesthat H89 incompletely blocked PKA or that therewere PKA-independent effects of ISO.45ÿ47

A small decrease in basal PLB phosphorylation (inthe absence of b-AR stimulation) in HG cells couldunderlie the effects described in this study, and wouldbe consistent with the signi®cantly lower PKAactivity found in HG cells under basal conditions.However, no signi®cant differences in steady-statephosphorylation of PLB were found between Nand HG cells, as determined by immunoblotsusing [16SerÿÿP]- and [17ThrÿÿP]-PLB antibodies.To address the concern that our method of proteinisolation was inadequate to preserve the state of PLBphosphorylation, we exposed myocytes to ISO and

measured the extent of PLB phosphorylation. ISOconsistently stimulated the phosphorylation of PLBat the 16Ser residue, but not the 17Thr residue, in bothN and HG myocytes [Fig. 4(c)]. This may be predictedsince the phosphorylation of the 17Thr site alsodepends on cytosolic Ca2�.33 Cytosolic Ca2� wouldnot be expected to increase in ISO-stimulated quies-cent cells, in fact, it would most likely decreasebecause of augmented SR Ca2� uptake.48 This sug-gests that either PLB may not be the protein target ofbasal PKA regulation in these cells, or that theduration of our treatment (i.e. 24 h) is insuf®cientto alter the extent of steady-state PLB phosphory-lation.5 An alternative mechanism may involveselective changes in compartmentalization of PKA-dependent phosphorylation.49,50

NCX is involved in controlling Ca2�i by extruding

Ca2� from the cytosol and facilitating trigger releaseof Ca2� from the SR in cardiac myocytes (reviewed byRef. 51). The pathophysiological consequence ofchanges in mRNA, protein content and functionof the NCX are not consistent in various models ofdiabetes.6,7,19,38,52 In the present study we showthat Ca2� extrusion by NCX and protein levels forNCX are normal in HG cells (Figs 3 and 4). Contrary toour ®ndings, Schaffer et al.38 found that NCX activity(measured by Ca2� uptake in isolated SL vesicles)was depressed in neonatal rat ventricular myocytescultured in HG (30 mM). These discrepancies may notbe mutually exclusive for several reasons. First, theregulation of NCX function may vary betweenneonatal and adult cardiomyocytes. Second, ourmethod for assessing NCX was quite different in thatwe used myocytes exposed to prolonged applicationsof caffeine. Third, and perhaps most interesting, isthat Schaffer et al.38 cultured cells in HG withoutinsulin, whereas our culture media contained highconcentrations of insulin (10ÿ7

M). Insulin de®ciencyalone depressed NCX in neonatal cells and acuteincubation with insulin normalized NCX in both theirN and HG myocytes. Furthermore, NCX was alsodepressed in SL vesicles isolated from insulin resistant(glucose intolerant) rats (without changes in NCXmRNA). Understanding the interplay between glu-cose toxicity and insulin signaling may shed light onthe regulation of NCX in diabetic cardiomyopathy.

In summary, depressed SERCA function (but notNCX) is a contributing factor to impaired relaxationin glucose-mediated cardiomyopathy. Furthermore,alterations in cytosolic regulatory enzyme(s) (e.g.diminished basal PKA activity) rather than changesin expression of Ca2� regulating proteins (e.g. SERCAor NCX) may be critical in the development of diabeticcardiomyopathy. Evaluating the signaling pathwaysassociated with cellular changes that precede overt

995Glucose-mediated Cardiomyopathy

changes in gene expression may provide insight intothe pathogenesis of a number of cardiomyopathies.

Acknowledgements

This research was supported by grants from theNational Institutes of Health, Bethesda, MD (R01HL60303 and HL66895 to AJD). The authors wishto thank Dr Mary L. Schwanke for her critical reviewof this manuscript.

References

1. GRUNDY SM, BENJAMIN IJ, BURKE GL, CHAIT A, ECKEL RH,HOWARD BV, MITCH W, SMITH SC, SOWERS JR. Diabetesand cardiovascular disease: a statement for health-care professionals from the American Heart Asso-ciation. Circulation 1999; 100: 1134ÿ1146.

2. GALDERISI M, ANDERSON KM, WILSON PWF, LEVY D.Echocardiographic evidence for the existence of adistinct diabetic cardiomyopathy (the FraminghamHeart Study). Am J Cardiol 1991; 68: 85ÿ89.

