Exchange Protein Directly Activated by cAMP (epac): A Multidomain cAMP Mediator in the Regulation of Diverse Biological Functions

1. Regulatory Role of Epac in Cardiac Calcium Homeostasis

Acute stimulation of β-adrenergic receptors and subsequent generation of cAMP regulate beneficial cardiac functions including cardiac contractility, relaxation, and heart rate (Rockman et al., 2002; Bers, 2008). Sustained elevation of cAMP, however, leads to hypertrophy and ventricular dysfunction and ultimately to the development of heart failure (Lohse et al., 2003; El-Armouche and Eschenhagen, 2009). Classically, most biologic effects of cAMP in the heart have been assigned to PKA, known to control excitation-contraction coupling through phosphorylation of L-type calcium channels, sarcoplasmic ryanodine receptors (RyR2), and the sarcoplasmic reticulum Ca²⁺-ATPase 2a regulator protein phospholamban, and thereby subsequently increases cardiac contractility (Rockman et al., 2002; Bers, 2007, 2008; Maier and Bers, 2007; Fig. 4). Increases in Ca²⁺ transients and cardiac contractility are generated upon enhanced myocardial relaxation and sacroplasmic reticulum loading through PKA-dependent phosphorylation of phospholamban, which in turn relieves phospholamban from its inhibition of sarcoplasmic reticulum Ca²⁺-ATPase 2a. In addition, myofilament relaxation and decreased Ca²⁺ sensitivity are accelerated through PKA-dependent phosphorylation of cardiac troponin I (cTnI) at Ser23/24 and cardiac myosin-binding protein C (cMyBPC) at Ser283 (Maier and Bers, 2007; Bers, 2008). Cardiac excitation-contraction coupling also encompasses activation of L-type calcium channels through cardiomyocyte depolarization, which in turn activates the sacroplasmic RyR2, resulting in cardiac contractility. Fine tuning of excitation-contraction coupling is achieved by Ca²⁺/calmodulin-dependent kinase II (CaMKII) also known to phosphorylate phospholamban and RyR2 and thereby to enhance sacroplasmic Ca²⁺ uptake and spontaneous Ca²⁺ release (Maier and Bers, 2007; Bers, 2008). From the three β-adrenergic subtypes expressed in the mammalian heart, regulation of cardiac functions is ascribed to the β₁- and β₂-adrenergic receptor subtypes (Lompré et al., 2010). Both receptor subtypes regulate distinct physiologic and pathophysiological responses. The β₁-adrenergic receptor, but not the β₂-adrenergic receptor, stimulates PKA-dependent phosphorylation of phospholamban and cardiac contractile proteins, induce hypertrophy upon moderate overexpression, and promote CaMKII-mediated cardiomyocyte apoptosis (Lompré et al., 2010; Fig. 4).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Epac and the heart. Epac provokes excitation-contraction coupling in response to β₁-adrenergic receptor stimulation through an Epac-Rap-PLC-ε-PKC-ε-CaMKII pathway and subsequent CaMKII-mediated phosphorylation of RyR2 and PLB. In addition, Epac-stimulated CaMKII directly phosphorylates cTnI and cMyBPC and thereby directly regulates myofilament function, a process supported by the induction of calmodulin. (A) Next to the accumulation of Cx43 through Epac at cardiac gap junctions, phosphorylation of Cx43 through PKCε enhances cardiac intercellular communication. Upon sustained cAMP elevation under cardiac stress, Epac induces gene transcription of prohypertrophic transcription factors including SkM α-actin, NFAT, MEF2, and ANF. Scaffolding of the Epac effector PLC-ε to the nuclear envelope-bound mAKAPβ induces ANF. Moreover, Epac induces NFAT, ANF, and SkM α-actin through Rap2B-PLC-ε-Rac-CaN and MEF2 through Rap2B-PLC-ε-Ras-CaMKII as well as nuclear export of HDAC. (B) Epac inhibits prohypertrophic gene transcription through Epac1-mAKAP-Rap1-mediated ERK5 inhibition and PDE4D3 activation. ANF, atrial natriuretic factor; β-AR, β-adrenergic receptors; CaN, calcineurin, Cx43, connexin43; I_ca, L-type calcium channels; MEF2, myocyte enhancer factor2; PDE4D3, phosphodiesterase 4D3; PLB, phospholamban; SERCA, sarco/endoplasmic reticulum Ca²⁺-ATPase; SkM α-actin, skeletal muscle α-actin; T-tubule, transverse tubule.

Given the importance of cAMP and Ca²⁺ in cardiac physiology and pathophysiology, several studies aim to define a function of Epac in excitation-contraction coupling. Initial studies in rat neonatal cardiomyocytes show that Epac-selective activation using 8-pCPT-2′-O-Me-cAMP triggers a dramatic increase in spontaneous Ca²⁺ oscillations, leaving the amplitude of the Ca²⁺ spikes unaffected (Morel et al., 2005). Adult ventricular cardiomyocytes from mice deficient for the Epac effector PLC-ε possess a decreased electrically evoked Ca²⁺ transient (intracellular Ca²⁺ concentration) in response to β-adrenergic receptor stimulation (Wang et al., 2005). Additional studies show that β-adrenergic receptor stimulation regulates sacroplasmic reticulum Ca²⁺ release through activation of a Epac-Rap- PLC-ε-protein kinase Cε-CaMKII pathway and subsequent CaMKII-dependent phosphorylation of RyR2 and phospholamban (Oestreich et al., 2007, 2009). In rat ventricular cardiac myocytes, however, activation of Epac by 8-pCPT-2′-O-Me-cAMP increase the Ca²⁺ oscillations despite a decrease in sacroplasmic reticulum Ca²⁺ load and a reduction of Ca²⁺ transients, processes being accompanied by CaMKII-dependent phosphorylation of RyR2 (Pereira et al., 2007; Fig. 4). Thus, although recent studies report on the identical CaMKII-dependent phosphorylation site on RyR2 (Ser 2815), the net outcome on Ca²⁺ transients remains controversial. Recent studies in adult rat ventricular myocytes indicate that excitation-contraction coupling rely on spatiotemporal compartmentalization of cAMP signaling through the β-adrenergic receptor and both PDE3 and PDE4 (Leroy et al., 2008), implicating that such mechanisms may contribute to distinct biologic net outcomes. In the murine heart, increases of spontaneous Ca²⁺ oscillations by the Epac activator 8-pCPT-2′-O-Me-cAMP results in ventricular arrhythmogenesis, a process depending on CaMKII (Hothi et al., 2008), indicating that elevation of spontaneous Ca²⁺ oscillations by Epac and subsequent CaMKII-dependent RyR2 phosphorylations profoundly alter Ca²⁺ homeostasis and thus excitation-contraction coupling.

Clearly, Epac controls cardiac excitation-contraction coupling through phosphorylation of the myofilament proteins cTnI and cMyBPC (Cazorla et al., 2009). In rat ventricular cardiac myocytes, Epac and PKA have opposite effects on Ca²⁺-activated myofilament force and myofilament Ca²⁺ sensitivity. Epac enhances myofilament Ca²⁺ sensitivity and phosphorylates cTnI and cMyBPC through phospholipase C (PLC)-protein kinase C (PKC)-CaMKII-dependent pathways; however, the specific Epac-dependent phosphorylation sites in cTnI and cMyBPC still have to be defined (Cazorla et al., 2009; Fig. 4A). In addition, Epac induces the expression of calmodulin (Ruiz-Hurtado et al., 2012), providing a feed-forward mechanisms to maintain phosphorylation by CaMKII. As cardiac myofilament function is altered in hypertrophy and heart failure (Layland et al., 2005), these findings suggest that Epac, in a PKA-independent fashion, contributes to these pathologies. Neoformation of cardiac gap junction is another prerequisite to maintain cardiac contractility and to control excitation-contraction coupling. Initial studies in rat neonatal cardiomyocytes show that Epac-selective 8-pCPT-2′-O-Me-cAMP induces accumulation of the predominant cardiac gap junction protein connexin43, whereas PKA enhances the gating of connexin43-bearing gap junctions. In addition, upregulation of the adherens junction protein N-cadherin by Epac promotes gap junction formation (Somekawa et al., 2005; Fig. 4A). Recent studies in rat neonatal ventricular cardiomyocytes confirm the potential link between Epac and connexin43. Activation of Epac by 8-pCPT-2′-O-Me-cAMP induces phosphorylation of connexin43 through PKCε and subsequently elevates gap junction intercellular communication studied by fluorescence recovery after photobleaching, providing evidence that Epac does not only alter subcellular localization of connexin43 but also its function (Duquesnes et al., 2010). In Langendorff-perfused isolated hearts submitted to global ischemia, however, activation of Epac does not induce preconditioning, a phenomenon known to induce cardioprotection (Duquesnes et al., 2010; Fig. 4A).

