FEBS 29097 FEBS Letters 579 (2005) 133–140

Phosphoinositide 3-kinase c controls autonomic regulation of themouse heart through Gi-independent downregulation of cAMP level

Giuseppe Alloattia,b, Andrea Marcantonia, Renzo Levia,b, Maria Pia Galloa, Lorenzo Del Sorboc,Enrico Patruccod, Laura Barberisd, Daniela Malana,b, Ornella Azzolinod, Matthias Wymanne,

Emilio Hirschd,*,1, Giuseppe Montrucchioc,1

a Dipartimento di Biologia Animale e dell�Uomo, Universita di Torino, 10123 Torino, Italyb Istituto Nazionale per la Fisica della Materia, Unita di Torino, Italy

c Dipartimento di Fisiopatologia Clinica, Universita di Torino, 10126 Torino, Italyd Dipartimento di Genetica, Biologia e Biochimica, Universita di Torino, 10126 Torino, Italy

e Centre of Biomedicine, Dept. Clinical-Biological Sciences, University of Basel, Basel, Switzerland

Received 18 October 2004; accepted 8 November 2004

Available online 30 November 2004

Edited by Veli-Pekka Lehto

Abstract Cardiac b-adrenergic and the muscarinic receptorscontrol contractility and heart rate by triggering multiple signal-ing events involving downstream targets like the phosphoinositide3-kinase c (PI3Kc). We thus investigated whether the lack ofPI3Kc could play a role in the autonomic regulation of the mouseheart. Contractility and ICaL of mutant cardiac preparations ap-peared increased in basal conditions and after b-adrenergic stim-ulation. However, basal and b-adrenergic stimulated heart ratewere normal. Conversely, muscarinic inhibition of heart ratewas reduced without alteration of the Gbc-dependent stimulationof IK,ACh current. In addition, muscarinic-mediated anti-adrener-gic effect on papillary muscle contractility and ICaL was signifi-cantly depressed. Consistently, cAMP level of PI3Kc-nullventricles was always higher than wild-type controls. Thus,PI3Kc controls the cardiac function by reducing cAMP concen-tration independently of Gi-mediated signaling.� 2004 Federation of European Biochemical Societies. Publishedby Elsevier B.V. All rights reserved.

Keywords: PI3Kc; Heart; b-Adrenergic receptor; Muscarinicreceptor; cAMP

1. Introduction

In the cardiovascular system, homeostatic regulation is

mainly controlled by the extracellular signals transduced by

G protein-coupled receptors (GPCRs). Sympathetic and para-

sympathetic transmitters act on cardiac GPCRs to modulate

heart rate and contractility [1]. In the heart, the predominant

receptors for the sympathetic agonists belong to the b-adrener-gic receptor (b-AR) family comprising three isoforms desig-

nated b1, b2 and b3. On the other hand, the parasympathetic

cholinergic action is mainly mediated by the M2 muscarinic

receptor [1]. Stimulation of these receptors leads to the disso-

ciation of trimeric G proteins into Ga and Gbc subunits that

in turn trigger a complex signaling cascade. Most b-ARs, like

*Corresponding author. Fax: +39 011 670 6547.E-mail address: emilio.hirsch@unito.it (E. Hirsch).

1 These authors contributed equally.

0014-5793/$22.00 � 2004 Federation of European Biochemical Societies. Pu

doi:10.1016/j.febslet.2004.11.059

the b1 receptor, are coupled to Gas, which increases heart rate

and contractility through the activation of adenylyl cyclase.

The increased intracellular concentration of cAMP activates

protein kinase A, which phosphorylates troponin I, the L-type

Ca2+-channel and phospholamban. These processes can be

contrasted by the antiadrenergic effect of the M2 muscarinic

receptor, which couples to Gai and thus promotes a decrease

of intracellular cAMP [2]. Concomitantly with the effects of

Gai, the activation of the M2 receptor also causes the release

of the Gbc subunits and their direct activation of the GIRK

potassium channels, that contribute to cardiomyocyte hyper-

polarization and reduction of heart rate [3].

In the heart, Gbc subunits are also thought to activate phos-

phoinositide 3-kinases (PI3Ks), a family of protein and lipid

kinases involved in multiple biological processes including cell

proliferation and survival, cytoskeletal remodeling and mem-

brane trafficking [4]. Based on their biochemical properties,

these enzymes are divided into three classes and at least three

members (a, b and c) of class I PI3Ks are expressed in the

mammalian heart [5]. While PI3Ka and b (class IA PI3Ks)

associate to p85-like regulatory proteins docking to phosphor-

ylated tyrosines in YXXMmotives, PI3Kc (the unique class IBPI3K) binds directly or in association with the p101 adaptor to

the bc subunits of G proteins and is thus activated by GPCRs

[6].

Indirect evidence suggests that the class IB PI3Kc can be

stimulated by Ca2+/PKC [7] or prototype cardiac GPCRs,

such as adrenergic and muscarinic receptors. In transfected

fibroblasts, PI3Kc is activated by b-ARs and contributes to

their desensitization through the membrane targeting of a

b-AR kinase (bARK)-PI3Kc cytoplasmic complex [8]. More-

over, ectopic expression of both M2 receptor and PI3Kc in

COS-7 cells causes a muscarinic agonist-dependent PI3Kcstimulation and the consequent activation of the protein ki-

nase B/Akt [9] and MAPK [10]. Interestingly, PI3Kc activity

appears to be upregulated in vivo in the mouse model of

pressure overload hypertrophy induced by transverse aortic

constriction [11].

We and others recently reported the generation of PI3Kc-deficient mice. These mutant mice are viable and fertile, but dis-

play defective GPCR-mediated responses to chemoattractants

in leukocytes [12–14], to ADP in platelets [15] and to adenosine

in mast cells [16]. On the other hand, cardiomyocytes derived

blished by Elsevier B.V. All rights reserved.

