Is Epac1 a good therapeutic strategy against cardiac remodelling?

This editorial refers to ‘Exchange protein activated by cyclic-adenosine monophosphate (Epac) regulates atrial fibroblast function and controls cardiac remodeling’ by S. Surinkaew et al., pp. 94–106. Over the last two decades, our knowledge on cardiac cAMP signalling has been constantly expending. During the late 90’s, the dogma of PKA as the single effector of cAMP in physiology and pathology was challenged by the discovery of Epac, a guanine nucleotide exchange protein directly activated by cAMP. Epac is now emerging as a new key factor in cardiac pathology (Figure 1).1 Cardiac cells express two Epac isoforms, Epac1 and Epac2, that share high structural homology with distinct cAMP-binding domain numbers. Epac1 contains one cAMP-binding domain whereas Epac2 contains two cAMP domains. Epac isoforms also differ in their subcellular localization and function.2 Epac1 is located at the nuclear envelope and seems to be part of the PKA anchoring protein (mAKAP) complex together with the phosphodiesterases 4D3. This specific subcellular localization of Epac1 is responsible, in ventricular cardiomyocytes, for the activation of autophagy3 and maladaptive hypertrophic signalling that involve the Ca/calcineurin/NFAT and the Ca/HDAC5/MEF2 pathways (Figure 1). Moreover, chronic activation of Epac1 increases the propensity for ventricular arrhythmia by up-regulation of calmodulin protein. Cardiac Epac2, mostly studied in ventricular cardiomyocytes, is located at the T-Tubules, where it may induce pro-arrhythmic events such as enhanced diastolic Ca release resulting from CaMKII-dependent phosphorylation of ryanodine receptors. The idea of a pathological role for Epac1 in ventricles meshes well with its up-regulation in human and animal models of heart failure.4,5 Epac1 is also involved in the regulation of collagen synthesis, fibroblast migration, and fibroblast adhesion, all part of the fibrotic process.6,7 However, in fibroblasts, the impact of Epac1 activation is not as clear as in ventricular cardiomyocytes. Indeed, in primary cultured rat adult ventricular fibroblasts or renal epithelial cells, pro-fibrotic agents, such as TGF-β and angiotensin II (AGII), down-regulate Epac1 mRNA expression6,7 suggesting an anti-fibrotic action whereas low cAMP concentration induces fibroblast migration via Epac1/Rap1 activation suggesting a pro-fibrotic action.7 Epac1 signalling pathway in ventricular cardiomyocyte and fibroblast. In ventricular cardiomyocyte, upon β adrenergic stimulation (β-AR) the concentration of cAMP is elevated resulting in the activation of Epac1 and Epac2. The activation of Epac1 activation, at the nuclear membrane, activates the CaMKII/HDAC/MEF2 and Ca2+/calcineurin (CN)/NFAT signalling pathway leading to maladaptive hypertrophy. The activation of Epac2 increases sarcoplasmic reticulum (SR) Ca2+ leak via the phosphorylation of the ryanodine receptor (RyR) by CaMKII participating, with Epac1-mediated over-expression of calmodulin (CaM) to the generation of ventricular arrhythmias. In ventricles, activation of both Epac isoforms leads to cardiac dysfunction. In fibroblasts, Epac1 activation by the β2-AR decreases the expression of collagen I and III (Col.I–III) resulting in an anti-fibrotic effect. Moreover, Epac1 expression is down-regulated by the α-AR, AGII-R and TGFβ1-R (via smad 3) leading to collagen I and III synthesis and deposition involved in the development of cardiac fibrosis and cardiac dysfunction. In fibroblasts, activation of Epac1 is anti-fibrotic and cardioprotective. Surinkaew et al.8 validate the cardioprotective and anti-fibrotic effect of Epac1 activation in atrial fibroblasts, but not in ventricular fibroblasts where Epac1 seems to be expressed at lower levels. Those findings dovetail exquisitely with previous data obtained in mice with Epac1 deletion. Indeed, genetic deletion of Epac1 in a pressure-overload mouse model prevents cardiac failure by decreasing myocytes fibrosis and apoptosis without affecting cardiac hypertrophy.9 Epac1 deletion also reduces the susceptibility for atrial fibrillation in mice with chronic isoproterenol infusion.10 However, the beneficial effect of Epac1 on