72 (2006) 292–302www.elsevier.com/locate/cardiores

Cardiovascular Research

Adenosine A2A receptors are expressed in human atrial myocytes andmodulate spontaneous sarcoplasmic reticulum calcium release

Leif Hove-Madsen a,⁎,1, Cristina Prat-Vidal a,1, Anna Llach a, Francisco Ciruela b,Vicent Casadó b, Carme Lluis b, Antoni Bayes-Genis a, Juan Cinca a, Rafael Franco b

a Cell Physiology Laboratory, Cardiology Department, Hospital de la Santa Creu i Sant Pau, Institut Catalá de Ciencies Cardiovasculars, UniversitatAutònoma de Barcelona, St Antoni M

a

Claret 167, 08025 Barcelona, Spainb Department of Biochemistry and Molecular Biology, Universidad de Barcelona, Av. Diagonal 645, 08028 Barcelona, Spain

Received 4 April 2006; received in revised form 25 July 2006; accepted 26 July 2006Available online 29 July 2006

Time for primary review 21 days

by guest on April 2, 2016

Dow

nloaded from

Abstract

Background: Alterations in the cyclic AMP-dependent regulation of the cardiac ryanodine receptor (RyR2) have been proposed to account forincreased spontaneous calcium release from the sarcoplasmic reticulum (SR) in patients with heart failure, ventricular tachyarrhythmias and atrialfibrillation. While the adenosine A2A receptor (A2AR) is known to regulate cyclic AMP levels, expression and function of this receptor in humancardiac myocytes has not been investigated.Methods: PCR, western blotting and immunofluorescence were used to identify the A2AR, and functional effects of A2AR stimulation weremeasured with confocal calcium imaging and patch-clamp technique.Results: The A2AR is expressed in the human right atrium and distributed in a banded pattern along the Z-lines, overlapping with theryanodine receptor. A2AR stimulation caused a protein kinase A dependent increase in spontaneous SR calcium release in isolated humanatrial myocytes. The A2AR agonist CGS21680 increased the frequency of calcium sparks from 0.12±0.03 to 0.31±0.08 sparks·μm min−1

( pb0.05) and calcium waves from 0.65±0.31 to 5.11±1.84 waves·min−1 ( pb0.03). Moreover, spontaneous Na–Ca exchange currents(INCX) increased from 1.19±0.17 to 2.50±0.42 min−1 ( pb0.001). In contrast, CGS21680 did not alter caffeine inducible calcium release(6.98±0.52 vs. 6.82±0.57 amol pF−1, p=0.6) or the spontaneous INCX amplitude (0.32±0.05 vs. 0.29±0.04 pA pF−1, p=0.2). Current–voltage relationship and amplitude of the L-type calcium current (1.62±0.18 vs. 1.80±0.18 pA pF−1) were not altered, but calcium releasedependent inactivation was faster with CGS21680 (13.4±0.7 vs. 15.8±1.0 ms, pb0.001).Conclusions: Adenosine A2A receptors are expressed in the human atrial myocardium and modulate the frequency of spontaneous calcium releasefrom the SR.© 2006 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.

Keywords: Adenosine; Arrhythmia (mechanisms); Calcium channel; e–c coupling; SR (function)

1. Introduction

Abnormal calcium release from the SR has been associatedwith heart failure [1–3] and with the genesis of after-depolarization induced triggered arrhythmias [4–8]. Two

⁎ Corresponding author. Tel.: +34 935565620; fax: +34 935565603.E-mail address: lhove@santpau.es (Leif Hove-Madsen).

1 These authors contributed equally to this work.

0008-6363/$ - see front matter © 2006 European Society of Cardiology. Publishdoi:10.1016/j.cardiores.2006.07.020

main mechanisms have been proposed to explain abnormalcalcium release from the SR: mutations and hyperphosphor-ylation of the SR calcium release channel (ryanodine receptor).

Several mutations in the SR calcium release channel havebeen shown to increase its opening probability in planar lipidbilayers [5,9,10] and these mutations are found in patientswith polymorphic ventricular tachycardia [11–13]. Hyper-phosphorylation of the ryanodine receptor mediated by cyclicAMP (cAMP)-dependent protein kinase A (PKA) has been

ed by Elsevier B.V. All rights reserved.

293L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

reported in patients with heart failure, exercise-inducedventricular arrhythmias, or atrial fibrillation [2,10,14].

It is therefore conceivable that pharmacological control ofreceptor mediated cAMP-dependent activation of the ryano-dine receptor may be a means to prevent or reduce abnormalcalcium release from the SR. In this regard, antagonists of theGs-protein coupled β-adrenergic receptor have been shown torestore SR calcium release channel function and improvecardiac muscle performance in human heart failure [15].Theoretically, other Gs-protein coupled receptors, such as theadenosine A2A receptor, may also exert a cAMP-dependentmodulation of intracellular calcium handling. Indeed, theA2AR has been shown to cause cAMP dependent inotropiceffects in some studies [16–18] and indirect pharmacologicaldata suggest that the A2AR is expressed in rat [16] and guineapig [19] ventricular myocytes. In contrast, other authors didnot find effects of A2AR stimulation [20], and A2AR mRNAwas not found in porcine cardiomyocytes [21].

The aim of the present study was to determine whether theA2AR is expressed in the human atrial myocardium and toinvestigate if this receptor modulates two main players in intra-cellular calcium handling: the L-type calcium current (ICa) andcalcium release from the sarcoplasmic reticulum through theryanodine receptor. Our results reveal the presence of adenosineA2A receptors in human right atrium and show that agoniststimulation of the A2AR induces a PKA-mediated increase inspontaneous SR calcium release in human atrial myocytes.