134 G. Alloatti et al. / FEBS Letters 579 (2005) 133–140

from PI3Kc-deficient embryonic stem cells show an impaired

reduction of the spontaneous Ca2+ spiking induced by ATP

[17]. Furthermore, both in vivo echocardiographic analysis

and in vitro studies on isolated myocardial cells indicate that

the lack of PI3Kc causes enhanced cardiac contractility [18].

This effect is correlated to an involvement of PI3Kc on cAMP

homeostasis: in fact, isolated PI3Kc-null cardiomyocytes in ba-

sal conditions and after stimulation with the b2-AR-specific

agonist zinterol produced significantly higher levels of cAMP

than wild-type controls.

Here, we further explored the role of PI3Kc in the auto-

nomic control of heart function and found that the lack of

PI3Kc not only affects b-adrenergic signaling but also musca-

rinic-dependent responses. In PI3Kc-deficient mice, while iso-

proterenol (Iso) caused a stronger enhancement of

contractility, carbachol (CCh) unexpectedly induced a weaker

chronotropic effect. Similarly, the anti-adrenergic activity

caused by cholinergic stimulation was reduced in mutant

cardiomyocytes. In PI3Kc-null papillary muscles, b-adrenergicreceptor desensitization was normal and, in ventricular myo-

cytes, the muscarinic regulation of potassium ion channels

was unaltered. Nevertheless, cAMP levels appeared higher in

mutant hearts than wild-type controls in all conditions tested.

Our data thus indicate that PI3Kc modulates cardiac function

by reducing cAMP concentration through a mechanism that is

independent of receptor internalization or Gi-dependent

signaling.

2. Materials and methods

2.1. MicePI3Kc-deficient mice (PI3Kc�/�) and ‘‘knock-in’’ mutants express-

ing a catalytically inactive PI3Kc (PI3KcKD/KD) mice were obtainedas previously described [12,19]. All experiments were performed withsex matched 8-week-old wild-type and PI3Kc�/� mice; both mutantand control animals (PI3Kc+/+) were inbred 129sv or littermates fromheterozygous crosses. Animal experiments and sacrifices were carriedout in compliance with institutional guidelines as well as with Italianand Swiss government regulations and under local ethical committeeguidance.

2.2. In vivo experimentsIn vivo analysis of cardiovascular responses to different agonists was

performed in Pentobarbital sodium (40 mg/kg in saline) anesthetizedmice. Animals were placed on a heating pad at 37 �C; heart rate wasanalyzed by standard electrocardiographic (ECG) techniques. Foragonist administration, a Teflon tubing was inserted in the left jugularvein and a bolus of 100 ll of agonist or saline alone was administeredwithin 30 s. Doses of isoproterenol (3 lg/kg) and carbachol (8 lg/kg)were chosen as previously described [20,21].

2.3. Analysis of spontaneous beating frequency in right atrium muscle

and contractility in isolated papillary muscleAfter sacrifice, heart was rapidly removed, and right atria or papil-

lary muscles from left ventricle were dissected free under microscopiccontrol. Atria and papillary muscles were perfused with a modifiedTyrode solution (154 mM NaCl, 4 mM KCl, 2 mM CaCl2, 1 mMMgCl2, 5.5 mM DD-glucose, and 5 mM HEPES, adjusted to pH 7.38with NaOH and gassed with 100% O2). Atria and papillary muscleswere equilibrated in modified Tyrode solution for at least 30 min;experiments were conducted at 37 �C. Isolated, spontaneously beatingright atria and papillary muscles were used to study the chronotropicand inotropic responses, respectively. Papillary muscles were drivenat a frequency of 240 beats/min with a pair of electrodes connectedto a 302 T Anapulse Stimulator via a 305-R Stimulus Isolator (W.P.Instruments, New Haven, CT), operating in constant current mode.

Isometric twitches were measured by a Harvard Isometric Transducer60–2997, recorded and analyzed by a Power Mac computer using cus-tom programmed Labview Software (National Instruments, Texas,USA). All drug-containing solutions were prepared fresh before theexperiments. Desensitization to repeated b-adrenergic stimulationswas studied in papillary muscles of both wild-type and PI3Kc-deficientmice according to the following protocol: after equilibration, papillarymuscles were stimulated three times with Iso (1 lM) for 5 min. Duringthe interval between each period of stimulation, papillary muscles werewashed with Tyrode solution for different times, lasting for 10 or 30min, respectively, between the first and the second, or the secondand the third period.

2.4. Isolation of cardiomyocytes and patch clamp analysisIsolated ventricular cells were obtained from adult mice by enzy-

matic dissociation after animal killing under anesthesia and heartexplantation. Briefly, after cannulation of ascending aorta, hearts wereperfused with Tyrode Calcium Free solution (containing, in mM: 135NaCl, 4 KCl, 1 MgCl2, 2 HEPES, 10 glucose, 10 butanedione monox-ime, and 5 taurine; pH 7.40, adjusted with NaOH) for 5 min (37 �C)and then with 10 ml of Tyrode Calcium Free with collagenase (0.3mg/ml) and protease (0.02 mg/ml). Hearts were then perfused andenzymatically dissociated with 30 ml of Tyrode solution containing50 lM CaCl2 and the same concentrations of enzymes, as mentionedbefore. Atria and ventricles were then separated and the ventricles werecut in small pieces and shaken for 10 min in 20 ml of Tyrode CalciumFree in the presence of 50 lMCaCl2, collagenase and protease. Whole-cell patch clamp experiments were performed at 30 �C. The whole-cellpatch clamp standard solution contained (mM): 133 CsCl; 5 EGTAfree acid; 5 Na2ATP; 5 Na2phosphocreatine; 5 HEPES; 3 MgCl2;and 0.4 Na2GTP; pH 7.3, adjusted with CsOH. Isolated ventricularcells were continuously bathed with Cesium Tyrode solution (contain-ing, in mM: 138 NaCl, 20 CsCl, 2 CaCl2, 1 MgCl2, 5.5 D-glucose, 1 4-aminopyridine, and 5 HEPES, pH 7.35, adjusted with NaOH); testsolutions were applied through fine silicon tubes. To measure ICaL,the cell was depolarized every 5 s from �100 to �40 mV for 50 ms,to inactivate both sodium current and T-type calcium current. Alldrug-containing solutions were prepared fresh before the experiments.Cell capacitance was measured using the standard procedures de-scribed in the Axopatch (Axon) manual.For the measure of acetylcholine-activated potassium currents, the

