The FASEB Journal express article 10.1096/fj.02-0292fje. Published online August 21, 2002. Augmentation of cardiac contractility with no change in L-type Ca2+ current in transgenic mice with a cardiac-directed expression of the human adenylyl cyclase type 8 (AC8) Marie Georget,* Philippe Mateo,* Grégoire Vandecasteele,* Jonas Jurevičius,*, Larissa Lipskaia, Nicole Defer, Jacques Hanoune, Jacqueline Hoerter,* and Rodolphe Fischmeister* *Laboratoire de Cardiologie Cellulaire et Moléculaire, INSERM U-446, Université Paris-Sud, Faculté de Pharmacie, F-92296 Châtenay-Malabry, France, and Laboratoire de Régulation des Gènes et Signalisation Cellulaire, INSERM U-99, Hôpital Henri-Mondor, F-94010 Créteil, France Corresponding author: Rodolphe Fischmeister, INSERM U-446, Université Paris-Sud, Faculté de Pharmacie, 5, Rue J.-B. Clément, F-92296 Châtenay-Malabry Cedex, France. E-mail: Fisch@vjf.inserm.fr Current address: Laboratory of Membrane Biophysics, Institute of Cardiology, Kaunas University of Medicine, 3007 Kaunas, Lithuania. E-mail: jojur@kmu.lt ABSTRACT The β-adrenergic cascade is severely impaired in heart failure (HF), in part because of a reduction in the activity of the two dominant cardiac adenylyl cyclase (AC) isoforms, AC5 and AC6. Hence, cardiac-directed AC overexpression is a conceivable therapeutic strategy in HF. In this study, we explored the consequences at the cellular and organ level of a cardiac-directed expression of the human AC8 in the transgenic mouse line AC8TG. Unlike AC5 and AC6, which are inhibited by intracellular Ca2+, AC8 is stimulated by Ca2+-calmodulin. Langendorff perfused hearts from AC8TG mice had a twofold higher left ventricular systolic pressure, a 40% faster heart rate, a 37% faster relaxation, and a 30% higher sensitivity to external Ca2+ than nontransgenic control mice (NTG). Cell shortening measured in isolated ventricular myocytes developed 22% faster and relaxed 43% faster in AC8TG than in NTG mice. Likewise, Ca2+ transients measured in fluo-3 AM-loaded myocytes were 30% higher and relaxed 24% faster in AC8TG compared with NTG mice. In spite of the large increase in Ca2+ transients and contraction, expression of AC8 had no effect on the whole-cell L-type Ca2+ current (ICa,L) amplitude. Moreover, ICa,L was unchanged even when AC8 was activated by raising intracellular Ca2+. Thus, cardiac expression of AC8 leads to an increase in cAMP that activates specifically Ca2+ uptake into the sarcoplasmic reticulum but not Ca2+ influx at the sarcolemma, suggesting a strong compartmentation of the cAMP signal. Key words: cAMP • isolated heart • calcium transient • cell shortening • compartmentation

A denylyl cyclase (AC) is one of the key proteins involved in the autonomic regulation of heart function. Noradrenaline, the main transmitter of the sympathetic nervous system, acting on β-adrenoceptors (βAR), activates Gs proteins, which stimulate AC activity,

leading to cAMP production and phosphorylation of several substrate proteins via activation of cAMP-dependent protein kinase (PKA) (26). Hence, phosphorylation via PKA of L-type Ca2+ channels (32), phospholamban (30), myofibrillar protein troponin I (TnI) (40), and ryanodine receptors (RyR) (35, 38) leads to positive inotropic, lusitropic, and chronotropic effects on the heart. The β-adrenergic cascade is severely impaired in virtually every form of experimental heart failure and in the human syndrome (36). Several studies have shown that AC5 and/or AC6 protein level, the dominant AC isoforms expressed in heart, is reduced in this pathology (10, 14), leading to a decrease in total AC activity (27), which contributes to a decrease in PKA activity and cardiac contraction. Recently, cardiac-directed expression of AC5 revealed an increase in basal AC activity and cardiac function that is independent of βAR activation (34). In addition, the cross-breed between these animals and transgenic mice with a depressed cardiac contractility (Gαq overexpressing mice) resulted in animals with a normalized cardiac function (33). Unlike these AC5 transgenic animals, transgenic mice with a cardiac overexpression of AC6 present a normal basal cardiac function but an increased response to βAR stimulation (11). The same result was obtained with intracoronary injection of an adenovirus expressing AC6 in pig heart (18). These data suggest that AC expression might provide a potential means to restore cardiac contractility during heart failure. Because AC5 and AC6 are the dominant cardiac AC isoforms, it was logical to examine at first how overexpression of these isoforms affects cardiac function. However, this does not exclude the possibility that expression of a noncardiac AC isoform might be beneficial to heart function. One possible advantage of introducing an alien AC gene is that the gene or the expressed protein might escape from the regulatory pathways affecting the activity of its endogenous congeners. Alternatively, the alien AC protein might be addressed to a different location compared with native AC isoforms, leading to a different function. With that in mind, a transgenic mouse expressing in the heart human AC type 8 (AC8) was created (22). As shown previously (2, 5), AC8 is essentially expressed in the brain. Also, in this model, AC8 activity is stimulated by Ca2+-calmodulin, unlike AC5 or AC6, which are inhibited by intracellular Ca2+ ([Ca2+]i) (2, 4, 5). During β-adrenergic stimulation, the resulting elevation of [Ca2+]i normally inhibits AC5 and AC6, and this supposedly acts as a negative feedback on L-type Ca2+ channel activity, [Ca2+]i, and cardiac contractility. However, in AC8TG mice, an increase in [Ca2+]i should stimulate AC8 activity, leading to a further increase in [Ca2+]i. Intuitively, such a positive feedback might generate Ca-overload and be deleterious to the animals. However, AC8TG mice present no phenotypic alteration or any sign of cardiomyopathy in spite of a clear increase in cardiac AC and PKA activities (22). Moreover, noninvasive assessment of cardiac function by echocardiography showed a similar basal cardiac function in AC8TG mice and their control littermates (NTG). However, hemodynamic evaluation after a bilateral vagotomy revealed a higher basal intrinsic contractility in AC8TG vs. NTG mice, indicating that AC8 is indeed functionally active in the heart of these transgenic animals (22).

