Recent Advances in the Discovery of Small Molecules Targeting Exchange Proteins Directly Activated by cAMP (EPAC)

Abstract

cAMP is a pivotal second messenger that regulates numerous biological processes under physiological and pathological conditions, including cancer, diabetes, heart failure, inflammation and neurological disorders. In the past, all effects of cAMP were initially believed to be mediated by PKA and cyclic nucleotide-regulated ion channels. Since the discovery of EPAC proteins in 1998, accumulating evidence has demonstrated that the net cellular effects of cAMP are also regulated by EPAC. The pursuit of the biological functions of EPAC has benefited from the development and applications of a growing number of pharmacological probes targeting EPAC proteins. In this Perspective, we seek to provide a concise update on recent advances in the development of chemical entities including various membrane-permeable analogues of cAMP and newly discovered EPAC-specific ligands from high throughput assays and hit-to-lead optimizations.

1. Introduction

Cyclic adenosine monophosphate (cAMP, cyclic AMP or 3′–5′-cyclic adenosine monophosphate) is a pivotal second messenger derived from its precursor adenosine triphosphate (ATP). A wide variety of extracellular ligands bind to G-protein coupled receptors (GPCRs), activate adenylate cyclases (ACs) to catalyze the conversion of intracellular ATP to pyrophosphate and cAMP.^1,2 cAMP regulates a number of key biological processes under physiological and pathological conditions, including neuronal signaling, gluconeogenesis, glycogenolysis, lipogenesis, cardiac and smooth muscle contraction, secretory processes, ion channel conductance, learning and memory.¹

The local concentration and distribution of intracellular cAMP is regulated by ACs and the cyclic nucleotide phosphodiesterases (PDEs). Generally, numerous extracellular signals trigger a series of the conformational changes of GPCRs on the cell surface. Typically, Gs protein stimulates ACs to increase cAMP production inside the cell, whereas Gi protein inhibits ACs and lowers the level of cAMP.^2–6 Meanwhile, the intercellular level of cAMP is degraded by PDEs which catalyze conversion of cAMP to 5′-AMP.⁷

In the past, all effects of cAMP were initially believed to be mediated by protein kinase A (PKA) and cyclic nucleotide-regulated ion channels.^8–11 In 1998, two independent groups reported their findings that PKA-independent mechanism of cAMP action was regulated by a family of guanine nucleotide exchange factors (GEFs) called cAMP-GEFs which are also named as exchange protein directly activated by cAMP (EPAC).^12,13 Since then, remarkable progress has been made on elucidating the molecular mechanism of EPAC proteins over the last fifteen years. Meanwhile, probing the biological functions of EPAC has been significantly facilitated by the development and applications of small-molecule EPAC ligands including various membrane-permeable analogues of cAMP and newly discovered EPAC-specific antagonists. Consequently, numerous additional biological functions of EPAC have been uncovered. This review briefly summarizes the structures of EPAC family members, EPAC signaling pathway and biological functions, and also provides a perspective on recent advances in the discovery of new chemical entities targeting EPAC proteins. In addition, these valuable pharmacological tools including cAMP analogues and EPAC antagonists have led to a greater understanding of the important role of EPAC proteins in different diseases, establishing EPAC proteins as novel molecular targets for new therapeutic strategies against various human diseases including cancer, diabetes, heart failure, inflammation and neurological disorders.

2. EPAC Family and EPAC2 Protein Structures

To date, two isoforms of EPAC have been identified, EPAC1 and EPAC2, which are also known as RAPGEF3 (cAMP-GEF-I) and RAPGEF4 (cAMP-GEF-II), respectively.^12–14 As depicted in Figure 1, each EPAC family member composes an auto-inhibitory amino-terminal regulatory region and a carboxyl-terminal catalytic region for activation of Rap GTPase.^14–18 The regulatory region contains a Dishevelled Egl-10 Pleckstrin (DEP) domain and at least one functional cyclic nucleotide binding domain (CNBD, one for EPAC1 and two for EPAC2). The carboxyl-terminal catalytic region consists of a Ras exchange motif (REM) domain and a Ras association (RA) domain as well as the CDC25-homology domain (CDC25-HD). The CDC25-homology domain is responsible for guanine nucleotide exchange activity and catalyzes the exchange of G-protein-bound GDP for GTP on the Ras-like small GTPases Rap1 and Rap2 isoforms.^19,20

Figure 1.

Domain structures of EPAC proteins. Each EPAC family member composes an auto-inhibitory amino-terminal regulatory region and a carboxyl-terminal catalytic region for activation of Rap GTPase. The regulatory region contains a Dishevelled Egl-10 Pleckstrin (DEP) domain and at least one functional cyclic nucleotide–binding domain (CNBD). The carboxyl-terminal catalytic region consists of a Ras exchange motif (REM) domain and a Ras association (RA) domain as well as the CDC25-homology domain (CDC25-HD). The CDC25-homology domain is responsible for guanine nucleotide exchange activity and catalyzes the exchange of G-protein-bound GDP for GTP on the Ras-like small GTPases Rap1 and Rap2 isoforms.