extracellular solution contained (mM): 120 NaCl; 20 KCl; 1 MgCl2;and 10 HEPES; pH 7.40, adjusted with NaOH. The intracellularsolution contained (mM): 100 Kaspartate; 40 KCl; 5 MgCl2;5Na2ATP; 2 EGTA; 0.01 GTP; and 5 HEPES; pH 7.40, adjusted withKOH [22].To measure inward rectifying potassium currents, the membrane po-

tential was held at �90 mV and then driven from �120 to +60 mV witha slope of 1 mV/ms. IK,ACh was computed as the difference current be-tween CCh treatment and control. The difference current was thenevaluated at �70 mV after subtraction of the current at its theoreticalreversal potential (�50 mV with the solutions described above) [22].

2.5. cAMP measurement in isolated heartsIsolated mouse hearts were mounted on a Langendorff perfusion

apparatus, and perfused at a constant flow rate using a peristalticpump with modified Tyrode solution warmed to 37 �C and gassed with100% O2. After stabilization, the hearts were subjected to stimulationwith 1lM Iso perfused alone or together with 1 lM CCh for 1 min.After stimulation, ventricles were removed and frozen in liquid nitro-gen. Tissue cAMP levels were measured with a radiolabeled competi-tive protein-binding assay (Cyclic AMP [3H]BiotrakTM AssaySystem, Amersham Biosciences, Piscataway, USA). For assessing therole of Gi in cAMP metabolism, mice were treated i.p. with 100 lg/kg of Pertussis Toxin (Sigma, St. Louis, MO) as previously described[19]. The effect of pertussis toxin was tested 16 hours later for its abilityto block the muscarinic chronotropic response by ECG measurement.

2.6. Statistical analysisData are expressed as means ± S.E.M. Comparison between groups

was performed using two-way ANOVA for repeated measures (in vivostudies) or one way ANOVA (in vitro measurements), followed byBonferroni post hoc test. Statistical significance was assigned to P va-lue <0.05.

G. Alloatti et al. / FEBS Letters 579 (2005) 133–140 135

3. Results

3.1. Increased positive inotropism, but no difference in the

chronotropic effect after b-adrenergic stimulation of

PI3Kc-null heartsIn the heart, stimulation of b-adrenergic receptors in-

creases contractility as well as beating frequency. We thus

evaluated whether the lack of PI3Kc could affect b-AR-

dependent enhancement of heart rate and force development

both in vivo and in isolated cardiac preparations. PI3Kc-deficient (PI3Kc�/�) and wild-type (PI3Kc+/+) mice showed

comparable basal values for heart rate (450 ± 8 and

409 ± 18 beats/min, n = 3 and n = 4 in PI3Kc+/+ and

PI3Kc�/�; respectively). The b-adrenergic-dependent positivechronotropism induced by intravenous administration of iso-

proterenol (single bolus dose, 3 lg/kg) showed an identical

increase in the heart rate of both wild-type and mutant ani-

mals (Fig. 1A). Similarly, we observed no significant differ-

ence in spontaneous beating frequency (203 ± 16 and

190 ± 40 beats/min; n = 4 in both cases) as well as in the po-

sitive chronotropic effect induced by Iso (1 lM) on both

PI3Kc+/+ and PI3Kc�/� isolated right atria (44.2 ± 15.9

and 36.1 ± 9.1% increase over basal values) (Fig. 1B).

Positive inotropism caused by b-adrenergic stimulation with

isoproterenol was next examined in electrically driven papil-

lary muscles. As expected from previous indications [18,19],

in basal conditions papillary muscles from mutant mice

showed an enhanced contractility in comparison with control

preparations (1.9 ± 0.3 vs. 3.3 ± 0.4 mN/g tissue wet weight

for wild-type and PI3Kc-deficient samples, respectively; n = 7

and 12; P < 0.05 by ANOVA; Fig. 2A and C). Moreover,

PI3Kc-deficient papillary muscles showed a more prominent

response to b-adrenergic stimulation, which was 80% stronger

than that detected in wild-type controls (Fig. 2B and D). As

shown in Fig. 2E, the inotropic effect of Iso (1lM) was signif-

icantly higher in PI3Kc-null papillary muscles (524.0 ± 65.5%

of basal value, n = 16), than in wild-type preparations

(337.5 ± 33.5% of basal value, n = 15; P < 0.01 by ANOVA).

Fre

quen

cy (

% B

asal

)

0

40

80

120

160B

Fre

quen

cy (

% B

asal

)

Time (s)

0 40 80 120

140

120

100

80

60

A

Fig. 1. The lack of PI3Kc has no effects on b-adrenergic chronotropicstimulation. (A) Heart rate after intravenous administration of 3 lg/kgisoproterenol in anesthetized animals. White diamonds, wild-type;black squares, PI3Kc�/�; values are expressed as % of basal heart rate.Isoproterenol was administered at time 0. (B) Right atrium beatingfrequency 30 s after administration of 1 lM isoproterenol. Bars (white,wild-type; black, PI3Kc�/�) represent the% of basal values. Basalvalues for both spontaneous beating frequency (203 ± 16 and 190 ± 40beats/min, wild-type and PI3Kc�/�, respectively; n = 4 in both cases)were not significantly different.

