Therapeutic potential for novel drugs targeting the type 1 cholecystokinin receptor

Abstract

Cholecystokinin (CCK) is a physiologically important gastrointestinal and neuronal peptide hormone, with roles in stimulating gallbladder contraction, pancreatic secretion, gastrointestinal motility and satiety. CCK exerts its effects via interactions with two structurally related class I guanine nucleotide-binding protein (G protein)-coupled receptors (GPCRs), the CCK₁ receptor and the CCK₂ receptor. Here, we focus on the CCK₁ receptor, with particular relevance to the broad spectrum of signalling initiated by activation with the natural full agonist peptide ligand, CCK. Distinct ligand-binding pockets have been defined for the natural peptide ligand and for some non-peptidyl small molecule ligands. While many CCK₁ receptor ligands have been developed and have had their pharmacology well described, their clinical potential has not yet been fully explored. The case is built for the potential importance of developing more selective partial agonists and allosteric modulators of this receptor that could have important roles in the treatment of common clinical syndromes.

This article is part of a themed section on Molecular Pharmacology of GPCR. To view the editorial for this themed section visit http://dx.doi.org/10.1111/j.1476-5381.2010.00695.x

The main purpose of this report is to explore the therapeutic potential of highly selective partial agonists and allosteric modulators of the type 1 cholecystokinin (CCK) receptor (CCK₁ receptor). To more fully appreciate this selective role, the normal physiology and pathophysiology of this receptor and of its natural, full agonist ligand, CCK, must first be reviewed.

Cholecystokinin is one of the earliest gastrointestinal hormones to be discovered, identified more than 90 years ago based on its ability to stimulate gallbladder contraction (Ivy and Oldberg, 1928). It was soon recognized to be identical to the factor responsible for stimulating pancreatic exocrine secretion (Harper and Raper, 1943). This hormone has subsequently been shown to have effects on enteric smooth muscle and on nerves at various locations in the peripheral and central nervous system (Rehfeld, 2004). One of its most important neural effects is post-cibal satiety (Kissileff et al., 1981; Smith and Gibbs, 1985; Muurahainen et al., 1988; Beglinger et al., 2001), a critically important role that could provide the basis of a highly useful treatment for obesity. In addition to its roles in stimulating smooth muscle cell contraction and exocrine cell secretion, CCK stimulates cell growth, energy production, gene expression and protein synthesis (Williams, 2001), processes that have substantial potential implications for agonist drug development.

It is well recognized that mammals synthesize CCK peptides in the I-cells of the small intestine, as well as the central nervous system (Rehfeld, 1978; Miller et al., 1984). CCK is found as variable length linear peptides, including 58, 39, 33 and eight residues, all of which share the pharmacophoric domain for the CCK₁ receptor, representing the carboxyl-terminal heptapeptide-amide (Rehfeld et al., 2001). CCK peptides also share their carboxyl-terminal tetrapeptide-amide with all molecular forms of gastrin, with this region providing the pharmacophoric domain for the type 2 CCK receptor (CCK₂ receptor) (Tracy and Gregory, 1964; Mutt and Jorpes, 1968). Critical post-translational modifications found on biologically active CCK peptides include sulfation of the tyrosine that is present seven residues from the carboxyl terminus of CCK, as well as amidation of the carboxyl-terminal phenylalanine residue (Solomon et al., 1984; Rehfeld, 1998).

CCK₁ receptor

Cholecystokinin exerts its physiological functions through the activation of two structurally related class I guanine nucleotide-binding protein (G protein)-coupled receptors (GPCRs) identified as the CCK₁ receptor (also known as the Type A CCK receptor) and the CCK₂ receptor (also known as the Type B CCK receptor) (consistent with IUPHAR and BJP GPCR Nomenclature). These receptors are approximately 67% conserved and are in close approximation on the phylogenetic tree of GPCRs, within the class I grouping (Kolakowski, 1994). This review will focus only on the CCK₁ receptor.

Localization and function of CCK₁ receptor

The activation of the CCK₁ receptor by CCK has been shown to be responsible for a broad variety of important physiological functions, including the stimulation of gallbladder contraction and pancreatic exocrine secretion, delay of gastric emptying, relaxation of the sphincter of Oddi, inhibition of gastric acid secretion and induction of post-cibal satiety (Solomon et al., 1984; Anagnostides et al., 1985; Kerstens et al., 1985; Schmitz et al., 2001). The activation of the CCK₁ receptor by CCK also stimulates trophic and proliferative effects in some target cells (Matozaki and Williams, 1989; Matozaki et al., 1990; Moralejo et al., 1998; Moralejo et al., 2001).

The CCK₁ receptor mRNA and protein have been localized in the human gastric mucosa within D cells, where CCK can stimulate somatostatin secretion, thereby affecting gastric acid secretion (Schmitz et al., 2001). The CCK₁ receptor is also located in the muscularis propria of human gastric antrum, fundus and pylorus (Reubi et al., 1997), where activation by CCK has been shown to modulate contractile ability and thus affect gastric emptying (Fried et al., 1991; Scarpignato et al., 1996; Zerbib et al., 1998).

One of the best described physiological roles of CCK activation of the CCK₁ receptor is stimulation of pancreatic enzyme secretion. In the exocrine pancreas, the CCK₁ receptor has been shown to be present with distinct cellular distributions in rodents, guinea-pigs and humans (Christophe et al., 1978; Deschodt-Lanckman et al., 1978; Jensen et al., 1980; Yu et al., 1987). The majority of the literature pertaining to the functional role of the CCK₁ receptor in the pancreas has been examined using the rodent pancreatic acinar cell model. Yet, until recently, the expression of the CCK₁ receptor in human pancreas has been difficult to demonstrate, possibly reflecting very low levels of this receptor in this species (Ji et al., 2001). Galindo et al. (2005) were finally able to apply qRT-PCR to multiple human specimens to clearly demonstrate expression of the CCK₁ receptor mRNA in human pancreas, but this did not establish cell of origin. Human pancreatic acinar cell expression of this receptor was finally supported by Murphy et al. (2008) when they demonstrated direct activation of isolated human pancreatic acinar cells at physiologic concentrations of CCK, with this hormone eliciting oscillatory increases in cytosolic calcium activation and stimulation of enzyme secretion. Substantial evidence also supports the presence of this receptor on intrapancreatic neurons and on abdominal branches of the vagus nerve in several species (Owyang and Logsdon, 2004). Indeed, CCK₁ receptors on vagal afferent fibres have been shown in vivo to mediate pancreatic enzyme secretion (Li et al., 1997).