2. Methods

2.1. Human samples

Atrial myocardial tissue samples were obtained from 72patients undergoing cardiac surgery. The average age was 64±2 years and 60% were male. All patients were in sinus rhythmand subjected to single or combined surgical interventions:coronary artery bypass surgery (28 patients), aortic valve re-placement (45 patients), mitral valve replacement (18 patients),other surgical procedures (10 patients). Sixteen patients receivedangiotensin converting enzyme inhibitors, 12 beta-blockers, 16diuretics, and 18 statins. Patients treated with calcium channelantagonists were excluded from the study. Specimen wasobtained from the right atrial appendage just prior to atrialcannulation for cardioplumonary bypass. Although the atrialtissue samples consisted of tissue that would normally be dis-carded during surgery, permission to be used in this study wasobtained from each patient. The study was approved by theEthical Committee of our institution and the investigationconforms with the principles outlined in the Declaration ofHelsinki. Patients treatedwith calcium channel antagonists wereexcluded from the study.

2.2. Myocyte isolation

Myocyteswere isolated fromhuman atrial tissue samples aspreviously described [6]. Briefly, small pieces of tissue were

incubated at 35 °C in aCa2+ free solution containing 0.5mg/mlcollagenase (Worthington type 2, 318 U/mg), and 0.5 mg/mlproteinase (Sigma type XXIV, 11 U/mg solid). After 45 min,the tissue was removed from the enzyme solution, and cellswere disgregated in Ca2+-free solution with a Pasteur pipette.The remaining tissue was digested for 3×15 min in a freshcalcium free solution containing 0.4 mg/ml collagenase. Onlyelongated cells with clear cross striations and without gra-nulation were used for experiments.

2.3. Spontaneous calcium release in isolated myocytes

Calcium sparks and calciumwaves were detected in fluo-3loaded cells using a laser-scanning confocal microscope(Leica TCS SP2 AOBS, Germany) as previously described[6]. The experimental solution contained (in mM): NaCl 136,KCl 4, NaH2PO4 0.33, NaHCO3 4, CaCl2 2, MgCl2 1.6,HEPES 10, Glucose 5, pyruvic acid 5, (pH=7.4). Fluores-cence emission was collected between 500 and 650 nm withthe excitation at 488 nm attenuated to 1–3%. Calcium sparksand calcium waves were measured at resting conditions.Calcium sparks were detected as an increase in the signalmass of a 3 μm section through the center of a calcium spark,without any detectable increase in an adjacent 3 μm section[6]. An increase in the signal mass in two or more adjacent3 μm sections were counted as calciumwaves. The amplitudeof each calcium spark, and its half-life were determined froman exponential fit of the decaying phase of the calcium sparktransient. The calcium spark frequency was determined foreach cell and normalized to the scanned cell length.

2.4. Patch-clamp

The transient inward Na–Ca exchange current associatedwith calcium waves was recorded in the perforated patchconfiguration using a software controlled patch-clamp am-plifier (EPC 10, HEKA, Germany). The pipette resistancewas 1.5–4 MΩ. Experiments were carried out at roomtemperature and began when the access resistance wasstable and had decreased to less than 5 times the pipetteresistance. The extracellular solution contained (in mM):NaCl 127, TEA 5, HEPES 10, NaHCO3 4, NaH2PO4 0.33,glucose 10, pyruvic acid 5, CaCl2 2, MgCl2 1.8, (pH=7.4).The pipette solution contained (in mM): aspartic acid 109,CsCl 47, Mg2ATP 3, MgCl2 1, Na2 phosphocreatine 5,Li2GTP 0.42, HEPES 10 and 250 μg/ml amphotericin B,(pH=7.2). The A2AR agonist CGS21680 and the proteinkinase A inhibitor H-89 were dissolved in DMSO and keptas 10 mM stock solutions. Ryanodine was kept as a 100 mMstock solution. CGS21680 was used at a final concentrationof 200 nm. At this concentration, CGS21680 selectivelystimulates the A2AR [22] and induces a robust stimulation ofcell shortening in rat ventricular myocytes [16,23]. L-typecalcium current amplitude and inactivation were measuredat steady state, using repetitive depolarizations from −80 to0 mV for 200 ms at a frequency of 0.5 Hz.

294 L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

The caffeine releasable SR calcium content was measuredby transiently exposing cells to 10 mM caffeine. The timeintegral of the resulting Na–Ca exchange current wasconverted to amoles (10−18 mol) of calcium released fromthe SR assuming a stoichiometry of 3Na+:1Ca2+ for the Na–Ca exchanger [6].

2.5. Total RNA extraction and PCR

Total RNA was isolated from 15 human right atrial tissuesamples using the QuickPrep Total RNA Extraction Kit(Amersham, Freiburg, Germany) according to the manufac-turer's instructions. First-strand cDNA was synthesized from1 μg of total RNA using a random hexamer primer andSuperScript II Rnase H− Reverse Transcriptase (Gibco BRL,Gaithersburg, MD, U.S.A.) according to the manufacturer'sprotocol. cDNAwas amplified using Taq DNA polymerase andthe following primers: for the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH), FGAPDH (5′-GCGGGGCTCTCCAGAACATCAT-3′) and RGAPDH (5′-GGTGGTCCAGGGGTCTTACTCC-3′) and for the human A2AR(hA2A), FhA2A (5′-GGCTGCCCCTACACATCATCAACT-3′) and RhA2A (5′-TGGGCCAGGGGGTCATCT-3′).