In this study, our aim was to examine the adaptive changes taking place within cardiac myocytes as a result of AC8 expression. To do this, we examined the functional properties of cardiac contraction in AC8TG and NTG mice measured at the level of the isolated heart and the isolated ventricular myocyte. Because we suspected that calcium signaling might be affected in this transgenic mouse model, we also measured [Ca2+]i transients and whole-cell L-type Ca2+ channel current (ICa,L) to examine how a cardiac expression of AC8 modifies the influx of Ca2+ and the sarcoplasmic reticulum (SR) function. MATERIALS AND METHODS The investigation conforms with our institution guidelines defined by the European Community guiding principles in the care and use of animals and the French decree n°87/848 of October 19, 1987. Authorizations to perform animal experiments according to this decree were obtained from the French Ministère de l'Agriculture, de la Pêche et de lAlimentation (n°7475, May 27, 1997). Ex vivo physiology Four- to five-month-old mice were anaesthetized by i.p. injection of pentothal (150 mg/kg). The heart was quickly removed, placed in oxygenated Krebs-Henseleit solution (95% O2 and 5% CO2, pH 7.35) containing low calcium concentration (0.4 mM) and heparin. The aorta was cannulated on a 20-gauge cannula and perfused by the Langendorff method with Krebs-Henseleit solution at constant pressure (75 mmHg). A small homemade latex balloon was inserted in the left ventricle (LV) chamber. The temperature was monitored online and maintained at 37±0.2°C throughout the experiments. All hearts were initially perfused with oxygenated Krebs-Henseleit solution at a perfusate Ca2+ concentration ([Ca2+]o) of 1.8 mM, and the LV balloon was progressively inflated to isovolumic condition of work corresponding to an end diastolic pressure (EDP) of 58 mmHg. The water-filled balloon was connected to a pressure transducer (Statham gauge Ohmeda, Bilthoven, Holland) for continuous measurement of LV systolic pressure (LVSP) on a paper recorder (Astro-Med, West Warrick, UK). Contractile parameters were measured online by programming a PC-compatible 486/50 microcomputer in Assembly language (Borland) to determine heart rate (HR), LVSP, rate-pressure product (RPP=LVSP×HR, used as an index of contractility), EDP, and the following kinetic parameters: time to peak of contraction (ttp), time for half relaxation (tR1/2), and the first derivatives of LVSP [+d(LVSP)/dt and d(LVSP)/dt]. To dissociate the specific changes in the kinetics of contraction and relaxation from those resulting from a modification of LVSP, these first derivatives were also expressed relative to LVSP [+d(LVSP)/dt/LSVP and d(LVSP)/dt/LSVP]. Part of the experiments was done under pacing conditions at a reduced [Ca2+]o (1 mM). In this case, the hearts were stimulated at 680 bpm via platinum wires (stimulator HSE, Hugo Sachs Electronics, March-Hugstetten, Germany). Ten minutes of equilibration in isovolumic working conditions were imposed before beginning the experiment. Preparation of mouse ventricular myocytes Ventricular myocytes were isolated from the hearts of NTG and AC8TG mice as previously described (37). Three- to four-month-old animals were anaesthetised by i.p. injection of

pentothal (150 mg/kg), and the heart was quickly removed and placed into a cold Ca2+-free Tyrodes solution. The aorta was cannulated, and the heart was perfused with an oxygenated Ca2+-free Tyrodes solution for 35 min, using retrograde Langendorff perfusion at 37°C. For enzymatic dissociation, the heart was perfused with Ca2+-free Tyrodes solution containing collagenase D and protease XXIV for 10 min. Then, the heart was removed and placed in a dish containing 0.2 mM Ca2+-Tyrodes solution. The ventricles were separated from atria, cut in small pieces, and triturated with a pipette to disagregate myocytes. Ventricular cells were filtered on gauze and allowed to sediment by gravity for 10 min. The supernatant was removed, and cells were suspended in 0.5 mM Ca2+-Tyrodes solution. The procedure was repeated, and the cells were suspended in 1 mM Ca2+-Tyrodes solution. The cells were stored at room temperature until use. Measurement of cell shortening Free-load cell shortening was measured in ventricular cardiomyocytes from NTG and AC8TG mice. The cells were electrically stimulated by external platinum electrodes placed in the bath around the cell. All images of the contracting myocyte were recorded at 200 frame/s, with an intensified high-speed Lhesa CCD camera (Lheritier SA, Cergy-Pontoise, France) and acquired on PC computer. The time course of the change in cell length was extracted off-line by tracing the cell ends in line-profile images. Myocyte shortening was calculated as the percent change in myocyte length at any given time with respect to the length at rest. In addition, the time to peak of contraction (tpeak), that is, the time interval between electrical stimulation and maximal shortening, and the time for half relaxation (t50), that is, the time interval between peak and half-maximal recovery, were calculated from the recordings. Measurement of [Ca2+]i transients To detect [Ca2+]i transients, we used the fluorescent calcium indicator fluo-3 acetoxymethyl ester (fluo-3 AM, Molecular Probes, Leiden, The Netherlands). Myocytes were incubated with 10 µM fluo-3 AM (dissolved in dimethyl sulfoxide [DMSO]) for 15 min. Then, the cells were washed in buffer solution for 30 min to allow sufficient time for deesterification. Subsequently, the cells were electrically stimulated by external platinum electrodes, and the fluorescent signal from the cell was monitored at a wavelength of 530 nm with an excitation wavelength of 488 nm. All images were acquired with the intensified high-speed Lhesa CCD camera and recorded on a PC microcomputer. [Ca2+]i transients were expressed as F/F0, where F is the mean fluorescence of the cell at any given time and F0 is the mean fluorescence at rest. L-type Ca2+ channel current measurement The whole-cell configuration of the patch-clamp technique was used to record L-type Ca2+ channel current (ICa,L). For routine ICa,L measurements, the cells were depolarized every 8 s from a holding potential of 50 mV to 0 mV for 400 ms. At this holding potential, fast Na+ current and T-type Ca2+ current were inactivated. External and internal (pipette) solutions contained Cs+ instead of K+ to eliminate K+ currents. The electrical resistance of the patch pipette varied between 0.5 and 1.2 MΩ. All experiments were done at room temperature. A challenger VM (Kinetic, Atlanta, GA) generated voltage-clamp protocols. The currents were measured by a