The CCK₁ receptor has also been detected in the endocrine pancreas in human insulin- and glucagon-secreting cells (Morisset et al., 2003), and CCK has been shown to stimulate the release of insulin from the pancreas (Karlsson and Ahren, 1989; 1991;). In the gallbladder, CCK₁ receptor expression on smooth muscle cells has been shown in humans and is responsible for stimulation of gallbladder contraction (Tokunaga et al., 1993).

It is well established that the CCK₁ receptor is present on distinct neuronal nuclei in the central nervous system, while the predominant CCK receptor present in the brain is the CCK₂ receptor (Innis and Snyder, 1980; Moran et al., 1987; Wank, 1995). Vagal afferent nerve expression of CCK₁ receptors has not only been shown to mediate pancreatic acid secretion, but has also been shown to be important in the response to gastric load (Schwartz et al., 1994) and eliciting satiety signals (Weatherford et al., 1993).

Pathophysiological role of CCK₁ receptor

The CCK₁ receptor has been implicated in common and important pathological disorders such as irritable bowel syndrome (IBS) that is characterized by abdominal pain and disturbed bowel habits (Zhang et al., 2008). Sjolund et al. (1996) have shown that CCK release may contribute to the intestinal dysmotility in IBS patients, and recently Zhang et al. (2008) have shown an increase in plasma CCK levels in IBS patients.

The CCK₁ receptor has also been implicated in the production of cholesterol gallstones. Cholesterol gallstone disease has been associated with reduced CCK binding to the CCK₁ receptor and with impaired gallbladder smooth muscle contraction in response to this hormone, while patients with pigment gallstones have been reported to have normal CCK binding and responsiveness. The impaired responses to CCK have been demonstrated to be reversed by removal of excess membrane cholesterol (Xiao et al., 1999).

One of the most interesting and relevant functions of the CCK₁ receptor stimulation by CCK is related to its role in modulating food consumption, which may provide the basis of a highly useful treatment for obesity. One of the first descriptions of the role of CCK as a food inhibitor was by Gibbs et al. in 1973, when they showed that intraperitoneal administration of CCK suppressed intake of both solid and liquid in a dose-dependent manner in rats (Gibbs et al., 1973). Since that initial observation, much research has been performed on the role CCK plays as a satiety factor. By definition, administration of a satiety factor at the start of a meal results in a smaller-than-normal meal being consumed. Satiety signals result from exogenously administered or endogenously produced mediators that activate specific receptors and result in cessation of eating. It is advantageous when this occurs early in the course of a meal. Studies in human have shown that intravenous infusion of CCK-8 and CCK-33 increase the perception of fullness, decrease hunger and reduce energy intake (Kissileff et al., 1981; Muurahainen et al., 1988; Lieverse et al., 1994; MacIntosh et al., 2001).

CCK₁ receptor signalling

In order to discuss the potential importance of developing more selective partial agonists and allosteric modulators of the CCK₁ receptor, we must first understand the broad spectrum of signalling initiated by activation with the natural full agonist peptide ligand, CCK. Here, we provide careful reference to signalling pathways that may have potential implications for agonist drug development, with these pathways depicted graphically in Figure 1.

Figure 1.

Schematic depiction of the signalling pathways stimulated by cholecystokinin (CCK) activation of the CCK₁ receptor. CCK₁ receptor activation by CCK is known to result in activation of G_q-initiated pathways with stimulation of classical second messengers (bold line) such as phospholipase C (PLC)/phosphatidylinositol 4,5-bisphosphate (PIP₂)/diacylglycerol (DAG) and protein kinase C (PKC)/inositol 3, 4, 5-triphosphate (IP₃) resulting in the release of intracellular calcium. Additionally, nicotinic acid adenine dinucleotide phosphate (NAADP) and cyclic ADP-ribose (cADPr) pathways have also been implicated in the release of calcium after ligand stimulation of the receptor. A number of other signalling pathways have been shown to be activated by ligand stimulation of the CCK₁ receptor. These include the class I phosphatidylinositol 3-kinase (PI3K) pathway involving Akt and stimulation of mammalian Target of Rapamycin (mTOR) that results in downstream activation of transcriptions factors, the NF-κB pathway as well as the G_s-initiated adenylate cyclase (AC) pathway. Three mitogen-activated protein kinase (MAPK) pathways have been shown to be stimulated by CCK₁ receptor activation, p38-MAPK, extracellular-regulated kinase (ERK) and c-Jun amino-terminal kinase (JNK). The activation of the CCK₁ receptor has also been shown to stimulate the phosphorylation of epidermal growth factor receptor (EGFR), resulting in activation of Ras that in turn has been shown to be important in the activation of the JNK and ERK pathways. While no direct evidence has shown CCK activation of arrestin family proteins, these are known to have various signalling roles in many GPCRs and bind and potentially regulate several subsets of proteins from various signalling pathways (dashed lines).

