Multiple phosphodiesterases (PDEs) hydrolyze cAMP in cardiomyocytes, but the functional significance of this diversity is not well understood. Our goal here was to characterize the involvement of three different PDEs (PDE2–4) in cardiac excitation–contraction coupling (ECC).
Sarcomere shortening and Ca2+ transients were recorded simultaneously in adult rat ventricular myocytes and ECC protein phosphorylation by PKA was determined by western blot analysis. Under basal conditions, selective inhibition of PDE2 or PDE3 induced a small but significant increase in Ca2+ transients, sarcomere shortening, and troponin I phosphorylation, whereas PDE4 inhibition had no effect. PDE3 inhibition, but not PDE2 or PDE4, increased phospholamban phosphorylation. Inhibition of either PDE2, 3, or 4 increased phosphorylation of the myosin-binding protein C, but neither had an effect on L-type Ca2+ channel or ryanodine receptor phosphorylation. Dual inhibition of PDE2 and PDE3 or PDE2 and PDE4 further increased ECC compared with individual PDE inhibition, but the most potent combination was obtained when inhibiting simultaneously PDE3 and PDE4. This combination also induced a synergistic induction of ECC protein phosphorylation. Submaximal β-adrenergic receptor stimulation increased ECC, and this effect was potentiated by individual PDE inhibition with the rank order of potency PDE4 = PDE3 > PDE2. Identical results were obtained on ECC protein phosphorylation.
Our results demonstrate that PDE2, PDE3, and PDE4 differentially regulate ECC in adult cardiomyocytes. PDE2 and PDE3 play a more prominent role than PDE4 in regulating basal cardiac contraction and Ca2+ transients. However, PDE4 becomes determinant when cAMP levels are elevated, for instance, upon β-adrenergic stimulation or PDE3 inhibition.
The cAMP pathway is critical for autonomic regulation of the heart. Sympathetic stimulation increases myocardial contractility mainly through stimulation of β-adrenergic receptors (β-ARs), elevation of intracellular cAMP ([cAMP]i), and activation of the cAMP-dependent protein kinase (PKA). PKA, in turn, phosphorylates key components of cardiac excitation–contraction coupling (ECC), such as the L-type Ca2+ channel (Cav1.2), ryanodine receptor type 2 (RyR2), phospholamban (PLB), troponin I (TnI), and myosin-binding protein C (MyBP-C). The phosphorylation of Cav1.2 and RyR2 leads to enhanced Ca2+ influx and sarcoplasmic reticulum (SR) Ca2+ release, contributing to enhanced Ca2+ transients. PLB phosphorylation increases SR Ca2+ uptake by the Ca2+ ATPase (SERCA2), thus accelerating cardiac relaxation but also increasing SR Ca2+ load and consequently SR Ca2+ release. The phosphorylation of TnI and MyBP-C reduces myofilament Ca2+ affinity and increases cross-bridge kinetics. Altogether, these events produce the typical inotropic and lusitropic effects of β-AR stimulation.1
Intracellular cAMP levels result from the balance between cAMP synthesis by adenylyl cyclases and cAMP degradation by cyclic-nucleotide phosphodiesterases (PDEs). PDEs are subdivided into 11 families among which four can hydrolyse cAMP in the heart: PDE1, which is activated by Ca2+/calmodulin; PDE2, which is stimulated by cGMP; PDE3, which is inhibited by cGMP; and PDE4.2 While PDE1 and PDE2 can hydrolyse both cAMP and cGMP, PDE3 preferentially hydrolyses cAMP and PDE4 is specific for cAMP. Another PDE, named PDE8A, which specifically hydrolyses cAMP, was shown recently to control ECC in mouse cardiac myocytes.3 All PDE isoforms except PDE8A4,5 are inhibited by xanthine derivatives such as 3-isobutyl-1-methylxanthine (IBMX), and a number of drugs have been developed as selective PDE inhibitors6,7: erythro-9-(2-hydroxy-3-nonyl)adenine8 and Bay 60-75509 for PDE2; milrinone, cilostamide (Cil), and other bipyridines for PDE37; rolipram and Ro 20-1724 for PDE4.7 There is currently no commercially available selective inhibitor of PDE1.