Since it has been suggested that PI3Kc could be involved in

b-adrenergic receptors internalization [5], we studied the re-

sponse of papillary muscle to repeated stimulations with Iso

(1 lM). No significant difference between control and mutant

papillary muscles was detected for the reduced response to Iso

10 min after a previous stimulation (50.3 ± 8.9 and 55.1 ± 6.2%

of the effect induced by the first stimulation with Iso, n = 4 and

5), as well as for the recovery after 30 min washout interval

(79.0 ± 5.1 and 81.4 ± 10.0% of the effect induced by the first

stimulation with Iso, n = 4 and 5), thus suggesting that the

absence of PI3Kc does not alter b-adrenergic receptors

desensitization.

The increased contraction in response to isoproterenol is

considered a consequence of b-AR-induced elevation of

cAMP, PKA activation and subsequent phosphorylation of

contraction-controlling proteins, including voltage-dependent

L-type Ca2+ channels. Phosphorylation of these channels pro-

motes Ca2+ influx that triggers the release of intracellular Ca2+

from the sarcoplasmic reticulum. The subsequent intracellular

Ca2+ transient finally activates the contraction of myofila-

ments. Patch clamp experiments revealed that, although no

significant difference in cell capacitance was found between

wild-type and PI3Kc�/� ventricular cells (189.77 ± 17.18 pF,

n = 34 and 180.73 ± 13.90 pF, n = 27 respectively), calcium

current density in the absence of stimulation was higher in

PI3Kc�/� cells (wild-type: 4.31 ± 0.51 pA/pF, n = 34; PI3Kc-null: 5.92 ± 0.53 pA/pF, n = 27; P < 0.05). Consistently,

PI3Kc�/� cardiac myocytes showed abnormally increased L-

type Ca2+ current in basal conditions (ICaL, Fig. 3A–D, arrows

#1 and 2) as well as in response to isoproterenol stimulation

(70% increase after 1lM isoproterenol, P < 0.01, Fig. 3E).

These data thus show that PI3Kc negatively regulates L-type

Ca2+ current before and after b-adrenergic stimulation. To

study if the higher basal values of calcium current were due

to cAMP-dependent phosphorylation, the PKA inhibitor

H89 (1 lM) was used. In this condition, the reduction of cal-

cium current density in cardiac cells from PI3Kc�/� mice ap-

peared stronger (74.47 ± 4.47%; n = 6) than that detected in

wild-type controls (39.40 ± 8.91%; n = 5). The calcium current

density, which in basal conditions was higher in PI3Kc�/� cells

(PI3Kc+/+: 3.65 ± 0.25 pA/pF (n = 5); PI3Kc�/�: 6.99 ± 1.10

pA/pF (n = 6); P < 0,05), became comparable in both geno-

types after H89 administration (PI3Kc+/+: 2.66 ± 0.25 pA/

pF; PI3Kc�/�: 3.98 ± 0.63 pA/pF).

3.2. Reduced muscarinic inhibition of heart rate in PI3Kc-nullmice

To study the role of PI3Kc in cholinergic signaling, heart

rate was evaluated in anesthetized PI3Kc-null and wild-type

mice before and after intravenous administration of carbachol

(8 lg/kg, single bolus dose). In basal conditions, wild-type and

PI3Kc-null mice did not show significant differences in heart

rate (432 ± 12 and 474 ± 27 beats/min, n = 5 for both cases).

Furthermore, intravenous administration of saline did not

cause different responses in mice of the two genotypes (not

shown). After infusion of carbachol into the jugular vein,

wild-type animals exhibited a dramatic drop in heart rate;

however, this effect was significantly reduced in mice lacking

PI3Kc. At 40 s after stimulation, when the effect of carbachol

reached its peak, heart rate of PI3Kc-null mice was 40% higher

than that of wild-type controls (150 ± 59 and 372 ± 42 bpm for

wild-type and PI3Kc�/� mice, respectively; P < 0.001; Fig. 4).

E

400

600

500

300

100

0

200

700

For

ce (

% B

asal

)

**

A

C

B

D

For

ce (

mN

/g)

For

ce (

mN

/g)

For

ce (

mN

/g)

For

ce (

mN

/g)

0 1 2 3 4

00 11 22 33 44

50 1 2 3 4 5

55

Time (s)

Time (s)Time (s)

Time (s)

Basal

Basal

Iso

Iso

12

1212

12

6

66

6

3

33

3

9

99

9

15

1515

15

0

00

0

Fig. 2. b-adrenergic inotropic effect on papillary muscle. Representative traces of the inotropic effect induced by isoproterenol (1 lM) on wild-type(A,B) and PI3Kc-deficient (C,D) papillary muscles. (E) Bar graph showing the maximal inotropic effect induced by Iso on wild-type (white bars) andPI3Kc�/� (black bars) papillary muscles. Bars represent the force of contraction as % of basal value (wild-type, n = 15; PI3Kc�/�, n = 16; P < 0.01 byANOVA).

C

ISO ISO+CCh

-1.8

-1.3

-0.8

13

2

A

D

3

2

1

-2.0

-1.5

-1.0

ISO ISO+CCh

Time (min)

0 1 2 3 4 5 6 7 8 9 1110

Time (min)

0 1 2 3 4 5 6 7 8 9

B E 100

80

60

40

20

0

*

∆I(%

Bas

al)

Ca

120

100

60

40

20

0

*

80

F

∆ I(%

Red

uctio

n)C

a

50 ms

21 3

0

-1

50 ms2

13

0

-1I(n

A)

Ca

I(n

A)

Ca

I(n

A)

Ca

I(n

A)

Ca

Fig. 3. Adrenergic and antiadrenergic effect of Iso alone and of its combination with CCh on isolated ventricular cells. Cells were stimulated with 1lM of each agonist. Panels A–D show typical experiments performed on wild-type (A,C) and PI3Kc�/� (B,D) mouse ventricular myocytes. In A andB, single current traces were recorded at the times indicated by the corresponding numbers in the time-courses illustrated on panels C and D. (E)Summary of the effects of 1 lM isoproterenol on the regulation of maximal ICaL in wild-type (white bar; n = 19) and PI3Kc�/� (black bar; n = 21)mouse ventricular myocytes. Bars represent the difference in% between the basal and the maximal isoproterenol-stimulated ICaL. Statistical differenceof wild-type vs. mutant as determined by Student�s t test: *P < 0.05. (F) Bar graph showing the maximal effect induced by CCh on ICaL on wild-type(white bar; n = 12) and PI3Kc�/� (black bars; n = 8) isolated cells pretreated with isoproterenol (1 lM). Bars represent the % reduction of theincrease induced by Iso (P < 0.05 by Student�s t test).