The two EPAC isoforms EPAC1 and EPAC2 are mostly expressed in both mature and developing tissues with different expression levels. EPAC1 is highly expressed in central nervous system, adipose tissue, blood vessels, kidney, ovary, and uterus, while EPAC2 is most detectable in the central nervous system, adrenal gland, and pancreas.²¹ EPAC2 has been found to exist as three splice variants. EPAC2A is expressed in cerebral cortex and pancreatic islets and has two cAMP-binding domains, whereas expression of EPAC2B is restricted to adrenal glands without cAMP-binding domain A. EPAC2C is found in the liver and lack of cAMP-binding domain A and DEP domain.²²

The X-ray crystal structure determinations and analysis of two different conformations of EPAC2 protein have revealed in atomic detail that cAMP binding causes conformation change allowing the catalytic region to be available for the binding of Rap (Figure 2).^15,16 NMR spectroscopy studies, peptide amide hydrogen/deuterium exchange mass spectrometry (DXMS) approach, small angle X-ray scattering, fluorescence resonance energy transfer (FRET) studies and molecular dynamics (MD) simulations further demonstrated the mechanism of autoinhibition and activation of EPAC2.^23–32 In the absence of cAMP, the apo-EPAC2 protein exists in an autoinhibited conformation (inactive state), in which the cyclic nucleotide–binding domains sterically hinder the access of down-regulating effectors to the catalytic domain, effectively inhibiting the guanine nucleotide exchange activity. Upon binding of cAMP, a sequence of structural reorganizations within the cyclic nucleotide–binding domains allows the regulatory domain to move to the back side of the catalytic region, exposing the Rap-binding site of the catalytic region. It is worth noting that while X-ray structural analysis reveals no significant conformational changes between the CDC25-HD in inactive and active stages,^15,16 DXMS studies indicate that in addition to relieving the steric hindrance imposed upon the catalytic lobe by the regulatory lobe, cAMP may also act as an allosteric modulator affecting the interaction between EPAC2 and Rap1 by inducing conformational changes at the ionic latch/hairpin structure, which is directly involved in Rap1 binding.³²

Figure 2.

The X-ray crystal structures of inactive EPAC2 (the full-length apo-EPAC2, PDB: 2BYV) and active EPAC2 (EPAC2Δ305:Sp-cAMPS:Rap1B complex, PDB: 3CF6). CNBD-A, DEP domain, CNBD-B, REM domain, RA domain, and CDC25-HD are colored in red, green, blue, yellow, cyan, and magentas, respectively. Upon binding of cAMP, a sequence of structural reorganizations within the cyclic nucleotide–binding domains allows the regulatory domain to open and move to the back side of the catalytic region. This conformational change leads to the exposure of the catalytic region for binding of Rap GTPase to catalyze the exchange of GDP for GTP.

3. EPAC Signaling Pathway and Biological Functions

Cellular cAMP levels are modulated by hormones and neuromediators with specific effects on the homeostasis of multicellular organisms. The elevation in intracellular concentration of cAMP after Gs protein-coupled receptors activation is sufficient to activate EPAC. As a part of the downstream targets of cAMP, EPAC proteins as guanine nucleotide exchange factors (GEFs) catalyze the exchange of GDP for GTP on the small GTPases Rap1 and Rap2.¹⁴ Rap1 and Rap2 as members of the Ras family are involved in the network of multiple proteins to regulate cell adhesion and cell junction. EPAC proteins act as the critical upstream effectors, among many other Rap GEFs, controlling the Rap-mediated biological processes of cAMP.^19,33 The signaling pathway of EPAC proteins is depicted in Figure 3. Accumulating evidence has demonstrated that EPAC signaling has a significant contribution in various biological processes, including cardiac function, insulin secretion, neuronal function, vascular function, inflammation, renal function, and airway function.^34,35

Figure 3.

Signaling pathway of EPAC proteins. EPAC1 only contains one functional cyclic nucleotide binding domain (CNBD-B), while EPAC2 contains two CNBDs (CNBD-A and CNBD-B).

3.1. EPAC and Cardiac Function

In a collection of specialized muscle cells called cardiac myocytes, cAMP modulates a series of cardiac functions such as heart rate, cardiac contractility and relaxation.^36,37 Sustained elevation of cAMP induces hypertrophy, cardiac dysfunction and eventually to the development of heart failure.^38–43 Given the critical role of cAMP in cardiac function, more and more studies focused on the cardiac function of EPAC proteins, providing plenty of evidence that EPAC is a new regulator in cardiac physiology and pathophysiology.^{37,38,44–49} Overexpression of EPAC1 in left ventricular hypertrophy of animal models and patients demonstrated that EPAC1 contributes to the progression of heart failure.^47,48 Very recent work using knockout mice for EPAC1, EPAC2, or both has revealed that β₁-adrenergic receptor (β₁-AR) activation causes EPAC2-mediated sarcoplasmic reticulum (SR) Ca²⁺ leak and arrhythmias via Ca²⁺/calmodulin-dependent protein kinase IIδ (CaMKIIδ)-dependent phosphorylation of ryanodine receptor 2 (RyR2)-S2814, indicating that this pathway contributes to β-AR-induced arrhythmias and reduced cardiac function.⁵⁰ Thus, EPAC proteins represent novel therapeutic targets for the treatment of cardiac disorders.

3.2. EPAC in Insulin Secretion

EPAC2 is highly expressed in the insulin-producing pancreatic β-cells.⁵¹ A growing body of biological studies have demonstrated that EPAC2/Rap1 signaling pathway is essential in the regulation of the dynamics of insulin secretion.⁵² The initial studies including in vivo experiments in EPAC2 knockout mice indicated that the traditional antidiabetic drugs sulfonylureas might interact with EPAC2 to activate Rap1.^53,54 Despite the controversial debate, this finding implicates that EPAC2 contributes to glucose-induced insulin secretion.^55–57 Additional studies suggest that several intracellular signaling pathways with relation to insulin secretion are linked with EPAC2 such as recent discovered pathway EPAC2-Rim2-Rab3.⁵⁸ The relevant genetic and pharmacological studies concerning EPACs’ involvement in glucose homeostasis and energy balance, through regulation of leptin and insulin signaling pathways, were discussed in a very recent review.⁵⁹ The accumulating evidence supports that EPAC may represent a novel target to provide therapeutic potential for the treatment of diabetes and obesity.