Direct monitoring of cAMP in living cardiac myocytes emphasized the importance of PDE isoforms 2–4 for the control of [cAMP]i. In the rodent heart, a predominance of PDE4 over other PDEs for the control of cAMP signals generated by stimulation of β-ARs and other GsPCRs was observed.10–14 PDE4 was shown to associate with β-ARs and ECC proteins, either directly or through scaffold proteins such as β-arrestins or A-kinase anchoring proteins.15,16 In the mouse heart, a specific PDE4 variant (PDE4D3) was shown to control the phosphorylation of RyR2 and regulates a diastolic SR Ca2+ leak.17 Similarly, a PDE4D variant was found to co-immunoprecipitate with SERCA2 in mouse and to control PLB phosphorylation and Ca2+ transient decay kinetics.18
As Cav1.2 is a well-characterized target of the cAMP/PKA pathway in the heart, the cardiac L-type Ca2+ channel current (ICa,L) has been frequently used as a functional index of the contribution of PDE isoforms to this pathway. Inhibition of PDE2, PDE3, and PDE4 was shown to increase ICa,L amplitude in a number of species.12,19–24 Recently, we observed in the mouse heart that a PDE4B variant associates with Cav1.2 and controls its activity upon β-AR stimulation.25 However, important differences exist among species as to whether PDE3 or PDE4 is predominant, and whether all three PDE isoforms control the ICa,L amplitude at basal or only upon β-AR stimulation. PDE3 inhibitors were also reported to increase SR Ca2+ uptake and SR Ca2+ content,26–28 an effect attributed to an increase in PKA-dependent phosphorylation of PLB.
While it is well established that PDE3 inhibition exerts a direct positive inotropic effect in the heart from large mammals,29–31 the contribution of other PDE families to the regulation of cardiac contractility remains less clear and often controversial.32 For instance, while selective inhibition of PDE2 or PDE4 was shown to increase cardiac contractility following β-AR stimulation in some studies,33–37 other studies failed to demonstrate an effect.38,39
In an attempt to reconcile these contradictory findings, we compared, in the same mammalian cardiac preparation (adult rat ventricular myocytes, ARVMs), the effect of a selective inhibition of PDE2, PDE3, and PDE4 on Ca2+ transients, sarcomere shortening, and phosphorylation of five key ECC proteins (TnI, Cav1.2, RyR2, PLB, and MyBP-C). Furthermore, we compared for each PDE isoform, the effects observed under basal condition and after sub-maximal β-AR stimulation with isoprenaline (Iso). Our study sheds new light on the respective contribution of the three major cardiac PDE isoforms in the regulation of cardiac ECC.
All experiments were carried out according to the European Community guiding principles in the care and use of animals (86/609/CEE, CE Off J n°L358, 18 December 1986), the local Ethics Committee (CREEA Ile-de-France Sud) guidelines, and the French decree n°87–848 of 19 October 1987 (J Off Rép Fr, 20 October 1987, pp. 12 245–12 248). Authorizations to perform animal experiments according to this decree were obtained from the Ministère Français de l'Agriculture, de la Pêche et de l'Alimentation (n° 92–283, 27 June, 141 2007).
Bay 60-7550 (Bay, 2-[(3,4-dimethoxyphenyl)methyl]-7-[(1R)-1-hydroxyethyl]-4-phenylbutyl]-5-methyl-imidazo[5,1-f][1,2,4]triazin-4(1H)-one) was from Alexis Biochemicals (Lausen, Switzerland): it blocks PDE2 with an IC50 value of 4.7 nM40 and was used here at a 100-nM concentration to fully block the enzyme.9 Cil was from Tocris Bioscience (Bristol, UK): it blocks PDE3 with an IC50 ranging from 541 to 27 nM42 and was used here at a 1-µM concentration. Ro 20-1724 [Ro, 4-(3-butoxy-4-methoxybenzyl)-2-imidazolidone] was from Calbiochem (Darmstadt, Germany): it blocks PDE4 with an IC50 value around 1 µM43,44 and was used here at 10 µM. At these concentrations, Bay, Cil, and Ro were shown to be selective for PDE2, PDE3, and PDE4, respectively (Supplementary material online, Table S1). IBMX and Iso were from Sigma-Aldrich (Saint-Quentin, France).