2. Epac and Cardiac Hypertrophy

Cardiac hypertrophy serves as a short-term adaptive process to increase the contractile mass of the heart; however, growth signals responsible for hypertrophic growth seem to induce cardiomyocyte apoptosis (Adams and Brown, 2001; Chen and Frangogiannis, 2010; Insel et al., 2012). Recent studies indicate that Epac induces apoptosis in several cell types including cardiomyocytes, albeit it acts cell type dependent in concert with PKA (Kwak et al., 2008; Suzuki et al., 2010). In H9c2 cells and neonatal rat cardiomyocytes, the PDE4 inhibitor roflumilast protects against nitric oxide-induced apoptosis through cAMP-PKA-CREB (cAMP response element-binding protein) and Epac-PKB/Akt pathways (Kwak et al., 2008); in primary cultures of neonatal mouse cardiomyocytes, the inability of Epac to induce apoptosis is correlated with the lack of expression of the apoptosis inducer Bim (Suzuki et al., 2010), suggesting that induction of apoptosis by Epac requires the engagement of complex signaling networks known to determine cell fate. Indeed, apoptosis seems to generate a functional network to autophagy, which sense under certain circumstances stress adaptation to avoid cell death and thus suppresses apoptosis, whereas in other cellular settings it promotes an alternative cell death pathway (Maiuri et al., 2007; Rubinsztein et al., 2007). Autophagy is regulated by mammalian targets of rapamycin (mTOR1 and mTOR2) through the action of phosphoinositide 3 kinase-dependent PKB/Akt signaling (Rubinsztein et al., 2007), the latter known to represent an effector of Epac. Intriguingly, recent studies indicate that Epac regulates autophagy in response to α-hemolysin (Misra and Pizzo, 2012) and that Epac is part of multiprotein complexes bearing mTOR1 and mTOR2 (Kelly et al., 2010), implicating that Epac regulates autophagy through mTOR1 and mTOR2 pathways. Excessive deposition of extracellular matrix during cardiac fibrosis contributes to cardiac dysfunction (Yokoyama et al., 2008c; Chen and Frangogiannis, 2010; Insel et al., 2012). In cardiac fibroblasts, Epac inhibits migration and synthesis of extracellular matrix components, including collagen I and III. Loss of Epac1 by fibrinogenic TGF-β1 accelerates collagen deposition and reduces Epac1-dependent migration of fibroblasts (Yokoyama et al., 2008c; Table 1). The fibrogenesis mediator TGF-β1 recruits and activates (myo)fibroblasts, which may derive from resident mesenchymal cells, circulating fibrocytes, and/or epithelial-to-mesenchymal transition, a process in which epithelial cells transdifferentiate into mesynchymal cells (Schmierer and Hill, 2007; Biernacka et al., 2011; Nieto, 2011; Kovacic et al., 2012; Small, 2012; Fig. 8). Next to fibroblast-myofibroblast transformation, recent studies indicate that Epac inhibits, either alone or in concert with PKA, the induction of epithelial-to-mesenchymal transition by TGF-β1 in Madin-Darby kidney cells. Activation of Epac reversed the inhibitory effect of TGF-β1 on the epithelial marker E-cadherin and normalized the expression of α-smooth muscle actin (Insel et al., 2012). Epithelial-to-mesenchymal transition has been proposed as a mechanism for tissue fibrosis in numerous settings including development and cancer (Krenning et al., 2010; Chaffer and Weinberg, 2011). Cancer associated fibroblasts are linked to mesenchymal to mesenchymal transition, bone marrow mesenchymal stem cell transition, and endothelial to mesenchymal transition (Cirri and Chiarugi, 2011). Recent studies link Epac to central biologic responses underlying such cell transition processes including cell migration, cell adhesion, cell homing, and (neo)vascularization (Carmona et al., 2008; Baljinnyam et al., 2009, 2010, 2011; Jang et al., 2012; Jia et al., 2012; Pullamsetti et al., 2012; Tang et al., 2012). The Epac effector Rap2 is involved in Wnt/β-catenin signaling by stabilizing β-catenin (Choi and Han, 2005) and could in this contribute to fibrotic processes. Epac strengthens tethering of catenin to adherens junctions and tight junctions (Netherton et al., 2007; Raymond et al., 2007; Noda et al., 2010; Rampersad et al., 2010; Suh and Han, 2010). Intriguingly, the secreted lysophospholipase D autotaxin promotes cancer invasion through a cAMP-Epac-Rac1 pathway in response to lysophosphatidic acid receptor 4 stimulation, a process supported by blood vessel formation through autotaxin (van Meeteren et al., 2006; Harper et al., 2010). Lysophosphatidic acid bound to its receptor can also elevate the phospholipase D product phosphatidic acid (Samadi et al., 2011), which has been recently shown to regulate subcellular localization and functions of Epac (Consonni et al., 2012; Gloerich et al., 2012; Fig. 2).

Expression of Epac is increased in patients with cardiac hypertrophy as well as in different animal models of myocardial hypertrophy and heart failure (Table 1); however, at present it is not known whether this is a cause or consequence of cardiac hypertrophy. Both Epac mRNAs are upregulated by chronic catecholamine infusion-induced hypertrophy, and only Epac1 is increased in pressure overload-induced hypertrophy (Ulucan et al., 2007). Expression of Epac1 protein is upregulated at the early phase of cardiac pressure overload-induced hypertrophy in rat left ventricular myocardium and in left ventricular samples from patients with heart failure (Métrich et al., 2008; Table 1). Activation of Epac by 8-pCPT-2′-O-Me-cAMP increases the cell surface area, protein synthesis, and expression of cardiac hypertrophy markers, including atrial natriuretic factor, c-fos, skeletal muscle α-actin, and nuclear factor of activated T cells (NFAT) (Morel et al., 2005; Métrich et al., 2008). Importantly, silencing of Epac1 expression strongly reduces the induction of the hypertrophic markers by β-adrenergic receptor stimulation (Métrich et al., 2008). Epac induces cardiac hypertrophic responses through activation of a Rap2B, Rac, H-Ras, PLC, the phosphatase calcineurin, and CaMKII (Morel et al., 2005; Métrich et al., 2010; Pereira et al., 2012; Fig. 4B). Sustained activation of Epac correlates with the induction of calmodulin (Ruiz-Hurtado et al., 2012); together with the induction of skeletal muscle α-actin Epac seems to fulfill central compensatory cardiac functions during the initiation of cardiac hypertrophy, favoring contraction but also risking the development of arrhythmias (Morel et al., 2005; Ruiz-Hurtado et al., 2012).

3. Epigenetic Regulation by Epac

Intriguingly, it is reported for the first time that Epac activation induces nuclear export of HDAC5, a process involving HDAC4 in a cell-type specific manner (Métrich et al., 2010) and subsequent induction of prohypertrophic transcription factors including NFAT and myocyte enhancer factor 2 in primary cardiac myocytes (Métrich et al., 2010; Pereira et al., 2012). These novel findings can be added to the growing list of nuclear-associated signaling properties of Epac, including nuclear translocation of the DNA damage-responsive DNA-PK (Huston et al., 2008), cell cycle-dependent localization of Epac to the mitotic spindle and centrosomes (Qiao et al., 2002), and targeting of Epac to the nuclear pore via Ran and/or RanBP2 (Liu et al., 2010a; Gloerich et al., 2011; Fig. 2). In addition, microtubules and RanBP2 seem to modulate the exchange activity of Epac at the nuclear envelope (Magiera et al., 2004; Gupta and Yarwood, 2005; Mei and Cheng, 2005; Borland et al., 2006; Gloerich et al., 2011), a process potentially being supported by components of the nucleoskeleton (Simon and Wilson, 2011; Jaspersen and Ghosh, 2012). Induction of skeletal muscle α-actin and calmodulin by Epac requires its interaction with myocardin-related transcription factors (Parmacek, 2007; Braun and Gautel, 2012; Small, 2012), a process being tightly controlled by the actin and microtubule cytoskeleton (Olson and Nordheim, 2010). The discovery that Epac regulates nuclear export of a specific subset of HDAC family members adds another puzzling aspect to its complex signaling. Although the detailed mechanisms of Epac-induced nuclear export of HDACs remain to be identified (Métrich et al., 2010; Pereira et al., 2012; Fig. 4B), the reports place Epac not only into the context of epigenetic regulation of gene expression (Arrowsmith et al., 2012), but also in post-transcriptional modification of cytosolic proteins via (de)acetylation. Histone modifications, such as acetylation and phosphorylation, play an important role in the regulation of gene transcription (Dekker and Haisma, 2009; Ghizzoni et al., 2011). Next to HDACs, histone acetylation is modulated by histone acetyltransferases. Whereas histone acetyltransferases promote the opening of the chromatin structure by acetylating lysine residues in histone proteins leading to gene activation, HDACs promote the reverse process. Interestingly, the class II HDAC family members HDAC4 and HDAC5 are reported to act as signal-responsive suppressors of the transcriptional program governing cardiac hypertrophy and heart failure (Zhang et al., 2002), which is in line with the previous reported role of Epac in the development of these diseases. Cardiac stress signals activate the phosphatase calcineurin and subsequently induce kinase activities directed at the class II HDACs including HDAC4 and HDAC5. Phosphorylation of HDAC4 and HDAC5 triggers their cytosolic interaction with the 14-3-3 scaffold proteins and thus promotes nuclear export. Due to the nuclear export of HDAC4 and HDAC5, transcription factors favor binding to HATs known to facilitate gene transcription (Zhang et al., 2002). In light of the growing number of Epac-interacting proteins future studies should focus on a potential link between Epac and the scaffold protein 14-3-3.