patch-clamp amplifier (Biologic RK 400, Claix, France), filtered at 3 kHz, and sampled at a frequency of 10 kHz. Cells were exposed to different drugs with a system of capillary tubings. The maximal amplitude of ICa,L was measured as the difference between the peak inward current at 0 mV and the current at the end of 400 ms pulse. Currents were not compensated for capacitive and leak currents. Two mV and 10 ms depolarizations from the holding potential were used to elicit capacitive currents, which were filtered at 10 kHz and sampled at a frequency of 20 kHz. Exponential analysis of these currents led to the estimate of the membrane capacitance and series resistance. Solutions and drugs The composition of the initial Krebs-Henseleit solution used for ex vivo experiments was (in mM) NaCl 113, NaHCO3 25, KCl 4.7, KH2PO4 1.2, MgSO4 1.2, CaCl2 1.8, glucose 11, and Na-pyruvate 10. To investigate the contractile response to external Ca2+ concentration, [Ca2+]o was varied from 0.5 to 2.5 mM. The ionic composition of Ca2+-free Tyrodes solution used for cell isolation was (in mM) NaCl 140, KCl 5.4, NaH2PO4 1, MgCl2 1.1, HEPES 5, and glucose 10. The pH was adjusted to 7.3 with NaOH. The enzymatic solution for the cardiomyocyte dissociation was composed of Ca2+-free Tyrodes solution to which was added 4.05 units collagenase D (Roche, Meylan, France), 17.8 units protease XXIV (Sigma, St Quentin Fallavier, France), 1 mg/ml bovine serum albumin (BSA; ICN Pharmaceuticals, Orsay, France), and 1.4% fetal calf serum (FCS; Life Technology, Eragny/Oise, France). The 0.2, 0.5, and 1 mM Ca2+-Tyrodes solutions were composed of Ca2+-free Tyrodes solution to which 1.4% FCS and 0.2, 0.5, and 1 mM Ca2+ were added, respectively. The composition of the control external solution for the patch-clamp experiment was (in mM) NaCl 107.1, CsCl 20, HEPES 10, glucose 5, NaHCO3 4, NaH2PO4 0.8, MgCl2 1.8, and CaCl2 1.8. The pH was adjusted to 7.4 with NaOH. Patch pipettes were filled with two types of internal solutions. The first one was a low Ca2+ solution (3 nM free Ca2+ concentration) and contained (in mM) CsCl 119.8, HEPES 10, EGTA 5, MgCl2 4, Na2CP 5, CaCl2 0.0062, Na2ATP 3.1, and Na2GTP 0.4. The second solution was designed to contain a higher free Ca2+ concentration (300 nM) and contained (in mM) CsCl 132.8, HEPES 10, EGTA 0.04, MgCl2 3.8, Na2CP 5, CaCl2 0.03, Na2ATP 3.1, and Na2GTP 0.4. The pH was adjusted to 7.3 with CsOH in both solutions. Data analysis The results are expressed as means ± SE. Students paired t test was used to test the effect of drugs on NTG and AC8TG hearts. Students t test for unpaired data or ANOVA followed by a Student-Newman-Keuls test were used to compare NTG and AC8TG mice. A value of P < 0.05 was considered to be statistically significant. RESULTS Characteristics of the animals For 2-month-old AC8TG mice (22), anatomical examination of 4- to 5-month-old AC8TG and NTG showed no obvious difference between the animals. Body weight (NTG, 30.9±0.4 g; AC8TG, 30.7±0.4 g), tibia length (NTG, 1.8±0.01 cm; AC8TG, 1.8±0.01 cm), dry heart weight

(NTG, 30.7±0.6 mg; AC8TG, 31.8±1.8 mg), heart-to-body weight ratio (NTG, 0.99±0.02 mg/g; AC8TG, 1.03±0.06 mg/g) and heart weight-to-tibia length ratio (NTG, 17.2±0.4 mg/cm; AC8TG, 18.0±1.0 mg/cm) were unchanged by transgene expression (NTG mice, n=6; AC8TG mice, n=8). Influence of AC8 expression on the spontaneous contractility of the heart Myocardial contractility was assessed at the peak of the pressure volume curve after progressive inflation of the LV balloon. The coronary flow was similar in NTG (2.8±0.2 ml/min, n=6) and AC8TG mice (3.4±0.5 ml/min, n=7) at [Ca2+]o=1.8 mM. Baseline spontaneous cardiac function was compared in 13 AC8TG and 12 NTG mice (Fig. 1). LVSP was twofold higher in AC8TG than in NTG (110± 9mmHg vs. 56±5 mmHg, P<0.001); spontaneous HR was significantly faster in AC8TG than in NTG hearts (481±21 bpm vs. 337± 15 bpm, P<0.001); and EDP was similar. As a result, the rate-pressure product (RPP) was nearly threefold higher in AC8TG compared with NTG mice, demonstrating that specific cardiac expression of AC8 enhanced basal spontaneous function of the isolated heart. Because AC8 activity is stimulated by Ca2+-calmodulin, we explored the influence of a change in [Ca2+]o on the contractility of both groups of animals. As shown in Figure 2, increasing [Ca2+]o from 0.5 to 2.5 mM resulted in a dose-dependent increase in LVSP in both AC8TG and NTG hearts (Fig. 2A) without alteration in HR (Fig. 2B). At [Ca2+]o concentrations up to 1.8 mM, LVSP remained significantly higher in AC8TG than in NTG. However the sensitivity to [Ca2+]o appeared enhanced in AC8TG hearts. Indeed, increasing [Ca2+]o from 1.8 to 2.5 mM did not significantly change the contractility in AC8TG hearts but still increased LVSP in NTG (from 61±8 to 90±11 mmHg, P<0.05). Moreover, the mean EC50 for [Ca2+]o, determined by curve fitting on individual hearts using the Hill equation (Fig. 2A), was significantly lower in AC8TG (1.05±0.07 mM) than in NTG mice (1.51±0.12 mM, P<0.05).