Cardiomyocyte isolation and culture
Male Wistar rats (250–300 g) were subjected to anaesthesia by intraperitoneal injection of pentothal (0.1 mg/g) and hearts were excised rapidly. Individual ARVMs were obtained by retrograde perfusion of the heart as previously described.23 For enzymatic dissociation, the hearts were perfused retrogradely at a constant flow of 6 mL/min at 37°C with a Ca2+-free Ringer solution containing (in mM): NaCl 117, KCl 5, NaHCO3 4.4, KH2PO4 1.5, MgCl2 1.7, d-glucose 11.7, Na-phosphocreatine 10, taurine 20, and 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES) 21, pH 7.1 during 5 min, followed by a perfusion at 4 mL/min for 1 h with the same solution containing 1 mg/mL of collagenase A (Boehringer, Mannhein, Germany) and 300 µM ethylene glycol tetraacetic acid (EGTA) and CaCl2 to adjust free Ca2+ concentration to 20 µM. The ventricles were then separated from atria, finely chopped, and gently agitated to dissociate individual cells. The resulting cell suspension was filtered on a gauze, and the cells were allowed to settle down. The supernatant was discarded and cells resuspended four more times with Ringer solution at increasing [Ca2+] from 20 to 300 µM. Freshly isolated cells were suspended in minimal essential medium (MEM: M 4780; Sigma, St Louis, MO USA) containing 1.2 mM [Ca2+] supplemented with 2.5% foetal bovine serum (FBS, Invitrogen, Cergy-Pontoise, France), 1% penicillin–streptomycin, 20 mM HEPES (pH 7.6), and plated on 35 mm, laminin-coated culture dishes (10 μg/mL) at a density of 104 cells per dish. After 1 h, the medium was replaced by 300 μL of FBS-free MEM. All experiments were performed on primary cell cultures 24 h after plating.
Measurements of Ca2+ transients and sarcomere shortening
Isolated cardiomyocytes were loaded with 5 µM Fura-2 AM (Invitrogen) at room temperature for 15 min and then washed with Ringer solution containing (in mM): NaCl 121.6, KCl 5.4, NaHCO3 4.013, NaH2PO4 0.8, HEPES 10, glucose 5, Na pyruvate 5, MgCl2 1.8, CaCl2 1.8, pH 7.4. The loaded cells were field stimulated (5 V, 4 ms) at a frequency of 0.5 Hz. Sarcomere length (SL) and Fura-2 ratio (measured at 512 nm upon excitation at 340 and 380 nm) were simultaneously recorded using an IonOptix System (IonOptix, Milton, MA, USA). Briefly, myocytes were imaged at 240 Hz using an IonOptix Myocam-S CCD camera. Images were displayed within the IonWizard acquisition software (IonOptix). The myocyte of interest was aligned horizontally, parallel to the field of view. The SL was measured from a user-defined region of interest that included at least seven sarcomeres to maximize the accuracy of the measurements. The IonWizard software determined a density trace corresponding to the alternating dark and light A- and I-bands of the cardiomyocyte, and analysed the periodicity in the density trace by calculation of a fast Fourier transform. This analysis allowed a direct real-time measurement of SL.
The measurements of Ca2+ transients and sarcomere shortening were performed on individual myocytes superfused for few minutes with control Ringer solution and then challenged with PDE inhibitors (single: Bay, 100 nM; Cil, 1 µM; Ro, 10 µM, or IBMX, 100 µM; or combinations: Bay + Ro, Bay + Cil or Cil + Ro) during 15 min or with 1 nM Iso during 10 min followed by 10 min treatment with PDE inhibitors (Bay, Cil, or Ro). SR Ca2+ load was measured by rapid application of caffeine (10 mM) to ARVMs after pacing at 0.5 Hz in control Ringer solution or after Cil or Cil + Ro application. Fractional Ca2+ release was calculated as the ratio between the Ca2+ transient amplitude and the caffeine transient amplitude (SR Ca2+ load) measured in the same ARVM.