136 G. Alloatti et al. / FEBS Letters 579 (2005) 133–140

To further explore the role of PI3Kc in the modulation of

chronotropism, we next tested whether the impaired musca-

rinic response detected in vivo could be confirmed in explanted

cardiac preparations. In unstimulated, spontaneously beating

isolated right atria, baseline frequency was similar in both

genotypes (264.2 ± 26.7 and 233.0 ± 23.2 beats/min, wild-type

and PI3Kc�/�, respectively; n = 4 in both groups). Carbachol

(1 lM), however, depressed heart rate less extensively in

PI3Kc-null than in wild-type right atria (88.2 ± 7.0% inhibi-

tion in wild-type, n = 4; 56.8 ± 9.3% in PI3Kc-null, n = 4;

P < 0.05, Fig. 5). These data thus indicate that control of mus-

carinic-induced negative chronotropism can be affected by the

lack of PI3Kc.These results suggested that the effect of the lack of PI3Kc

on cardiac cAMP levels could be responsible for the reduced

responsiveness to muscarinic agonists. As the expression in

Fre

quen

cy (

% B

asal

) 120

80

40

0

20 40 60 80 100

120

140

Time (s)

0

Fig. 4. In vivo chronotropic effect of carbachol. Wild-type, emptydiamonds; PI3Kc�/�, filled squares. CCh (8 lg/kg) was administeredi.v. at time 0. Heart rate was obtained from ECG recording. Basalvalues for heart rate (432 ± 12 beats/min, wild-type, n = 5 and 474 ± 27beats/min, PI3Kc�/�, n = 5) were superimposable. Statistical differenceof wild-type vs. mutant: * P < 0.05; ** P < 0.01 by ANOVA withBonferroni pair-wise comparison.

G. Alloatti et al. / FEBS Letters 579 (2005) 133–140 137

PI3KcKD/KD mice of a catalytically inactive PI3Kc rescues the

increased cAMP levels and contractility seen in PI3Kc-nullheart [19], we tested the negative chronotropic effect of

CCh on PI3KcKD/KD hearts. Isolated right atria from

PI3KcKD/KD mice showed a spontaneous beating frequency

(245.8 ± 13.3, n = 5) and a negative chronotropic effect of

CCh (86.5 ± 2.7% inhibition), which were fully comparable

to those found in WT preparations. This finding indicates that

the reduced response to muscarinic modulation is not due to

the kinase activity of PI3Kc.

3.3. Stimulation of IK,ACh after cholinergic stimulation is not

affected by the lack of PI3KcHeart rate is known to be negatively controlled mainly by

the muscarinic receptor-dependent activation of the IK,ACh

potassium current. To investigate whether effects on IK,ACh

were involved in the reduced muscarinic inhibition of heart

rate observed in PI3Kc-null mice, we studied the muscarinic

modulation of IK,ACh in isolated atrial cells. The absence of

PI3Kc did not affect the entity of the potassium current in-

duced by CCh (1 lM) measured at �70 mV (58.7 ± 4.2 vs.

E

BA

D

-0.3

For

ce (

a.u.

)

0 1 2 3 4 5

Time (s)

1.2

0.6

0.9

0.3

0

-0.3

1.2

0.6

0.9

0.3

0

0 1 2 3 4 5

Time (s)

For

ce (

a.u.

)

0 1 2 3 4 5

Time (s)

0 1 2 3 4 5

Time (s)

Fig. 5. Muscarinic chronotropic effects on isolated atria. (A–F) Represenspontaneously beating right atria, before (A,D) and 30 s after the administratG: summary of CCh-mediated effects on beating frequency of right atria. Barvalues before addition of CCh (*P < 0.05 by Student�s t test). Basal values fomin, wild-type and PI3Kc�/�, respectively; n = 4 in both groups) and condifferent.

73.3 ± 12.2 pA/pF, n = 7 and 5 in wild-type and PI3Kc-defi-cient cells, respectively), thus indicating that the lack of PI3Kcdoes not alter Gi activation and that PI3Kc controls heart rate

independently of potassium current and Gbc-dependent signal

transduction events.

3.4. The muscarinic antiadrenergic effect in papillary muscles

and isolated cardiomyocytes is reduced in the absence of

PI3KcIn the mouse, muscarinic stimulation is known to be unable

to affect contractility of ventricular muscle, in the absence of a

previous enhancement of cAMP intracellular concentration

[2]. In agreement, we did not observe significant effects of car-

bachol (1 lM) on contractile force developed by papillary

muscles of both genotypes in basal conditions (Table 1). Since

it has been shown that muscarinic agonists exert an indirect

depression of ventricular muscle contractility after b-adrener-gic stimulation [23], we studied the antiadrenergic effect of

CCh (1 lM) in the presence of Iso (1 lM). In these conditions,

the lack of PI3Kc significantly reduced the negative inotropic

response to muscarinic stimulation (Table 1). In agreement

with the results obtained in papillary muscle, we observed that

CCh (1 lM) had not significant effect on baseline values of L-

type calcium current in cells from both PI3Kc+/+ and PI3Kc�/�

mice, and that the antiadrenergic effect of CCh (1 lM) was sig-

nificantly reduced in PI3Kc-null ventricular cells (96.3 ± 16.9

and 47.6 ± 12.0% of the enhancement induced by Iso in wild-

type and PI3Kc-deficient, respectively; n = 12 and 8, P < 0.05

by ANOVA; Fig. 3A–D).