Influence of AC8 expression on the contractility of the paced heart Because cardiac expression of AC8 results in a strong acceleration of HR, which in itself might affect the contractile properties of the hearts, the following experiments were performed under pacing at 680 bpm in six NTG and six AC8TG mice. The measurements were performed at 1 mM [Ca2+]o. At this concentration, cardiac work was not yet at maximum in both types of mice. Under this condition, the coronary flow was 2.9 ± 0.4 ml/min and 3.3 ± 0.3 ml/min in NTG and AC8TG mice, respectively. Under pacing conditions, AC8TG mice still developed a significantly higher LVSP than did NTG mice (Table 1). A detailed analysis of the kinetics of contraction and relaxation revealed that the maximal rate of contraction, measured as the maximal value of +d(LVSP)/dt, was significantly higher in AC8TG mice (3613±483 mmHg/s) compared with NTG mice (1392±38 mmHg/s, P<0.01). However, this difference solely resulted from the increase in LVSP. When normalized to LVSP, both the maximal rate of contraction [+d(LVSP)/dt/LVSP] and the time to peak of contraction (ttp) were similar in both groups of mice (respectively: 52±1 s1 and 26±1 ms in NTG; 56±3 s1 and 24±1 ms in AC8TG mice). All relaxation parameters were markedly

affected by AC8 expression. Indeed, the maximal rate of relaxation [d(LVSP)/dt] was about threefold higher in AC8TG mice (2508±318 mmHg/s) compared with NTG mice (788±73 mmHg/s). However, by contrast to the onset of contraction, the rate of relaxation was still ∼26% higher when normalized to LVSP [d(LVSP)/dt/LVSP: 39±3 s1 in AC8TG mice; 29±1 s1 in NTG mice]. Likewise, the half-time of relaxation (tR1/2) was 37% shorter in AC8TG mice (20±1 ms) compared with NTG mice (32±2 ms). Thus, cardiac AC8 expression induced a marked acceleration of relaxation.

Measurement of cell shortening To determine whether AC8 cardiac expression also modified basal contractility of ventricular myocytes, we measured cell shortening in isolated ventricular myocytes from AC8TG and NTG mice. Figure 3A shows that the amplitude of cell shortening was not statistically different between NTG and AC8TG mice (18.7±2%, n=8 in NTG mice and 21.3±0.9%, n=11 in AC8TG mice). This was because the myocytes were not attached at the bottom of the dish, so that the cells contracted free of load and shortening was maximal in both cases. However, the time to peak (tpeak) was significantly shorter (159±10 ms, n=8, in NTG mice and 126±7 ms, n=11, in AC8TG mice, P<0.01) and the time for half relaxation (t50) was much faster in transgenic myocytes (160±24 ms, n=8, in NTG mice and 108±10 ms, n=11, in AC8TG mice, P<0.05) (see Fig. 3B and 3C). On average, t50 and the tpeak were reduced by 43% and 22%, respectively, in AC8TG mice.

Measurement of [Ca2+]i transients Because contraction and relaxation kinetics are largely dependent on Ca2+ release from and uptake into the SR, we measured [Ca2+]i transients in AC8TG and NTG myocytes, using the fluorescent Ca2+ indicator fluo-3 AM. As shown in Figure 4A, [Ca2+]i transients were larger in amplitude in AC8TG than in NTG mice (peak F/Fo was 2.1±0.1, n=16, and 2.7±0.2, n=24, respectively, in NTG and AC8TG mice, P=0.02). Moreover, a kinetic analysis showed that [Ca2+]i transient developed with a similar time course in AC8TG and NTG mice but relaxed substantially faster in AC8TG (Fig. 4B and 4C). Indeed, half maximal relaxation of [Ca2+]i transient (t50) was reduced from 173.8±13.8 ms, n=16, in NTG to 140±9.6 ms, n=24, in AC8TG mice (P<0.05).

Calcium current measurements Because the amplitude of cardiac contraction is largely dependent on the activity of L-type Ca2+ channels, which initiate the calcium-induced calcium-release mechanism, we examined how cardiac expression of AC8 affects ICa,L in isolated ventricular myocytes. After breaking the membrane patch and reaching the whole-cell configuration, cell membrane capacitance (Cm) was measured in AC8TG and NTG myocytes. Both mice had a similar Cm (177±9 pF, n=16, in NTG; 175±7 pF, n=18, in AC8TG), indicating that on average the cells had a similar size and confirming the absence of cell hypertrophy in AC8TG mice. The cells were then voltage-clamped at a holding potential of 50 mV, and ICa,L was recorded in control solution during 400 ms depolarizations to 0 mV. Figure 5A shows two representative traces of ICa,L recorded from NTG and AC8TG ventricular myocytes. Averaged ICa,L amplitudes were 766±59 pA, n = 18, in