3. ZARICH SW, NESTO RW. Diabetic cardiomyopathy.Curriculum Cardiol 1989; 118: 1000ÿ1012.

4. CHATHAM JC, FORDER JR, MCNEILL JH. The Heart inDiabetes. Norwell, MA: Kluwer Academic Publishers;1996.

5. ZHONG Y, AHMED S, GRUPP IL, MATLIB MA. Altered SRprotein expression associated with contractile dys-function in diabetic rat hearts. Am J Physiol 2001;281: H1137ÿH1147.

6. KASHIHARA H, SHI ZQ, YU JZ, MCNEILL JH, TIBBITS GF.Effects of diabetes and hypertension on myocardialNa�ÿCa2� exchange. Can J Physiol Pharmacol 2000;78: 12ÿ19.

7. CHATTOU S, DIACONO J, FEUVRAY D. Decrease insodiumÿcalcium exchange and calcium currents indiabetic rat ventricular myocytes. Acta Physiol Scanda1999; 166: 137ÿ144.

8. LAGADIC-GOSSMANN DL, BUCKLER KJ, LE PRIGENT K,FEUVRAY D. Altered Ca2� handling in ventricularmyocytes isolated from diabetic rats. Am J Physiol1996; 270: H1529ÿH1537.

9. SHIMONI Y, FIREK L, SEVERSON D, GILES W. Short-termdiabetes alters K� currents in rat ventricular myo-cytes. Circ Res 1994; 74: 620ÿ628.

10. JOURDON P, FEUVRAY D. Calcium and potassiumcurrents in ventricular myocytes isolated fromdiabetic rats. J Physiol 1993; 470: 411ÿ429.

11. FLARSHEIM CE, GRUPP IL, MATLIB MA. Mitochondrialdysfunction accompanies diastolic dysfunction indiabetic rat heart. Am J Physiol 1996; 271:H192ÿH202.

12. HOFMANN PA, MENON V, GANNAWAY KF. Effects ofdiabetes on isometric tension as a function of [Ca2�]and pH in rat skinned cardiac myocytes. Am J Physiol1995; 269: H1656ÿH1663.

13. PIERCE GN, DHALLA NS. Heart mitochondrial functionin chronic experimental diabetes in rats. Can J Cardiol1985; 1: 48ÿ54.

14. KOYA D, KING GL. Protein kinase C activation andthe development of diabetic complications. Diabetes1998; 47: 859ÿ866.

15. GILES TD, OUYANG J, KERUT EK, GIVEN MB, ALLEN GE,MCILWAIN EF, GREENBERG SS. Changes in proteinkinase C in early cardiomyopathy and in gracilismuscle in the BB/Wor diabetic rat. Am J Physiol(Heart Circ Physiol) 1998; H295ÿH307.

16. MALHOTRA A, GEENEN DL, REICH D, BUTTRICK PM.Cardiac myocytes from diabetic animals displaypersistant translocation of protein kinase C epsilonand phosphorylation of troponin I: role of angio-tensin receptor blockade. Circ Res 1997; 81:1027ÿ1033.

17. XIANG H, MCNEILL JH. Protein kinase C activity isaltered in diabetic rat hearts. Biochem Biophys ResComm 1992; 187: 703ÿ710.

18. TAMADA A, HATTORI Y, HOUZEN H, YAMADA Y, SAKUMA I,KITABATAKE A, KANNO M. Effects of beta-adrenoceptorstimulation on contractility, [Ca2�]i, and Ca2�

current in diabetic rat cardiomyocytes. Am J Physiol1998; 274: H1849ÿH1857.

19. GOLFMAN L, DIXON IMC, TAKEDA N, CHAPMAN D,DHALLA NS. Differential changes in cardiac myo®-brillar and sarcoplasmic reticular gene expression inalloxan-induced diabetes. Mol Cell Biochem 1999;200: 15ÿ25.

20. REN J, DAVIDOFF AJ. Diabetes rapidly induces con-tractile dysfunctions in isolated ventricular myocytes.Am J Physiol 1997; 272: H148ÿH158.

21. SHIMONI Y, EWART HS, SEVERSON D. Type I and IImodels of diabetes produce different modi®cations ofK� currents in rat heart: role of insulin. J Physiol1998; 507: 485ÿ496.