Cell extracts and western blot analysis
Prior to cell homogenization and biochemical analysis, intact ARVMs were treated at room temperature for 15 min with PDE inhibitors (single: Bay, 100 nM; Cil, 1 µM; Ro, 10 µM, or IBMX, 100 µM; or combinations: Bay + Ro, Bay + Cil or Cil + Ro), or for 10 min with 1 nM Iso followed by 10 min treatment with PDE inhibitors (Bay, Cil, or Ro). To prepare whole ARVM cell lysates, cells were lysed in cold HEPES, NaCl, triton, glycerol lysis buffer containing (in mM): HEPES 50 (pH 7.5), NaCl 400, NaF 100, Na-pyrophosphate 10, MgCl2 1.5, EGTA 1, 10% glycerol, 1% Triton X-100 supplemented with 1 mM Na3VO4, and a protease inhibitor mixture [1 mM 4-(2-aminoethyl) benzenesulphonyl fluoride hydrochloride, 1 mM benzamidine, 10 μg/mL of soybean trypsin inhibitor, 10 μg/mL of aprotinin, 1 μg/mL of leupeptin, 1 μg/mL of antipain, and 1 μg/mL of pepstatin]. The lysates were centrifuged at 10 000 g for 10 min and assayed for protein concentration using the bicinchoninic acid-protein reagent (Sigma-Aldrich). For immunoblotting assays, proteins were resolved by 4–12% or 8–16% SDS–PAGE (Lonza Verviers, Verviers, Belgium), transferred onto nitrocellulose membranes, and stained with 3% Ponceau S (Sigma-Aldrich). The membranes were saturated with 3% bovine serum albumin. Phosphorylation of PLB at Ser16 and cardiac TnI at Ser23/24 were detected using specific antibodies from Cell Signaling Technology (Beverly, MA, USA). Phosphorylation of RyR2 at Ser2809 was detected using a specific antibody kindly provided by Dr A.R. Marks (Columbia University, USA). L-type Ca2+ channel pore-forming subunit (Cav1.2) phosphorylation at Ser1928 was detected using the anti-CH3-P antibody.45 Phosphorylation of MyBP-C at Ser282 was detected using antibodies from Alexis Biochemicals (San Diego, CA, USA). Immunoreactive proteins were revealed using a 1 : 30 000 dilution of anti-rabbit or anti-mouse IgG peroxidase conjugate antibodies (Promega Corp., Madison, USA) and were visualized with the Supersignal West Pico reagent (Perbio Science, Vigneux, France). The membranes were stripped and reprobed with specific antibodies against calsequestrin (CSQ) (Affinity BioReagents, Ozyme, Saint-Quentin-Fallavier, France) used as a loading control. Each sample was normalized to CSQ (phosphorylated proteins/CSQ ratio) and double-normalized to the control ratio. Densitometric analyses of western blots were performed using the NIH ImageJ software.
Cell contraction was assessed by the percentage of sarcomere shortening, which is the ratio of twitch amplitude (difference of end-diastolic and peak systolic SLs) to end-diastolic SL. Ca2+ transient amplitude was assessed by the percentage of variation in the Fura-2 ratio, by dividing the twitch amplitude (difference of end-diastolic and peak systolic ratios) to the end-diastolic ratio. Relaxation was assessed by measuring the time-to-50% relaxation from the time-to-peak shortening, and the Ca2+ transient decay was evaluated by measuring the time-to-50% decay of the Fura-2 ratio from the time-to-peak ratio. These parameters were obtained by analysing the 10 last contractions and calcium transients before addition of the next drug. All parameters were calculated offline on a dedicated software (IonWizard 6×, IonOptix). All results are expressed as mean ± SEM. The GraphPad Prism software (GraphPad software, Inc., La Jolla, CA, USA) was used for statistical analysis. To determine statistical significance with multiple groups, we used a one-way analysis of variance (ANOVA) followed by a Newman–Keuls test for multiple comparisons. Differences with P-values of <0.05 were considered as statistically significant. The number of independent experiments performed is indicated in the figure legends.