4. The Cardiac Epac Signalosome

Despite the compelling evidence that links Epac to the induction of cardiac hypertrophy, the current studies remain controversial. In neonatal rat cardiomyocytes, initial studies report that multiprotein complex composed of nuclear envelope-associated mAKAP, PKA, PDE4D3, and Epac1 control cardiomyocyte hypertrophy through regulation of ERK5, a process tightly regulated by cAMP (Dodge-Kafka et al., 2005). It is shown that high cAMP attenuates cardiac hypertrophy through Epac1-Rap1-dependent inhibition of ERK5 and subsequent activation of PDE4D3; however, low cAMP enhances cardiac hypertrophy through ERK5-mediated inactivation of PDE4D3 and subsequent increased PKA-mediated RyR2 phosphorylation as well as NFAT induction (Dodge-Kafka et al., 2005; Fig. 4B). Likewise, mice deficient for the Epac effector PLC-ε show increased cardiac hypertrophy in response to chronic stress (Wang et al., 2005), further substantiating the notion that Epac inhibits cardiac hypertrophy. However, earlier studies demonstrate that RyR2 binds to mAKAP and induce cardiac hypertrophy in response to β-adrenergic receptor stimulation (Pare et al., 2005), suggesting that CaMKII-dependent phosphorylation of RyR2 by Epac also modulates cardiac hypertrophy (Morel et al., 2005; Oestreich et al., 2007, 2009). Indeed, in neonatal rat ventricular cardiomyocytes depletion of the Epac effector PLC-ε using siRNA reduces hypertrophic growth and induction of atrial natriuretic factor. PLC-ε binds to the striated myocytes exclusively expressed mAKAPβ, which promotes scaffolding of PLC-ε to the nuclear envelope (Zhang et al., 2011; Fig. 4B). Thus, studies that potentially link Epac to cardiac hypertrophy are still under debate.

To add even more complexity, earlier studies report that the PLD product phosphatidic acid activates PDE4D3 (Grange et al., 2000) and that PLD expression is upregulated in response to ventricular pressure-overload hypertrophy (Peivandi et al., 2005). These findings implicate that alterations in activity and/or expression of factors that are part of cAMP-sensing multiprotein complexes, such as PDE4D3 (Dodge-Kafka et al., 2005), or factors that determine the subcellular localization of Epac, such as phosphatidic acid (Consonni et al., 2012), exhibit different signaling properties and therefore alter the ability of Epac to act either anti- and/or pro-hypertrophic in the heart. In vascular smooth muscle Epac1 enhances the expression of the α_2C-adrenergic receptor through a Rap1-RhoA signaling pathway (Jeyaraj et al., 2012), suggesting that Epac may even alter the expression profile of β₁- and/or β₂-adrenergic receptors known to be linked to cardiac functions. In addition, a β-arrestin-CaMKII-Epac1 complex is formed in response to the stimulation of the β₁-adrenergic receptor, but not of the β₂-adrenergic receptor (Mangmool et al., 2010). Ligand-directed or -biased agonism through β-arrestin has been reviewed in the context of heart failure (Whalen et al., 2011; Reiter et al., 2012), suggesting that β-arrestin-Epac1 complex formation may contribute to cardiac hypertrophy. As outlined above cardiac hypertrophy may lead to heart failure (Lohse et al., 2003; El-Armouche and Eschenhagen, 2009), a process predominantly relying on cardiac signaling in response to the β₁-adrenergic receptor. Indeed, recent studies show that heart failure changes primarily compartmentalized cAMP signaling in response to the β₂-adrenergic receptor, but not in response to the β₁-adrenergic receptor (Nikolaev et al., 2010). By use of nanoscale live scanning ion conductance combined with fluorescence resonance energy transfer techniques, it is shown that the β₁-adrenergic receptor is distributed throughout the entire cardiomyocyte cell surface encompassing the transverse T-tubules and the cell crest, whereas the β₂-adrenergic receptors exclusively localized in transverse T-tubules (Nikolaev et al., 2010). These are novel findings that perfectly reflect their functional diversity in the heart. Nevertheless, the relative contribution of the adrenergic receptor subtypes to the apparent link between Epac and cardiac hypertrophy remains to be studied.

1. Role of Epac in Neuronal Differentiation and Regeneration

Until very recently, most biologic effects of cAMP known to regulate molecular mechanisms of a wide variety of neuronal processes, including development, growth, differentiation, neurotransmitter release, excitability, remodeling, plasticity, learning, and memory, have been assigned to PKA (Tronson and Taylor, 2007; Abel and Nguyen, 2008; Lee and Silva, 2009). Several studies, however, clearly show that Epac, acting either alone or in concert with PKA, regulates neuronal physiology through its signaling properties to a plethora of distinct effectors and contributes to the pathophysiology of several diseases (Grandoch et al., 2010). Epac regulates neuronal differentiation, neurite outgrowth, and axon regeneration, suggesting that Epac plays a key role in the development and the maintenance of the nervous system. In PC12 cells, initial findings demonstrated that the Epac activator 8-pCPT-2′-O-Me-cAMP induces neurite outgrowth (Christensen et al., 2003). Subsequent studies show that, depending on the duration of signaling, Epac converts cAMP from a proliferative into a differentiation signal (Kiermayer et al., 2005). In retinal ganglion cells, Epac links, both in vitro and in vivo, ganglion cell survival and axon growth to PKA and soluble AC (Corredor et al., 2012). As Epac regulates the phosphorylation of PKB/Akt in primary cortical neurons (Nijholt et al., 2008), this Epac signaling route probably supports neuronal survival. In human SH-SY5Y neuroblastoma cells, Epac regulates neuronal differentiation in response to the pituitary adenylate cyclase-activation polypeptide (PACAP)-38 (Monaghan et al., 2008). Of note, neuronal differentiation in response to PACAP-38 involves activation of Rit by Epac independent of Rap (Shi et al., 2006; Fig. 5). As Rit is implicated in neurotrophin signaling (Shi and Andres, 2005), these findings indicate that Epac uses a non-classic signaling property to create a novel molecular link to neuronal development and regeneration.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Epac and the brain. Epac is linked to diverse neuronal processes including neuronal differentiation, neurite outgrowth, axon regeneration, remodeling, neurotransmitter release, pain perception, as well as learning and memory. Next to physiologic neuronal processes linked to Epac, Epac is implicated in brain disorders, such as anxiety and depression, autism, schizophrenia, Alzheimer’s disease, and Huntington’s disease. Identified signaling routes are indicated. LTP, long-term potentiation; MAPK, mitogen-activated protein kinase; miR-124, microRNA-124; PKB/Akt, phosphoinositide 3-kinase-dependent protein kinase B/Akt; sAC, soluble adenylyl cyclase.

Although few studies indicate that developmental and/or regenerative neuronal responses do not rely on Epac (Emery and Eiden, 2012; Xu et al., 2012), a growing list of recent studies point to a central function of Epac. Together with PKA, Epac regulates polarized responses of neuronal growth cone attraction and repulsion. The key axon guidance cues, however, from attraction to repulsion require a switch from Epac to PKA activation (Murray et al., 2009a). Next to Epac-induced neurite outgrowth in rat dorsal root ganglia, Epac enhances neurite regeneration on adult spinal cord in vitro (Murray and Shewan, 2008; Fig. 5), suggesting that Epac represents a target to induce axon regeneration in vivo. In neurons within the pre-Bötzinger complex, Epac induces neurite outgrowth and reinforces neuronal bursting activity (Mironov et al., 2011; Mironov and Skorova, 2011), suggesting that Epac signaling properties compensate for neurodevelopmental deficits characteristic for the Rett syndrome.

2. Regulation of Neuronal Signaling by Epac

In glutamatergic synapses, Epac enhances neurotransmitter release from the rat brain calyx of Held (Sakaba and Neher, 2003) and in the crayfish neuromuscular junction (Zhong and Zucker, 2005). In dorsal root ganglion, Epac activates PKCε translocation and contributes to perception of inflammatory pain in response to PGE₂-induced P₂X₃ receptor responses (Hucho et al., 2005; Fig. 5; Table 1); the upregulated Epac1 expression in inflamed neurons is expected to sensitize pain perception (Wang et al., 2007). Intriguingly, recent studies provide new insights into the role of Epac in inflammatory hyperalgesia. It is shown that the low nociceptor G protein-coupled receptor kinase 2 (GRK2) leads to prolonged inflammatory hyperalgesia through biased signaling from PKA to Epac-Rap1, ERK/PLCε (Eijkelkamp et al., 2010). In cultured mouse cerebral neurons, Epac activates neuronal excitability upon modulation of Ca²⁺-dependent K⁺-channels through Rap and p38 mitogen activated protein kinase (Ster et al., 2007, 2009; Fig. 5). In the suprachiasmatic nuclei of the hypothalamus, Epac is part of the core component of the mammalian circadian pacemaker (O'Neill et al., 2008). In spine synapses, Epac2 promotes dynamic remodeling and depression of spiny synapses through the postsynaptic adhesion molecule neuroligin 3, the scaffold protein PSD-95, and Rap (Woolfrey et al., 2009; Penzes et al., 2011).

3. Learning and Memory Retrieval

Several recent studies indicate that Epac regulates distinct steps in processes required for synaptic plasticity, learning, and memory. In the CA1 area of mouse hippocampal slices, Epac enhances long-term potentiation known to be one characteristic of synaptic plasticity through ERK1/2 signaling (Gelinas et al., 2008). In the hippocampus, Epac mediates PACAP-dependent long-term depression through Rap, p38 mitogen-activated protein kinase, and PSD-95/Disc-large/ZO-1 (PDZ) homology domains (Ster et al., 2009). Independent of PKA, Epac enhances memory consolidation in the hippocampus (Ma et al., 2009). In dopamine-β-hydroxylase-deficient mice, coapplication of selective activators for both PKA and Epac rescues memory impairments in contextual fear conditioning (Ouyang et al., 2008). Subsequent studies indicate that β₂-adrenergic receptors through coupling to G_i/o and β₁-adrenergic receptors through coupling to G_s have opposing effects on hippocampus-dependent emotional memory (Schutsky et al., 2011), suggesting that the beneficial effects of Epac on memory retrieval are also controlled by the relative input of adrenergic receptor subtypes. Intrahippocampal injection of the Epac activator 8-pCPT-2′-O-Me-cAMP demonstrates that Epac2 specifically enhances time-limited memory retrieval and contextual fear conditioning (Ostroveanu et al., 2010). Recent studies in Epac(−/−) double knockout mice confirmed these initial findings. The double knockout mice of Epac1 and Epac2 exhibit severe defects in long-term potentiation, spatial learning, and social interactions through a Rap1-dependent suppression of miR-124 and subsequent Zif268 translation (Yang et al., 2012; Fig. 5). These research findings offer several novel concepts to the scientific community. First, as both Epac1(−/−) and Epac2(−/−) are phenotypically normal (Yang et al., 2012), the concept that both Epac1 and Epac2 bear functional redundancy is further supported. Second, it is the first report that links Epac to the mechanisms of post-transcriptional regulation by microRNAs. Currently, individual microRNAs are believed to act as pivotal modulators of mammalian functions during both development and disease (Filipowicz et al., 2008). Additional studies in Epac2-deficient mice indicate that Epac2 depletion is sufficient to induce impairments in social interactions, ultrasonic vocalization, and cortical structures (Srivastava et al., 2012a), suggesting that both Epac1 and Epac2 control a distinct subset of neuronal functions with potential alterations under diseased conditions.