CCK₁ receptor G_q signalling pathways

Over the years, it has become clear that multiple affinity states of the CCK₁ receptor can be observed that are most clearly related to its status of coupling with its heterotrimeric G protein (Jensen et al., 1980; Sankaran et al., 1980). There is general agreement that the agonist-occupied CCK₁ receptor that is coupled with G_q represents a high affinity state of this receptor. However, as noted below, multiple G proteins are capable of coupling with this receptor, yet the specific functional effects on ligand binding of each are not clear. Similarly, it is well established that the cellular environment in which this receptor resides can affect its specific binding characteristics. This might reflect differences in the lipid composition of the plasma membrane, as well as differences in the relative stoichiometry of receptor and individual G protein expression and in the cellular concentrations of guanine nucleotides. Ultimately, it will be key to determine the molecular and structural basis for observed affinity states and for the initiation of particular signalling pathways. Unfortunately, at the present time, these levels of knowledge and understanding are absent.

The signalling pathways involved in the spectrum of action of CCK at the CCK₁ receptor have become an area of great interest. Stimulation of this receptor by CCK is known to result in the proximal initiation of several different vectorial signalling pathways. These include coupling with G_q and G_s, as well as the possibility of activating non-G protein-mediated signalling pathways. The predominant pathway studied for CCK activation of the CCK₁ receptor includes the G_q family of G proteins, G_q, G₁₁ and G₁₄ (Xu et al., 1998; Yule et al., 1999). Stimulation of the CCK₁ receptor by an agonist such as natural CCK results in a conformational change in this receptor, including its third intracellular loop where G proteins often couple. This exposes a region of the receptor that binds G_q and results in the exchange of GTP for bound GDP at the G protein, with the dissociation of the GTP-bound α-subunit from its β and γ subunits. The GTP-bound α-subunit then activates phospholipase C (PLC). PLC is known to cleave phosphatidylinositol 4, 5-bisphosphate (PIP₂) to form inositol 3, 4, 5-triphosphate (IP₃) and diacylglycerol (DAG) (Trimble et al., 1987; Matozaki and Williams, 1989). IP₃ moves to the endoplasmic reticulum where it stimulates the release of calcium stores (Muallem et al., 1985). The diacylglycerol activated by PLC then activates various protein kinase C (PKC) isoforms.

In the rodent pancreatic acinar cell model, physiological concentrations of CCK binding to the high affinity state of this receptor have been shown to generate transient calcium oscillations (Matozaki et al., 1990). In contrast, stimulation of the CCK₁ receptor with high concentrations of CCK that would be expected to saturate both high and low affinity states of the receptor has been shown to cause a rapid transient elevation of intracellular calcium that peaks within seconds, followed by a smaller sustained increase (plateau) in intracellular calcium without oscillations. IP₃ has been shown to be responsible for the calcium responses to stimulation of this receptor by high concentrations of CCK in rodent pancreatic acinar cells, yet low concentrations of CCK are reported to result in no measurable IP₃ response (Matozaki et al., 1990). The latter may reflect the relative insensitivity of the assay.

It has been postulated that the initial release of calcium in response to physiological concentrations of CCK is from IP₃-sensitive stores (Wakui et al., 1990). This subsequently may result in the calcium-induced release of calcium from other IP₃-insensitive stores (Wakui et al., 1990). ADP-ribose (cADPr) may be involved in one of these IP₃-insensitive pathways, acting on ryanodine receptors (Galione et al., 1991; Meszaros et al., 1993; Lee et al., 1997). Both IP₃- and cADPr-sensitive calcium release channels have been shown to be involved in CCK-stimulated calcium spiking, with blockade of cADPr resulting in enhancement of the IP₃-induced calcium spiking (Thorn et al., 1994). Other work in pancreatic acinar cells has suggested that CCK could elicit calcium spiking via either the cADPr or IP₃ pathways, with the level of intracellular glucose determining which of the two pathways would be predominant (Cancela et al., 1998).

Along with the cADPr and IP₃ pathways, the NAADP (nicotinic acid adenine dinucleotide phosphate) pathway has been implicated in the release of calcium from intracellular stores. NAADP is a metabolite of NADP⁺ that is also synthesized by ADPr cyclase and releases calcium from stores that are physically separate from IP₃ and cADPr (Lee and Aarhus, 1995; Lee et al., 1997). Yamasaki et al. (2005) produced some of the first evidence linking CCK₁ receptor stimulation to NAADP production, as well as demonstrating a difference in time course between NAADP and cADPr responses to physiological concentrations of CCK. They also postulated that physiological concentrations of CCK could initially enhance NAADP production and stimulate a localized increase in calcium release that would then activate cADPr-sensitized receptors. This, in turn, would allow the maintenance of calcium spiking for the duration of agonist stimulation of pancreatic acinar cells.

Mitogen-activated protein kinases (MAPKs) are serine-threonine-directed kinases that are activated by a variety of stimuli including various hormones. MAPKs have been shown to regulate a number of important cellular processes such as gene transcription, protein translation, metabolism and cytoskeletal function. Three MAPK pathways, representing the ERK (extracellular-regulated kinase), JNK (c-Jun amino-terminal kinase) and p38 MAPK pathways, have been described in pancreatic acinar cells in response to CCK stimulation of the CCK₁ receptor. ERK1 and ERK2 are activated by mitogens in all cells and serve an important role in cell growth and mitogenic proliferation, as well as being implicated in the enhancement of cell cycle entry (Robinson and Cobb, 1997). Isolated rat pancreatic cells have been used to identify increases in ERK1 and ERK2 activities after stimulation with CCK, implicating both the PKC and tyrosine kinase pathways (Duan and Williams, 1994). CCK stimulation has also been shown to activate Ras, which then can activate Raf family members (RafA, RafB and cRaf1) (Dabrowski et al., 1997). This group went on to show that CCK stimulation of the CCK₁ receptor activated MEK1/2 (MAP kinase or ERK kinase) that then could activate ERK1/2. CCK stimulation of pancreatic acinar cells has also been shown to result in the formation of a complex of Shc (Src homology/collagen related)-Grb2 (growth factor receptor-bound protein 2)-SOS (Drosophila homolog, Son of Sevenless) using a PKC-dependent mechanism. This group proposed that this may provide the link between G_q-coupled CCK receptor stimulation and Ras (Dabrowski et al., 1996b). Using the pancreatic acinar carcinoma cell line AR42J, it was shown that CCK activation can cause tyrosine phosphorylation of EGFR (epidermal growth factor receptor), resulting in the subsequent formation of the Shc-Grb2-SOS complex (Piiper et al., 2003). Additionally, it has been shown that CCK-stimulated activation of these signalling events can be abolished by transfection with a dominant negative Ras construct (Nicke et al., 1999). Downstream of ERK1/2, the 90-kDa ribosomal S6 protein kinase (p90^RSK) is activated in a time- and dose-dependent manner (Bragado et al., 1997b). ERKs and p90^RSK have also been shown to activate transcription factors in other cell systems, but little is known about the specific transcription factors activated by CCK stimulation of the CCK₁ receptor.

