Serotonin 5-HT₇ receptor agents: structure-activity relationships and potential therapeutic applications in central nervous system disorders

Abstract

Since its discovery in the 1940s in serum, the mammalian intestinal mucosa, and in the central nervous system, serotonin (5-HT) has been shown to be involved in virtually all cognitive and behavioral human functions, and alterations in its neurochemistry have been implicated in the etiology of a plethora of neuropsychiatric disorders. The cloning of 5-HT receptor subtypes has been of importance in enabling them to be classified as specific protein molecules encoded by specific genes. The 5-HT₇ receptor is the most recently classified member of the serotonin receptor family. Since its identification, it has been the subject of intense research efforts driven by its presence in functionally relevant regions of the brain. The availability of some selective antagonists and agonists, in combination with genetically modified mice lacking the 5-HT₇ receptor, has allowed for a better understanding of the pathophysiological role of this receptor. This paper reviews data on localization and pharmacological properties of the 5-HT₇ receptor, and summarizes the results of structure-activity relationship studies aimed at the discovery of selective 5-HT₇ receptor ligands. Additionally, an overview of the potential therapeutic applications of 5-HT₇ receptor agonists and antagonists in central nervous system disorders is presented.

1. Introduction

Serotonin (5-hydroxytryptamine; 5-HT) is a major neurotransmitter involved in a large number of processes in the central nervous system, including the regulation of feeding behavior, aggression, mood, perception, pain, and anxiety (Bradley et al., 1986). In the periphery, 5-HT is involved in the regulation of the autonomic nervous sytem, cardiovascular function, vascular and nonvascular smooth muscle contraction, platelet aggregation, uterine smooth muscle growth, and gastrointestinal functioning (Hoyer et al., 1994; Ramage, 2006; Ramage and Villalón, 2008). To mediate these functions, a family of receptors, divided into seven main families comprising at least fourteen different receptors, has evolved (Hoyer et al., 1994). The 5-HT₇ receptor was the last addition to the large serotonin subfamily of G-protein coupled receptors. Considering the number of papers published in the serotonin receptor subtypes field, the 5-HT₇ receptor is not the most studied, being surpassed by 5-HT_1A, 5-HT_2A, 5-HT₃, and 5-HT₄ receptors. This might be explained by the lack for a long time period of selective 5-HT₇ agents that would help researchers to shed light on the pathophysiological role of this receptor. Thus, since antagonists and, to some extent, agonists became available as well as 5-HT₇ receptor-knockout mice, our understanding of the role of the 5-HT₇ receptor in normal and pathological processes has improved.

After a general introduction to the 5-HT₇ receptor (molecular structure, functional expression, distribution), this review will illustrate the medicinal chemistry studies aimed to the identification of selective 5-HT₇ ligands. The structure-activity relationships for each class of compounds will be delineated and the pharmacophore models proposed for agonists and antagonists will be also discussed. A second part of this review will give details on studies that used the identified compounds as well as other techniques to gain insight into the involvement of the 5-HT₇ receptor in normal and pathological processes of the central nervous system

1.1. Characteristics of the 5-HT₇ receptor

1.1.1. Structure

The serotonin (5-HT) 5-HT₇ receptor belongs to the large family of G protein-coupled receptors (GPCR). In contrast to, for example, the 5-HT_1A or 5-HT_2A receptors that were identified as distinct receptors for 5-HT on the basis of pharmacological and physiological differences, the 5-HT₇ receptor was discovered by targeted analysis of cDNA libraries based on conserved sequences of known receptors. The primary amino acid sequence of the 5-HT₇ receptor of human, rat, mouse, guinea pig, and pig has been elucidated. Comparison of the amino acid sequence of the transmembrane regions of 5-HT₇ receptor from various species indicated the highest homology with the 5-HT_dro1 receptor. Lower homologies in the transmembrane domains were found with other 5-HT receptors (Table 1).

Table 1.

Percentages of amino acid homologies in the transmembrane domains between the 5-HT₇ receptor and other cloned 5-HT receptors.

Receptor	Humana	Ratb	Guinea Pigc	Moused
5-HTdro1	57	56	48	42
5-HT1A		45	40
5-HT1B		50	43	36
5-HT1D		47	43	36
5-HT1E		45	42	36
5-HT1F		48	42
5-HT2A		39	35
5-HT2C			34
5-HT5		46		32
5-HT6		44

^aBard et al. (1993)

^bShen et al. (1993)

^cTsou et al. (1994)

^dPlassat et al. (1993)

The 5-HT₇ receptor has been cloned from pig by Bhalla and coworkers (2002). The nucleotide sequence of porcine 5-HT₇ receptor cDNA encoded a protein of 447 amino acids, which showed high homology (92–96%) with the sequence of the 5-HT₇ receptor from the other species.