4. The Epac Signalosome in Brain Disorders

Indeed, recent studies link specific defects in neuronal Epac1 and/or Epac2 signaling to brain disorders (Table 1). In human subjects diagnosed for autism, an Epac2 rare coding variant controls basal dendritic morphology through a Ras-Eapc2-Rap pathway (Srivastava et al., 2012b). In addition, reduced signaling of Gα_s is associated with schizophrenia symptoms through defects in Epac signals (Kelly et al., 2009). Initial studies associate Epac-1 gene variants with anxiety and depression (Middeldorp et al., 2010). Additional studies demonstrate that the corticotropin releasing factor receptor modulates anxiety and depression through a wide range of Epac-associated signals including Rap-PLCε-ERK (Hauger et al., 2009; Fig. 5). In human brain regions associated with Alzheimer’s disease, upregulation of Epac1 mRNA correlates with a downregulation of Epac2 mRNA (McPhee et al., 2005). Initial studies in primary neurons show that Epac increases the activity of α-secretase and subsequently the secretion of the neuroprotective and memory-enhancing soluble amyloid precursor protein through both Rap1 and Rac (Maillet et al., 2003; Robert et al., 2005; Fig. 5). Studies in primary cortical neurons indicate that Epac, primarily Epac2, induces phosphorylation of PKB/Akt known to support neuronal survival and memory processes (Nijholt et al., 2008), suggesting that reduced Epac2 functioning in patients with Alzheimer’s disease promotes loss of neuroprotective PKB/Akt signaling. Next to the neurodegenerative disorder Alzheimer’s disease, Epac is linked to polyglutamine disorders, including Huntington’s disease (Sarkar et al., 2009). It is shown that Epac regulates autophagy independent of mTOR through cAMP-Rap2B-PLCε- and thereby supports Ca²⁺-calpain signaling (Williams et al., 2008; Fig. 5).

These findings indicate that the link of Epac, acting either alone and/or in concert PKA, to several brain disorders requires its localization to specific brain regions and signaling to distinct effectors in a time- and space-limited manner. In murine primary cortical neurons and HT-4 cells, Epac2 enhances PKB/Akt phosphorylation through Rap1 complexed to AKAP79/150 (Nijholt et al., 2008). In addition, Epac promotes neuronal survival and neurite growth through a nuclear complex composed of PKA and sAC (Corredor et al., 2012), suggesting that distinct cAMP microdomains maintained by sAC profoundly alter the signaling properties of Epac. The ability of Epac2 to promote remodeling of spiny synapses requires a distinct complex composed of neuroligin-3/PSD95/Rap complex (Woolfrey et al., 2009). Earlier reports indicate that the scaffold PSD-95 also interacts with the Ras-GAP p135 SynGAP (Chen et al., 1998), indicating that PSD-95 might represent the molecular link between Epac2 and Ras in neurons. Finally, GRK2 binds to Epac1 and thereby promotes biased signaling through Rap1-PLC-ε-ERK (Eijkelkamp et al., 2010), indicating that new G protein-coupled receptor signaling paradigms modulate the neuronal signaling properties of Epac. In conclusion, these findings indicate that next to complex formation of Epac with members of the AKAP family scaffolding through PSD-95 and GRK2 may alter its impact on neuronal processes both under physiological and pathophysiological settings.

1. Regulation of Insulin Secretion by Epac

The metabolic disease diabetes is expected to be a leading cause of disability, morbidity, and premature mortality by the year 2030, which will severely affect both developed and developing countries (Rydén et al., 2007; Chalmers and Cooper, 2008; Danaei et al., 2011). The complications of diabetes primarily affect the eyes, kidneys, nerves, and the cardiovascular system, causing several microvascular-associated complications, including kidney failure. Current treatment of diabetes, particularly type 2 diabetes, focuses on reducing glucose levels by elevating the plasma levels of insulin through several distinct molecular mechanisms, including inhibition of ATP-sensitive K⁺-channels by sulfonylureas on pancreatic β-cells (Doyle and Egan, 2007; Verspohl, 2012). Insulin is produced in the islet of Langerhans in the pancreatic β-cells. Glucose induces a pulsatile insulin secretion, and it has been proposed that the observed initial and late secretion of insulin point to functionally distinct granule pools (Tengholm and Gylfe, 2009). The incretin hormone glucagon-like peptide-1 (GLP-1) and gastric inhibitory peptide are produced in the gut upon duodenum entry of food containing fat, peptides, and/or glucose. In pancreatic β-cells, glucose-induced insulin secretion is by the cAMP signaling pathway following GLP-1 receptor stimulation (Holz et al., 2006; Doyle and Egan, 2007; Fig. 6). By use of evanescent-wave fluorescence imaging as a new technique for single-cell measurements of cAMP, it is reported that glucose-induced cAMP oscillations are linked to the pulsatile insulin secretion in the pancreatic MIN6 and in primary mouse β-cells (Dyachok et al., 2008). Several studies indicate that the effect of cAMP acts via activation of Epac—partly in cooperation with PKA. Thus, silencing studies have assigned Epac2 as the key player in glucose-induced insulin secretion through pulsatile cAMP oscillations (Idevall-Hagren et al., 2010). As described in detail in section I.D, Epac2 is allosterically activated through sulfonylureas, which are widely used as antidiabetic drugs. These findings further link Epac to insulin secretion as sulfonylureas potentiate glucose-induced insulin secretion upon binding to the SUR1 subunit of ATP-sensitive K⁺-channels, upon direct interaction of sulfonylurea with arginine 447 in Epac2 (Zhang et al., 2009; Herbst et al., 2011). The initial findings, however, ignite a controversial debate among experts in the field (Tsalkova et al., 2011; Parnell et al., 2012; Rehmann, 2012; Fig. 6). Nevertheless, the studies directly link Epac2 to both sulfonylureas and SUR1, implicating that Epac2 captures a rather central function to fine tune the SUR1-dependent effects of the antidiabetic sulfonylureas (Doyle and Egan, 2007; Verspohl, 2012).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Epac and the pancreas. In pancreatic β-cells, glucose-induced insulin secretion is potentiated through Epac2-Rim2-Piccolo signaling in response to stimulation of the GLP-1 receptor. Closure of ATP-sensitive K⁺-channels (K_ATP) is promoted through Epac2-SUR1 interaction and Epac2-Rap-dependent stimulation of PLC-ε, the latter known to increase channel ATP sensitivity through hydrolysis of PIP₂. Subsequent membrane depolarization induces opening of voltage-dependent Ca²⁺-channels (VDCC) and facilitates insulin secretion through fusion of insulin granules with the plasma membrane. Interaction of Epac2 with Rim2 and Piccolo promotes insulin granule fusion and thereby insulin secretion, such as the activation of calcium channels on the endoplasmic reticulum (ER). Epac2 interacts in a Rim2-Rab3-dependent manner with glucokinase and thereby enhances mitochondrial metabolism of glucose. IP₃R, inositol-1,4,5-trisphosphate receptor; ROS, reactive oxygen species.

Epac2, which is the predominantly expressed Epac isoform in pancreatic β-cells, is required for the potentiation of insulin secretion by incretins (Kashima et al., 2001). Additional studies show that Epac2 binds to Rim (Rab3-interacting molecule) and to Rim2, a novel regulator of fusion of vesicles to the plasma membrane (Ozaki et al., 2000; Shibasaki et al., 2004). Potentiation of Ca²⁺-dependent insulin granule excocytosis in response to GLP-1 relies, at least in part, on Epac2 (Kang et al., 2001, 2003). In pancreatic β-cells, the Ca²⁺ sensor Piccolo stabilizes the Epac2-Rim2 complex through heterodimerization (Fujimoto et al., 2002). Likewise, Epac2 directly interacts with the SUR1 subunit of ATP-sensitive K⁺ channels and thereby regulates priming of pancreatic β-cells (Eliasson et al., 2003; Fig. 6). Recent findings indicate that GLP-1 potentiates glucose-induced insulin secretion by enhancing glucokinase activity in pancreatic β-cells through Epac2-Rim2-Rab3 (Park et al., 2012). Interestingly, GLP-1 potentiates both Ca²⁺-dependent insulin granule excocytosis and glucose-induced insulin secretion through Epac2-Rim2-Rab3 signaling. In primary pancreatic β-cells from Epac2 knockout mice, the initial phase of insulin secretion is severely reduced (Shibasaki et al., 2007), further substantiating that Epac2 is of central importance for the release of insulin from pancreatic β-cells and thus for the regulation of blood glucose. Interestingly, GLP-1 enhances glucokinase activity through Epac2-Rim2-Rab3 (Park et al., 2012; Fig. 6). Although the studies link glucokinase to Epac2, these findings indicate that Epac not only regulates insulin secretion but also profoundly alters the metabolism of glucose because glucokinase controls mitochondrial consumption of glucose. In conclusion, several molecular mechanisms that regulate glucose-mediated insulin secretion upon engagement of a sophisticated network of scaffold proteins and ion channels have identified Epac2 as a key player.