Stress and CCK can activate the p38 MAPK pathway in pancreatic acinar cells, which can lead to the activation of MAPKAP kinase-2, resulting in an increase in Hsp27 phosphorylation (Schafer et al., 1998). It has been shown that p38 MAPK activation is maximal 5 to 10 min after stimulation, which is similar to that seen with ERK1/2 stimulation. Additionally, it has been found that the CCK concentration needed to activate p38 MAPK is approximately in the range of 300 pM to 1 nM, similar to that needed for ERK1/2 activation and closer to the physiological range than that needed to obtain JNK stimulation. Higher concentrations of CCK could lead to a decrease in p38 MAPK activity. Other experiments showed that CCK stimulation of the CCK₁ receptor resulted in activation of the p38 MAPK pathway via G_q-coupled signalling (Schafer et al., 1998). Other than Hsp27, the specific transcription factors that are activated by CCK-induced p38 MAPK activation are not known.

Another MAPK pathway that has been shown to be activated by CCK is that of JNK. It was identified and named after its ability to phosphorylate the amino terminus of the c-Jun transcription factor (Hibi et al., 1993; Derijard et al., 1994; Kallunki et al., 1994). JNK has been implicated in mitogenic signalling and activation of JNK has been shown to induce apoptosis (Xia et al., 1995; Ip and Davis, 1998). Early work in isolated pancreatic acinar cells stimulated with CCK described activation of two forms of JNK first described as protein 46 (p46) and protein 55 (p55) based on their apparent molecular weights, with the maximal activation produced after 30 min (Dabrowski et al., 1996a). This work also demonstrated that the CCK concentration needed to stimulate JNK was ten times higher than that of ERK1/2 (Dabrowski et al., 1996a). This group noted that in their cell model JNK was shown to be independent of PKC, calcium and cAMP, and was dependent on Ras (Dabrowski et al., 1996a). While little is known about the role the JNK pathway plays in physiologic CCK₁ receptor activation, it has been suggested that JNK activation by high concentrations of CCK may induce pancreatitis (Dabrowski et al., 1996a).

The class I phosphatidylinositol 3-kinases (PI3K) signalling pathway has been implicated in many different aspects of cell biology including vesicle trafficking, cell growth, DNA synthesis, as well as regulation of apoptosis and cytoskeletal changes (Vanhaesebroeck et al., 2001). The molecular mechanism for PI3K activation by CCK is not well understood. However, it is able to phosphorylate and to activate Akt, which is known to have several targets including mammalian Target of Rapamycin (mTOR) (Vanhaesebroeck et al., 2001). In acinar cells, mTOR activation has been related to the translational control of protein synthesis. Specifically, ribosomal protein S6 has been shown to be phosphorylated by p70^S6K in response to CCK stimulation of pancreatic cells (Bragado et al., 1997a). Another target for mTOR that has been described in pancreatic acinar cells stimulated by CCK is the binding protein for eukaryotic initiation factor 4E (eIF4E-BP). In isolated acinar cells and in vivo, CCK has been shown to stimulate eIF4E-BP phosphorylation, resulting in the release of eIF4E and the formation of an eIF4E-eIF4G complex (Bragado et al., 1998).

The rat pancreatic acinar cell model has implicated the PI3K pathway in some potentially pathologic responses of pancreatic acinar cells, such as activation of the NF-κB transcription factor and intracellular activation of trypsinogen (Gukovsky et al., 2004). These authors suggested that the PI3K pathway may play a role in acute pancreatitis induced by stimulation with supramaximal concentrations of CCK (Gukovsky et al., 2004).

Recently, further investigation of the PI3K-Akt-mTOR pathway by Berna et al. (2009) identified some important distinctions between the Akt responses with high concentrations of CCK compared with that of low concentrations of this hormone. This group demonstrated that physiological concentrations of CCK could stimulate a ‘biphasic’ Akt response, with an increase in p85 phosphorylation and Akt activation by a Src-dependent mechanism. In contrast, high concentrations of CCK, possibly acting through a low affinity state of the CCK₁ receptor, were shown to be dependent on PLC and independent of Src or ERK (Berna et al., 2009). In addition to this, it was found that high concentrations of CCK lead to a decrease in p85 tyrosine phosphorylation, PI3K activity, Akt activation and translocation to the membrane (Berna et al., 2009).

An additional signalling pathway of interest includes activation of transcription factor NF-κB. This has been implicated in inflammatory and cell death pathways (Gukovsky et al., 1998; Steinle et al., 1999). Supramaximal concentrations of cerulein have been shown to activate NF-κB in the isolated pancreas, and in vivo studies of experimental pancreatitis showed that NF-κB activation is one of the earliest events in pancreatic inflammation (Grady et al., 1997; Gukovsky et al., 1998). The hypothesis put forward by Gukovsky et al. (1998) was that NF-κB activation resulted in upregulation of cytokines associated with pancreatitis that mediate acinar cell death and inflammation (Gukovsky et al., 1998). In rat pancreatic acinar cells, the translocation of PKC isoforms δ and ε is necessary for CCK-8-induced NF-κB activation (Satoh et al., 2004). Activation has been shown to be regulated through calcium signalling (Gukovsky et al., 2004) and to involve both phosphoinositol (PI)-specific PLC and phosphocholine (PC)-specific PLC (Satoh et al., 2004).