3.3. Neuronal Function of EPAC

Both EPAC1 and EPAC2 are predominant in the brain including the hippocampus, striatum, and prefrontal cortex, cerebral cortex, cerebellum, olfactory bulb, thalamus, habenula, and pituitary.^13,60 Compelling evidence has demonstrated that EPAC proteins have a significant contribution in the development and function of the nervous system.^40,61 In addition, a recent study using EPAC knockout mice provided the direct in vivo evidence that EPAC1 and EPAC2 proteins synergistically regulated neurological functions in the brain for processing spatial learning and social interactions.⁶² Additional studies also showed that EPAC2-deficient mice exhibited robust deficits in social and communication behaviors, indicating potential contributions in brain disorders.⁶³ Thus, targeting EPAC signaling pathway can be considered as a promising strategy for the treatment of various neurological disorders such as autism, anxiety and depression, schizophrenia and Alzheimer’s disease.^63–73

3.4. EPAC and Vascular Function

The vascular endothelium regulates the extravasation of solutes, macromolecules, and cells between the circulating blood and the adjacent tissues. Accumulating evidence suggests that the EPAC signaling pathway controls cAMP-mediated hormones on endothelial function including cell adhesion, endothelial barrier function and gap junction formation in vascular endothelial cells.^74–86 In addition, EPAC1-Rap signaling has been reported to regulate vascular permeability, intracellular responses, extracellular matrix adhesion and migration and inflammatory processes.^87–89

3.5. EPAC in Inflammation

As mentioned above, EPAC1 modulates inflammatory processes by regulating the immune responses. During inflammation, leukocytes extravasate from the blood into the extravascular space to regulate leukocyte and transendothelial migration and integrin-mediated adhesion. Several studies have shown that EPAC1 activation modulates numerous inflammatory mediators in various leukocytes.^33,79 In addition, EPAC1 enhances leukocyte adhesion and migration as a proinflammatory mediator.^{34,71,83,90–94} Intriguingly, recent studies using two mouse models of hyperalgesic priming provide new insights into the role of EPAC in chronic inflammatory pain. It has demonstrated that increasing the expression of GPCR kinase 2 (GRK2) in nociceptors or decreasing EPAC1 inhibited chronic hyperalgesia, suggesting that therapies targeted at tuning the balance between GRK2 and EPAC1 levels may have promise to prevent and treat chronic pain.⁹⁵

3.6. Renal Function of EPAC

EPAC proteins are also involved in many physiological processes in the kidney including ion transport and cellular proliferation. EPAC1 and EPAC2 are highly expressed by all three segments of the proximal tubules and are enriched at the brush border membrane.⁹⁶ Recent studies indicate that EPAC1 and/or EPAC2 represent a novel therapeutic target for diabetes-related diseases, particularly kidney failure. Although the exact mechanism of EPAC proteins -especially the contribution of EPAC1 and EPAC2 - to these diseases remains to be elucidated, the development of EPAC-specific activators or inhibitors as pharmacological probes will expedite the discovery process and biological characterization.^97–101

3.7. Airway Function of EPAC

Because of the critical role of cAMP in modulating airway functions, it was anticipated that cAMP-induced regulation of these processes including airway smooth muscle contraction and proliferation could be mediated by EPAC. Recent studies in human airway smooth muscle cells showed that cigarette smoke extract reduced EPAC1 expression.⁷¹ Interestingly, the reduction of EPAC1 expression was also found in lung tissue from patients with chronic obstructive pulmonary disease (COPD).⁷¹ More and more findings have demonstrated that EPAC-mediated signaling represents a novel key pathway in airway fibroblasts and airway smooth muscle cells.^102–104

4. Small Molecules Targeting EPAC

Since the discovery of the EPAC proteins in 1998, to better understand their functions and elucidate the EPAC-mediated signaling pathways under both physiological and pathological conditions, there is an increasing need to develop EPAC-specific ligands including agonists and antagonists as pharmacological probes and potential molecular therapeutics for human diseases. Several EPAC-selective and PKA-selective cAMP analogues have been identified as powerful pharmacological tools to distinguish the signaling pathways of EPAC and PKA in numerous biological processes.^105,106 However, cAMP analogues could not discriminate the functions of EPAC1 and EPAC2 as two major isoforms share extensive sequence homology. Although some non-nucleotide small molecules such as brefeldin A were shown to inhibit EPAC function, there was lack of evidence to suggest that these compounds directly interact with EPAC.^65,107–109 Early in 2012, Cheng’s research group reported a robust high throughput screening (HTS) assay by using a fluorescent cAMP derivative to identify EPAC-specific antagonists from compound libraries^110–112. After that, Lezoualc’h and colleagues reported their validated fluorescence-based HTS assay on the basis of the ability of EPAC to regulate the nucleotide exchange activity of Rap1 for screening EPAC ligands¹¹³. Recently, by collaborating with Cheng, our team made substantial medicinal chemistry effort on hit-to-lead optimizations.^114–116. The structure-activity relationships of cAMP analogues and newly discovered EPAC-specific antagonists, as well as the potential applications of these chemical probes for new therapeutic strategies will be discussed herein.

4.1. cAMP analogues as EPAC ligands

Early systematic studies focusing on structure-activity relationships of cAMP derivatives suggest that modification on 3′- and 5′-positions of the cyclophosphate ring of cAMP (1) with nitrogen or sulfur diminishes the affinity of cAMP analogues towards PKA (Figure 4). Further introduction of various substituents at these positions was found to be inappropriate.^117,118 The cAMP binding to PKA is sensitive to most modifications of the cyclic phosphate ring, while modifications on the exocyclic phosphate-oxygen result in several more interesting findings. For example, replacement of the exocyclic phosphate-oxygen by sulfur generated two diastereoisomers, Sp-cAMPS (2) and Rp-cAMPS (3), respectively (Figure 4).^119–121 Compound 2 is an activator of both PKA and EPAC, but has an affinity about 10-fold lower than cAMP. Compound 3 is a weak partial activator of PKA that acts as a competitive inhibitor at physiological PKA concentrations. It binds weakly to EPAC, where it acts as an inhibitor.¹²² Taken together, these results indicate that modification on the cyclic nucleotide is very sensitive for the activation properties.^122–126 Further modification on 2 and 3 at the position 2 or 8 of the adenine base with different spacers has provided a set of valuable tools for chemical proteomics approaches.¹²³

Figure 4.