As L-type Ca2+ channels are controlled by PKA, we mea-

sured the levels of cAMP in wild-type and PI3Kc-deficienthearts before and after stimulation with Iso and the combina-

tion of Iso/CCh. In agreement with previous reports [18], the

basal levels of cAMP were significantly higher in mutant than

in controls (Fig. 6). Nonetheless, this difference was still detect-

able after both Iso and Iso/CCh stimulation, where cAMP in

PI3Kc-null hearts was 40% and 50% higher than controls,

respectively (Fig. 6). Remarkably, the percentage inhibition

of cAMP concentration induced by CCh was significantly

*

F

C

0 1 2 3 4 5

Time (s)

0

0 1 2 3 4 5

Time (s)

G

40

50

30

10

0

20

60

Fre

quen

cy(%

Bas

al)

tative traces of the rate of wild-type (A–C) and PI3Kc�/� (D–F)ion of 1 lM CCh (B, E) or 5 min after removal of the agonist (C,F). Ins (wild-type, empty; PI3Kc�/�, filled) represent the percentage of basalr spontaneous beating frequency (264.2 ± 26.7 and 233.0 ± 23.2 beats/tractile force (4.1 ± 0.57 and 4.3 ± 0.68 mN/g) were not significantly

Table 1Direct and antiadrenergic effects of carbachol on papillary musclecontraction

Wild-type PI3Kc�/�

Direct Effecta

CCh 1 lM 110.1 ± 10.3 (4) 108.3 ± 5.9 (4)

Antiadrenergic Effectb

CCh 1 lM 73.7 ± 5.6 (4) 13.2 ± 12.1 (4)*

Values are represented as means ± SEM.aValues are expressed as percentage of basal contractility.bValues are expressed as percentage reduction of isoproterenol (1 lM)effect.*P < 0.01 (Wild-type vs. PI3Kc�/�). The number of individual samplesanalyzed is indicated in brackets.

0.0

0.2

***

**

*

****

0.4

0.6

0.8

1.0

1.2

Basal Iso Iso+CCh

cAM

P(p

mol

/mg)

Fig. 6. Total cAMP concentration in wild-type and mutant hearts.Levels of cAMP (wild-type, white bars; PI3Kc�/�, black bars)measured by standard radioactive assay in Langendorff-perfusedhearts before and after 1 min of stimulation with Iso (1 lM) or CCh(1 lM) or their combination (for each bar n = 6, *P < 0.05, **P < 0.01,***P < 0.001 by ANOVA).

138 G. Alloatti et al. / FEBS Letters 579 (2005) 133–140

higher in wild-type than in PI3Kc�/� hearts (39 ± 2% vs.

23 ± 4%, respectively; n = 6, P < 0.001 by Students t test).

Stimulation with Iso/CCh after overnight pretreatment of

hearts with 100 lg/kg of pertussis toxin did not cause any

change in the levels of cAMP detected with Iso alone (data

not shown). These data thus show that the reduced antiadren-

ergic response, detected in mutant hearts, is due to an abnor-

mally elevated cAMP concentration and that PI3Kc controls

cAMP levels independently of Gi-mediated signaling.

4. Discussion

Here, we reported that the effect of b-adrenergic stimulation

on spontaneous heart rate was similar in both wild-type and

PI3Kc-null mice and isolated atria. However, the b-adrener-gic-mediated enhancement of contractile force induced by Iso

was significantly higher in PI3Kc-deficient papillary muscles.

In murine cardiac myocytes, the major receptors for catechol-

amines are the b1 and b2-ARs [1,2]. Being the b1-AR the un-

ique responsible for modulation of pacemaker activity [20],

this finding suggests that the loss of PI3Kc might not affect

b1-AR signal transduction. As previously indicated [18], our

data further suggest that the lack of PI3Kc might unmask

the activity of b2-AR. It has been indeed reported that the

b2-AR specific agonist zinterol, which is inactive on wild-type

cardiac cells, causes increased enhancement of cAMP produc-

tion and phospholamban phosphorylation in PI3Kc-deficientcardiomyocytes [18]. To control cAMP levels, PI3Kc associ-

ates with the PDE3B phosphodiesterase and positively modu-

lates its activity in kinase independent way [19]. Our data

extend these results providing the evidence that the increase

in ventricular cAMP concentration perfectly correlates with

the increase in calcium current density. Although PI3Kc was

previously reported to activate L-type Ca2+ channels in vascu-

lar smooth muscle cells by the activation of PKC [24,25], our

observations strongly suggest that the higher contractility of

PI3Kc hearts is most likely due to abnormally high levels of

cAMP that through the activation of PKA cause increased

open probability of L-type channels.

An alternative explanation of the increased positive inotro-

pic effect observed in isoproterenol-stimulated mutant papil-

lary muscles is based on a defective receptor desensitization.

It has been indeed suggested that PI3Kc is involved in these

processes and favors b-AR desensitization by internalization

[26]. However, our data, showing no significantly different re-

sponses between wild-type and PI3Kc-deficient papillary mus-

cles to repeated stimulation with Iso, suggest that the in vivo

absence of PI3Kc does not influence b-AR desensitization.

This is in agreement with the observation that the lack of

PI3Kc or the alteration of PtdIns(3,4,5)P3 production does

not modify baseline expression of cardiac b-ARs [18] even

after prolonged exposure to isoproterenol [27]. Furthermore,

it has been reported that other class IA PI3K, like PI3Kacan associate with b ARK and specifically compensate for

the lack of PI3Kc [28]. Overall, these data as well as our obser-

vations of increased cAMP, ICaL, and ventricular contractility

in basal conditions suggest, on the other hand, that PI3Kc af-

fects cardiac inotropism primarily through its action on cAMP

metabolism. Finally, the finding that inhibition of PKA abro-

gates the difference in baseline ICaL density strongly suggests

that this difference is due to cAMP-PKA dependent phosphor-

ylation of the channels, rather than an increased channel

expression.