Regulation of ECC by PDE2, PDE3, and PDE4 in ARVMs
To investigate the functional consequences of individual PDE inhibition on basal and β-AR-stimulated ECC in ARVMs, Ca2+ transients and sarcomere shortening were simultaneously recorded in Fura-2-loaded ARVMs. In quiescent myocytes, the average diastolic SL was 1.68 ± 0.01 µm (n = 95) (Supplementary material online, Table S2). Upon pacing at 0.5 Hz, the average peak amplitude of Ca2+ transients was 22 ± 0.1% above the diastolic Fura-2 ratio (Supplementary material online, Table S3), whereas SL decreased by 0.78 ± 0.05% (n = 95) (Supplementary material online, Table S1). The relaxation kinetics of both parameters, as estimated by the time-to-50% decay (t1/2 off), were on average 0.43 ± 0.01 s for Ca2+ transients and 0.49 ± 0.02 s for cell contraction (n = 95). On average, application of 1 nM Iso increased the amplitude of Ca2+ transients by 257 ± 20.3% and sarcomere shortening by 1033.8 ± 61.2%, (n = 62, P < 0.001 vs. control). Iso also significantly accelerated the relaxation phases of both signals, decreasing t1/2 off for Ca2+ decay by 38.4 ± 1.6%, and t1/2 off for contraction by 56.3 ± 1.3% (n = 62, P < 0.001 vs. control). As shown in Supplementary material online, Figure S1, the Iso response on Ca2+ transients and cell shortening was stable during the 10-min incubation period, suggesting that β-AR stimulation by 1 nM Iso does not induce desensitization of the receptors, in our experimental conditions.
In a first series of experiments, the effects of PDE2 inhibition were tested (Figure 1 and Supplementary material online, Tables S2 and S3) using the selective inhibitor Bay. As shown on the representative traces in Figure 1A and on the summary data in Figure 1B, 100 nM Bay increased basal Ca2+ transient amplitude and basal cell contraction ∼two-fold. Bay also accelerated the decay kinetics of both signals, as shown by a significant decrease in the t1/2 off values for both parameters (Figure 1C). These effects were small when compared with that of 1 nM Iso. To examine whether PDE2 modulates ECC under β-AR stimulation, 100 nM Bay was applied on the top of Iso. As shown in Figure 1A and Supplementary material online, Tables S2 and S3, the Iso effect was potentiated upon PDE2 inhibition. However, the decay kinetics of the signals were unchanged compared with those obtained under Iso stimulation (Figure 1C).
We next studied the consequences of a selective PDE3 inhibition on Ca2+ transients and cell contraction using Cil (Figure 2 and Supplementary material online, Tables S2 and S3). Application of 1 µM Cil alone induced a significant increase in Ca2+ transient and sarcomere shortening amplitudes by ∼two- and three-folds, respectively (Figure 2A and B). Cil also significantly accelerated the decay kinetics of both signals (Figure 2C). This was accompanied by an increase in SR Ca2+ load with no change in SR fractional Ca2+ release (Figure 5). PDE3 inhibition also strongly potentiated the effect of Iso on both parameters, enhancing their amplitude by ∼30% for Ca2+ transient and 70% for cell shortening (Figure 2B), and accelerating their relaxation kinetics (Figure 2C).
In contrast to the results obtained with PDE2 or PDE3 inhibitors, PDE4 inhibition by 10 µM Ro had no effect on Ca2+ transients and sarcomere shortening under basal conditions (Figure 3A and B and Supplementary material online, Tables S2 and S3). However, Ro strongly potentiated the effects of Iso (1 nM), further increasing the Ca2+ transient amplitude and sarcomere shortening by ∼60% (Figure 3B). PDE4 inhibition also significantly accelerated the decay kinetics of both signals (Figure 3C).