AC8TG myocytes and 911±128 pA, n = 16, in NTG myocytes and were not statistically different (see Fig. 5B). Likewise, when normalized to cell membrane capacitance, ICa,L current density was similar in AC8TG and NTG ventricular myocytes: 5.2±0.4 pA/pF, n=18, in AC8TG vs. 4.2±0.4 pA/pF, n=16, in NTG (see Fig. 5C). Thus, AC8 expression did not modify basal ICa,L amplitude and current density at 0 mV. In addition, as shown in Figure 5D, AC8 expression did not modify the voltage dependence of peak ICa,L. In these experiments, ICa,L was measured with a patch pipette containing a high (5 mM) EGTA concentration together with 6.2 µM total Ca2+, so that free [Ca2+]i was estimated to be ∼3 nM. Because the activity of AC8 is stimulated by Ca-calmodulin, one could imagine that AC8 was inactive under our experimental conditions, because [Ca2+]i was too low. To test this hypothesis, we designed another series of experiments in which the cells were dialyzed with a much lower EGTA concentration (40 µM) together with 30 µM total Ca2+, so that free [Ca2+]i was estimated to be ∼300 nM. Reducing the Ca2+ buffering capacity of the pipette solution led to an increase in basal ICa,L amplitude. However, the current amplitudes were still similar in AC8TG and NTG mice (1170.5±123.9 pA, n=20, for NTG vs. 1405.3±140.1 pA, n=22, for AC8TG). In another set of experiments, the effect of increasing the stimulation frequency under voltage-clamp conditions was tested on ventricular myocytes from NTG and AC8TG hearts. The rationale behind these experiments was that increasing the stimulation frequency should lead to a progressive increase of [Ca2+]i in a submembrane compartment, and this may result in a stimulation of cAMP synthesis through Ca2+ activation of AC8 in AC8TG mice, but not in NTG mice. The cells were depolarized at 0 mV during 40 ms from a 50 mV holding potential, and the stimulation frequency was progressively increased from 0.125 Hz (control frequency) to 8 Hz in twofold steps, with 10 depolarizations applied at each frequency. As shown in Figure 6, increasing the stimulation frequency elicited a similar negative staircase phenomenon on ICa,L amplitude, with no significant difference between the two animal models. Altogether, these experiments demonstrate that cardiac-directed expression of AC8 leads to enhanced cardiac function that does not involve an increase in ICa,L. DISCUSSION A transgenic mouse with a cardiac-specific expression of human AC8 isoform has been created (22). AC8 is normally present in the brain; absent in the heart; and unlike endogenous cardiac AC5 and AC6, which are inhibited by Ca2+, at least in vitro; AC8 stimulated by Ca2+-calmodulin (2, 5, 22). In this study, we show that 1) cardiac-specific expression of AC8 is associated with a strong enhancement of ex vivo basal cardiac function and an increased sensitivity of cardiac contraction to [Ca2+]o; 2) at the level of the isolated ventricular myocyte, cell shortening developed faster and relaxed faster in AC8TG than in NTG mice; 3) [Ca2+]i transient amplitude was enhanced and its decay was faster in AC8TG vs. NTG mice, which indicates a faster re-uptake of Ca2+ by the SR; and 4) in spite of the large increase in Ca2+ transients and contraction, ICa,L amplitude was unchanged in AC8TG, even when AC8 was activated by raising [Ca2+]i. Myocardial overexpression of endogenous or exogenous AC isoforms leads to different cardiac phenotypes. In AC6 transgenic mice, cAMP production is not altered under basal conditions but is largely enhanced upon β-adrenergic stimulation (11), whereas in AC8TG (22) and AC5

transgenic mice (34), basal AC activity is increased. In AC8TG mice, echocardiography assessment showed no modification in cardiac function (22), whereas basal contraction was enhanced under in vivo (22) and ex vivo recordings (as shown by this study). Increasing AC6 levels are associated with an enhancement of in vivo basal cardiac contraction and HR (measurement performed with a bilateral vagotomy), which are not modified under echocardiography and ex vivo recordings (11). The heterogeneity of these results suggests that cardiac function does not simply depend on the global AC activity and cAMP level but that each AC isoform could provide a specific modulation of myocardial contraction, for instance, different localization at the cell membrane leading to different interaction with subcellular compartments (16). Myocardial contractility is critically dependent on Ca2+ influx via the L-type Ca2+ channels (32), which suggests a possible alteration of ICa,L to explain the strong increase in cardiac function observed in AC8TG mice. To our surprise, ICa,L current amplitude was similar in ventricular myocyte of AC8TG and NTG mice. Moreover, increasing intracellular Ca2+ concentration by either reducing the Ca2+ buffering capacity of the pipette solution or by increasing the stimulation frequency did not generate any difference between ICa,L amplitude in NTG and AC8TG mice. This finding led us to conclude that the observed increase in contractility in AC8TG mice is not due to a stimulation of L-type Ca2+ channel activity. In fact, it seems as if there was no crosstalk between AC8 and Ca2+ channels in AC8TG mice: cAMP production by AC8 does not lead to Ca2+ channel phosphorylation, and Ca2+ influx via ICa,L does not lead to AC8 stimulation. It is likely that adaptive changes occurred in AC8TG mice to avoid a positive feedback between AC8 and intracellular Ca2+, hence preserving cardiac myocytes from Ca2+ overload. Although cAMP produced by AC8 appears not to be accessible to Ca2+ channels, it clearly is available for other elements of the EC coupling. Similar results have been reported for the transgenic mouse (TG4) overexpressing the β2-adrenoreceptor (β2AR) (25, 29, 4143). The TG4 mouse heart expresses a large amount of constitutively active β2AR, which leads to an increase in basal AC activity. However, the cardiac contractility is enhanced without any alteration in basal ICa,L (43), suggesting a compartmentation of β2AR-mediated signaling (16, 31, 43). If the increased cardiac contractility in AC8TG mice is not due to an increase in Ca2+ influx via L-type Ca2+ channels, what other mechanism(s) might be involved? For several reasons, we believe that an enhanced SR function might explain the functional alterations observed in AC8TG mice. First, in vivo hemodynamic studies showed an acceleration in the cardiac contractile parameters in AC8TG vs. NTG mice (22). Second, in this study, using the ex vivo Langendorff perfused heart model, we found a marked acceleration in the contractile relaxation in AC8TG mice compared with NTG mice. Third, at the level of the single ventricular myocyte, both cell shortening and calcium transient relaxed much faster in AC8TG than in NTG mice. Thus, cardiac-directed expression of AC8 might specifically accelerate Ca2+ re-uptake into the SR without affecting Ca2+ influx into the cell. This could be achieved by an increase in the total amount of SR Ca2+-ATPase (SERCA), an alleviated inhibitory effect of phospholamban (PLB), its key regulator protein (30), due to PKA phosphorylation, or both. Interestingly, knockout mice lacking PLB develop a very similar phenotype to the AC8TG mice: an increased cardiac contraction and relaxation (13, 23, 24) and, at the cell ventricular level, a shorter time to peak, a