Chemical structures of cAMP molecule (1) and cAMP analogues with modified exocyclic oxygen atom (2 and 3).

More recently, there has been remarkable progress in understanding the molecular mechanism of PKA and EPAC activation. In particular, X-ray crystal structure determinations of PKA and EPAC2 have revealed the structural information of the amino acid residues around the cAMP binding site. From comparison of the primary amino acid sequences of the different cAMP-binding domains between PKA and EPAC, it was noticed that there is a highly conserved glutamate residue only present in the cAMP binding domains of the PKA.^71,127,128 Insight into the structural basis indicated that 2′-OH on the ribose ring is essential for the binding to the conserved glutamate residue in the cAMP binding domain of PKA.

Several systematic structure–activity relationship investigations were carried out on 2′-OH on the ribose ring, the N⁶-position and the 8-position of the adenine moiety of cAMP. A variety of new cAMP analogues displayed the ability to selectively activate EPAC or PKA. 2′-deoxy-cAMP (4) with the 2′-OH moiety replaced by hydrogen selectively activated EPAC without any effect towards PKA (Figure 5).^122,129 Despite compound 4 exhibiting some selectivity towards EPAC versus PKA, it is not an ideal EPAC activator due to a binding affinity >1000-fold less than cAMP itself.⁷¹ Further detailed evaluation showed that 2′-O-alkyl-modified cAMP analogues such as 2′-O-Me-cAMP (5) displayed more potent EPAC1 activation than cAMP and were unable to activate PKA, and these compounds displayed about 10 to 100-fold improved EPAC/PKA binding selectivity (Figure 5).⁶⁴ These findings further suggest that the 2′-OH group is important for binding affinity and selectivity.^{64,106,127,128,130}

Figure 5.

Chemical structures of 2′-deoxy-cAMP (4), 2′-O-Me-cAMP (5), 6-Bnz-cAMP (6) and N⁶-phenyl-cAMP (7).

In contrast, modifications on the N⁶-position of cAMP abolished the ability of cAMP analogues to activate EPAC but not PKA. For instance, 6-Bnz-cAMP (6) and N⁶-phenyl-cAMP (7) are full PKA activators and inefficient EPAC activators (Figure 5).^64,122 This implies that suitably 6-substituted cAMP analogues may differentiate their selectivity between PKA and EPAC. It is worth mentioning that the generally low efficiency of the 6-modified analogues is likely not ascribed to the steric hindrance resulting from the introduction of bulky substituents. Some analogues with small groups, such as an oxygen atom at 6-position (like that in cGMP), are even less bulkier than that of cAMP and still have a lower affinity to EPAC.⁶¹

Systematical modifications on the 8-position of cAMP revealed that 8-substituted cAMP analogues possess differences in charge accommodation, hydrogen bonding and adaptability between cAMP homologous binding sites.^129,131,132 In addition, the lipophilicity of cAMP analogues with the hydrophobic substitutions at 8-position has a major impact on their cell membrane permeability.¹⁰⁶ As shown in Figure 6, several 2′-O-Me-cAMP analogues with modifications on the 8-position have been synthesized and their abilities to activate EPAC and PKA were compared. 8-Br-2′-O-Me-cAMP (8) and 8-pCPT-2′-O-Me-cAMP (9, a.k.a. 007) have been found not to activate PKA with an enhanced EPAC/PKA binding selectivity about three orders of magnitude. Detailed SAR studies have revealed that the S-phenyl ring on 8-position with para-substituted chloro (9), methoxy (10), or hydroxyl (11) can improve the binding affinity selectively towards EPAC (Figure 6).

Figure 6.

Chemical structures of 8-Br-2′-O-Me-cAMP (8), 8-pCPT-2′-O-Me-cAMP (9), 8-pMeOPT-2′-O-Me-cAMP (10) and 8-pHPT-2′-O-Me-cAMP (11).

EPAC-selective cAMP analog 9 was administered in several animal models, demonstrating its biological applications under physiological circumstances. For instance, EPAC activation by intraplantar injection of 9 was found to induce hyperalgesia via a PKCε-dependent route.^95,133 Intrarenal administration of 9 reduced renal failure in a mouse model for ischemia-reperfusion injury.¹³⁴ Intrahippocampal injection of 9 improved fear memory retrieval in contextual fear conditioning through regulating hippocampal EPAC2 signaling.¹³⁵ Although 9 was used as a handy pharmacological tool to elucidate the critical role of EPAC proteins in cAMP signaling pathway, the relatively low membrane permeability limited its further biological applications due to its negatively charged phosphate group. Introduction of an acetoxymethyl ester to 9 to mask the negatively charged, singly bonded oxygen on the phosphate group generated the labile ester 12 which consists of the equatorial and the axial isomers (Figure 7).¹³⁶ This prodrug-like molecule is capable of penetrating cell membranes more efficiently and can be intracellularly hydrolyzed directly by cellular esterases to release the biologically active parent compound 9. 8-pCPT-2′-O-Me-cAMP-AM (12) has thus become a powerful tool in EPAC and PKA related research. Compound Sp-8-pCPT-2′-O-Me-cAMPS (13) has been further developed as a highly specific EPAC activator with high lipophilicity and membrane permeability, while it is not metabolized by phosphodiesterases (Figure 7).^68,136,137 Despite significant progress on development of cAMP analogues as EPAC ligands, most of the selectivity studies with these cAMP analogues were based on EPAC1, none of which has been known to distinguish between the isoforms EPAC1 and EPAC2. In addition, the off-target effects of cAMP analogues including 9 to other targets such as PDEs are also an issue that needs to be addressed.^4,7,137

Figure 7.