We next examined the functional consequences of combinations of selective PDE inhibitors on cell contraction under basal conditions (Figure 4 and Supplementary material online, Table S2). As shown in Figure 4A and B, simultaneous inhibition of PDE2 and PDE4 by a combination of 100 nM Bay and 10 µM Ro (Bay + Ro) increased Ca2+ transients and sarcomere shortening, and slightly accelerated the relaxation kinetics of the twitch. However, comparison with Figure 1 shows that these effects were similar to that of individual PDE2 inhibition under basal conditions. In contrast, concomitant inhibition of PDE2 and PDE3 by 100 nM Bay and 1 µM Cil strongly increased the amplitude of Ca2+ transients and sarcomere shortening, and these effects were significantly larger than when either PDE2 or PDE3 alone were inhibited (P < 0.05 vs. Bay and Cil alone). The relaxation kinetics were significantly accelerated when compared with control (Figure 4C), and t1/2 off values of sarcomere shortening were significantly smaller for Bay + Cil when compared with Bay alone (compare Figures 1C and 4C), but not when compared with Cil alone (compare Figures 2C and 4C). This suggests that PDE3 has a more prominent role in controlling cell relaxation than PDE2 under basal conditions. Concomitant inhibition of PDE3 and PDE4 by Cil + Ro had a major impact on ECC, increasing the amplitude of Ca2+ transients 2.5-fold, and sarcomere shortening 15-folds (Figure 4B). This was accompanied by a major acceleration of Ca2+ transients and cell shortening decay phases, which was significantly larger than with Bay + Cil (Figure 4C). Furthermore, Cil + Ro doubled SR Ca2+ load and drastically increased fractional Ca2+ release from 40% in Ctrl or after PDE3 inhibition, to 80% when both PDE3 and PDE4 were inhibited (Figure 5). Finally, application of the broad spectrum PDE inhibitor IBMX (100 µM) had similar functional effects as a concomitant inhibition of PDE3 and PDE4.
Regulation of the phosphorylation status of key ECC proteins by PDEs
To get some insights into the molecular mechanisms by which individual PDE families regulate ECC in ARVMs, we examined the consequences of PDE inhibition on the phosphorylation of key ECC proteins. For this, we performed western blot analysis with phospho-specific antibodies directed against TnI, MyBP-C, PLB, RyR2, and Cav1.2 under basal conditions (Figure 6) or after a non-maximal β-AR stimulation by Iso (1 nM) (Figure 7). In line with ECC measurements, Figure 6 shows that global PDE inhibition with IBMX or concomitant inhibition of PDE3 and PDE4 by Cil + Ro led to a major increase in the phosphorylation of TnI (Figure 6A), MyBP-C (Figure 6B), PLB (Figure 6C), RyR2 (Figure 6D), and CaV1.2 (Figure 6E). Individual inhibition of PDE2 and PDE3, as well as simultaneous inhibition of PDE2 and PDE3 or PDE2 and PDE4, slightly but significantly increased the phosphorylation of TnI, whereas inhibition of PDE4 alone did not (Figure 6A). In contrast, MyBP-C phosphorylation was significantly increased by selective inhibition of PDE2, PDE3, or PDE4 and by their dual inhibition, especially in the case of PDE3 and PDE4 (Figure 6B). Selective PDE3 inhibition by Cil or simultaneous inhibition of PDE2 and PDE3 also increased significantly PLB phosphorylation, whereas individual or associated PDE2 and PDE4 inhibition did not (Figure 6C). The phosphorylation of CaV1.2 and RyR2 was only increased by simultaneous blockade of PDE3 and PDE4 or global PDE inhibition with IBMX (Figure 6D and E).
Figure 7 shows that, after a non-maximal β-AR stimulation by Iso, we observed a significant increase in the phosphorylation of TnI, MyBP-C, PLB, RyR2, and CaV1.2. A maximal level of phosphorylation was obtained by treatment of cells by Iso followed by IBMX (Figure 7). The effect of Iso on the phosphorylation of all of these proteins was potentiated by the selective inhibition of PDE4, in agreement with the results obtained on Ca2+ transients, sarcomere shortening, and relaxation kinetics. Similar results were obtained upon PDE3 inhibition, except that Cil only tended to increase the level of phosphorylation of TnI. In contrast, PDE2 inhibition only increased the phosphorylation of MyBP-C and CaV1.2 in the presence of Iso.