GLP-1-induced potentiation of Ca²⁺-dependent insulin granule excocytosis is mediated by Epac2 (Kang et al., 2001, 2003). Likewise, the Epac activator 8-pCPT-2′-O-Me-cAMP induces a ryanodine-sensitive Ca²⁺-mobilization in pancreatic β-cells (Kang et al., 2001, 2003; Holz et al., 2006), suggesting that Epac2 also contributes to calcium-induced calcium release from the endoplasmic reticulum. Earlier studies indicate that Epac induces inositol-1,4,5-trisphosphate-sensitive Ca²⁺-mobilization through Rap2B-dependent stimulation of PIP₂-hydrolyzing PLC-ε (Schmidt et al., 2001), a mechanism being expected to contribute to Ca²⁺-dependent insulin granule excocytosis. In addition, the hydrolysis of PIP₂ through Epac2 is expected to increase ATP-sensitive K⁺-channels responsiveness to ATP and subsequently to promote their closure (Holz et al., 2006), leading to insulin secretion. Indeed, recent studies show that PLC-ε links Epac2 to the potentiation of insulin secretion in mouse islets of Langerhans (Dzhura et al., 2011; Fig. 6). In transgenic mice that lack the Epac effector PLC-ε, the Epac activator 8-pCPT-2′-O-Me-cAMP loses the ability to potentiate glucose-induced insulin secretion. Importantly, however, these mice also lose their ability to increase calcium in pancreatic β-cells (Dzhura et al., 2011), indicating that a Rap1-regulated PLC-ε also links Epac2 to Ca²⁺-dependent insulin granule excocytosis. Since a recent study in cardiomyocytes reported on scaffolding of PLC-ε to mAKAP-β (Zhang et al., 2011), it is tempting to speculate that cellular PLC-ε scaffolding contributes to its ability to potentiate glucose-induced insulin secretion as well.

2. The Epac Signalosome in Diabetes-Related Diseases

Recent studies link signaling properties of Epac1 and/or Epac2 to diabetes-related diseases, particularly kidney failure. Activation of Epac by 8-pCPT-2′-O-Me-cAMP induces protection of tubular epithelial cells from cisplatin-induced apoptosis (Qin et al., 2012). In a mouse model of ischemia-induced kidney failure, the Epac activator 8-pCPT-2′-O-Me-cAMP preserves the epithelial barrier function during hypoxia in vitro (Stokman et al., 2011). Importantly, intrarenal administration of 8-pCPT-2′-O-Me-cAMP reduces renal failure in vivo through a reduced expression of the tubular stress marker clusterin-α and lateral expression of β-catenin, which is indicative for sustained tubular function (Stokman et al., 2011). In accordance with the expression profile of Epac1 and Epac2 in the kidney (Ulucan et al., 2007; Li et al., 2008), silencing studies indicate that the renal protective effects require Epac1 (Stokman et al., 2011; Qin et al., 2012). Recent studies indicate that the GLP-1 receptor agonist exendin-4 suppresses the production of detrimental reactive oxygen species through Epac (Mukai et al., 2011), which may contribute to the protection against renal ischemia-reperfusion injury through Epac1. Reactive oxygen species (Fig. 6) also intermingle with the mitochondrial glucose consumption controlled by glucokinase, in particular under disease conditions such as diabetic nephropathy (Singh et al., 2008; Kanwar et al., 2011). Complex mechanisms obviously contribute to glucose homeostasis through the signaling properties of Epac. Recent studies indicate that, in contrast to the positive effects of Epac2, Epac1 mediates high glucose-induced renal proximal tubular cell hypertrophy through Akt and the cyclin-dependent kinase inhibitor p21 (Sun et al., 2011). In addition, Epac induces the resistance to leptin through an impairment of leptin-induced Rap1-SOCS-3 signaling and subsequent elevation of JAK-STAT responses (Fukuda et al., 2011). Thus, Epac bears, in addition to its beneficial effects, rather deleterious signaling properties under diabetes-related disorders. In this context it is unclear whether the finding that Epac1 is increased in the cortical tubules of the kidney in a mouse model of streptozotocin-induced diabetes (Sun et al., 2011; Table 1) is a cause or a consequence. Interestingly, the level of induction was correlated with the level of glucose in the blood of these animals (Sun et al., 2011). A causal relation between increased glucose levels and increased expression of Epac2 was demonstrated by the finding that d-glucose, but not l-glucose, induces Epac1 expression in human kidney HK-2 cells (Sun et al., 2011; Table 1). Future studies on the relative contributions of Epac1 versus Epac2 in regulating glucose homeostasis are warranted. Studies in Epac1 and Epac2 transgenic mice under disease conditions would help to clarify whether Epac2 indeed is beneficial, whereas Epac1 is detrimental. Clearly, this research will also greatly benefit from subtype selective pharmacological activators and inhibitors (see section I.C).

1. Epac and the Endothelial Barrier Function

The vascular endothelium covers the inner lining of blood vessels and serves as an interface between circulating blood and surrounding tissue. The semipermeable endothelial barrier dynamically regulates exchange of ions, solutes, and cells both through and between cells or the endothelium (Mehta and Malik, 2006; Vestweber, 2007; Giepmans and van Ijzendoorn, 2009; Pannekoek et al., 2009). During acute inflammation, leukocytes transmigrate through the permeable endothelial barrier; however, chronic inflammation also leads to excessive extravasation of blood components and fluid accumulation into the extravascular space, causing edema and vascular dysfunction (Hirase and Node, 2012; Vestweber, 2012). The barrier is maintained by tight junctions and adherens junctions, creating strong cell-cell contacts. Tight junctions act as size-selective barriers and maintain cell polarity due to their apical-lateral cell surface localization (Chiba et al., 2008; Anderson and Van Itallie, 2009). Adherens junctions are located at the more basal cell side and express (V)E-cadherin and catenin (Giepmans and van Ijzendoorn, 2009; Pannekoek et al., 2009; Vestweber et al., 2009). (V)E-cadherin interacts with α-catenin and/or β-catenin and thereby connects the adherens junctions to the circumferential belt of actin filaments to strengthen cell-cell contacts. Barrier disrupting agents, including tumor necrosis factor-α, thrombin, and the bacterial endotoxin lipopolysaccharide, alter the dynamic actin-microtubule network architecture of tight and adherens junctions (Niessen, 2007; Vestweber, 2007; Vestweber et al., 2009, 2010; Sayner, 2011; Fig. 7).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Epac and the vasculature. Barrier disrupting agents, such as TNF-α, thrombin, and LPS, signal to Rho and subsequently reduce the VE-cadherin-dependent endothelial barrier through a reduced stability of cortical actin and microtubules as well as an increased formation of actin stress fibers. Epac reduces endothelial barrier leakage through skewing the balance of RhoA/Rac1 activation toward Rac1, thereby leading to a reduction of the phosphorylation of the Rho-Rho-kinase target myosin light chain and subsequent reduction of actin stress fiber formation. In addition, Epac promotes microtuble growth. Epac directly interacts with PDE4D at VE-cadherin-bearing cell-cell contacts, a process further supporting its endothelial barrier stabilizing function. Interaction of Epac with AKAP9 represents a direct link between VE-cadherin and integrins. The Epac effector Rap1 stabilizes β-catenin-bearing cell-cell contacts through Krit. Targeting of Epac to integrins at junctional sites seems to be supported through ERM-Epac interactions. Epac limits proinflammatory IL-6 actions through induction of the suppressor of IL-6 receptor signaling SOCS-3 and subsequent inhibits JAK-STAT signaling through Rap-PKC-ε-ERK1/2-C/EBPs. C/EBP, CCAAT/enhancer-binding protein; JAK, Janus kinase; Krit, Krev1 interaction trapped gene; LPS, lipopolysaccharide; PDE4D, phosphodiesterase 4D; STAT, signal transducer and activator of transcription; TNF-α, tumor necrosis factor-α.

Reduction of the barrier leakage is achieved by cAMP elevation in response to β₂-agonists, prostanoids, and forskolin and has been linked to the inhibition of actin-microtubule dynamics linked to tight and adherens junctions under tight control of members of the Ras family, in particular Rho GTPases and Rap. Several studies indicate that Epac, acting either alone and/or in concert with PKA, reduces barrier leakage predominantly through its signaling properties to Rho via Rap on the actin-microtubule network. Notably, Epac1 modulates the endothelial barrier independent from the Rho GTPases and Rap (see below). As Rap1 promotes cell-cell junction formation through signaling to E-cadherin-catenin and integrin-extracellular matrix complexes (Retta et al., 2006; Pannekoek et al., 2009; Noda et al., 2010; Suh and Han, 2010; Fig. 7), reduction of endothelial permeability through Rap1 activation by Epac1 most likely reflects its signaling property to the actin-microtubule network. Several studies indicate that Epac promotes microtubule elongation and thereby enhances the endothelial barrier through VE-cadherin and integrins (Sehrawat et al., 2008, 2011). As earlier studies indicate that microtubulin enhances the exchange activity of Epac toward Rap1 (Magiera et al., 2004; Gupta and Yarwood, 2005; Borland et al., 2006, 2009), the interaction of Epac with the actin-microtubule network seems to be bidirectional in nature. In addition, interaction of Epac with ERM proteins seems to serve as an additional molecular link to promote cell adhesion (Gloerich et al., 2010; Ross et al., 2011), a process known to rely on the communication between integrins and extracellular matrix components (Pannekoek et al., 2009; Fehon et al., 2010). ERM proteins require for their full activation PIP₂, which is produced by phosphatidyl-4-phosphate 5-kinases in a phosphatidic acid-sensitive manner (Oude Weernink et al., 2000, 2007; Weernink et al., 2004). Recent findings indicate that subcellular localization of Epac1 is determined by its interaction with phosphatidic acid (Consonni et al., 2012; Gloerich et al., 2012). The phosphatidic acid-producing PLD1 acts as an Epac effector and represents a direct molecular link to the actin cytoskeleton through phosho-cofilin (López de Jesús et al., 2006; Han et al., 2007). Thus, interaction of Epac with the actin-microtubule network might be rather diverse in nature and therefore contribute to the sophisticated mechanisms leading to proper endothelial barrier function and cell adhesion.