faster relaxation phase of contraction, and a faster decay of Ca2+ transient (19, 20). In TG4 mice, in vivo cardiac relaxation is enhanced and associated with a reduction of PLB (28). Consistent with these studies, AC8 cardiac expression could affect the content and/or phosphorylation of PLB, leading to an enhanced cardiac contraction and relaxation as observed in our model. AC8TG mice might thus provide a unique model of cAMP compartmentation, whereby cAMP produced by AC8 specifically activates Ca2+ uptake into the SR without increasing Ca2+ influx at the sarcolemma. However, other mechanisms cannot be excluded, such as an enhancement of myofilament Ca2+ sensitivity due to PKA-phosphorylation of troponin I (26, 39, 40), an overexpression of SERCA2a, the cardiac SERCA isoform, or βARK inhibitor, which have been shown independently to increase cardiac contractility (12, 17). Several transgenic models overexpressing a protein involved in the β-adrenergic signaling cascade develop a cardiomyopathy (6, 36). Overexpression of β2AR leads to a constitutive increase in cardiac contractility in adult TG4 mice (1, 25), but old TG4 mice with a high level of β2AR expression develop a cardiomyopathy (8, 21). Moreover, ischemia and heart failure have more deleterious consequences in TG4 than in NTG animals (3, 7). Increasing β1-receptor or Gαs protein levels leads to an increase in heart rate, a decrease in cardiac contractility, and a progressive development of hypertrophy and heart failure (9, 15). Surprisingly, AC8TG mice do not show any detrimental consequences of cardiac expression of AC8, at least up to the age of 14 months (N. Defer, unpublished data). Unlike AC6 transgenic mice, which need to be activated by the β-adrenergic pathway to improve cardiac function (11), AC8TG mice develop an enhanced cardiac function under basal conditions. Thus, cardiac expression of AC8 might provide a means to increase heart function even when the β-adrenergic signaling cascade is down-regulated at the receptor level, as during heart failure. Additional experiments, such as gene transfer in adult animal models of heart failure, are required to examine whether cardiac-directed expression of AC8 could provide a potential means for the treatment of heart failure. ACKNOWLEDGMENTS We thank Patrick Lechêne for skillful programming of the automated data analysis and Dr. Ingrid L. Grupp from the Department of Pharmacology and Cell Biophysics at the University of Cincinnati for expert teaching of the mouse Langendorff preparation. REFERENCES

1. Bittner, H. B., Chen, E. P., Milano, C. A., Lefkowitz, R. J., and Van Trigt, P. (1997) Functional analysis of myocardial performance in murine hearts overexpressing the human β2-adrenergic receptor. J. Mol. Cell. Cardiol. 29, 961967

2. Cali, J. J., Zwaagstra, J. C., Mons, N., Cooper, D. M., and Krupinski, J. (1994) Type VIII adenylyl cyclase. A Ca2+/calmodulin-stimulated enzyme expressed in discrete regions of rat brain. J. Biol. Chem. 269, 1219012195

3. Cross, H. R., Steenbergen, C., Lefkowitz, R. J., Koch, W. J., and Murhy, E. (1999) Overexpression of the cardiac β2-adrenergic receptor and expression of a β-adrenergic receptor kinase-1 (βARK1) inhibitor both increase myocardial contractility but have differential effects

on susceptibility to ischemic injury. Circ. Res. 85, 10771084

4. Defer, N., Best-Belpomme, M., and Hanoune, J. (2000) Tissue specificity and physiological relevance of various isoforme of adenylyl cyclase. Am. J. Physiol. 279, F400F416

5. Defer, N., Marinx, O., Stengel, D., Danisova, A., Iourgenko, V., Matsuoka, I., Caput, D., and Hanoune, J. (1994) Molecular cloning of the human type VIII adenylyl cyclase. FEBS Lett. 351, 109113

6. Du, X. (2001) Sympathoadrenergic mechanisms in functional regulation and development of cardiac hypertrophy and failure: findings from genetically engineered mice. Cardiovasc. Res. 50, 443453

7. Du, X. J., Autelitano, D. J., Dilley R.J., Wang, B., Dart, A. M., and Woodcock, E. A. (2000) β2-adrenergic receptor overexpression exacerbates development of heart failure after aortic stenosis. Circulation 101, 7177

8. Du, X. J., Gao, X. M., Wang, B., Jennings, G. L., Woodcock, E., and Dart, A. M. (2000) Age-dependent cardiomyopathy and heart failure phenotype in mice overexpressing β2-adrenergic receptors in the heart. Cardiovasc. Res. 48, 448454

9. Engelhardt, S., Hein, L., Wiesmann, F., and Lohse, M. J. (1999) Progressive hypertrophy and heart failure in β1-adrenergic receptor transgenic mice. Proc. Natl. Acad. Sci. USA 96, 70597064

10. Espinasse, I., Iourgenko, V., Richer, C., Heimburger, M., Defer, N., Bourin, M. C., Samson, F., Pussard, E., Giudicelli, J. F., Michel, J. B., Hanoune, J., and Mercadier, J. J. (1999) Decreased type VI adenylyl cyclase mRNA concentration and Mg2+-dependent adenylyl cyclase activities and unchanged type V adenylyl cyclase mRNA concentration and Mn2+-dependent adenylyl cyclase activities in the left ventricle of rats with myocardial infarction and longstanding heart failure. Cardiovasc. Res. 42, 8798

11. Gao, M. H., Lai, N. C., Roth, D. M., Zhou, J. Y., Zhu, J., Anzai, T., Dalton, N., and Hammond, H. K. (1999) Adenylylcyclase increases responsiveness to catecholamine stimulation in transgenic mice. Circulation 99, 16181622

12. He, H., Giordano, F. J., Hilal-Dandan, R., Choi, D. J., Rockman, H. A., McDonough, P. M., Bluhm, W. F., Meyer, M., Sayen, M. R., Swanson, E., and Dillmann, W. H. (1997) Overexpression of the rat sarcoplasmic reticulum Ca2+ ATPase gene in the heart of transgenic mice accelerates calcium transients and cardiac relaxation. J. Clin. Invest. 100, 380389

13. Hoit, B. D., Khoury, E. G., Ball, N., and Walsh, R. A. (1995) In vivo echocardiographic detection of enhanced left ventricular function in gene-targeted mice with phospholamban deficiency. Circ. Res. 77, 632637

14. Ishikawa, Y., Sorota, S., Kiuchi, K., Shannon, R. P., Komamura, K., Katsushika, S.,

Vatner, D. E., Vatner, S. F., and Homcy, C. J. (1994) Downregulation of adenylylcyclase types V and VI mRNA levels in pacing-induced heart failure in dogs. J. Clin. Invest. 93, 22242229

15. Iwase, M., Uechi, M., Vatner, D. E., Asai, K., Shannon, R. P., Kudej, R. K., Wagner, T. E., Wight, D. C., Patrick, T. A., Ishikawa, Y., Homcy, C. J., and Vatner, S. F. (1997) Cardiomyopathy induced by cardiac Gsα overexpression - Rapid Communication. Am. J. Physiol. 41, H585H589