Additional G protein-coupled pathways

The CCK₁ receptor has also been shown to be capable of coupling to G_S, with stimulation by high concentrations of CCK leading to increases in cAMP (Sjodin and Gardner, 1977; Yule et al., 1993). Wu et al. (1997) have shown that CCK₁ receptors on HEK-293 cells are coupled to adenlyl cyclase (AC) exclusively through G_S. Furthermore, the specific residue, Asn82 in the first intracellular loop of the CCK₁ receptor, has been shown to be necessary for the cAMP responses observed after agonist stimulation (Wu et al., 1999).

Another G protein that can be activated by CCK stimulation of this receptor is G₁₃. In NIH3T3 cells, stimulation of the CCK₁ receptor by CCK has been shown to activate G₁₃ and to stimulate RhoA, resulting in the induction of actin stress fibres (Le Page et al., 2003).

Other signalling pathways

Additional signalling molecules that have been shown to be stimulated after activation of the CCK₁ receptor include protein kinase D1 (PKD1), p125 focal adhesion kinase (125^FAK) and Src. PKD1 is a serine/threonine kinase (Valverde et al., 1994) that has been shown to be activated by PKC family members (Waldron et al., 2001). In rat pancreatic acini, the most relevant PKC isoform for this response was believed to represent PKC-δ (Berna et al., 2007b). Activation of PKD1 may be of substantial importance, as it has been implicated in cell proliferation and invasion (Bowden et al., 1999; Zhukova et al., 2001).

The focal adhesion tyrosine kinase (FAK) p125^FAK and the closely related proline-rich kinase 2 (PYK2) are cytoplasmic protein kinases that have been implicated in important processes such as cellular growth, motility, adhesion and changes to the cellular cytoskeletal system (Schaller, 2001). Pace et al. (2003) have shown that activation of the CCK₁ receptor results in rapid tyrosine phosphorylation of p125^FAK and PYK2 in rat pancreatic acini. An additional protein implicated in the FAK signalling pathway is Lyn, a member of the Src family of kinases (SFK). Activation of the CCK₁ receptor by CCK has been shown to activate Lyn, resulting in association with PKC-δ, SHC, p125^FAK and PYK2 in pancreatic acinar cells (Pace et al., 2006).

Non-G protein-coupled signalling pathways

In addition to the classical G protein-mediated signalling pathways, it is now accepted that many GPCRs may also function through an additional pathway that involves arrestin family proteins. Studies with rhodopsin and the β-adrenergic receptor first described the role of arrestins in GPCR desensitization (Kohout and Lefkowitz, 2003). This mechanism involves the phosphorylation of agonist-stimulated GPCRs by one or more of seven G protein-coupled receptor kinases (GRK) that leads to the recruitment of one of four arrestin family members that impairs signalling by uncoupling of the receptor from its respective G protein (Freedman and Lefkowitz, 1996; Lefkowitz and Shenoy, 2005; Tobin et al., 2008). It is now recognized that arrestin may also be involved in signal transduction, thought to be produced by the conformational change elicited when arrestin binds to GPCRs. Arrestin family members have been shown to bind to a number of signalling proteins and have been implicated in signalling via various pathways such as JNK, ERK and p38 MAPK (Shenoy and Lefkowitz, 2003). While no direct studies of this phenomenon have been performed with the CCK₁ receptor, the CCK₂ receptor has been shown to recruit and bind arrestin after agonist stimulation (Barak et al., 2003), making it likely that arrestin plays a similar role in CCK₁ receptor activation.

Peptidic analogues

Two particularly useful peptide analogues of CCK have been identified as JMV180 and OPE (Galas et al., 1988; Gaisano et al., 1989). The JMV180 analogue corresponds to the carboxyl-terminal heptapeptide of CCK in which the carboxyl-terminal amidated phenylalanine residue has been replaced by a phenylethyl ester moiety and the amino terminus is blocked with a t-BOC moiety (t-BOC-Tyr(SO₃H)-Nle-Gly-Trp-Nle-Asp-O-phenylethyl ester) (Galas et al., 1988). The OPE analogue (_DTyr-Gly-Asp-Tyr(SO₃H)-Nle-Gly-Trp-Nle-Asp-O-phenylethyl ester) differs in its amino terminus, where a Gly spacer and a _DTyr site for direct radioiodination have been provided (Gaisano et al., 1989; Gaisano and Miller, 1992).

The OPE and JMV180 peptide analogues of CCK have been shown to bind to the CCK₁ receptor in manner similar to CCK, yet they elicit only a subset of the biological responses as those stimulated by the natural hormone (Gaisano et al., 1989; Sato et al., 1989; Gaisano and Miller, 1992). Hence, they represent partial agonists. The biological responses stimulated by these agents correlate with the responses attributed to the high affinity state of the receptor, while they have been described as low affinity CCK₁ receptor antagonists, even reversing CCK-induced supramaximal inhibition of enzyme secretion (Gaisano et al., 1989). JMV180 and OPE are fully efficacious stimulants of amylase secretion and do not show supramaximal inhibition with high concentrations as is typical of CCK (Gaisano et al., 1989; Sato et al., 1989; Gaisano and Miller, 1992). These analogues stimulate intracellular calcium oscillations with spiking similar to that stimulated by low concentrations of CCK (pM) (Galas et al., 1988; Gaisano and Miller, 1992). Both JMV180 and OPE do not stimulate measureable IP₃ responses, similar to that observed with low concentrations of CCK. At supramaximal concentrations of these agents, a low level of IP₃ response has been observed, but it is significantly lower than that observed after stimulation with equivalent concentrations of CCK (Gaisano et al., 1994). Additional pathways have been investigated for JMV180 and OPE, such as pancreatic growth and the phospholipase A₂ signalling pathways. JMV180 has been shown to stimulate pancreatic growth, but its effect was shown to be 1000 times less potent than CCK (Rivard et al., 1994). The intracellular calcium response elicited by JMV180 stimulation of the CCK₁ receptor was investigated further, and it was found that the calcium oscillations could be inhibited by a phospholipase A₂ (PLA₂) inhibitor, and a 2.5-fold increase in the intracellular levels of arachidonic acid (AA) was obtained in a biphasic manner (Tsunoda and Owyang, 1995). The authors of this study concluded that the high affinity CCK₁ receptor is coupled to the PLA₂ pathway to produce AA that mediates calcium oscillations and monophasic amylase secretion in rat pancreatic acinar cells. The authors did note that the intracellular calcium pool that is sensitive to the PLA₂ pathway is the same as or similar to the IP₃-sensitive pool (Tsunoda and Owyang, 1995).