2. Other Endothelial Functions Regulated by Epac

In vascular endothelial cells, Epac1 induces the transcriptional activity of CCAAT/enhancer-binding protein through protein kinase Cα and ERK1/2 (Yarwood et al., 2008; Borland et al., 2009; Woolson et al., 2009) and thereby promotes the expression of SOCS-3. Epac1-dependent induction of SOCS-3 inhibits the IL-6 receptor trans-signaling complex and subsequently limits proinflammatory actions of IL-6 in vascular endothelial cells (Parnell et al., 2012; Wiejak et al., 2012; Fig. 7). It is noteworthy that Epac1 thereby protects the vascular endothelium from IL-6-induced dysfunction, at least in part through inhibition of the JAK-STAT pathway. In the human promonocytic cell line U937, fibrinogenic TGF-β reduced the expression of Epac1 transcript and thereby reduced the transmigratory capacity of monocytes (Basoni et al., 2005), suggesting that Epac achieves limitations of proinflammatory signals through elaborated effects on both leukocytes and endothelial cells.

In addition, modulation of actin-myosin contractility by Epac can also be envisioned. Myosin IIA and IIB are primarily expressed at cell-cell contacts (De La Cruz and Ostap, 2004), and myosin, as part of the circumferential belt and bounded to actin, can control the shape of the cell via this belt. In MCF7 breast epithelial cells (Smutny et al., 2010; Fig. 8), junctional localization of myosin IIA requires next to E-cadherin adhesion, Rho-Rho-kinase, and myosin light chain-kinase activation, and thereby subsequently increase the contractile force of circumferential belt and tight junction integrity. Myosin IIB, via Rap1A, supports myosin IIA-Rho-Rho-kinase signaling to E-cadherin and to myosin light chain kinase and thereby also subsequently induces the stabilization of the apical ring structure and enhancement of the junctional integrity (Smutny et al., 2010; Fig. 8). In support of the findings in endothelial cells, transformation of the human pancreatic carcinoma epithelial-like cell line PANC-1 with constitutively active Rac1 (V12) redistributes E-cadherin-β-catenin complexes to cell-cell contacts through IQ motif containing (IQ)GAP1, whereas cell transformation with dominant negative Rac1 (N17) exerts opposing effects (Hage et al., 2009). Together with the finding that IQGAP1 binds the Epac effector Rap1 (Jeong et al., 2007), Epac-dependent compartmentalized cAMP signaling might require scaffolding proteins including IQGAPs. Thus, Epac, by activation of Rap, may regulate barrier function of endothelial as well as epithelial cells. Several recent findings point also to a proper RhoA versus Rac1 signal balance (Lawrence et al., 2002; Lorenowicz et al., 2007a,b; Baumer et al., 2008a,b; Sayner, 2011). Thus, Rap1 and Rho GTPases RhoA and Rac1 capture key functions to preserve structural and functional properties of tight and adherens junctions (Fig. 7). Confusingly, Rac1 promotes both endothelial barrier stabilization and destabilization (Pullar et al., 2004; Netherton et al., 2007; Raymond et al., 2007; Baumer et al., 2008b; Rampersad et al., 2010; Spindler et al., 2011; Spindler and Waschke, 2011). Distinct barrier protection properties of Rac1, however, in microvascular versus macrovascular endothelial cells seem to be reasonable (Spindler et al., 2011).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 8.

Epac and the lung. In airway epithelium, elevation of cAMP in response to β₂-agonists and prostanoids activates Rap through Epac and thereby promotes signaling through Rho and Rac via yet unidentified mechanisms. Rac destabilizes the IQGAP-β-catenin complex and releases β-catenin to E-cadherin-bearing cell-cell contacts, a process known to act in concert with Rap-dependent E-cadherin recruitment and subsequently to stabilize the epithelial barrier. Rho, in concert with myosin IIB, enhances the stability of actin filaments and cortical actin through phosphorylation of myosin light chain. PDE4D creates a functional cAMP diffusion domain and thereby promotes cystic fibrosis transmembrane conductance regulator (CFTR) channel function, a process supported by the cortical actin (A). Elevation of cAMP in response to β₂-agonists and prostanoids activates both PKA and Epac and thereby induces airway smooth muscle relaxation, inhibits airway smooth muscle proliferation, and modulates cytokine secretion. Epac induces airway smooth muscle relaxation through inhibition of RhoA and activation of Rac1, a process that may involve the dual-specific AKAP ezrin known to modulate Rho-Rho-kinase. PKA induces airway smooth muscle relaxation through signaling to MLCK and MLCP. Epac and PKA inhibit airway smooth muscle proliferation and cytokine secretion through distinct signaling to PKB/Akt, p70S6K, ERK1/2, and NF-κB (B). Activation of PKA and Epac differentially inhibits specific fibroblast functions; activation of PKA inhibits airway fibroblast collagen synthesis, whereas activation of Epac inhibits fibroblast proliferation. Both PKA and Epac contribute to the inhibition of fibroblast transdifferentiation to myofibroblasts in response to β₂-agonists and prostanoids. PKA inhibits collagen synthesis by inhibiting the profibrotic kinase PKC-δ, a process supported by AKAP9. Epac activates Rap1, which subsequently inhibits fibroblast proliferation. Both PKA and Epac inhibit transdifferentiation and thus α-smooth muscle actin (α-SMA) expression. The molecular mechanisms involved, however, remain to be determined. C/EBP, CCAAT/enhancer-binding protein; CFTR, cystic fibrosis transmembrane conductance regulator; ECM, extracellular matrix; EMT, epithelial-to-mesenchymal transition; Krit, Krev1 interaction trapped gene; MLC, myosin light chain; MLCK, myosin light chain kinase; MLCP, myosin light chain phosphatase; PDE4D, phosphodiesterase 4D; PKB/Akt, protein kinase B/Akt

3. Epac and Vascular Smooth Muscle Functioning

Epac seems to be to able alleviate symptoms associated with vascular dysfunctions, including atherosclerosis and (pulmonary) hypertension (Yokoyama et al., 2008a; Frumkin, 2012; Hirase and Node, 2012), through targeting vascular smooth muscle functions. In rat aortic smooth muscle cells, Epac inhibits ATP-sensitive K⁺-channels through Ca²⁺-dependent calcineurin (Purves et al., 2009), suggesting that Epac, in particular Epac1 being primarily expressed in vascular smooth muscle, controls vascular tone and blood flow. Indeed, specific activation of Epac induces relaxation of rat aorta (Sukhanova et al., 2006) and rabbit and rat pulmonary aorta (Murray et al., 2009b; Zieba et al., 2011) through downregulation of RhoA-dependent myosin light chain phosphatase phosphorylation (Zieba et al., 2011; Fig. 8B). Thus, Epac alters the Rho/Rac balance and thereby regulates biologic responses that rely on actin-myosin signals. Interestingly, the mRNA and protein expression of both Epac1 and Epac2 are downregulated in patients with pulmonary hypertension as well as in rats with monocrotalin-induced pulmonary hypertension (Murray et al., 2009; Table 1). The functional consequence of the reduced Epac expression is clear from the reduced ability of the Epac activator 8-pCPT-2′-O-Me-cAMP to relax pulmonary arteries from monocrotalin-treated mice (Murray et al., 2009). Intriguingly, Epac-mediated arterial smooth muscle relaxation also involves activation of nitric oxide synthase in the endothelium (Garcia-Morales et al., 2012). Loss of Epac1 expression in the endothelium due to hypoxia (Garcia-Morales et al., 2012), which may induce pulmonary hypertension by inducing vasoconstriction, could contribute to the development of pulmonary hypertension.

In addition to regulating vascular tone, Epac induces adhesion of microvascular smooth muscle cells and promotes vascular smooth muscle migration (Yokoyama et al., 2008a; Eid, 2012). Thus, specific activation of Epac as well as overexpression of Epac1 induces migration of rat aortic smooth muscle cells via Epac1-mediated activation of Rap1 (Yokoyama et al., 2008a). By contrast, specific activation of PKA inhibits cell migration. Interestingly, the AC activator forskolin induces migration at lower concentrations and inhibits migration at higher doses (Yokoyama et al., 2008a), indicating cAMP effector selectivity at different cAMP concentrations. The importance of Epac1 in cell migration is supported by the finding that Epac1, but not Epac2, is upregulated during neointima formation following vascular injury in mice (Yokoyama et al., 2008a; Table 1). Furthermore, Epac inhibits vascular smooth muscle proliferation either alone and/or concert with PKA through JNK-ERK (Hewer et al., 2011; Mayer et al., 2011). Together with the finding that Epac induces the early gene nuclear receptor subfamily 4, which seems to act anti-atherogenic (Mayer et al., 2011), these findings indicate that Epac may modulate phenotypical modulation of vascular smooth muscle cells. Conversely, Epac also bears the capacity to mediate procontractile functions in the vasculature, as Epac1 enhances the expression of the α_2C-adrenergic receptor through a Rap1-RhoA signaling pathway (Jeyaraj et al., 2012). In conclusion, these findings indicate that Epac exhibits several distinct biologic responses in vascular smooth muscle cells, the relative net outcome may even differ from the vessel type.