16. Jurevičius, J., and Fischmeister, R. (1996) cAMP compartmentation is responsible for a local activation of cardiac Ca2+ channels by β-adrenergic agonists. Proc. Natl. Acad. Sci. USA 93, 295299

17. Koch, W. J., Rockman, H. A., Samama, P., Hamilton, R. A., Bond, R. A., Milano, C. A., and Lefkowitz, R. J. (1995) Cardiac function in mice overexpressing the β-adrenergic receptor kinase or a βARK inhibitor. Science 268, 13501353

18. Lai, N. C., Roth, D. M., Gao, M. H., Fine, S., Head, B. P., Zhu, J., McKirnan, M. D., Kwong, C., Dalton, N., Urasawa, K., Roth, D. A., and Hammond, H. K. (2000) Intracoronary delivery of adenovirus encoding adenylyl cyclase VI increases left ventricular function and cAMP-generating capacity. Circulation 102, 23962401

19. Li, L., Desantiago, J., Chu, G. X., Kranias, E. G., and Bers, D. M. (2000) Phosphorylation of phospholamban and troponin I in β-adrenergic-induced acceleration of cardiac relaxation. Am. J. Physiol. 278, H769H779

20. Li, L., Chu G., Kranias E.G., and Bers, D.M. (1998) Cardiac myocyte calcium transport in phospholamban knockout mouse: relaxation and endogenous CaMKII effects. Am. J. Physiol. 43, H1335H1347

21. Liggett, S. B., Tepe, N. M., Lorenz, J. N., Canning, A. M., Jantz, T. D., Mitarai, S., Yatani, A., Dorn, G. W. (2000) Early and delayed consequences of β2-adrenergic receptor over expression in mouse heart. Circulation 101, 17071714

22. Lipskaia, L., Defer, N., Esposito, G., Hajar, I., Garel, M. C., Rockman, H. A., and Hanoune, J. (2000) Enhanced cardiac function in transgenic mice expressing a Ca2+-stimulated adenyl cyclase. Circ. Res. 86, 795801

23. Lorenz, J. N., and Kranias, E. G. (1997) Regulatory effects of phospholamban on cardiac function in intact mice. Am. J. Physiol. 273, H2826H2831

24. Luo, W., Grupp, I. L., Harrer, J., Ponniah, S., Grupp, G., Duffy, J. J., Doetschman, T., and Kranias, E. G. (1994) Targeted ablation of the phospholamban gene is associated with markedly enhanced myocardial contractility and loss of β-agonist stimulation. Circ. Res. 75, 401409

25. Milano, C. A., Allen, L. F., Rockman, H. A., Dolber, P. C., Mcminn, T. R., Chien, K. R., Johnson, T. D., Bond, R. A., and Lefkowitz, R. J. (1994) Enhanced myocardial function in

transgenic mice overexpressing the β2-adrenergic receptor. Science 264, 582586

26. Rapundalo, S. T. (1998) Cardiac protein phosphorylation: functional and pathophysiological correlates. Cardiovasc. Res. 38, 559588

27. Reithmann, C., Reber, D., Kozlik-Feldmann, R., Netz, H., Pilz, G., Welz, A., and Werdan, K. (1997) A post-receptor defect of adenylyl cyclase in severely failing myocardium from children with congenital heart disease. Eur. J. Pharm. 330, 7986

28. Rockman, H. A., Hamilton, R., Jones, L. R., Milano, C. A., Mao, L., and Lefkowitz, R. J. (1996) Enhanced myocardial relaxation in vivo in transgenic mice overexpressing the β2-adrenergic receptor is associated with reduced phospholamban protein. J. Clin. Invest. 97, 16181623

29. Rockman, H. A., Koch, W. J., and Lefkowitz, R. J. (1997) Cardiac function in genetically engineered mice with altered adrenergic receptor signaling. Am. J. Physiol. 41, H1553H1559

30. Simmerman, H. K. B., and Jones, L. R. (1998) Phospholamban: Protein structure, mechanism of action, and role in cardiac function. Physiol. Rev. 78, 921947

31. Steinberg, S. F., and Brunton, L. L. (2001) Compartmentation of G protein-coupled signaling pathways in cardiac myocytes. Annu. Rev. Pharmacol. Toxicol. 41, 751773

32. Striessnig, J. (1999) Pharmacology, structure and function of cardiac L-type Ca2+ channels. Cell. Physiol. Biochem. 9, 242269

33. Tepe, N. M., and Liggett, S. B. (1999) Transgenic replacement of type V adenylyl cyclase identifies a critical mechanism of β-adrenergic receptor dysfunction in the Gαq overexpressing mouse. FEBS Lett. 458, 236240

34. Tepe, N. M., Lorenz, J. N., Yatani, A., Dash, R., Kranias, E. G., Dorn, G. W. 2., and Liggett, S. B. (1999) Altering the receptor-effector ratio by transgenic overexpression of type V adenylyl cyclase: enhanced basal catalytic activity and function without increased cardiomyocyte β-adrenergic signalling. Biochemistry 38, 1670616713

35. Valdivia, H. H., Kaplan, J. H., Ellisdavies, G. C. R., and Lederer, W. J. (1995) Rapid adaptation of cardiac ryanodine receptors: Modulation by Mg2+ and phosphorylation. Science 267, 19972000

36. Wang, X., and Dhalla, N. S. (2000) Modification of β-adrenoceptor signal transduction pathway by genetic manipulation and heart failure. Mol. Cell. Biochem. 214, 131155

37. Wolska, B. M., and Solaro, R. J. (1996) Method for isolation of adult mouse cardiac myocytes for studies of contraction and microfluorimetry Am. J. Physiol. 40, H1250H1255

38. Yoshida, A., Takahashi, M., Imagawa, T., Shigekawa, M., Takisawa, H., and Nakamura, T.

(1992) Phosphorylation of ryanodine receptors in rat myocytes during β-adrenergic stimulation. J. Biochem. 111, 186190

39. Zhang, R., Zhao, J., Mandveno, A., and Potter, J. D. (1995) Cardiac troponin I phosphorylation increases the rate of cardiac muscle relaxation. Circ. Res. 76, 10281035