It is well established that cardiac ECC is regulated by the cAMP/PKA pathway. Initiation of the pathway takes place at the sarcolemmal membrane by cAMP synthesis through the activity of adenylyl cyclases; termination of the pathway involves the activity of cyclic-nucleotide PDEs that are responsible for the hydrolysis of cAMP into 5′-AMP, and phosphatase activity that is responsible for PKA substrate dephosphorylation. At least five PDE families are reported to degrade cAMP in the heart.2,46 Comparison of the functional role of individual PDE families is possible by using pharmacological agents that selectively inhibit their activity. Since selective inhibitors of PDE1 and PDE8 are still lacking, we focused our attention on PDE2, PDE3, and PDE4 for which selective inhibitors exist. Using Bay to inhibit PDE2, Cil to inhibit PDE3, and Ro to inhibit PDE4, as well as the broad spectrum PDE inhibitor (IBMX), we explored the role of these three PDEs in three different sets of experimental conditions (at basal, upon β-AR stimulation, and upon dual inhibition), and on nine different parameters: PKA phosphorylation level of TnI, MyBP-C, PLB, RyR2, and Cav1.2; amplitude and relaxation kinetics of sarcomere shortening and Ca2+ transients. It is known that cAMP signals induced by β-AR stimulation also activate the Ca2+/calmodulin-dependent kinase II (CaMKII), which phosphorylates some of the main PKA substrates.47 Thus, it is conceivable that some of the effects of the PDE inhibitors are due to the activation of CaMKII. However, inhibition of this kinase by KN-93 (1 µM) did not alter the positive inotropic effects of IBMX (Supplementary material online, Figure S2), suggesting that PKA is the main effector of the increased Ca2+ transients and sarcomere shortening observed upon PDE inhibition.
Our results indicate that only PDE2 and PDE3 appear to regulate ECC under basal conditions. However, upon β-AR stimulation, individual inhibition either of PDE2, PDE3, or PDE4 potentiates the β-AR response of Ca2+ transient and cell contraction to a similar extent. This suggests that PDE2 and PDE3 are active in basal conditions, whereas PDE4 becomes activated upon β-AR stimulation or PDE3 inhibition.
In ARVMs, biochemical and functional measurements indicate that PDE3 and PDE4 are the major PDEs degrading cAMP.12,14,23 Accordingly, experiments shown here indicate that PDE3 and PDE4 dominate. This is especially true for the control of the SR Ca2+ load (Figure 5), which determines Ca2+ transient amplitude.48 Interestingly, while PDE3 inhibition alone does not increase fractional release (probably because it affects only SR Ca2+ load), the combination Cil and Ro have a tremendous effect on this parameter. This can be explained by the increased ICa,L amplitude14 and RyR2 open probability, leading to an increase of the gain of ECC49 occurring only when both PDEs are inhibited. Hence, inhibition of these two PDEs induces a massive phosphorylation of ECC proteins, and a large increase in Ca2+ transients and sarcomere shortening (Figures 4 and 6). Other dual combinations of PDE inhibition had much smaller effects on protein phosphorylation and ECC, indicating that PDE3 and PDE4 can compensate for each other, and that their function is partially redundant.