cardiac fibrosis could be model-dependent since Epac1 deletion has no impact on fibrosis in cardiomyopathy induced by radiotherapy.11 The originality of the work of Surinkaew et al.8 resides in their comparative study between freshly isolated atrial and ventricular fibroblasts, in addition to primary cultured fibroblasts. Using a ventricular tachy-paced dog model, they found down-regulation of Epac1, as previously descripted in cultured cells,6,7 which was only significant in atrial, and not in ventricular fibroblasts. This new awareness of cell specificity in Epac1 down-regulation suggests a regionalization of Epac1 function in fibrosis. In cultured left atrial fibroblasts, the authors found that Epac1 activation, with the non-isoform selective Epac activator 8-pCPT, dramatically reduces the expression of the fibrotic collagen 1A1 whereas Epac1 inhibition with ESI-09 increases collagen 1A1 production. Moreover, treatment with the pro-fibrotic TGF-β1,12 known to increase in heart failure, down-regulates Epac1 expression (Figure 1) as seen in previous studies on adult and neonatal rat ventricular fibroblasts.13,14 Interestingly, downstream effectors seem cell-type specific, with Smad3 involved in atrial fibroblast, as shown here by the authors, and MAPK/Smad2 in ventricular fibroblasts of neonate rats.13 In addition, pharmacological modulation of the predominant β2-adrenergic receptor clearly reproduced the relationship between Epac1 expression and collagen production mediated by PI3K in a PKA-independent way in ventricular adult rat cultured fibroblasts exposed to angiotensin.14 However, the question whether the Epac1 down-regulation in ventricular tachy-paced dogs is associated to a decrease in collagen expression and activation of β2-AR and Smad3 remains unanswered. One striking aspect of Surinkaew et al.8 article lies in the in vivo study. In mice with myocardial infarction, 2 weeks treatment with Sp-8CPT, which activates both Epac1 and Epac2, not only reduces fibrotic tissue but also reduces infarct size and decreases post-MI remodelling and cardiac dysfunction. Those results are quite intriguing knowing the pro-hypertrophic and pro-arrhythmogenic effect of Epac1 and Epac2 activation in ventricular cardiomyocytes and raise some interesting questions. Is ventricular function and signalling altered in the Surinkaew et al. animal model? Is the presumed deleterious (or beneficial) effect of Epac1 activation on heart failure dependent on the animal model? Will reduced fibrosis via Epac1 activation be able to restore cardiac function in early stage of cardiomyopathy whereas activation in later stage will induce cardiac hypertrophy and propensity to cardiac arrhythmia?15 Thus, further detailed studies of Epac1 signalling pathway are needed. One limitation of this work lies in the use of pharmacological Epac activator Sp-8-CPT for the in vivo study. Sp-8CPT can activate both Epac1 and Epac2. Even though Epac2 is less expressed in fibroblast its role is not clear. In addition, Sp-8-CPT can also bind to phosphodiesterases that could indirectly change cAMP and/or cGMP levels and activate other downstream effectors such as PKA and PKG.1 This work would have beneficiated from the use of a transgenic mice over-expressing Epac1 or, as proposed by the authors, a si-RNA knockdown of Epac1, to exclude any potential non-specific effects. The article of Surinkaew et al.8 found Epac1 as an anti-fibrotic factor in atria with cardioprotective effect. This interesting work raises a question about Epac1 targeting for the treatment of cardiomyopathy due to its opposite function on ventricular cardiomyocytes and atrial cardiac fibroblasts: whereas the former may be detrimental, the latter may be beneficial for cardiac function. Thus, the challenge ahead will be to find cardiac cell-specific targeting of Epac1 activator or inhibitor, as well as to improve the selectivity and reduce the toxicity of the available pharmacological Epac1 modulators. The author would like to thank Drs A.M. Gomez and H.H. Valdivia for careful reading of this editorial. Conflict of interest: none declared. This work was supported by the grant ANR-11-1DEX-0003-02 to L.P. as members of the Laboratory of Excellence LERMIT.