Chemical structures of 8-pCPT-2′-O-Me-cAMP-AM (12), Sp-8-pCPT-2′-O-Me-cAMPS (13) and 8-(2-[7-nitro-4-benzofurazanyl]aminoethylthio)adenosine-3′, 5′-cyclic monophosphate (14, 8-NBD-cAMP).

4.2. Non-nucleotide Small Molecules as EPAC Antagonists

As discussed above, although the EPAC-selective cAMP analog 9 as an EPAC activator has proven very useful for the elucidation of EPAC-mediated signaling pathways, there remains an urgent need to develop non-nucleotide small molecules as additional EPAC-selective and isoform-specific ligands (e.g. antagonists) for expediting the discovery of new biological roles and functions of EPAC in human diseases as well as potential therapeutic strategies.

Cheng and colleagues first reported their sensitive and robust HTS assay for identifying antagonists that directly compete with a fluorescent cAMP derivative 8-NBD-cAMP (14, Figure 7) in binding to purified full-length EPAC2.¹¹⁰ This assay was based on the finding that the fluorescent signal changed in a dose-dependent manner after binding of 8-NBD-cAMP to EPAC2. In addition, more than 100-fold increase of the fluorescent signal was altered under near saturating EPAC2 concentration. Furthermore, the addition of excess unlabelled cAMP decreased the fluorescent signal in a dose-dependent manner. Consistent with previous findings,¹³⁸ alteration of the magnitude of fluorescence signal was only modest with a limited increase upon 8-NBD-cAMP binding to EPAC1 protein. Nevertheless, the significant reversible fluorescence change made it a remarkable readout for developing a sensitive and robust HTS assay for screening EPAC antagonists with high reproducibility.¹¹⁰

The pilot screening was preformed simply and effectively in 96-well format by using the NCI DTP (Developmental Therapeutics Program) diversity set library. Three compounds, NSC45576 (15), NSC119911 (16) and NSC686365 (17) (Figure 8) have been identified to significantly decrease the fluorescence signal with apparent IC₅₀ values of 1.7 μM, 3.8 μM and 7.9 μM, respectively. Further validation demonstrated that these compounds not only inhibited EPAC2 GEF activity but also inhibited EPAC1-mediated Rap1 nucleotide exchange at 25 μM.¹¹⁰ The counter-screening assays measuring type I and II PKA holoenzyme activities were performed to test the specificity of these compounds. The results revealed that 15 and 17 do not block cAMP mediated PKA activation suggesting that these two compounds are EPAC specific antagonists that selectively inhibit cAMP-induced EPAC activation.

Figure 8.

Chemical structures of identified EPAC inhibitors 15–17.

Further studies through a more extensive screen of the Maybridge Hitfinder compound library (14,400 compounds) under a 384-well format led to the discovery of seven HTS hits 18–24 (ESI-04 to ESI-10) that are able to inhibit EPAC2 activity completely at 25 μM in the presence of the equal concentration of cAMP.¹¹¹

As ESI-06 (20) and ESI-08 (22) shared the same fragment of 5-cyano-6-oxo-1,6-dihydro-pyrimidine (Figure 9), further exploration of the structural determinants on this scaffold have been investigated.¹¹⁴ Extensive structure–activity relationship (SAR) analysis has resulted in the identification of two more potent EPAC antagonists HJC0197 (25) and HJC0198 (26) (Figure 10). Compound 26 with a cyclopropyl moiety and 25 with the 6-cyclopentyl moiety instead of the 6-cyclohexyl group displayed enhanced activity with the IC₅₀ values of 4.0 μM and 5.9 μM, respectively. Further evaluation revealed that 25 also inhibits EPAC1-mediated Rap1-GDP exchange activity, while 26 is a little more specific for EPAC2.¹¹⁴

Figure 9.

Chemical structures of HTS hits 18–24.

Figure 10.

Chemical structures of 25 and 26.

Molecular docking studies using the AutoDock Vina algorithm^114,139,140 revealed that the hydrophobic groups of C-6 position and 2-position are critical for the binding affinity towards EPAC2. The SAR results and the representative docking data also indicated that optimization on C-6 position may lead to identification of more EPAC2-specific antagonists. As depicted in Figure 11A, the cyclopentyl group at the C-6 position interacts with the hydrophobic residues of Phe367, Ala415 and Ala416, while the hydrophobic S-benzyl moiety at 2-position forms interactions with Leu406 and Leu449. Further evaluation of 25 and 26 showed that both compounds did not alter cAMP-induced type I and II PKA holoenzymes activation at 25 μM, suggesting that both compounds are EPAC-specific inhibitors. The results in HEK293/EPAC1 and HEK293/EPAC2 cells have demonstrated that 25 and 26 can suppress EPAC1 and EPAC2 function in vitro.¹¹⁴

Figure 11.

Predicted binding mode and molecular docking of 25, and 23 into the cAMP binding domain B (CNBD-B) of EPAC2 protein. Important residues are drawn in sticks. Hydrogen bonds are shown as dashed green lines. (A) Binding mode of 25 (orange). The cyclopentyl group at the C-6 position interacts with the hydrophobic residues of Phe367, Ala415 and Ala416, while the hydrophobic S-benzyl moiety of 2-position forms interactions with Leu406 and Leu449. (B) Binding mode of 23 (pink). The tert-butylisoxazolyl moiety forms a hydrogen bond with the residue Gly404 and interacts with the hydrophobic residues of Phe367, Leu406, Ala407, and Ala415. Meanwhile, 3-chlorophenyl fragment forms hydrophobic interactions with residues Val386, Val394 and Leu397.