2.6. Gel electrophoresis and immunoblotting

Membranes from human right atrial tissue or from tran-siently transfected HEK cells were treated with SDS-PAGEsample buffer (8 M urea, 2% SDS, 100 mM DTT, 375 mMTris, pH 6.8) by heating at 37 °C during 2 h and resolved bySDS-polyacrylamide gel electrophoresis in 10% gels. Pro-teins were transferred to PVDF membranes using a semi-drytransfer system and immunoblotted with the indicated anti-body and then horseradish-peroxidase (HRP)-conjugatedgoat anti-rabbit IgG (1/60,000). The immunoreactive bandswere developed using a chemiluminescent detection kit(SuperSignal, West Pico Chemiluminescent substrate,Pierce).

2.7. Immunostaining

Human atrial tissue samples used for immunohistochemis-try were embedded in OCT and frozen in liquid nitrogen-cooled isopentane. Eight micrometer sections were cut on acryostat cooled to −18 °C. Sections were collected ontoSuperFrost Plus (BDH) slides, air dried and stored at −70 °C.For immunofluorescence, sections were blocked for 30 min in10% donkey serum in Tris-buffered saline (TBS) (150 mMNaCl/50 mM Tris–HCl, pH 7.5). Slides were incubated withaffinity purified anti-A2AR (VC21, 25 μg/ml) for one hour atroom temperature and then washed three times for 10 min inTBS. Alexa Fluor® 488 conjugated goat anti-mouse IgG andTexas Red®-conjugated goat anti-rabbit IgG were applied inblocking solution at a dilution of 1:2000. Sections were rinsedand mounted with Vectashield immunofluorescence medium(Vector Laboratories Inc., Burlingame, CA, U.S.A.).

Freshly isolated human cardiac myocytes used for immu-nocytochemistry were plated in 24 mm glass coverslipscoated with poly-D-lysine. After adhesion, cells were rinsedin phosphate-buffered saline and were fixed with 2%paraformaldehyde in PBS for 20 min at room temperature.Cells were washed with PBS and were incubated with 0.1 Mglycine for 5 min to quench the aldehyde groups. Afterwashing with PBS, myocytes were permeabilized with 0.2%Triton-X-100 in PBS for 15 min. Cells were then washedwith PBS, blocked for 30 min with 10% horse serum at roomtemperature and incubated with rabbit polyclonal anti-adenosine A2A receptor (10 μg/ml), and either mouse anti-α-actinin (1:500 dilution), mouse anti-myosin (1:500dilution), mouse anti-connexin-43 (1:150 dilution), ormouse anti-ryanodine receptor (1:500 dilution) for onehour at room temperature, washed three times with PBS, andstained with anti-mouse Alexa Fluor 488 antibody and anti-rabbit Texas Red. Finally cells were rinsed again for threetimes and mounted with Vectashield immunofluorescencemedium (Vector Laboratories Inc.).

For these studies a Leica TCS 4D confocal laser scanningmicroscope was used (Leica Lasertechnik GmbH, Heidelberg,Germany).

2.8. Data analysis and statistics

Ca2+ sparks and Ca2+ waves recorded in cells from thesame patient were averaged. Unless otherwise stated,average values from each patient were used for statisticalanalysis and expressed as mean±s.e.m. Data sets were testedfor normality. Student's t-test was used to assess significantdifferences when testing a specific effect. ANOVAwas usedfor comparison of multiple effects and Student–Newman–Keuls post test was used to evaluate the significance ofspecific effects. Differences were considered statisticallysignificant when pb0.05.

3. Results

3.1. Expression and localization of A2AR in human rightatrium

Fig. 1a and b show that mRNA coding for A2AR waspresent in right atrium and that both monomeric anddimeric species of the receptor are expressed. The A2ARwas expressed in a transversal-banded pattern along themyocardial fiber. A2AR immunostaining was specific sinceit disappeared when the primary A2AR antibody wasomitted or when the A2AR antibody was pre-incubated withthe peptide used to raise it (Fig. 1c). The A2AR co-distributed with the cytoskeletal-associated protein α-ac-tinin at the level of the Z line in the sarcomer and itsdistribution overlapped with that of the ryanodine receptor(Fig. 2). In contrast, A2AR did not co-localize with the H-band marker myosin or with the gap junction proteinconnexin-43.

Fig. 1. Expression and localization of A2AR in human right atrium. (a)Human A2AR (hA2AR) and glyceraldehyde-3-phosphate dehydrogenase(hGAPDH) total mRNA expression with hA2AR primers in the absence ofany cDNA (negative control, C−), in HUVEC cells (positive control, C+)and in human atrial tissue (Atrium). An experiment representative of sixdifferent individuals is shown. (b) Immunoblotting with an anti-A2ARantibody in cell membranes from transiently transfected HEK cells withhuman A2AR (positive control, C+) and human atrial membranes. (c)Cryosections from human right atrium were stained with anti-A2ARantibody. Immunofluorescent images with and without anti-A2AR antibodyare shown in the left and right panels respectively.