4. Regulatory Role of the Epac Signalosome in the Vasculature

In human dermal microvascular endothelial cells, cAMP-dependent Rac1 activation requires both vasodilator-stimulated phosphoprotein and AKAP-anchored PKA (Baumer et al., 2008b; Spindler et al., 2010, 2011; Spindler and Waschke, 2011). Rac1 activation through AKAP-anchored PKA is sensitive to the PKA-binding blocking peptide st-Ht31 (Schlegel and Waschke, 2009). PKA-dependent phosphorylation of vasodilator-stimulated phosphoprotein contributes to the barrier protection (Bogatcheva et al., 2009). In HUVECs, PDE4D-bearing VE-cadherin-based multiprotein complexes control the vascular permeability, a process requiring Epac1 (Raymond et al., 2007; Rampersad et al., 2010; Fig. 7). In human microvascular endothelial cells, Epac induces activation of Rac1, enhancement of the barrier function, and redistribution of VE-cadherin to cell-cell contacts (Baumer et al., 2008b; Spindler et al., 2011). On the molecular level, mechanisms leading to Rac1 activation through Epac1 are solved in human pulmonary artery endothelial cells in vitro and in ventilator-induced lung injury in vivo (Birukova et al., 2007, 2008, 2009, 2010; Xing and Birukova, 2010). The cAMP elevating compounds PGE₂, prostacyclin I₂, and the atrial natriuretic peptide activate Rac1 through Epac1, Rap1, the Rac-specific exchange factors Tiam1-Vav2 signals, and p115 Rho-GEF-dependent RhoA inhibition, processes acting in concert with PKA (Fig. 7). Thus, protection of the endothelial barrier by Epac1 resembles mechanisms known to induce smooth muscle relaxation (Roscioni et al., 2011b; Zieba et al., 2011; Fig. 8). A reduction of the phosphorylation of the Rho-Rho-kinase target myosin light chain skews the balance of RhoA/Rac1 activation toward Rac1.

Earlier studies reported on the first molecular link of Epac and the actin-microtubule and its impact on the regulation of the barrier function in HUVECs and human pulmonary aortic endothelial cells (Cullere et al., 2005; Fukuhara et al., 2005; Sehrawat et al., 2008; Noda et al., 2010). Recent studies demonstrate that Krit1 (Krev1 interaction trapped gene) is required for the stabilization of β-catenin-bearing cell-cell contacts by the Epac effector Rap1 and that the Epac-Rap1 effector Krit1 is required for the maintenance of the endothelial barrier (Glading et al., 2007; Stockton et al., 2010). Loss of Krit1, known to account for the endothelial junction disease cerebral cavernous malformations (Stockton et al., 2010), induced destabilization of the endothelial barrier by increasing the phosphorylation of the Rho-Rho-kinase target myosin light chain (Fig. 7). Although the molecular link to Epac has still to be studied, these findings implicate that Epac-bearing multiprotein complexes are of utmost importance for endothelial barrier maintenance.

Recent studies in HUVECs and human dermal microvascular endothelial cells, indeed, indicate that AKAP9 (also known as AKAP450) complexes with Epac1, promotes microtubule growth to coordinate integrin signaling at lateral cell borders, and thereby enhances endothelial barrier properties of Epac1 (Sehrawat et al., 2011; Fig. 7). Surprisingly, AKAP9 does not alter Epac1-dependent GTP-loading of Rap1, reorganization of cortical actin, and de novo VE-cadherin adhesion (Sehrawat et al., 2011), suggesting that AKAP9 integrates functions of VE-cadherin and integrins at junctional sites to enhance the endothelial barrier in response to activation of Epac (Fig. 7). Thus, several recent studies indicate that compartmentalization of cAMP signaling in endothelial cells (as reviewed by Feinstein et al., 2012) might turn out to be of key importance for the maintenance of central endothelium-associated signaling properties including a proper and tight control of the endothelial barrier.

1. Epac and the Airway Epithelium

The airway epithelium constitutes protection through its barrier, mucus production, ciliary beating, antimicrobial production, and immune responses to deleterious agents (Knight and Holgate, 2003; Lai and Rogers, 2010). Airway epithelial barrier disruption leads to epithelium remodeling, goblet cell metaplasia, and mucus gland hypertrophy in asthma and COPD (Fahy and Dickey, 2010). The cAMP signaling pathway is thought to regulate the barrier function (see also section II.D), although the molecular mechanisms that regulate the epithelial barrier are less characterized. In a mouse model of allergic asthma, chronic application of β₂-adrenoceptor inverse agonists reduces airway epithelial barrier dysfunctions (Callaerts-Vegh et al., 2004; Lin et al., 2008; Nguyen et al., 2008, 2009; Dickey et al., 2010; Walker et al., 2011), suggesting that ligand-directed signaling or biased agonism is operational in the airway epithelium. In polarized human Calu-3 airway epithelial cells, a functional cAMP diffusion barrier maintained by PDE4D determines the function of cystic fibrosis transmembrane conductance regulator chloride channel (Barnes et al., 2005), indicating that functional responses to cAMP in airway epithelial cells rely on cAMP microdomains (Fig. 8A). Indeed, in human airway epithelial cells, regulation of the cystic fibrosis transmembrane conductance regulator requires next to the subcortical cytoskeleton compartmentalized cAMP signaling (Monterisi et al., 2012). Different ACs and PDEs are expressed in airway epithelium (Defer et al., 2000; Hanoune and Defer, 2001; Small et al., 2003; Tantisira et al., 2005; Pierre et al., 2009), which may have the potential to generate cAMP-sensing multiprotein complexes exhibiting a distinct composition to maintain specific airway epithelium functions. In human bronchial epithelial cells, the COPD inducer cigarette smoke disrupts the epithelial barrier and specifically downregulates AKAP9 (Oldenburger et al., 2011). As recent studies report on the maintenance of the endothelial barrier through AKAP9 (Sehrawat et al., 2011; Fig. 7), these findings indicate that AKAP9 is a key regulator of both the endothelial and epithelial barrier. The role of Epac in these processes is currently unknown. However, Epac1 and Epac2 as well as the Epac effectors Rap1 and Rap2, which regulate epithelial integrity in Caenorhabditis elegans (Pellis-van Berkel et al., 2005), are expressed in human bronchial epithelial cells (S.S. Roscioni, C.R.S. Elzinga, I.S.T. Bos, M.H. Menzen, F.J. Warnders, H. Maarsingh, A.J. Halayko, H. Meurs, M. Schmidt, unpublished observations) and may therefore be involved in the cAMP-mediated effects on the epithelial barrier. Taken together, compartmentalized cAMP signaling seems also to account for the proper functioning of the epithelial barrier.

2. Regulation of Airway Smooth Muscle Tone and Secretory Function by the Epac Signalosome

Airway smooth muscle cells control airway caliber and contribute to bronchoconstriction, local inflammation, wound healing, and remodeling (Hirst, 2003; Halayko et al., 2008; Damera and Panettieri, 2011; Billington et al., 2012). Classically, the bronchodilating effect of β₂-agonists was assumed to be mediated by PKA, which reduces contractile force through inhibition of myosin light chain phosphorylation (Pfitzer, 2001; Giembycz and Newton, 2006; Penn, 2008; Billington et al., 2012). However, it has been shown that additional PKA-independent mechanisms contribute to airway smooth muscle relaxation (Spicuzza et al., 2001). Recent studies in airway smooth muscle preparations show that Epac, presumably by downstream activation of Rap1B, inhibits myosin light chain phosphorylation through RhoA inhibition and Rac1 activation causing relaxation (Roscioni et al., 2011a; Zieba et al., 2011; Fig. 8B). Regulation of smooth muscle contraction requires a tightly controlled network of the actin-microtubule cytoskeleton (Olson and Nordheim, 2010), most likely acting in concert with members of the ERM family known to regulate cytoskeleton driven processes, including migration, motility, and contraction (Niggli and Rossy, 2008; Fehon et al., 2010). In airway smooth muscle, a recent study report on the expression of the dual-specific AKAP protein ezrin and its impact on compartmentalized cAMP signaling (Horvat et al., 2012). Ezrin interacts with Rho-Rho kinase signaling (Bretscher et al., 2002; Niggli and Rossy, 2008; Gloerich et al., 2010), suggesting that ezrin contributes to the regulation of airway smooth muscle contraction by regulating calcium sensitivity. Future studies should address the potential role of ezrin in mediating cAMP-induced airway smooth muscle relaxation and its relation with Epac.