Unexpectedly, however, our results clearly indicate that the

lack of PI3Kc also affects the muscarinic-dependent cardiac re-

sponses. Indeed, we observed that muscarinic inhibition of

heart rate was reduced both in vivo and in isolated right atria

from PI3Kc-deficient mice. Moreover, the anti-adrenergic ef-

fect of muscarinic stimulation on contractility and ICaL was

significantly depressed.

Reduction of heart rate (negative chronotropism) and coun-

teraction of the enhanced contractility induced by adrenergic

agonists (indirect negative inotropism) are mainly controlled

by the Gai-coupled muscarinic receptor of the M2 subtype

[1,2]. The report of PI3Kc being downstream M2 receptor acti-

vation [10,29] suggested that this enzyme might play a role in

the muscarinic modulation of cardiac activity. In the mouse

heart, the muscarinic chronotropic response is primarily med-

iated by the Gi-coupled M2 receptor present in the sino-atrial

node [1,2,30]. This receptor can reduce heart rate by modulat-

ing at least two major signaling pathways: it can induce the

opening of a type of inward rectifying potassium current

G. Alloatti et al. / FEBS Letters 579 (2005) 133–140 139

(IK,ACh) and it can block cAMP synthesis causing the inhibi-

tion of the hyperpolarization-activated current (If). Our obser-

vation that muscarinic stimulation of IK,ACh current appeared

identical in mutant and wild-type atrial cardiomyocytes sug-

gests that PI3Kc is not involved in the muscarinic modulation

of this current. This finding is consistent with the fact that

IK,Ach is directly activated by binding of the bc subunits of

Gi to the GIRK potassium channels, while If, like L-type cal-

cium current, is controlled by cAMP levels [3,31]. In agreement

with this view, our data showed increased cAMP contents in

PI3Kc-null samples concomitantly stimulated with Iso and

CCh. Therefore, a likely explanation for the reduced musca-

rinic effect is that an increase in the response to b-adrenergicstimulation has partially overcome the inhibitory effect of mus-

carinic receptor activation. This would indeed be expected be-

cause of the competitive nature of the interaction between the

two signaling pathways involved. Nevertheless, basal and

b-adrenergically stimulated heart rate are not altered in

PI3Kc-null mice, thus indicating that the increase in cAMP

concentration is compartmentalized and, in these conditions,

does not influence the pacemaker activity. The finding that

muscarinic anti-adrenergic effect on cAMP levels of PI3Kc-null cardiac samples is reduced indicates that PI3Kc influences

cAMP concentration independently of the action of Gi. As the

lack of PI3Kc does not affect adenylate cyclase activity [19],

these further support the view that PI3Kc modulates cAMP

destruction through its positive effect on PDE3B activity. It

is thus possible that the Gai-mediated inactivation of adenylate

cyclase could unmask a slower rate of cAMP destruction that

finally affects frequency.

Taken together, our results confirm and extend the notion

that PI3Kc is involved in the b-adrenergic-dependent controlof positive inotropism but not of positive chronotropism,

and suggest that this enzyme exerts its function independently

of receptor desensitization. Finally, our results indicate that

PI3Kc controls muscarinic autonomic cardiac regulation and

the muscarinic anti-adrenergic effect, independently of Gi-med-

iated signaling and through its ability to downregulate cAMP

concentration.

Acknowledgment: This work was supported by Human Frontier Sci-ence Project and the European Union Fifth Framework ProgramQLG1-2001-02171 to EH and MPW; MIUR Cofin 2002 to GA, EHand GM; MIUR 60% 2003 to GA and GM, INFM to GA and RL;Compagnia di San Paolo to MPG.

References

[1] Rockman, H.A., Koch, W.J. and Lefkowitz, R.J. (2002) Seven-transmembrane-spanning receptors and heart function. Nature415, 206–212.

[2] Brodde, O.E. and Michel, M.C. (1999) Adrenergic and muscarinicreceptors in the human heart. Pharmacol. Rev. 51, 651–690.

[3] Yamada, M., Inanobe, A. and Kurachi, Y. (1998) G proteinregulation of potassium ion channels. Pharmacol. Rev. 50, 723–760.

[4] Cantley, L.C. (2002) The phosphoinositide 3-kinase pathway.Science 296, 1655–1657.

[5] Prasad, S.V., Perrino, C. and Rockman, H.A. (2003) Role ofphosphoinositide 3-kinase in cardiac function and heart failure.Trends Cardiovasc. Med. 13, 206–212.

[6] Vanhaesebroeck, B., et al. (2001) Synthesis and function of 3-phosphorylated inositol lipids. Annu. Rev. Biochem. 70, 535–602.

[7] Tornay, S., Bjorklof, K., Laffargue, M., Kuenzi, P., Reuther,G.W., Hirsch, E., Finan, P., Wymann, M.P., 2003. In: Biochem-ical Society Focused Meeting Pi3-Kinase in Signalling andDisease, pp. Abstract P023, Horsham, London, UK.

[8] Naga Prasad, S.V., Barak, L.S., Rapacciuolo, A., Caron, M.G.and Rockman, H.A. (2001) Agonist-dependent recruitment ofphosphoinositide 3-kinase to the membrane by beta-adrenergicreceptor kinase 1. A role in receptor sequestration. J. Biol. Chem.276, 18953–18959.

[9] Murga, C., Laguinge, L., Wetzker, R., Cuadrado, A. andGutkind, J.S. (1998) Activation of Akt/protein kinase B by Gprotein-coupled receptors. A role for alpha and beta gammasubunits of heterotrimeric G proteins acting through phosphati-dylinositol-3-OH kinasegamma. J. Biol. Chem. 273, 19080–19085.