CCK₁ receptor peptide analogues of CCK

This review will highlight some of the CCK₁ receptor agonists identified that have shown the most promise as anti-obesity therapeutic agents; however, this study will only briefly explore these compounds, as excellent detailed reviews have been published describing the chemistry of development of specific CCK₁ receptor agonists (Szewczyk and Laudeman, 2003; Berna et al., 2007a; Garcia-Lopez et al., 2007).

Tilley et al. published several studies in which they modified Ac-CCK-7 to change its functional profile (Tilley et al., 1991; 1992b,c; Danho et al., 1992). The modifications that produced functional changes were the replacement of the dipeptide Met28 and Gly29 with analogues incorporating 3-aminobenzoic acid and (1S)-trans-2-aminocyclopentanecarboxylic acid. These were high affinity, potent anorectic agents (Tilley et al., 1992b). Modification of Asp32 with (R)-5,5-dimethylthiazolidine-4-carboxylic acid produced a peptide that exhibited high affinity, very good CCK₁ receptor/CCK₂ receptor selectivity and more effective inhibition of food intake than that of Ac-CCK-7 (Tilley et al., 1992a).

A CCK-7 heptapeptide analogue, A71378 (des-NH₂-Tyr(SO₃H)-Nle-Gly-Trp-Nle-(NMe)Asp-Phe-NH₂), was described in the literature in 1990 and was found to be 700-fold more selective for pancreatic CCK₁ receptor and a full agonist for stimulation of intracellular calcium (Lin et al., 1990). Another class of potent and selective CCK₁ receptor agonists characterized was based on the replacement of Met31 in CCK-4 tetrapeptide with a substituted lysine. The two most potent compounds identified were A71623 and A70874, both without the tyrosine residue thought to be required for potent CCK₁ receptor activity in larger peptides (Liddle, 1989). A71623 (Boc-Trp-Lys(o-tolylaminocarbonyl)-Asp-MePhe-NH₂) and A70874 (Boc-Trp-Lys(p-hydroxycinnamoyl)-Asp-(NMe)Phe-NH₂) are 1200- to 140-fold selective for pancreatic CCK₁ receptors over brain CCK₂ receptors respectively. Functionally, A71623 is a fully efficacious agonist in producing PI hydrolysis, similar to CCK-8, and shows supramaximal inhibition of amylase secretion. In contrast, A70874 is only partially efficacious in stimulating PI hydrolysis and does not show supramaximal inhibition of amylase secretion, as well as reversing the inhibition produced by supramaximal concentrations of CCK-8. The amylase secretion profile of A70874 is very similar to JMV180 and OPE, except A70874 possesses greater efficacy to stimulate amylase secretion (Lin et al., 1991).

Further modifications to CCK-8 resulted in FPL14294 (also known as ARL142941 and ARR142941) (Hpa(SO3H)-Met-Gly-Trp-Met-Asp-(Me)Phe-NH2) (Simmons et al., 1994) and then the improved analogue ARR15849 (also known as ARL15849) (4-hydroxyphenylacetyl(SO₃H)-Nle-Gly-Trp-Nle-(Me)Asp-Phe-NH₂) (Pierson et al., 1997). ARR15849 was synthesized by moving the N-methyl group from Phe to Asp, which resulted in a decreased affinity to the CCK₂ receptor, without affecting its affinity for the CCK₁ receptor, giving a 6600-fold selectivity for the CCK₁ receptor (Pierson et al., 1997). ARR15849 was shown to inhibit food intake with nanomolar potency following intraperitoneal administration in rat and intranasal administration in beagle dogs (Simmons et al., 1994; Pierson et al., 1997), supporting a CCK receptor-mediated effect.

Non-peptidyl CCK₁ receptor agonists

Peptide ligands have limited therapeutic utility due particularly to limited bioavailability and substantial metabolism. For this reason, there is much more enthusiasm for the development of non-peptidyl ligands that can be administered orally. One such approach was the characterization of the non-peptidyl tetra-substituted tryptophan derivative, PD149164, that was shown to have full agonist activity at the CCK₁ receptor, but lower potency than natural CCK (Hughes et al., 1996). This research group took the approach of designing a non-peptidyl agonist, PD149164, by appending a key feature of the selective peptide agonist A71623, described above. PD149164 was shown to have a higher affinity for the CCK₂ receptor than for the CCK₁ receptor, but the enantiomer, PD151932, was shown to have a much higher binding affinity for the CCK₁ receptor; however, it was found to represent an antagonist. Functional studies of the PD149164 agonist showed that it elicited a maximal amylase response both in vitro and in vivo, and produced intracellular calcium oscillations similar to CCK-8 (Hughes et al., 1996). The same group made further modifications to PD149164 to improve its affinity and potency at the CCK₁ receptor and the resulting compound was called PD170292 (Bernad et al., 2000). The pharmacological profile was noted by the authors to be similar to that of JMV180 (Galas et al., 1988).