Although PDE2 represents only a few per cents of total hydrolytic activity in ARVMs,12 we show here that it regulates basal and β-AR-stimulated ECC. This is consistent with previous studies, showing that PDE2 modulates cAMP levels and β-AR responses in rodent cardiac myocytes.23,33 In particular, PDE2 modulates the β-AR regulation of the L-type Ca2+ current (ICa,L) in cardiac myocytes,23 thus providing a possible mechanism for the effects of PDE2 inhibition on Iso-stimulated Ca2+ transients and contraction. In contrast, the mechanism by which PDE2 regulates basal ECC is less clear since PDE2 inhibition has no effect on basal ICa,L in ARVMs.23 MyBP-C could participate in the effects of PDE2 on basal ECC by increasing cross-bridge cycling and myofilament Ca2+ sensitivity.50
In contrast to PDE4, selective PDE3 inhibition increased the amplitude and relaxation kinetics of Ca2+ transients and sarcomere shortening under basal conditions (Figure 2), and this was associated with enhanced phosphorylation of PLB, TnI, and MyBP-C but not Cav1.2 or RyR2 (Figure 6). These data are in line with previous findings, showing a subcellular localization of PDE3 in SR-enriched membrane fraction51,52 and the positive effects of PDE3 inhibitors on SR Ca2+ uptake and SR Ca2+ content.26–28 This is also consistent with the lack of effect of selective PDE3 inhibition on basal ICa,L in ARVMs.23
In the present study, we established that PDE2, PDE3, and PDE4 have distinct roles in controlling ECC. In the absence of hormonal stimulation, when intracellular cAMP concentration is low, PDE2 and especially PDE3 are dominant to control cardiac contraction. PDE3 contribution to the degradation of basal cAMP could be explained by its high affinity for cAMP (Km <1 µM).53 For PDE2, the scheme must be different because of its low affinity for cAMP. Nonetheless, under basal conditions, localized production of cGMP may occur to activate PDE2 by binding to its regulatory cGMP-specific phosphodiesterase, adenylyl cyclases, and FhIA-B domain.54 PDE2 activation mediated by cGMP could exacerbate its role in controlling cAMP levels regulating ECC as recently reported in cardiomyocytes.55 When cAMP levels are increased, PDE4 turns on and becomes crucial to control global cAMP homeostasis and ECC. However, despite a modest effect on the cAMP level, inhibition of PDE2 and PDE3 also enhances β-AR inotropic and lusitropic effects, demonstrating that PDE2, PDE3, and PDE4 act in concert to modulate β-AR stimulation of cardiac contraction. Interestingly, although cardiotonic drugs such as milrinone produce beneficial inotropic effects upon acute administration,56 these agents increase mortality in patients with dilated cardiomyopathy upon chronic treatment.57 A challenge of future research will be to determine the molecular mechanisms that underlie long-term detrimental effects of PDE inhibitors in heart failure in order to design new therapeutic approaches to enhance the beneficial effects and to suppress detrimental long-term effects of PDE inhibitors.
The present study focused on the role of three different PDE families in controlling the ECC in isolated ARVMs. It is known that the expression level of these PDEs varies among different species and is dependent on the development stage and on the cardiac territory investigated.32 However, PDE expression is relatively conserved between murine models and human.58 In the mammalian heart, PDE3 and PDE4 remain the two major enzymes degrading cAMP and controlling cardiac ECC. Thus, even if our results might not be fully transposable to any cellular cardiac models or human heart, we unveil how cardiac function is finely tuned by these PDE families. Furthermore, their expression and distribution are altered under pathophysiological conditions.59 Therefore, the relative contribution of PDE2, PDE3, and PDE4 to compartmentalize cAMP signals and to maintain specific PKA phosphorylation of the key ECC proteins is most probably differently affected in various cardiac diseases. Future studies investigating their participation in pathological conditions are necessary to design new therapeutic strategies targeting PDEs.
Conflict of interest: none declared.
This work was supported by Leducq cycAMP grant 06CVD02 (to R.F.), by European Union contract LSHM-CT-2005-018833/EUGeneHeart (to R.F.), the Fondation de France (to G.V.), the Agence Nationale de la Recherche 2010 BLAN 1139-01 (to G.V.), the Fondation pour la Recherche Médicale (to R.F.), the Société Française de Cardiologie (to J.L.), the Investment for the Future program ANR-11-IDEX-0003-01 within the LABEX ANR-10-LABX-0033 (to R.F.) and a PhD grant from Ile de France CORDDIM (to P.B.).
The authors wish to thank Florence Lefebvre and Aurore Germain for technical assistance with cardiomyocyte isolation, and Dr Andrew R. Marks (Columbia University, USA) for kindly providing the RyR2 phosphospecific antibody.
- cardiac myocyte
- western blotting
- myocardial contraction
- 3',5'-cyclic-nucleotide phosphodiesterase
- phosphoric diester hydrolase
- beta-adrenergic receptor
- ryanodine receptor calcium release channel
- cyclic amp
- protein c
- troponin i
- diethylstilbestrol monophosphate
- excitation-contraction coupling
- ventricular myocyte
- international organization for standardization
- cyclic phosphodiesterase, type 2