[10] Lopez-Ilasaca, M., Crespo, P., Pellici, P.G., Gutkind, J.S. andWetzker, R. (1997) Linkage of G protein-coupled receptors to theMAPK signaling pathway through PI 3-kinase gamma. Science275, 394–397.

[11] Naga Prasad, S.V., Esposito, G., Mao, L., Koch, W.J. andRockman, H.A. (2000) Gbetagamma-dependent phosphoinosi-tide 3-kinase activation in hearts with in vivo pressure overloadhypertrophy. J. Biol. Chem. 275, 4693–4698.

[12] Hirsch, E., et al. (2000) Central role for G protein-coupledphosphoinositide 3-kinase gamma in inflammation. Science 287,1049–1053.

[13] Li, Z., Jiang, H., Xie, W., Zhang, Z., Smrcka, A.V. and Wu, D.(2000) Roles of PLC-beta2 and -beta3 and PI3Kgamma inchemoattractant-mediated signal transduction. Science 287, 1046–1049.

[14] Sasaki, T., et al. (2000) Function of PI3Kgamma in thymocytedevelopment, T cell activation, and neutrophil migration. Science287, 1040–1046.

[15] Hirsch, E., Bosco, O., Tropel, P., Laffargue, M., Calvez, R.,Altruda, F., Wymann, M. and Montrucchio, G. (2001) Resistanceto thromboembolism in PI3Kgamma-deficient mice. FASEB J.15, 2019–2021.

[16] Laffargue, M., Calvez, R., Finan, P., Trifilieff, A., Barbier, M.,Altruda, F., Hirsch, E. and Wymann, M.P. (2002) Phosphoino-sitide 3-kinase gamma is an essential amplifier of mast cellfunction. Immunity. 16, 441–451.

[17] Bony, C., Roche, S., Shuichi, U., Sasaki, T., Crackower, M.A.,Penninger, J., Mano, H. and Puceat, M. (2001) A specific role ofphosphatidylinositol 3-kinase gamma. A regulation of autonomicCa(2)+ oscillations in cardiac cells. J. Cell Biol. 152, 717–728.

[18] Crackower, M.A., et al. (2002) Regulation of myocardial con-tractility and cell size by distinct PI3K-PTEN signaling pathways.Cell. 110, 737–749.

[19] Patrucco, E., et al. (2004) PI3Kgamma modulates the cardiacresponse to chronic pressure overload by distinct kinase-depen-dent and -independent effects. Cell. 118, 375–387.

[20] Chruscinski, A.J., Rohrer, D.K., Schauble, E., Desai, K.H.,Bernstein, D. and Kobilka, B.K. (1999) Targeted disruption of thebeta2 adrenergic receptor gene. J. Biol. Chem. 274, 16694–16700.

[21] Gehrmann, J., Meister, M., Maguire, C.T., Martins, D.C.,Hammer, P.E., Neer, E.J., Berul, C.I. and Mende, U. (2002)Impaired parasympathetic heart rate control in mice with areduction of functional G protein betagamma-subunits. Am. J.Physiol. Heart Circ. Physiol. 282, H445–H456.

[22] Bunemann, M., Brandts, B., zu Heringdorf, D.M., van Koppen,C.J., Jakobs, K.H. and Pott, L. (1995) Activation of muscarinicK+ current in guinea-pig atrial myocytes by sphingosine-1-phosphate. J. Physiol. 489, 701–777.

[23] Mery, P.F., Abi-Gerges, N., Vandecasteele, G., Jurevicius, J.,Eschenhagen, T. and Fischmeister, R. (1997) Muscarinic regula-tion of the L-type calcium current in isolated cardiac myocytes.Life Sci. 60, 1113–1120.

[24] Steinberg, S.F. (2001) PI3King the L-type calcium channelactivation mechanism. Circ. Res. 89, 641–644.

[25] Viard, P., Exner, T., Maier, U., Mironneau, J., Nurnberg, B. andMacrez, N. (1999) Gbetagamma dimers stimulate vascular L-typeCa2+ channels via phosphoinositide 3-kinase. FASEB J. 13, 685–694.

[26] Naga Prasad, S.V., Laporte, S.A., Chamberlain, D., Caron,M.G., Barak, L. and Rockman, H.A. (2002) Phosphoinositide 3-kinase regulates beta2-adrenergic receptor endocytosis by AP-2

140 G. Alloatti et al. / FEBS Letters 579 (2005) 133–140

recruitment to the receptor/beta-arrestin complex. J. Cell Biol.158, 563–575.

[27] Oudit, G.Y., et al. (2003) Phosphoinositide 3-kinase gamma-deficient mice are protected from isoproterenol-induced heartfailure. Circulation. 108, 2147–2152.

[28] Nienaber, J.J., Tachibana, H., Naga Prasad, S.V., Esposito, G.,Wu, D., Mao, L. and Rockman, H.A. (2003) Inhibition ofreceptor-localized PI3K preserves cardiac beta-adrenergic recep-tor function and ameliorates pressure overload heart failure. J.Clin. Invest. 112, 1067–1079.

[29] Wang, Y.X., Dhulipala, P.D., Li, L., Benovic, J.L. andKotlikoff, M.I. (1999) Coupling of M2 muscarinic receptorsto membrane ion channels via phosphoinositide 3-kinasegamma and atypical protein kinase C. J. Biol. Chem. 274,13859–13864.

[30] Harvey, R.D. and Belevych, A.E. (2003) Muscarinic regulation ofcardiac ion channels. Br. J. Pharmacol. 139, 1074–1084.

[31] Biel, M., Schneider, A. and Wahl, C. (2002) Cardiac HCNchannels: structure, function, and modulation. Trends Cardio-vasc. Med. 12, 206–212.