More recently the 5-HT₇ receptor has been described in several additional species, including the nematode Caenorhabditis elegans (Hobson et al., 2003), the yellow fever mosquito Aedes aegypti (Lee and Pietrantonio, 2003), the honey bee Apis mellifera (Schlenstedt et al., 2006) and the fresh water snail Helisoma trivolvis (Mapara et al., 2008).

1.1.2. Splice Variants

Three splice variants of the 5-HT₇ receptor have been identified in humans (designated 5-HT_7a, 5-HT_7b, and 5-HT_7d), which encode proteins that differ in their carboxy terminals (Heidmann et al., 1997). The human 5-HT_7a is the full length receptor (445 amino acids), while the human 5-HT_7b is truncated at amino acid 432 due to an alternative splice donor site. The human 5-HT_7d is a distinct isoform of the receptor: the retention of an exon cassette in the region encoding the carboxyl terminal results a 479-amino acid receptor with a C-terminus markedly different from the human 5-HT_7a. A 5-HT_7c splice variant is detectable in rat tissue but is not expressed in humans. Conversely, rats do not express a splice variant homologous to the human 5-HT_7d, as the rat 5-HT₇ gene lacks the exon necessary to encode this isoform. Drug binding affinities are similar across the three human splice variants (Krobert et al., 2001). All the three human splice variants constitutively activate adenylyl cyclase but inverse agonist efficacies appear to differ between the splice variants (Krobert and Levy, 2002). Additionally, there appears to be a difference in agonist mediated internalization between the 5-HT_7d variant and the 5-HT_7a and 5-HT_7b variants (Guthrie et al., 2005). All three splice variants have been associated with a protein known as PLAC-24 that might regulate transportation and stabilization of newly synthesized receptor protein (De Martelaere et al., 2007). Also the porcine 5-HT₇ receptor gene contains two introns, but the possibility of alternative splicing has not been investigated (Bhalla et al., 2002).

In humans a pseudogene for the 5-HT₇ receptor has been identified (Lassig et al., 1999; Olsen and Schechter, 1999). It is transcribed, but an in-frame stop codon is present in the fifth trans-membrane region (Olsen and Schechter, 1999). The pseudogene has also been described in rhesus monkeys, but is not found in rats or mice (Olsen and Schechter 1999).

Screening of patients with psychiatric disorders has revealed three base pair exchanges in the 5-HT₇ receptor gene (Erdmann et al., 1996). One of them is silent, but the other two lead to amino acid exchanges. The mutations were equally prevalent (1%) in both the patient and control groups suggesting that they were not relevant for pathogenesis. Nevertheless, the Thr92Lys and Pro279Leu variants are of pharmacological interest. It has been demonstrated that the Thr92Lys exchange leads to decreased agonist, but not antagonist, binding (Brüss et al., 2005). The Pro279Leu variant has been associated with a virtually abolished signal transduction, most likely due to an altered coupling between the third intracellular loop and the G protein (Kiel et al., 2003).

1.1.3. Signaling

5-HT₇ receptors are coupled to stimulation of adenylyl cyclase. The EC₅₀ values for the reference agonists 5-CT, 5-HT, and 8-OH-DPAT are listed in Table 2. As can be noted, the order of agonist potency in assays of adenylyl cyclase activity varies between species. Also, 8-OH-DPAT behaves as full or partial agonist depending on species.

Table 2.

Stimulation of cAMP accumulation by cloned 5-HT₇ receptor from various species

Agonist	Species	Cell Line	EC50 (nM)	Antagonists
5-HTa	Rat	HEK-293	63	methiothepin, clozapine
5-HTb	Rat	CHO	1.2	mianserin
5-HTc	Rat	HeLa	13	--
5-HTd	Human	Cos-7	992	methiothepin
5-HTe	Mouse	Cos-7	45	methiothepin, methysergide, ergotamine
5-CTc	Rat	HeLa	156	mesulergine, ritanserin, NAN-190
5-CTf	Guinea Pig	CHO-K1	3.2	methiothepin, spiperone
5-CTe	Mouse	Cos-7	5.0	methiothepin, methysergide, mesulergine

^aShen et al. (1993)

^bRuat et al. (1993)

^cLovenberg et al. (1993)

^dBard et al. (1993)

^ePlassat et al. (1993)

^fTsou et al. (1994)