295L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

3.2. A2AR stimulation and L-type calcium current

To determine if the A2AR modulates L-type calcium cur-rent, ICa amplitude and inactivation kinetics were measuredbefore, during, and after exposure of right atrial myocytes to200 nm of the selective A2AR agonist CGS21680 (Fig. 3a).Fig. 3b shows that a 10 min exposure to CGS21680 had noeffect on the ICa amplitude and Fig. 3c reveals that the A2ARagonist significantly shortened the fast, calcium release-dependent, ICa inactivation time constant (tau) from 15.8±1.0 to 13.4±0.7 ms (pb0.001, n=27). Inhibition of calciumrelease from the SR with 400 μM ryanodine (not shown)increased tau from 13.2±0.9 to 23.5±1.6 ms (n=7,pb0.001) confirming that tau is SR calcium release de-pendent. Moreover, CGS21680 did not alter tau when addedin the presence of ryanodine (23.4±1.4 ms, p=0.92, n=7).Fig. 3d shows the effect of CGS21680 on the current–voltage relationship for ICa.

3.3. A2AR activation and spontaneous SR calcium release

To examine if A2AR agonists modulate SR calcium releasethe effect of CGS21680 on spontaneous SR calcium releasein isolated human right atrial myocytes was analyzed. Bothlocal non-propagating calcium release (calcium sparks) andpropagating calcium release from the SR (calcium waves)were analyzed. To determine whether A2AR stimulationmodifies SR calcium loading or its rate of refilling, weanalyzed respectively the amplitude and the decay of 254

calcium sparks from 14 patients. The left panel of Fig. 4ashows a fluo-4 loaded atrial myocyte with the scan lineindicated in red. The middle panel shows calcium sparksrecorded with this scan line and the right hand panel showsthe corresponding changes in the intracellular [Ca2+] for thecalcium sparks at the black arrows. Fig. 4b shows that neitherthe calcium spark amplitude (left panel) nor the decay of thecalcium sparks (right panel) were altered by incubation withCGS21680. Fig. 5a shows consecutive line-scan images froma control cell and a cell preincubated with 200 nm CGS21680and Fig. 5b shows that this A2AR agonist significantlyincreased both the number of calcium sparks (left panel) andcalcium waves (right panel), corresponding to an increase inthe calcium wave frequency from 0.65±0.31 to 5.11±1.84waves min−1.

To determine whether the observed effect of CGS21680on spontaneous SR calcium release is secondary to an effecton the membrane potential, the patch-clamp technique wasused to hold the membrane potential at −80 mV. This voltageclamping did, however, not prevent CGS21680 fromreversibly increasing the number of spontaneous calciumwaves measured as the Na–Ca exchange current (INCX)activated by them (Fig. 6a). The right hand panel shows thatthere was no difference in the amplitude of the spontaneousINCX. Fig. 6b shows the time course of the stimulatory effectof CGS21680 and the right hand panel summarizes thestimulatory effect of CGS21680 on INCX (pb0.001, n=27).The effect of CGS21680 was reversed by the selective A2ARantagonist ZM241385 (Fig. 6c). Assessment of the SRcalcium loading by rapid exposure of myocytes to 10 mMcaffeine at a holding potential of −80 mV gave comparableSR calcium loading before and after agonist activation of theA2AR (7.0±0.5 amol pF− 1 and 6.8±0.6 amol pF− 1

respectively (p=0.62, n=16, paired t-test).Together, these results suggest that the observed effects

of CGS21680 are due to a direct effect on the ryanodinereceptor. To test this further, the effect of CGS21680 onspontaneous INCX and the caffeine releasable SR calciumcontent was examined in the presence of tetracaine andryanodine, two compounds known to inhibit calciumrelease through the ryanodine receptor. Tetracaine abolishedspontaneous INCX and moreover, exposure to CGS21680 inthe presence of tetracaine did not induce any spontaneousINCX. Washing of tetracaine induced a transient overshoot inINCX amplitude and frequency (see supplementary figurepanel A). This was observed in cells from 5 patientsexposed to tetracaine. Application of 400 μM ryanodine tocells from 7 patients irreversibly abolished spontaneousINCX, which did not reappear in the presence of CGS21680.Moreover, ryanodine prevented 10 mM caffeine fromreleasing calcium from the SR with and withoutCGS21680, suggesting that caffeine could not overcomethe inhibitory effect of ryanodine. In contrast, the timeintegral of the caffeine induced INCX was increased 3-foldin the presence of 250 μM tetracaine, showing thattetracaine is able to inhibit spontaneous INCX in spite of a

Fig. 2. Distribution of the A2AR in isolated human atrial myocytes. Isolated human atrial myocytes were stained with anti-A2AR antibody, and with either anti-myosin antibody, anti-α-actinin antibody, anti-connexin-43, or anti-ryanodine receptor antibody. Superimposed images (merge) and magnification of the areawithin the white rectangles revealed an overlapping distribution of A2AR with α-actinin and with the ryanodine receptor (RyR) but not with connexin-43 ormyosin. Scale bars, 10 μm (merge) and 5 μm (magnification).

296 L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

3-fold higher SR calcium load. Simultaneous exposure ofthe cell to tetracaine and CGS21680 did not further increasethe caffeine induced INCX (see supplementary figure panelB and C). These findings further support that A2AR-stimulation increases spontaneous calcium release throughthe ryanodine receptor.

3.4. Protein kinase A dependency of A2AR-mediatedspontaneous calcium release

Since the A2AR is coupled to Gs proteins [24], agonist-mediated receptor activation should increase cAMP levelsand this in turn could activate protein kinase A (PKA)-dependent phosphorylation of phospholamban and/or theryanodine receptor. To test whether A2AR exerts its effect

through PKA activation, the selective PKA inhibitor H-89was employed. As shown in the representative current tracesof Fig. 7a, H-89 eliminated the stimulatory effect ofCGS21680 in isolated myocytes. Fig. 7b summarizes theaverage effect of CGS21680 on spontaneous calcium releasein the absence and the presence of H-89. In accordance withthese results, confocal calcium imaging showed that the sparkfrequency was significantly lower in myocytes incubatedwith H-89+CGS21680 than in myocytes incubated withCGS21680 alone (0.018 ± 0.012 versus 0.754 ± 0.260sparks·μm−1 min−1, p=0.01, n=5). Moreover, spontaneouscalcium waves were abolished in myocytes incubated withCGS21680+H-89. These results suggest that the A2AR-dependent stimulation of spontaneous SR calcium release isin fact mediated by PKA activation.