Thiozole derivatives: (SR146131)

The biological activity in vitro of the thiazole derivative, SR14613, was described by Bignon et al. in 1999 and it was shown to be a potent full agonist for intracellular calcium responses and IP₁ formation at the CCK₁ receptor, but only a partial agonist for MAPK activation and early gene expression (Bignon et al., 1999b). The in vivo effects of SR14613 investigated by the same group showed it was able to inhibit gastric emptying and reduction in gallbladder volume in mice after administration of low oral doses (Bignon et al., 1999a). SR14613 was also shown to reduce food intake in rodents as well as in one primate species (Bignon et al., 1999a).

1,5-benzodiazepine derivatives

In 1996, Aquino et al. explored a company registry for novel leads towards the development of non-peptidyl CCK₁ receptor agonists. Compounds were selected based on structural features of CCK₁ receptor agonist peptides that had been described, in particular the features present in tetrapeptide, A71623, and then were evaluated for contractibility in isolated guinea pig gallbladder. The group identified a series of 1,5-benzodiazepines as potential CCK₁ receptor agonist candidates (Aquino et al., 1996). The efficacy of the compounds in this series was shown to be modulated by variation of substituents on the N1-anilinoacetamide moiety. Further studies of the series of 1,5-benzodiazepine compounds were shown to have decreased affinity for the CCK₁ receptor compared with CCK-8, but some showed equipotent anorectic effects in rats following intraperitoneal administration (Aquino et al., 1996). Since the first reports of the 1,5-benzodiazepine series exhibiting CCK₁ receptor agonist effects, the GSK compound GI181771 progressed to phase II clinical trials for the treatment of obesity (Szewczyk and Laudeman, 2003). Two clinical trials in humans have been reported investigating the role of GI18177X as a novel oral CCK₁ receptor agonist (Castillo et al., 2004; Jordan et al., 2008). The first study carried out by Castillo et al. in 2004 looked in particular at the effects of GI18177X on gastric emptying of solids, gastric accommodation and postprandial symptoms. The study looked at the effects of a four dose regime and a placebo on gastric functions and postprandial symptoms in a group of 61 healthy men and women (for more details of the study parameters refer to Castillo et al., 2004). The results showed GI18177X treatment resulted in delays in gastric emptying of solids, increased fasting gastric volumes and an acceptable safety profile. In 2008, Jordan et al. extended the clinical studies by investigating whether GI18177X could induce weight loss in obese patients. Unfortunately, GI18177X did not reduce the body weight of obese patients and had no effects on waist circumference or cardiometabolic risk markers. The authors noted that mild tolerability-limiting gastrointestinal events, including vomiting, were observed and a small number of study subjects showed treatment-related events of cholecystitis and gallstone-related pancreatitis (Jordan et al., 2008).

While the clinical studies of the 1,5-benzodiazepine, GI18177X, have showed varied results, this class of compound has lead to interesting developments in relation to drug discovery for a CCK₁ receptor agonist. Our group identified the binding sites for active enantiomers in racemic mixtures of structurally related benzophenone derivatives of 1,5-benzodiazepine for both agonist and antagonists to the CCK₁ receptor (Hadac et al., 2006). We have also carefully examined the molecular basis of binding of other benzodiazepine ligands for this receptor using both receptor mutagenesis and classical analysis of allosterism (Gao et al., 2008). This work has clearly established the allosteric nature of the binding of such ligands within the helical bundle, in a site totally distinct from the natural peptide-binding site that resides within the extracellular loop and tail regions at the surface of the lipid bilayer.

The therapeutic potential of drugs that act on CCK₁ receptor

To this point, we have described some of the important roles of the CCK₁ receptor, both physiologically and pathophysiologically, as well as some of the key signalling pathways initiated after stimulation with the natural ligand, CCK. We have also described some of the most promising CCK₁ receptor agonists developed to date and the findings of the clinical trials with these candidate agonists. It is clear from reviewing the literature that there is a need for the development of a more selective and safer CCK₁ receptor agonist that could have key roles in the treatment of common clinical syndromes, in particular obesity. There are three potential concepts that could be explored to increase specificity in relation to the activation of the CCK₁ receptor by an agonist. These concepts are allosteric modulation, partial agonism and biased agonism.

Researchers have long been interested in cell surface receptors such as GPCRs as targets for drug development. Until recently, the focus of drug design for GPCRs has been on the development of agonists acting at the orthosteric site of a specific GPCR, but many of these have been shown to have a high incidence of tolerance and dependence. Therefore, the concept of an allosteric modulator that has been defined by the International Union of Pharmacology Committee on Receptor Nomenclature and Drug Classification in 2003 (Neubig et al., 2003) as ‘a ligand that increases or decreases the action of an agonist or antagonist by combining with a distinct site on the receptor macromolecule’ has become an area of great interest in the pharmaceutical industry with regard to GPCRs.

The development of allosteric modulators to GPCRs is of benefit as such agents may not exhibit excessive activation of a specific GPCR. As stimulation of the receptor with an allosteric modulator would be dependent on the presence of both the natural agonist ligand and the allosteric modulator simultaneously, the enhanced effect would only occur during the brief post-cibal periods when the peptide hormone, having a very short half-life, would be in the circulation. Thus, an allosteric modulator theoretically might decrease the risk of desensitization and tolerance that could be observed with a ligand acting at the orthosteric site. Secretion of the endogenous hormone CCK that occurs after ingestion of food has a major role in the digestion of food, as well as in stimulating the feeling of satiety, therefore an allosteric modulator that might enhance the effects of CCK in patients in whom it does not have strong effects, could return towards a more physiologic state. Additionally, the enhancement elicited by the allosteric modulator would be dependent on the number of receptors occupied by both the allosteric ligand and the endogenous hormone, perhaps limiting the possibility of overdose compared with an orthosteric drug (Christopoulos, 2002).