40. Zhang, R, Zhao, J, and Potter, J. D. (1995) Phosphorylation of both serine residues in cardiac troponin I is required to decrease the Ca2+ affinity of cardiac troponin C. J. Biol. Chem. 270, 3077330780

41. Zhou, Y. Y., Cheng, H. P., Bogdanov, K. Y., Hohl, C., Altschuld, R., Lakatta, E. G., and Xiao, R. P. (1997) Localized cAMP-dependent signaling mediates β2-adrenergic modulation of cardiac excitation-contraction coupling. Am. J. Physiol. 42, H1611H1618

42. Zhou, Y. Y., Cheng, H. P., Song, L. S., Wang, D. J., Lakatta, E. G., and Xiao, R. P. (1999) Spontaneous β2-adrenergic signaling fails to modulate L-type Ca2+ current in mouse ventricular myocytes. Mol. Pharmacol. 56, 485493

43. Zhou, Y. Y., Song, L. S., Lakatta, E. G., Xiao, R. P., and Cheng, H. P. (1999) Constitutive β2-adrenergic signalling enhances sarcoplasmic reticulum Ca2+ cycling to augment contraction in mouse heart. J. Physiol. 521, 351363

Received April 25, 2002; accepted June 27, 2002.

Table 1 Ex vivo basal cardiac function in NTG and AC8TG mice under pacing conditionsa

NTG (n=6)

AC8TG (n=6)

P

LVSP (mmHg)

+dLVSP/dt (mmHg.s1)

dLVSP/dt (mmHg.s1)

+dLVSP/dt /LVSP (s1)

ttp (ms)

dLVSP/dt/LVSP (s1)

tR1/2 (ms)

27 ± 1

1392 ± 38

788 ± 73

52 ± 1

26 ± 1

29 ± 1

32 ± 2

64 ± 7

3613 ± 483

2508 ± 318

56 ± 3

24 ± 1

39 ± 3

20 ± 1

0.002

0.005

0.002

NS

NS

0.02

0.001

aThe hearts were perfused with oxygenated Krebs-Henseleit solution containing 1 mM [Ca2+]o, and all parameters were measured under stimulated condition at 680 bpm. LVSP: left ventricular systolic pressure; +d(LVSP)/dt and d(LVSP)/dt): the first derivatives of LVSP; +d(LVSP)/dt/LSVP and d(LVSP)/dt/LSVP: first derivatives expressed relative to LVSP; ttp: time to peak of contraction; tR1/2: time for half relaxation; NS, not significant.

Fig. 1

Figure 1. Ex vivo basal cardiac function in NTG and AC8TG mice under spontaneous beating conditions. The hearts were perfused with oxygenated Krebs-Henseleit solution containing 1.8 mM [Ca2+]o. A) left ventricular systolic pressure (LVSP); B) heart rate (HR) and C) rate-pressure product (RPP) were measured in NTG (empty bars) and AC8TG mice (filled bars) when all parameters were stable and under spontaneous beating conditions. The bars are means ±SE. NTG, n=13; AC8TG, n=12. ***P <0.0001.

Fig. 2

Figure 2. Ex vivo cardiac response of NTG and AC8TG mice to [Ca2+]o. The calcium concentration in the perfusate ([Ca2+]o) was progressively increased from 0.5 to 2.5 mM while continuously measuring LVSP (A) and HR (B) in NTG () and AC8TG mice (). The points are means ±SE. NTG, n=7; AC8TG, n=5. *P<0.05, **P<0.01, ***P<0.001 vs. NTG.

Fig. 3

Figure 3. Contractile measurements in isolated ventricular myocytes from NTG and AC8TG mice. A) Cell contraction is represented as the percentage of cell shortening relative to the maximal length (when the cell is at rest) in NTG () and AC8TG mice (). The cells were electrically stimulated by external platinum electrodes placed into the bath close to the cell (see Materials and Methods). The mean values of time to peak (tpeak) and time of half relaxation of cell shortening (t50) in NTG (empty bars) and AC8TG (filled bars) cardiomyocytes are shown in B and C, respectively. The points in A and the bars in B and C are means ±SE. NTG, n=8 cells from 4 hearts; AC8TG, n=11 cells from 5 hearts. *P<0.05 vs. NTG.

Fig. 4

Figure 4. [Ca2+]i transients in ventricular myocytes from NTG and AC8TG mice. A) [Ca2+]i transients were determined using fluo3-AM in NTG () and AC8TG mice (). [Ca2+]i transients are expressed as F/F0, where F is the mean fluorescence of the cell at any given time and F0 is the mean resting fluorescence. B and C represent the time to peak fluorescence (tpeak) and the time to 50% fluorescence decay (t50), respectively, in NTG (empty bars) and AC8TG (filled bars). The points in A and the bars in B and C are means ±SE. NTG, n=16 cells; AC8TG, n=24 cells. *P<0.05 vs. NTG.

Fig. 5

Figure 5. Basal L-type Ca2+ current (ICa,L) properties in NTG and AC8TG ventricular myocytes. A) Representative whole-cell ICa,L current traces recorded in NTG (left) and AC8TG myocytes (right). The currents were elicited every 8 sec by a 400 ms depolarization to 0 mV from a holding potential of 50 mV. B, C) The bars indicate the mean ±SE ICa,L basal amplitude (B) and current density (C) measured in 16 NTG (empty bars) and 18 AC8TG myocytes (filled bars). D) Current-voltage (I-V) relationships of peak ICa,L in NTG () and AC8TG myocytes (). ICa,L was normalized to the cell capacitance to obtain current densities (pA/pF). Data points are means ±SE from five NTG and four AC8TG cells.

Fig. 6

Figure 6. Effect of increasing the stimulating frequency on basal ICa,L in NTG and AC8TG ventricular myocytes. Whole-cell ICa,L were recorded in NTG and AC8TG myocytes at six increasing stimulation frequencies: 0.25, 0.5, 1, 2, 4, and 8 Hz. The currents were elicited upon a 40 ms depolarization to 0 mV from a holding potential of 50 mV. At each frequency, 10 successive pulses were applied before applying a twofold higher frequency. The ICa,L amplitudes were normalized to the amplitude of the current at the control stimulation frequency (0.125 Hz). The points indicate the means ±SE from three NTG () and three AC8TG myocytes ().