297L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

4. Discussion

This study establishes for the first time that the A2AR isexpressed in the human atrial myocardium and demonstratesthat this receptor modulates calcium release from the SR inisolated human right atrial myocytes.

4.1. Expression of the A2AR in human right atrium

PCR-technique identified the mRNA of the A2AR andwestern blotting confirmed the presence of both monomericand homodimeric forms of the A2AR in the human atrium.Our results confirm previous indirect evidence for thepresence of the A2AR in mammalian ventricle [16–19,23].The striated pattern of the A2AR at the level of the Z-line andits co-distribution with α-actinin may be explained by ascaffolding action of α-actinin because α-actinin interactsdirectly with the cytoplasmatic C-terminal region of humanA2AR and plays a role in the traffic of this receptor [25]. Theobserved overlap in the distribution of the A2AR and theryanodine receptor and the positive coupling of A2AR withadenylate cyclase reported in rat ventricular myocytes[18,19], also suggest that an A2AR mediated cAMP-dependent modulation of the ryanodine receptor is feasible.

4.2. Function of the A2AR on L-type calcium current

To address potential functional effects of the A2AR onintracellular calcium handling, we first determined the effectof the A2AR agonist CGS21680 on ICa amplitude. Our resultsshowed no significant effect of CGS21680 on ICa amplitude.

Fig. 3. Stimulation of the A2AR does not alter L-type calcium current amplitude. (0 mV before (control), during (200 nM CGS), and after (wash) exposure of a huamplitude (p=0.2, n=27). (c) Fast time constant for ICa inactivation with CGS21voltage relationship for the ICa amplitude with (CGS) and without (CON) A2AR s

A previous study failed to find functional effects of A2ARstimulation in a mammalian ventricular preparation [20]whereas others have shown that A2AR stimulation induces apositive inotropic effect [16–18]. The observation of apositive inotropic effect of CGS21680 in the absence of aconcurrent stimulation of ICa as observed here, are notmutually exclusive as long as another calcium source, such asthe SR, is activated by CGS21680.

4.3. Function of the A2AR on SR calcium release

In this study we found that A2AR stimulation increasesspontaneous calcium release from the SR and reduces the fasttime constant for ICa inactivation. Since a reduction in the fasttime constant for ICa inactivation is observed with larger SRcalcium release [26] our results suggest that A2AR stimula-tion also triggers calcium induced calcium release from theSR. The stimulatory action of CGS21680 on SR calciumrelease may occur at the ryanodine receptor or it may besecondary to an effect of CGS21680 on the SR calciumcontent or on the ICa amplitude, since both of these factorshave been shown to regulate SR calcium release [27–30]. Ourresults suggest that CGS21680 acts on the ryanodine receptorsince neither the amplitude of spontaneous INCX, the caffeinereleasable SR calcium content, nor the ICa amplitude werealtered by CGS21680. Moreover, a stimulatory effect of thiscompound on SR calcium uptake or SR calcium loadingwould be expected to speed up the decay and increase theamplitude of calcium spark transients respectively [28,29].However, CGS21680 elevated the frequency of calciumsparks without changing their amplitude or decay. Finally, the

a) L-type calcium currents elicited by a 200 ms depolarization from −80 toman atrial myocyte to CGS21680. (b) Average effect of CGS21680 on ICa680 (CGS) and without (CON). ⁎⁎⁎ denote pb0.001, n=27. (d) Current–timulation.

Fig. 5. A2AR activation increases spontaneous SR calcium release. (a) Five consecutive line-scan images from a myocyte incubated with the A2AR agonistCGS21680 and without (Control) showing local increases in the intracellular [Ca2+] reflected by a more intense green color. Changes in intracellular calcium in3 μm sections at the level of the blue and red arrows are shown below the confocal images in the corresponding colors. The scale bars on the left indicate a 1-unitincrease in the F/F0 ratio. (b) CGS2160 (CGS) significantly increased the frequency of calcium sparks (left panel) and calcium waves (right panel) in right atrialmyocytes from 14 patients. Significant differences between control and CGS21680 are indicated with ⁎ (pb0.05) and ⁎⁎ (pb0.01).

Fig. 4. Calcium spark amplitude and kinetics are not changed by A2AR stimulation. (a) Bidimensional confocal image of a fluo-3 loaded human atrial myocyte withindication of the scan line used to obtain the line-scan image in the middle panel. The right panel shows the changes in the fluorescence intensity (F/F0) of twocalcium sparks occurring at the level of the two black arrows. (b) Agonist stimulation of the A2ARwith CGS21680 had no effect on calcium spark amplitude (F/F0) oron the time constant (tau) for the decay of calcium sparks (p=0.36 respectively p=0.08, n=14).