Further evaluation of cAMP-mediated EPAC1 GEF activity of seven hits (Figure 9) has revealed that compounds ESI-05 (19) and ESI-07 (21) are unable to inhibit EPAC1 GEF activity. Additional investigation on the specificity of 19 and 21 showed that neither of them displayed significant effects towards PKA, indicating that they are specific targeting EPAC2 protein. It is important to note that ESI-05 and ESI-07 inhibited EPAC2 activity with IC₅₀ values of 0.4 ± 0.1 μM and 0.6 ± 0.1 μM, respectively, whereas cAMP showed an IC₅₀ value of 40 ± 1 μM. Further studies demonstrated that 19 and 21 selectively modulated EPAC activation in living cells. Moreover, EPAC- and PKA-based fluorescence resonance energy transfer (FRET) sensor in living cells and Rap1-GTP pull-down assays further confirmed these results.¹¹¹ Recent detailed biophysical study further confirmed 19 as a direct and selective inhibitor of EPAC2 with a binding affinity about 20-fold higher than cAMP.¹⁴¹

As 19 and 21 are exclusively EPAC2-specific antagonists with no effect toward EPAC1 protein, the potential mechanism of these small chemical probes was further investigated to analyze the effect of 21 binding to EPAC2 protein using deuterium exchange mass spectrometry (DXMS) approach. The resulting data indicated that EPAC2-specific antagonists may bind to the interface of CNBD-A and CNBD-B and lock the EPAC2 protein in its autoinhibitory conformation.¹¹¹ Although other possible binding modes might not be excluded, the identification and characterization of these pharmacological probes capable of selectively targeting the EPAC2 protein represent a major milestone in the exploration of the complex cAMP signaling pathway. These EPAC2-specific antagonists have provided a new platform for further development.

Previous SAR studies and the molecular docking results demonstrated that appropriate hydrophobic moieties of the molecules play an essential role for the specificity and potency towards EPAC2 protein. Based on the scaffold of 19, the 2,4,6-trimethylphenyl moiety was retained as a privileged fragment in the preliminary medicinal chemistry effort for hit-to-lead optimization. Consistent with the previous SAR studies, the 2,4,6-trimethylphenyl ring as an appropriate hydrophobic pharmacophore is important for the potency. For instance, compound 27 with such a privileged fragment and another appropriate hydrophobic group was also found to be a potent antagonist with an IC₅₀ of 0.7 μM, while the methoxy compound 28 exhibited only moderate potency (Figure 12).¹¹⁵

Figure 12.

Chemical structures of representative diarylsulfones 27, 28 and N,N-diarylamines 29–32, and arylsulfonamides 33–35.

Based on the above SAR findings and the hit ESI-10 (24), a series of N,N-diarylamine derivatives have been designed, synthesized and characterized.¹¹⁵ Compound 29 showed an enhanced IC₅₀ of 3.8 μM in comparison with 24. It is worth highlighting that 30 and HJC0338 (31) with the 3,5-dichloro or 2,5-dichloro displayed high potency with IC₅₀ values of 0.9 and 0.4 μM, respectively (Figure 12). However, compound 32 containing a thiazole ring was found inactive with a dramatic loss of activity. SAR analysis of these N,N-diarylamine analogues further demonstrated that an appropriate hydrophobic phenyl moiety is preferable for competing with 8-NBD-cAMP binding to EPAC2.¹¹⁵

On the basis of the structural features of the preliminary SAR studies and the hit ESI-05, a series of arylsulfonamide derivatives as a novel chemical entity of EPAC2 antagonists have been designed and characterized (Figure 12).¹¹⁵ Compound 33 with an indole ring displayed a moderate potency with an IC₅₀ of 1.2 μM. Compound 34 with the 2-ethylpyrrole moiety as a hydrophobic fragment showed higher potency with an IC₅₀ of 0.5 μM, while 2,4-dimethylpyrrole derivative HJC0350 (35) displayed a significantly enhanced activity with an IC₅₀ of 0.3 μM. The systematic and extensive SAR results have also revealed that appropriate hydrophobic pharmacophores are important for the potency of these EPAC ligands. Further validation studies have demonstrated that these three series including diarylsulfones, N,N-diarylamines, and arylsulfonamides (Figure 12) are EPAC2-specific antagonists. In addition, live cell imaging studies using EPAC1, EPAC2, or PKA FRET reporters also confirmed that compound 35 is an effective EPAC2-specific antagonist.¹¹⁵

3-(5-tert-Butyl-isoxazol-3-yl)-2-[(3-chlorophenyl)-hydrazono]-3-oxo-propionitrile (ESI-09, 23) has also been indentified and characterized as a novel non-cyclic nucleotide EPAC2 antagonist with an IC₅₀ of 7.0 μM (Figure 9).¹¹² Meanwhile, 23 inhibited cAMP-mediated EPAC1 GEF activity with an IC₅₀ of 3.2 μM but no effect towards PKA, indicating that it is an EPAC-specific antagonist. Molecular docking studies based on the scaffold of 23 provided useful insight to design more potent antagonists. As shown in Figure 11B, the tert-butylisoxazolyl moiety forms a hydrogen bond with the residue Gly404 and interacts with the hydrophobic residues of Phe367, Leu406, Ala407, and Ala415. Meanwhile, 3-chlorophenyl fragment forms hydrophobic interactions with residues Val386, Val394 and Leu397. Further biological functional investigation showed that 23 specifically suppressed Akt phosphorylation, EPAC-mediated Rap1 activation and insulin secretion in pancreatic β cells. Using hit 23 as a pharmacological probe, the functional roles of EPAC1 in pancreatic cancer migration and invasion have been discovered, indicating that EPAC1 may represent an attractive therapeutic target for pancreatic cancer.¹¹²

Previous studies have revealed that activation of cAMP-EPAC pathway in hypothalamus induced multiple indices of leptin resistance, suggesting that EPAC may represent a novel pharmacological target for obesity.¹⁴² A more recent study demonstrated that resistance to high fat diet in EPAC1 knockout mice improved glucose tolerance, heightened leptin signaling and induced obesity. Compound 23 enhanced leptin signaling in organotypic hypothalamic slice culture system. Furthermore, administration of 23 in wild-type mice significantly reduced plasma leptin. These findings support the notion that EPAC1 may represent a novel therapeutic target for diabetes or obesity.¹⁴³