One of the more important issues that need to be raised in relation to drug design of a CCK₁ receptor agonist is its selectivity between CCK₁ and CCK₂ receptors. A desirable drug needs to have minimal or no biological activity at the CCK₂ receptor, as this receptor may mediate panic attacks and has other CNS actions (Dauge and Lena, 1998) that may be undesirable. Having a therapeutic drug that has either no stimulation of the CCK₂ receptor or is a CCK₂ receptor antagonist would be important, due to the chronic nature of obesity and due to the need for such a drug to be administered over a long term (Szewczyk and Laudeman, 2003). Another theoretical benefit of an allosteric ligand is that allosteric sites are generally less conserved than orthosteric sites in a particular receptor family, thus potentially enabling even greater specificity (Birdsall and Lazareno, 2005). Additionally, targeting of an allosteric modulator may be more effective than the orthosteric site as topographically it may be difficult for a small molecule to gain access (Jensen and Spalding, 2004). Of interest is that recently the benzodiazepine antagonist devazepide has been shown to bind to an allosteric site within the helical bundle region of the CCK₁ receptor (Hadac et al., 2006; Gao et al., 2008), thus identifying a potential site for further development of an allosteric modulator that might act as an enhancer of CCK stimulation of the CCK₁ receptor and might increase specificity.

One of the first descriptions of ‘partial agonism’ was made by Ariens and de Groot in 1954, and since then a partial agonist has been defined by the IUPHAR Compendium of Receptor Characterization and Classification (1998) as an agonist that ‘cannot elicit as large an effect … as can a full agonist acting through the same receptors’. Partial agonists have been shown to have several benefits, one of these being the reduced propensity to elicit side effects or to desensitize their target receptors (Clark et al., 1999). This adds to their therapeutic potential. Additionally, partial agonist ligands acting at the orthosteric site of a receptor can also reduce the effects of the natural full agonist by competing for its binding (Ohlsen and Pilowsky, 2005). In regard to the CCK₁ receptor, the partial agonists OPE and JMV180 have been shown to stimulate a subset of the biological actions of CCK. OPE and JMV180 are partial agonists to the CCK₁ receptor and do not exhibit supramaximal stimulation of amylase secretion (Gaisano et al., 1989; Sato et al., 1989; Gaisano and Miller, 1992), and they stimulate calcium oscillations and IP₃ levels that are similar to those seen with low concentrations of the natural agonist, CCK (Galas et al., 1988; Gaisano and Miller, 1992; Gaisano et al., 1994). It is, therefore, conceivable that an allosteric partial agonist exhibiting similar functional characteristics to those of OPE and JMV180 could be of substantial therapeutic benefit.

Until recently, it has been thought that an agonist was a ligand that would bind to its receptor in an inactive state and would induce a conformational change promoting the active state of the receptor, thus initiating second messenger signalling that might ultimately be limited by desensitization. It was also thought that the binding of an agonist stimulated all receptor functions to an approximately equal extent due to the conformational change from an inactive to the active state (Benovic et al., 1988). Therefore, in previous years, the determination of a ligand's efficacy was generally determined by its ability to stimulate one second messenger pathway. Over the past decade, however, evidence has been produced showing that the original explanation of agonist stimulation is more complex than had been believed. The terms ‘biased agonists’ or ‘functionally selective agonists’ were first described in relation to the ability of a GPCR to selectively interact with various G protein signalling pathways (Kenakin, 1995; Lawler et al., 1999). These terms have now been extended to encompass other signalling pathways such as the G protein-independent, arrestin-mediated signalling which has been extensively described in relation in relation to GPCRs (Lefkowitz, 2004; Terrillon and Bouvier, 2004; Lefkowitz and Shenoy, 2005; Luttrell, 2005; Lefkowitz et al., 2006). The concept of biased agonism may be extrapolated to the development of a CCK₁ receptor agonist that is biased towards the signalling pathways responsible for satiety, but that does not simulate pathways implicated in trophic effects on the pancreas that also may induce pancreatitis and other side effects that have hindered the development of a therapeutically useful CCK₁ receptor agonist. In relation to CCK₁ receptor, biased agonist pathways that may be selected for are those of the G_q second messenger pathway shown to be responsible for the secretion of amylase, in particular having a bias towards agonists that do not exhibit supramaximal stimulation of amylase and the signalling pathways that are associated with it.

In summary, there have been substantial increases in our knowledge of the CCK₁ receptor, its physiology and its potential role in the treatment of obesity. Additionally, several of the potentially important signalling pathways that are activated in response to both physiological and pathological levels of CCK have been described. To date, several peptide and non-peptidyl CCK₁ receptor agonists have been developed with only a few progressing to clinical trials, and these having had properties that precluded their further development. Therefore, future development of useful CCK₁ receptor-active drugs may include allosteric modulators, biased agonists and partial agonists.

Glossary

Abbreviations:

AA

arachidonic acid

Ac

acetylated

cADPr

cyclic ADP-ribose

CCK

cholecystokinin

CCK₁ receptor

type 1 cholecystokinin receptor

CCK₂ receptor

cholecystokinin type 2 receptor

EGFR

epidermal growth factor receptor

eIF4E-BP

eukaryotic initiation factor 4E-binding protein

ERK

extracellular-regulated kinase

GPCR

guanine nucleotide-binding protein-coupled receptor

Grb2

growth factor receptor-bound protein 2

IBS

irritable bowel syndrome

IP₃

inositol 3, 4, 5-triphosphate

JNK

c-Jun amino-terminal kinase

MAPK

mitogen-activated protein kinase

MEK

MAP kinase or ERK kinase

mTOR

mammalian Target of Rapamycin

NAADP

nicotinic acid adenine dinucleotide phosphate

NO

nitric oxide

p90^RSK

90-kDa ribosomal S6 protein kinase

PI3K

class I phosphatidylinositol 3-kinase

PIP₂

phosphatidylinositol 4,5-bisphosphate

PKC

protein kinase C

PLA₂

phospholipase A₂

PLC

phospholipase C

SAR

drug structure-activity relationship

Shc

Src homology/collagen related

SOS

Drosophila homolog, Son of Sevenless

Conflict of interest

No specific conflicts recognized.