298 L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

Fig. 6. Reversibility of A2AR mediated SR calcium release. (a) Spontaneous INCX were recorded with the membrane potential clamped at −80 mV before(control), during (CGS) and after (wash) exposure of a myocardiocyte to 200 nM CGS21680. Vertical scale bars correspond to 100 pA. The right hand panelshows the average INCX amplitude in the presence (CGS) and the absence (CON) of CGS21680 (p=0.16, n=22). (b) Time course of the stimulatory effect ofCGS21680 on the number of spontaneous Na–Ca exchange currents (counted as events on each 30 s sweep). The thick line above bars indicates the period ofCGS21680 application. The effect of CGS on spontaneous INCX frequency is shown on the right (⁎⁎⁎ denotes pb0.001, n=27). (c) The A2AR antagonistZM241385 (1 μM) was able to revert the stimulatory effect of 200 nM CGS21680 on spontaneous INCX. Vertical scale bars correspond to 100 pA. Averagevalues of 5 cells from 3 patients are shown on the right. Spontaneous INCX frequency was significantly higher (pb0.05, asterisk) with CGS-treatment than incontrol (CON) or with CGS+ZM.

299L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

effect of CGS21680 was similar in cells with the restingpotential clamped at −80 mV and in unclamped cells,discarding that A2AR mediated effects are secondary tochanges in the resting membrane potential, another importantmodulator of the spontaneous SR calcium release [28,31].

Inositol-triphosphate (IP3) receptors are also expressed inmammalian atrial myocytes and co-localize with theryanodine receptor [32]. Indeed, a cross-talk between theIP3 receptor and the ryanodine receptor has been shown tomodulate spontaneous calcium release from the SR throughthe ryanodine receptor in mammalian atrial myocytes [33–35]. Theoretically, the stimulatory effect of CGS21680 onspontaneous calcium release could therefore be due to aneffect of this compound on IP3-dependent calcium release.However, CGS21680 has been shown to inhibit IP3 pro-duction [36]. Moreover, PKA-dependent phosphorylation ofthe IP3 receptor has been shown to reduce type-3 IP3

dependent calcium release [37] and the type-2 IP3 receptor,which predominates in the heart, is a poor substrate for PKA-dependent phosphorylation [38]. Finally, we found thatinhibition of SR calcium release with 400 μM ryanodine or250 μM tetracaine abolished spontaneous INCX as well as theeffects of CGS21680 on spontaneous INCX and fast ICainactivation, in spite of a three-fold larger SR calcium loadwith 250 μM tetracaine. Taken together, this suggests thatactivation of the A2AR with CGS21680 stimulates theryanodine receptor itself.

4.4. Protein kinase A dependent activation of SR calciumrelease

The A2AR is positively coupled to adenyate cyclasethrough Gs proteins and their stimulation has been reported toincrease cAMP levels in rat ventricular myocytes [18,19].

300 L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

Stimulation of SR calcium release by the A2AR agonistCGS21680 could therefore result from a PKA-dependentstimulation of the ryanodine receptor in a manner similar tothat reported for Gs protein-coupled β-adrenergic receptors[3,15] in patients with failing hearts. In these patients, a PKA-dependent phosphorylation of the ryanodine receptor hasbeen proposed to cause an increased calcium leak from theSR [2,15]. While this proposal is controversial [39,40] ourfinding that H-89 abolished the stimulatory effect ofCGS21680 on spontaneous SR calcium release does confirma positive coupling between the A2AR, PKA, and the ryano-dine receptor in human atrial myocytes. Moreover, ourfindings suggest that the A2AR mediates a PKA-dependentregulation of the ryanodine receptor itself, since A2AR stimu-lation had no effect on ICa amplitude, spontaneous INCXamplitude, decay of calcium sparks, or the amount of calciumthat could be released from the SR by a rapid caffeineapplication.

Fig. 7. A2AR mediated SR calcium release is PKA-dependent. (a) Representative c(CGS) or CGS21680+H-89. (b) Addition of 10 μMH-89 abolished the stimulatoryon ranks showed a significant effect of the treatment on the wave frequency (p=0.00control (⁎), for CGS vs. control (⁎⁎), and for CGS vs. CGS+H-89 (⁎⁎⁎).

The lack of effect of CGS21680 on SR calcium loadingand calcium spark decay is also in accordance with a studyshowing that A2AR stimulation in guinea pig ventricle in-creases the cellular cAMP content but not phospholambanphosphorylation [19]. Furthermore, there is growing evi-dence that cAMP-mediated pathways are locally controlled[41–43] and a local A2AR-mediated control of cAMP couldexplain why CGS21680 stimulates SR calcium release with-out affecting SR calcium loading, calcium spark decay, or ICaamplitude.

In view of the reported PKA-dependent hyperphosphor-ylation of the ryanodine receptor in heart failure [2,3,15] andin catecholaminergic polymorphic ventricular tachycardia[4], it will be important to determine whether the A2AR isexpressed and modulates spontaneous SR calcium release inthe diseased ventricular myocardium.

In summary, we show that the adenosine A2A receptor isexpressed in the human atrial myocardium and has the

urrent traces recorded before (control) and during application of CGS21680effect of 200 nM CGS21680 in 6 patients. Kruskal–Wallis one way ANOVA2) and Student–Newman–Keuls post test gave a pb0.05 for CGS+H-89 vs.

by guest on April 2, 2016

Dow

nloaded from

301L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

by guest on April 2, 2016

Dow

nloaded from

potential to modulate SR calcium release in human atrialmyocytes. Therefore, pharmacological control of A2AR acti-vation may offer a novel approach to control calcium releasefrom the SR.

Acknowledgements

This work was supported by Ramon y Cajal Grants to L.Hove-Madsen and F. Ciruela, and grants from the SpanishMinistry of Science and Technology (SAF2001-3474), Redde investigación cooperativa de las enfermedades cardiovas-culares del Instituto de Salud Carlos III (C03/01), FundacióMutua Universal, and Sociedad Española de Cardiología.