Very recent studies showed that genetic EPAC1 knockout mice protected them from fatal rickettsiosis, nearly completely blocked rickettsial attachment and/or invasion into the endothelial cells.¹⁴⁴ Importantly, pharmacological inhibition of EPAC1 in vivo using 23 completely recapitulates the EPAC1 knockout phenotype. Treatment with 23 in wild-type mice significantly protected them against rickettsial infection with much milder disease manifestations and dramatically improved survival, indicating that EPAC1 is a potential target for the prevention and treatment of fatal rickettsioses. 23 as the selective pharmacological probe has proven successful in unraveling the in vivo functions of EPAC1 to overcome potential limitations of knockout mouse models, and may provide potential novel therapeutics.¹⁴⁴ Meanwhile, 23 showed excellent tolerability during the in vivo studies, indicating its low toxicity to animals.¹⁴⁴ In addition, 23 displays no significant inhibitory effects on PDEs, as well as very weak inhibitory activities towards hERG and CYP450 enzymes (unpublished data). All these combined observations support that such non-nucleotide small molecules may have more advantages in terms of off-target effects and toxicities than cAMP analogs, although more extensive preclinical ADMET assessments remain to be explored for the clinical development.

Since compound 23 was identified as a novel EPAC-specific pharmacological probe to discern the physiologic functions of EPAC in vitro and in vivo, a practical and efficient method to readily prepare 23 at large scale was developed including a one-pot synthesis of isoxazole synthon, modified protocol for cyanomethyl ketone and a coupling step.¹¹⁶ The development of this efficient synthetic route for 23 will greatly facilitate the ongoing medicinal chemistry efforts of lead optimization and SAR studies in searching for new molecules with improved DMPK profiles. This is particularly important in light of a recent report by Rehmann suggesting that compounds 23 and 25 exhibit non-specific effects on protein stability when used at high concentrations above 50 μM.¹⁴¹ While more detailed characterizations are required to sort out the actual causes and mechanisms of such effects, the ability of 23 to recapitulate the genetic EPAC knockout phenotypes in multiple biological systems both in vitro and in vivo suggests that this class of compounds is specific to EPAC at lower concentrations or appropriate doses.^112,143,144 Therefore, it is imperative to further optimize these lead compounds through rational drug design approaches to develop advanced candidates for IND-enabling studies.

Lezoualc’h and colleagues reported another functional fluorescence-based high throughput screening assay to identify EPAC inhibitor compounds from chemical libraries.¹¹³ This variant of the assay is on the basis of the ability of EPAC to catalyze the nucleotide exchange activity of Rap1.¹⁴⁵ Tetrahydroquinoline CE3F4 (36) has been identified as EPAC1 inhibitor that blocked EPAC1-induced Rap1 activation both in cell-free systems and in intact cells and did not influence protein kinase A holoenzyme activity (Figure 13).¹¹³ Preliminary structure-activity study revealed that the formyl moiety on position 1 (37) and the bromo group on position 5 (38) of this class of analogues were essential for their activity targeting the EPAC1 (Figure 13). Although the additional data revealed that 36 and its analogues could also inhibit EPAC2, 36 and related compounds may serve as new chemical leads for further structural optimizations and useful pharmacological tools for determining the biological functions of EPAC1 in a wide range of biological processes.^113,146 Further investigation on the activity of the (R)- and (S)-enantiomers of 36 revealed that (R)-CE3F4 (39) is a more potent cAMP antagonist than the racemic 36 and (S)-enantiomer (Figure 13). (R)-enantiomer 39 displays 10-fold selectivity for EPAC1 over EPAC2. The biochemical evaluation of tetrahydroquinoline analogues is very helpful in elucidating the structural features that are essential for high EPAC1 inhibitory activity.¹⁴⁶

Figure 13.

Chemical structures of tetrahydroquinoline analogues 36–39.

5. Conclusions and Future Directions

Since the discovery of EPAC proteins as the novel cAMP mediators in 1998, significant progress has been made by many pioneers in discriminating between PKA- and EPAC-mediated effects under physiological and pathophysiological circumstances.¹⁴⁷ X-ray crystal structure determinations and NMR spectroscopy techniques have provided novel insights into the mechanism of cAMP activation of cyclic nucleotide binding domains. In particular, the development of the EPAC-selective cAMP analog 9 (8-pCPT-2′-O-Me-cAMP, 007) through rational design based on structure/sequence alignment analysis contributed to identification of new biological functions. Moreover, its prodrug 12 (8-pCPT-2′-O-Me-cAMP-AM, 007-AM) with improved cell permeability further delivered the reliable insights into the multiplicity of the biologic properties of EPAC. However, the potency, selectivity and specificity, as well as the off-target effects of cAMP analogues remain essential for further improvement.

Recent studies on development of the sensitive and robust HTS assays provided highly valuable non-nucleotide small molecules as EPAC-specific ligands (e.g. antagonists) for elucidating the signaling pathways and dissecting the physiological functions. Meanwhile, it is important to screen the specificity of new EPAC modulators using cells with deletion of EPAC1, EPAC2 or both, as well as their off-target effects to other proteins such as PDEs. Thus, developing new methodologies of chemical biology in this regard is also urgently needed. It is worthy to mention that the combination of genetic knockout animal models and small molecular EPAC-specific modulators has proven very useful to reveal the in vivo physiological functions of EPAC. Treatment with the EPAC-specific ligands with favorable in vivo bioavailability in wild-type animal models may overcome the potential shortcomings of the genetic knockout manipulations such as the secondary effects due to the complex signaling pathways and the possible physiological compensation.