22. REN J, GINTANT GA, MILLER RE, DAVIDOFF AJ. Highextracellular glucose impairs cardiac EÿC couplingin a glycosylation-dependent manner. Am J Physiol1997; 273: H2876ÿH2883.

23. DAVIDOFF AJ, REN J. Low insulin and high glucoseinduce abnormal relaxation in cultured adult ratventricular myocytes. Am J Physiol 1997; 272:H159ÿH167.

24. DIPLA K, MATTIELLO JA, MARGULIES KB, JEEVANANDAM V,HOUSER SR. The sarcoplasmic reticulum and theNa�/Ca2� exchanger both contribute to the Ca2�

transient of failing human ventricular myocytes. CircRes 1999; 84: 435ÿ444.

25. WICKENDEN AD, KAPRIELIAN R, KASSIRI Z, TSOPORIS JN,TSUSHIMA R, FISHMAN GI, BACKX PH. The role of actionpotential prolongation and altered intracellular cal-cium handling in the pathogenesis of heart failure.Cardiovasc Res 1998; 37: 312ÿ323.

26. SAH R, RAMIREZ RJ, BACKX PH. Modulation of Ca2�

release in cardiac myocytes by changes in repolar-ization rates. Role of phase-1 action potentialrepolarization in excitationÿcontraction coupling.Circ Res 2002; 90: 165ÿ173.

27. DUTTA K, PODOLIN DA, DAVIDSON MB, DAVIDOFF AJ.Cardiomyocyte dysfunction in sucrose-fed rats isassociated with insulin resistance. Diabetes 2001;50: 1186ÿ1192.

28. YAO A, SU Z, NOAKA A, ZUBAIR I, LU L, PHILIPSON KD,BRIDGE JHB, BARRY WH. Effects of overexpression ofthe Na�ÿCa2� exchanger on [Ca2�]i transients inmurine ventricular myocytes. Circ Res 1998; 82:657ÿ665.

996 K. Dutta et al.

29. YAO A, SPITZER KW, ITO N, ZANIBONI M,LORELL BH, BARRY WH. The restriction of diffusionof cations at the external surface of cardiac myocytesvaries between species. Cell Calcium 1997; 22:431ÿ438.

30. MCCALL E, GINSBURG KS, BASSANI RA, SHANNON TR,QI M, SAMAREL AM, BERS DM. Ca ¯ux, contractility,and excitationÿcontraction coupling in hypertroph-ic rat ventricular myocytes. Am J Physiol 1998; 274:H1348ÿH1360.

31. CALA SE, MILES K. Phosphorylation of the cardiacisoform of calsequestrin in cultured rat myotubesand rat skeletal muscle. Biochim Biophys Acta 1992;1118: 277ÿ287.

32. BERS DM, BERLIN JR. Kinetics of [Ca]i decline incardiac myocytes depend on [Ca]i. Am J Physiol1995; 268: C271ÿC277.

33. WEGENER AD, SIMMERMAN HK, LINDEMANN JP, JONES LR.Phospholamban phosphorylation in intact ventri-cles. Phosphorylation of serine 16 and threonine 17in response to beta-adrenergic stimulation. J BiolChem 1989; 264: 11468ÿ11474.

34. REN J, DOMINGUEZ LJ, SOWERS JR, DAVIDOFF AJ. Metfor-min but not glyburide prevents high glucose-inducedabnormalities in relaxation and intracellular Ca2�

transients in adult rat ventricular myocytes. Diabetes1999; 48: 2059ÿ2065.

35. PIERCE GN, RUSSELL JC. Regulation of intracellularCa2� in the heart during diabetes. Cardiovasc Res1997; 34: 41ÿ47.

36. SCHAFFER SW. Cardiomyopathy associated withnoninsulin-dependent diabetes. Mol Cell Biochem1991; 107: 1ÿ20.

37. MISRA T, GILCHRIST JSC, RUSSELL JC, PIERCE GN. Cardiacmyo®brillar and sarcoplasmic reticulum functionare not depressed in insulin-resistant JCR:LA-cprats. Am J Physiol 1999; 45: H1811ÿH1817.