Appendix A. Supplementary data

Supplementary data associatedwith this article can be found,in the online version, at doi:10.1016/j.cardiores.2006.07.020.

References

[1] Pogwizd SM, Qi M, Yuan W, Samarel AM, Bers DM. Upregulation ofNa(+)/Ca(2+) exchanger expression and function in an arrhythmogenicrabbit model of heart failure. Circ Res 1999;85:1009–19.

[2] Marx SO, Reiken S, Hisamatsu Y, Jayaraman T, Burkhoff D, RosemblitN, et al. PKA phosphorylation dissociates FKBP12.6 from the calciumrelease channel (ryanodine receptor): defective regulation in failinghearts. Cell 2000;101:365–76.

[3] Reiken S, Gaburjakova M, Guatimosim S, Gomez AM, D'Armiento J,Burkhoff D, et al. Protein kinase A phosphorylation of the cardiaccalcium release channel (ryanodine receptor) in normal and failinghearts. Role of phosphatases and response to isoproterenol. J Biol Chem2003;278:444–53.

[4] WehrensXH, Lehnart SE, Huang F, Vest JA, Reiken SR,Mohler PJ, et al.FKBP12.6 deficiency and defective calcium release channel (ryanodinereceptor) function linked to exercise-induced sudden cardiac death. Cell2003;113:829–40.

[5] Jiang D, Xiao B, Yang D, Wang R, Choi P, Zhang L, et al. RyR2mutations linked to ventricular tachycardia and sudden death reducethe threshold for store-overload-induced Ca2+ release (SOICR). ProcNatl Acad Sci U S A 2004;101:13062–7.

[6] Hove-Madsen L, Llach A, Bayes-Genis A, Roura S, Rodriguez Font E,Aris A, et al. Atrial fibrillation is associated with increased spontaneouscalcium release from the sarcoplasmic reticulum in human atrial myocytes.Circulation 2004;110:1358–63.

[7] Schlotthauer K, Bers DM. Sarcoplasmic reticulum Ca(2+) releasecauses myocyte depolarization. Underlying mechanism and thresholdfor triggered action potentials. Circ Res 2000;87:774–80.

[8] Burashnikov A, Antzelevitch C. Reinduction of atrial fibrillationimmediately after termination of the arrhythmia is mediated by latephase 3 early afterdepolarization-induced triggered activity. Circula-tion 2003;107:2355–60.

[9] Jiang D, Xiao B, Zhang L, Chen SR. Enhanced basal activity of a cardiacCa2+ release channel (ryanodine receptor) mutant associated withventricular tachycardia and sudden death. Circ Res 2002;91:218–25.

[10] Lehnart SE, Wehrens XH, Laitinen PJ, Reiken SR, Deng SX, Cheng Z,et al. Sudden death in familial polymorphic ventricular tachycardiaassociated with calcium release channel (ryanodine receptor) leak.Circulation 2004;109:3208–14.

[11] Postma AV, Denjoy I, Hoorntje TM, Lupoglazoff JM, Da Costa A,Sebillon P, et al. Absence of calsequestrin 2 causes severe forms ofcatecholaminergic polymorphic ventricular tachycardia. Circ Res2002;91:e21–6.

[12] Priori SG, Napolitano C,MemmiM, Colombi B, Drago F, GaspariniM,et al. Clinical and molecular characterization of patients with cate-cholaminergic polymorphic ventricular tachycardia. Circulation2002;106:69–74.

[13] Laitinen PJ, BrownKM, Piippo K, SwanH,Devaney JM,Brahmbhatt B,et al. Mutations of the cardiac ryanodine receptor (RyR2) gene in familialpolymorphic ventricular tachycardia. Circulation 2001;103:485–90.

[14] Vest JA, Wehrens XH, Reiken SR, Lehnart SE, Dobrev D, Chandra P,et al. Defective cardiac ryanodine receptor regulation during atrialfibrillation. Circulation 2005;111:2025–32.

[15] Reiken S, Wehrens XH, Vest JA, Barbone A, Klotz S, Mancini D, et al.Beta-blockers restore calcium release channel function and improvecardiac muscle performance in human heart failure. Circulation2003;107:2459–66.

[16] Dobson Jr JG, Fenton RA. Adenosine A2 receptor function in ratventricular myocytes. Cardiovasc Res 1997;34:337–47.

[17] Monahan TS, Sawmiller DR, Fenton RA, Dobson Jr JG. Adenosine A(2a)-receptor activation increases contractility in isolated perfusedhearts. Am J Physiol Heart Circ Physiol 2000;279:H1472–81.

[18] Xu H, Stein B, Liang B. Characterization of a stimulatory adenosineA2a receptor in adult rat ventricular myocyte. Am J Physiol 1996;270:H1655–61.

[19] Boknik P, Neumann J, Schmitz W, Scholz H, Wenzlaff H.Characterization of biochemical effects of CGS 21680C, an A2-adenosine receptor agonist, in the mammalian ventricle. J CardiovascPharmacol 1997;30:750–8.

[20] Kilpatrick EL, Narayan P, Mentzer Jr RM, Lasley RD. Cardiacmyocyte adenosine A2a receptor activation fails to alter cAMP orcontractility: role of receptor localization. Am J Physiol Heart CircPhysiol 2002;282:H1035–40.