It is the opinion of these authors that further HTS efforts to discover new chemical entities as EPAC-specific activators, and further medicinal chemistry efforts to optimize the currently available antagonists with enhanced potency, specificity to each of the isoforms, and drug-like properties, as well as the in vivo pharmacological evaluation in relevant disease models of lead candidates are highly anticipated to pave the way for the ultimate development of clinical applications. Given that EPAC proteins have been implicated to play important roles in major human pathological conditions such as cancer, neurological disorders, infectious disease, heart disease and diabetes, we believe that it is of great value in the development of pharmacological tools to probe the functions of EPAC in disease processes, and it is the hope that such efforts will eventually lead us to the development of advanced compounds into a clinical setting as potential mechanism-based therapies for those diseases.

ABBREVIATIONS USED

cAMP

3′–5′-cyclic adenosine monophosphate or cyclic AMP

ATP

adenosine triphosphate

GPCRs

G-protein coupled receptors

ACs

adenylate cyclases

PDEs

phosphodiesterases

PKA

protein kinase A

GEFs

guanine nucleotide exchange factors

EPAC

exchange protein directly activated by cyclic AMP

CNBD

cyclic nucleotide binding domain

DEP

Dishevelled Egl-10 Pleckstrin

REM

Ras exchange motif

RA

Ras association

CDC25-HD

CDC25-homology domain

Rap

Ras related protein

DXMS

deuterium exchange mass spectrometry

FRET

fluorescence resonance energy transfer

MD

molecular dynamics

β₁-AR

β₁-adrenergic receptor

SR

sarcoplasmic reticulum

CaMKIIδ

Ca²⁺/calmodulin-dependent protein kinase IIδ

RyR2

ryanodine receptor 2

GRK2

GPCR kinase 2

COPD

chronic obstructive pulmonary disease

HTS

high throughput screening

SAR

structure-activity relationship

PKCε

Protein kinase Cε

CYP450

cytochrome P450 enzymes

hERG

the human ether-à-go-go-related gene

8-NBD-cAMP

8-(2-[7-Nitro-4-benzofurazanyl]aminoethyl-thio) adenosine-3′, 5′-cyclic monophosphate

DTP

Developmental Therapeutics Program

ADMET

absorption, distribution, metabolism, excretion and toxicity

DMPK

drug metabolism and pharmacokinetics

IND

Investigational New Drug

Biographies

Haijun Chen received a dual degree (B.S. in chemistry and B.E. in bioengineering) from Fuzhou University, China in 2003. He earned his Ph.D. in medicinal chemistry from Shanghai Institute of Materia Medica, under the supervision of Dr. Fajun Nan in 2008. After graduation, he worked at ShangPharm for two years and then did his postdoctoral training at University of Texas Medical Branch (UTMB) under the direction of Dr. Jia Zhou. Currently, Dr. Chen is an Associate Professor at Fuzhou University. His research focuses on the rational drug design and synthesis of novel agents for the diagnosis and treatment of cancer and CNS disorders.

Christopher Wild studied biology and chemistry at the California State University, Northridge, where he received his Bachelor and Master of Science degrees under the direction of Dr. Gagik Melikyan. Subsequently, he worked as a program chemist for ChemicoMays, CA and a research chemist at Celenese, TX where he was a member of the acetyl catalyst development team under the tutelage of Dr. Michael Nutt. Currently, Christopher is a member of the chemistry faculty at San Jacinto College, TX, and is pursuing a Ph.D. as a Keck Research Fellow at UTMB under the supervision of Dr. Jia Zhou. His research focus is on the design and synthesis of small molecule inhibitors of EPAC and allosteric modulators of 5-HT_2C receptor.

Xiaobin Zhou received his B.S. in pharmacy from South-Central University for Nationalities, China in 2012. He is currently pursuing his Ph.D. degree in medicinal chemistry at the College of Chemistry and Chemical Engineering, Fuzhou University. His main interest focuses on the design and synthesis of small molecules targeting EPAC proteins.

Na Ye earned her Ph.D. in medicinal chemistry from Shanghai Institute of Materia Medica, Chinese Academy of Sciences, under the supervision of Dr. Ao Zhang in 2013. She is currently pursuing her postdoctoral training under the direction of Dr. Jia Zhou at UTMB. Her research topics focus on the target-based drug design and chemical synthesis of bioactive small molecules including EPAC-specific inhibitors as pharmacological probes and drug candidates for the treatment of CNS disorders, cancer, and other human diseases.

Xiaodong Cheng received a PhD in biochemistry and molecular biology from UTMB in 1994. After completing his postdoctoral fellowship at the University of California, San Diego in 1999, he returned to UTMB as a faculty member in the Department of Pharmacology and Toxicology. Currently, he is a Professor in the Department of Integrative Biology and Pharmacology, the University of Texas Health Science Center at Houston. His research has focused on applying multidisciplinary approaches to understand the structure and function of exchange protein directly activated by cAMP (EPAC) and oncogene KRAS. Recently, his research group has discovered novel small molecule inhibitors for EPAC proteins, established animal disease models, and is in the process of evaluating the therapeutic potential of targeting EPAC proteins.

Jia Zhou received his Ph.D. in organic chemistry in 1997 from Nankai University, China. Then he joined the chemistry faculty in the same university and was promoted to Associate Professor there. In 1999, he started his postdoctoral training in organic chemistry with Dr. Sidney M. Hecht at the University of Virginia. After further postdoctoral training in medicinal chemistry with Dr. Alan P. Kozikowski at Georgetown University, he has conducted research at Acenta Discovery, and PsychoGenics, Inc. as a Senior Principal Scientist for 7 years. Dr. Zhou is currently an Associate Professor (tenured) at the Chemical Biology Program, Department of Pharmacology and Toxicology at UTMB, leading a drug discovery research group. He is an author of more than 70 papers and an inventor of 9 patents.