38. SCHAFFER SW, BALLARD-CROFT C, BOERTH S, ALLO SN.Mechanisms underlying depressed Na�/Ca2�

exchanger activity in the diabetic heart. CardiovascRes 1997; 34: 129ÿ136.

39. GANDO S, HATTORI Y, AKAISHI Y, NISHIHIRA J, KANNO M.Impaired contractile response to beta adrenoceptorstimulation in diabetic rat hearts: alterations in betaadrenoceptors-G protein-adenylate cyclase systemand phospholamban phosphorylation. JPET 1997;282: 475ÿ484.

40. KASHIWAGI A, NISHIO Y, SAEKI Y, KIDA Y, KODAMA M,SHIGETA Y. Plasma membrane-speci®c de®ciencyin cardiac b-adrenergic receptor in streptozocin-diabetic rats. Am J Physiol 1989; 257: E127ÿE132.

41. DINCË ER UD, BIDASEE KR, GUÈ NER S, TAY A, OÈ ZCË ELIKAY AT,ALTAN VM. The effect of diabetes on expression of

b1-, b2-, and b3-adrenoreceptors in rat hearts.Diabetes 2001; 50: 455ÿ461.

42. LITWIN SE, ZHANG D, BRIDGE JHB. DyssynchronousCa2� sparks in myocytes from infarcted hearts. CircRes 2000; 87: 1040ÿ1047.

43. LIM CC, LIAO R, VARMA N, APSTEIN CS. Impairedlusitropy-frequency in the aging mouse: role of Ca2�-handling proteins and effects of isoproterenol. Am JPhysiol (Heart Circ Physiol 46) 1999; 274:H2083ÿH2090.

44. GOÂ MEZ AM, VALDIVIA HH, CHENG H, LEDERER MR,SANTANA LF, CANNELL MB, MCCUNE SA, ALTSCHULD RA,LEDERER WJ. Defective excitationÿcontraction coup-ling in experimental cardiac hypertrophy and heartfailure. Science 1997; 276: 800ÿ806.

45. HUSSAIN M, DRAGO GA, BHOGAL M, COLYER J.Effects of the protein kinase A inhibitor H-89 onCa2� regulation in isolated ferret ventricular myo-cytes. P¯uÈgers ArchÿÿEur J Physiol 1999; 437:529ÿ537.

46. YATANI A, TAJIMA Y, GREEN SA. Coupling of beta-adrenergic receptors to cardiac L-type Ca2� chan-nels: preferential coupling of the beta1 versus beta2receptor subtype and evidence for PKA-independentactivation of the channel. Cell Signal 1999; 11:337ÿ342.

47. KUSCHEL M, ZHOU YY, SPURGEON HA, BARTEL S,KARCZEWSKI P, ZHANG SJ, KRAUSE EG, LAKATTA EG,XIAO RP. b2-adrenergic cAMP signaling is uncoupledfrom phosphorylation of cytoplasmic proteins incanine heart. Circulation 1999; 99: 2458ÿ2465.

48. YU Z, QUAMME GA, MCNEILL JH. Depressed [Ca2�]i

responses to isoproterenol and cAMP in isolatedcardiomyocytes from experimental diabetic rats. Am JPhysiol 1994; 266: H2334ÿH2342.

49. ZAKHARY DR, MORAVEC CS, STEWARD RW, BOND M.Protein kinase A (PKA)-dependent troponin-I phos-phorylation and PKA regulatory subunits aredecreased in human dilated cardiomyopathy. Circu-lation 1999; 99: 505ÿ510.

50. KENTISH JC, MCCLOSKEY DT, LAYLAND J, PALMER S,LEIDEN JM, MARTIN AF, SOLARO RJ. Phosphorylation oftroponin I by protein kinase A accelerates relaxationand crossbridge cycle kinetics in mouse ventricularmuscle. Circ Res 2001; 88: 1059ÿ1065.

51. BLAUSTEIN MP, LEDERER WJ. Sodium/Calciumexchange: Its physiological implications. PhysiolReviews 1999; 79: 763ÿ854.

52. HATTORI Y, MATSUDA N, KIMURA J, ISHITANI T,TAMADA A, GANDO SKO, KANNO M. Diminishedfunction and expression of the cardiac Na�ÿCa2�

exchanger in diabetic rats: implication in Ca2�

overload. J Physiol 2000; 527: 85ÿ94.