[21] Hein TW, Wang W, Zoghi B, Muthuchamy M, Kuo L. Functional andmolecular characterization of receptor subtypes mediating coronarymicrovascular dilation to adenosine. JMolCell Cardiol 2001;33:271–82.

[22] Murphree LJ, Marshall MA, Rieger JM, MacDonald TL, Linden J.Human A(2A) adenosine receptors: high-affinity agonist binding toreceptor-G protein complexes containing Gbeta(4). Mol Pharmacol2002;61:455–62.

[23] Woodiwiss AJ, Honeyman TW, Fenton RA, Dobson Jr JG. AdenosineA2a-receptor activation enhances cardiomyocyte shortening via Ca2+-independent and -dependent mechanisms. Am J Physiol 1999;276:H1434–41.

[24] Fredholm BB, Arslan G, Halldner L, Kull B, Schulte G, WassermanW.Structure and function of adenosine receptors and their genes. Naunyn-Schmiedeberg's Arch Pharmacol 2000;362:364–74.

[25] Burgueno J, Blake DJ, Benson MA, Tinsley CL, Esapa CT, Canela EI,et al. The adenosine A2A receptor interacts with the actin-bindingprotein alpha-actinin. J Biol Chem 2003;278:37545–52.

[26] Sham JS, Cleemann L, Morad M. Functional coupling of Ca2+channels and ryanodine receptors in cardiac myocytes. Proc Natl AcadSci U S A 1995;92:121–5.

[27] Satoh H, Blatter LA, Bers DM. Effects of [Ca2+]i, SR Ca2+ load, andrest on Ca2+ spark frequency in ventricular myocytes. Am J Physiol1997;272:H657–68.

[28] Cannell MB, Cheng H, Lederer WJ. The control of calcium release inheart muscle. Science 1995;268:1045–9.

[29] Santana LF, Kranias EG, Lederer WJ. Calcium sparks and excitation–contraction coupling in phospholamban-deficient mouse ventricularmyocytes. J Physiol 1997;503(Pt 1):21–9.

[30] Lukyanenko V, Viatchenko-Karpinski S, Smirnov A, Wiesner TF,Gyorke S. Dynamic regulation of sarcoplasmic reticulum Ca(2+)content and release by luminal Ca(2+)-sensitive leak in rat ventricularmyocytes. Biophys J 2001;81:785–98.

[31] Cheng H, Lederer WJ, Cannell MB. Calcium sparks: elementaryevents underlying excitation–contraction coupling in heart muscle.Science 1993;262:740–4.

[32] Moschella MC, Marks AR. Inositol 1,4,5-trisphosphate receptorexpression in cardiac myocytes. J Cell Biol 1993;120:1137–46.

302 L. Hove-Madsen et al. / Cardiovascular Research 72 (2006) 292–302

[33] Mackenzie L, Bootman MD, Laine M, Berridge MJ, Thuring J,Holmes A, et al. The role of inositol 1,4,5-trisphosphate receptors in Ca(2+) signalling and the generation of arrhythmias in rat atrial myocytes.J Physiol 2002;541:395–409.

[34] Lipp P, Laine M, Tovey SC, Burrell KM, Berridge MJ, Li W, et al.Functional InsP3 receptors that may modulate excitation–contractioncoupling in the heart. Curr Biol 2000;10:939–42.

[35] Zima AV, Blatter LA. Inositol-1,4,5-trisphosphate-dependent Ca(2+)signalling in cat atrial excitation–contraction coupling and arrhyth-mias. J Physiol 2004;555:607–15.

[36] Abebe W, Mustafa SJ. Effects of adenosine analogs on inositol 1,4,5-trisphosphate production in porcine coronary artery. Vascul Pharmacol2002;39:89–95.

[37] Giovannucci DR, Groblewski GE, Sneyd J, Yule DI. Targetedphosphorylation of inositol 1,4,5-trisphosphate receptors selectivelyinhibits localized Ca2+ release and shapes oscillatory Ca2+ signals.J Biol Chem 2000;275:33704–11.

[38] Wojcikiewicz RJ, Luo SG. Phosphorylation of inositol 1,4,5-trispho-sphate receptors by cAMP-dependent protein kinase. Type I, II, and III

receptors are differentially susceptible to phosphorylation and arephosphorylated in intact cells. J Biol Chem 1998;273:5670–7.

[39] Stange M, Xu L, Balshaw D, Yamaguchi N, Meissner G. Character-ization of recombinant skeletal muscle (Ser-2843) and cardiac muscle(Ser-2809) ryanodine receptor phosphorylation mutants. J Biol Chem2003;278:51693–702.

[40] Xiao B, Jiang MT, Zhao M, Yang D, Sutherland C, Lai FA, et al.Characterization of a novel PKA phosphorylation site, serine-2030,reveals no PKA hyperphosphorylation of the cardiac ryanodinereceptor in canine heart failure. Circ Res 2005;96:847–55.

[41] Fraser ID, Tavalin SJ, Lester LB, Langeberg LK, Westphal AM, DeanRA, et al. A novel lipid-anchored A-kinase Anchoring Protein facilitatescAMP-responsive membrane events. EMBO J 1998;17:2261–72.

[42] Jurevicius J, Fischmeister R. cAMP compartmentation is responsiblefor a local activation of cardiac Ca2+ channels by beta-adrenergicagonists. Proc Natl Acad Sci U S A 1996;93:295–9.

[43] Zaccolo M, Pozzan T. Discrete microdomains with high concentrationof cAMP in stimulated rat neonatal cardiac myocytes. Science2002;295:1711–5.

by guest on April 2, 2016

Dow

nloaded from