Airway smooth muscle cells synthesize and release multiple mediators, including extracellular matrix proteins and cytokines-chemokines (Hirst, 2003; Damera and Panettieri, 2011). In airway smooth muscle, the cAMP-elevating compounds PGE₂ and β₂-agonists increase the release of the chemokines IL-6 and IL-8 in response to TNF-α and bradykinin (Huang et al., 1998; Pang and Knox, 1998; Ammit et al., 2002; Roscioni et al., 2009). Recent studies indicate that both Epac1 and Epac2 as well as PKA enhance bradykinin-induced IL-8 release through Rap1 and ERK1/2 (Roscioni et al., 2009). Thus, Epac bears the potential to potentiate inflammation in airway smooth muscle cells. Conversely, Epac, acting in concert with PKA, inhibits IL-1β-induced release of GM-CSF (granulocyte-macrophage colony-stimulating factor), RANTES (regulated on activation, normal T cell expressed, and secreted) and eotaxin, and reduced IL-1β-induced ICAM1 expression (Hallsworth et al., 2001; Ammit et al., 2002; Clarke et al., 2004; Kaur et al., 2008). Recent studies in human airway smooth muscle show that both Epac1 and Epac2 limit the release of IL-8 in response to cigarette smoke extract through inhibition of NF-κB, whereas the inhibitory effect of PKA is mediated via inhibition of ERK1/2 (Oldenburger et al., 2012; Fig. 8B). Importantly, the expression of Epac1, but not Epac2, is reduced in cigarette smoke extract-exposed airway smooth muscle cells and lung tissue from COPD patients (Oldenburger et al., 2012; Table 1), suggesting that cigarette smoke may reduce the anti-inflammatory effects of cAMP by downregulation of Epac1. The cigarette smoke-induced downregulation of Epac1 resembles the effect of TGF-β on the Epac1 expression level (Yokoyama et al., 2008c); however, the underlying molecular mechanisms still have to be defined. The notion that, depending on the stimulus, Epac may act both proinflammatory and anti-inflammatory is rather exciting. The recent finding that airway smooth muscle cells also express AKAP family members, including ezrin, AKAP79 (AKAP5), and AKAP250 (AKAP12) (Horvat et al., 2012; Poppinga et al., 2012) suggests that cAMP compartmentalization through distinct AKAPs determines inflammatory properties of Epac (Fig. 8B).

3. Epac as a Target to Treat Airway Remodeling

Airway wall remodeling in asthma is characterized by increased airway smooth muscle mass through both hyperplasia and hypertrophy and airway fibrosis (Ebina et al., 1993; Kuwano et al., 1993; Jeffery, 2001, 2004; Benayoun et al., 2003; Woodruff et al., 2004; Dekkers et al., 2009) and leads to lung function decline and airway hyperresponsiveness (Lambert et al., 1993; Saetta et al., 1998; Hogg et al., 2004; Oliver et al., 2007). Exposure to mitogenic stimuli switches the airway smooth muscle phenotype from a (normo)contractile to a proliferative, hypocontractile state characterized by increased expression of proliferative markers and decreased contractile responses associated with decreased expression of contractile proteins, a process reversible in nature (Halayko and Amrani, 2003, Dekkers et al., 2009). Mitogen-induced airway smooth muscle cells proliferation is reduced in response to β₂-agonists and even more pronounced in response to PGE₂ (Tomlinson et al., 1994; Stewart et al., 1999; Lee et al., 2001; Kassel et al., 2008; Roscioni et al., 2011a; Yan et al., 2011). PGE₂ may act antimitogenic or promitogenic via distinct EP receptor subtypes (Yan et al., 2011). Inhibition of airway smooth muscle proliferation does not correlate with the generation of cAMP in response to β₂-agonists and PGE₂ (Kassel et al., 2008; Yan et al., 2011), pointing to the contribution of compartmentalized cAMP signaling. In primary human lung fibroblasts, PGE₂ and the PDE inhibitors roflumilast and piclamilast diminish fibroblast proliferation and myofibroblast differentiation (Selige et al., 2010, 2011). The PDE4 inhibitor rolipram also reduced epithelial-mesenchymal transition (Kolosionek et al., 2009). PGE₂-induced cAMP elevation exerts antifibrotic properties in human and mouse lung fibroblasts, and defects in the synthesis of PGE₂ correlate with the degree of airway remodeling in mice (Bauman et al., 2010; Okunishi et al., 2011; Stumm et al., 2011). PGE₂ also inhibits collagen deposition in fibrotic lung fibroblasts through a PGE₂-sensing AKAP450 (AKAP9)-PKA-protein phosphatase 2A multiprotein complex (Okunishi et al., 2011; Fig. 8C). The extracellular mediator plasmin profoundly alters the AKAP complex signaling properties, restoring PGE₂ sensitivity and subsequent inhibition of activated fibroblasts (Okunishi et al., 2011). Interestingly, a distinct subset of AKAPs family members directs PGE₂ toward the enhancement of toll-like receptor signaling (Kim et al., 2011) and may therefore be important for asthma and COPD pathophysiology. As AKAP9 has been reported to complex with PDE4 (Tasken et al., 2001), the presence of PDE4 in the PGE₂-sensing AKAP9 complex in lung fibroblasts awaits additional investigations. The distinct composition of the cAMP-responsive multiprotein complexes is expected to generate and maintain local cAMP gradients in a cell-type specific fashion.

Several recent studies indicate that next to PKA (Misior et al., 2008; Yan et al., 2011), Epac contributes to the inhibition of cell proliferation (Kassel et al., 2008; Roscioni et al., 2011a). Epac-induced antiproliferative effects associated with the normalization of mitogen-induced hypocontractility and maintenance of normal expression of contractile protein (Roscioni et al., 2011a,c). Intriguingly, Epac inhibits phenotype switching by inhibition of ERK1/2, whereas PKA inhibits ERK1/2 and phosphoinositide 3-kinase (Roscioni et al., 2011a, Fig. 8B). As mentioned above, expression of a distinct subset of AKAPs in human airway smooth muscle (Horvat et al., 2012; Poppinga et al., 2012) and bronchial epithelial cells (Oldenburger et al., 2011; Schmidt et al., 2011) maintains compartmentalized cAMP signaling of PKA and Epac and modifies functional responses. Collectively, Epac inhibits mitogen-induced phenotypic switching of airway smooth muscle cells, which may potentially be regulated AKAP-containing complexes. Alterations in spatiotemporal dynamics of cAMP may therefore profoundly alter biologic responses of airway smooth muscle. Future studies should define the precise impact of AKAP-Epac-PKA interactions on airway smooth muscle function and responses to pharmacological interventions.

Airway fibrosis encompasses increased deposition of extracellular matrix proteins in response to abnormal wound repair through excessive recruitment and activation of (myo)fibroblasts from resident mesenchymal cells, circulating fibrocytes, and/or epithelial-to-mesenchymal transition (Postma and Timens, 2006; Salazar and Herrera, 2011; Wynn, 2011). Resident fibroblasts are transformed upon “activation” into the more contractile, proliferative, and secretory active myofibroblasts, a process importantly driven by TGF-β (Bartram and Speer, 2004; Leask and Abraham, 2004; Meneghin and Hogaboam, 2007). In human lung fibroblasts, cAMP-elevating compounds, such as PGE₂, EP₂ receptor agonists, β-adrenoceptor agonists, the direct AC activator forskolin, and PDE inhibitors, suppress the transdifferentiation into myofibroblasts as indicated by a reduced expression of α-smooth muscle actin (Kolodsick et al., 2003; Liu et al., 2004; Baouz et al., 2005; Dunkern et al., 2007; Thomas et al., 2007; Lamyel et al., 2011). The PKA agonist 6-Bnz-cAMP and the Epac agonist 8-CPT-2′-O-Me-cAMP also reduce the expression of α-smooth muscle actin (Lamyel et al., 2011), suggesting that fibroblast transformation is regulated by PKA and Epac (Fig. 8C).

In MRC-5 lung fibroblasts, expression of both Epac1 and Epac2 mRNA is shown (Haag et al., 2008), although protein expression was only observed for Epac1 (Huang et al., 2007, 2008; Haag et al., 2008). Unfortunately, our current knowledge about the fibroblast-specific expression profile of Epacs is predominantly based on studies in human fetal lung fibroblasts, and only a rather limited amount of studies point to a (functional) role of Epacs in primary adult human lung fibroblasts (Huang et al., 2007, 2008). In (primary) human lung fibroblasts PKA and Epac differentially regulate fibroblast proliferation and collagen synthesis. Epac agonists only inhibit fibroblast proliferation, presumably through Rap1, to a similar extent as β₂-agonists and prostanoids, whereas PKA activation only inhibits collagen synthesis, presumably via inhibition of PKC-δ (Zhang et al., 2004; Huang et al., 2007, 2008; Haag et al., 2008; Fig. 8C). The effects of Epac are most likely mediated via Epac1, as downregulation of Epac1, but not Epac2, largely prevented the inhibitory effects of butaprost and 8-pCPT-2′-O-Me-cAMP on fibroblast proliferation (Huang et al., 2007, 2008; Haag et al., 2008). Interestingly, relative to Epac2, expression of Epac1 becomes dominant in the adult lung compared with fetal tissue (Ulucan et al., 2007), suggesting that Epac’s function in the lung may alter during development. On the basis of our current knowledge of the expression profile of Epac1 and Epac2 in lung fibroblasts, it is tempting to assign Epac-dependent functional responses to Epac1. In this respect it is important to note that TGF-β reduces the expression of Epac1 in rat lung fibroblasts (Yokoyama et al., 2008c; Table 1). Because a direct interaction between Epac1 and TGF-β1 has been observed with subsequent inhibition of Smad-dependent TGF-β signaling (Conrotto et al., 2007), it is tempting to speculate that loss of Epac1 expression by TGF-β promotes the transformation of resident fibroblasts into active myofibroblasts due to a dysfunction in the dynamic signaling features of the endogenous suppressor cAMP.

In conclusion, Epac plays an important role in regulating airway functions, such as inflammatory cytokine release, cell proliferation, airway smooth muscle contraction, and the production of extracellular matrix proteins. In vivo studies using transgenic mice and/or pharmacological agents targeting the Epac pathway are needed to further clarify the role of Epac in the pathophysiology and treatment of airways diseases, such as asthma and COPD.