As mentioned above, 5-HT_7a, 5-HT_7b, and 5-HT_7c splice variants constitutively activate adenylyl cyclase. Therefore, 5-HT₇ receptor antagonists display either full or partial inverse agonist activity. In an early study, Thomas et al. (1998) reported potencies of various compounds in antagonizing 5-CT-induced stimulation of adenylyl cyclase activity in comparison with their inverse agonist behavior on 5-HT_7a (Table 3). A few years later, Krobert and Levy (2002) published a detailed report of the inverse agonist properties of several antagonists on all three human 5-HT₇ receptor splice variants (Table 4). No differences in inverse agonist potencies (pIC₅₀) were observed between the splice variants. The rank order of potencies was in agreement and highly correlated with antagonist potencies (pK_B) determined by antagonism of 5-HT-stimulated adenylyl cyclase activity (methiothepin > metergoline > mesulergine ≥ clozapine ≥ spiperone ≥ ritanserin > methysergide > ketanserin). It has been demonstrated that the constitutive activity of the 5-HT₇ receptor is regulated by palmitoylation of its C-terminal domain (Kvachnina et al., 2009). Interestingly, it has been found that the inverse agonist properties of 5-HT₇ receptor ligands can lead to heterologous desensitization of other G protein-coupled receptor families (Krobert et al., 2006). In addition to being coupled to G_s, the 5-HT₇ receptor has also been shown to be coupled to the G₁₂-protein which activates small GTPases of the Rho family in turn leading to enhanced neurite outgrowth (Kvachnina et al., 2005). In cell lines, it has been demonstrated that the 5-HT₇ receptor not only activates the adenylyl cyclase AC5 normally linked to G_s, but also cyclases AC1 and AC8 that are activated by intracellular calcium (Baker et al., 1998). Coexpression of the 5-HT₇ receptor, AC1, and AC8 is present in the hippocampus, and might thus be important for hippocampal function, which is discussed further below. At least when expressed in HEK-293 cells, the receptor is tightly associated with the G protein, regardless of agonist binding (Bruheim et al., 2003). This suggests the existence of a complex between inactive receptor and the G-protein. The 5-HT₇ receptor has been found to activate the extracellular signal-related kinase (ERK) (Errico et al., 2001) through a mechanism dependent on a Ras G protein (Norum et al., 2003). This activation of ERK might be also of importance for hippocampal function and depression. Although cAMP activation is normally believed to be mediated through protein kinase A (PKA), it has been demonstrated that the 5-HT₇ receptor can stimulate ERK through a PKA-independent pathway, possibly by using the cAMP-guanine nucleotide exchange factor Epac (Lin et al., 2003).

Table 3.

Drug effects on basal and 5-CT-stimulated adenylyl cyclase activity^a

Antagonist	pKBb	pIC50c	% Inhibition of Basald
Risperidone	8.3 (0.1)	7.6	86
Methiothepin	8.1 (1)	7.3	82
Mesulergine	7.5 (1)	7.1	49
Clozapine	7.2 (1)	6.6	79
Olanzepine	6.3 (10)	6.3	88
Ketanserin	6.1 (10)	6.0	77

^aThomas et al. (1998)

^bpK_B (−log antagonist equilibrium dissociation constant) calculated from the antagonism of 5-CT-stimulated adenylyl cyclase activity. The concentration of drug tested (µM) is shown in parentheses

^cpIC₅₀ (−log IC₅₀) for inhibition of basal adenylyl cyclase activity

^d% inhibition of basal adenylyl cyclase activity for antagonists, measured in the absence of added agonist

Table 4.

Inverse agonist potency (pIC₅₀) and efficacy at human 5-HT₇ receptor splice variants determined from inhibition of constitutive adenylyl cyclase activity.^a

	pIC50b			Efficacy (% of methiothepin)
Antagonist	5-HT7a	5-HT7b	5-HT7c	5-HT7a	5-HT7b	5-HT7c
Methiothepin	9.04	9.07	9.18	100	1000	100
Metergoline	7.96	8.22	8.49	26	36	20
Mesulergine	NAc	7.73	NA	0	47	−10
Clozapine	7.47	7.69	7.68	96	91	96
Spiperone	7.75	7.73	7.79	103	101	97
Ritanserin	7.62	7.37	7.64	110	100	93
Methysergide	7.36	7.35	7.21	18	19	−2
Ketanserin	6.59	6.42	6.61	106	95	93

^aKrobert & Levy (2002)

^bAdenylyl cyclase activity measured in the presence of increasing concentrations of antagonist in membranes of HEK-293 cells stably expressing either human 5-HT_7a, 5-HT_7b or 5-HT_7d receptors, and potency and efficacy of antagonist inhibition of constitutive adenylyl cyclase activity was determined. Inverse agonist efficacy is reported as a percentage of the inhibition of basal adenylyl cyclase activity produced by 10 µM methiothepin

^cNot applicable

1.1.4. Pharmacological properties

Several compounds known to be ligands for other 5-HT receptor subtypes were found to also have high affinity for recombinant 5-HT₇ receptors from various species (human, rat, mouse, guinea-pig), as shown in Table 5. All four species homologues of the 5-HT₇ receptor have high affinity for 5-CT (pK_i 9.0–9.4) and comparable affinities for the endogenous ligand 5-HT (pK_i 8.1–9.0). One of the most important revelations was that 8-OH-DPAT had relatively high affinity (Table 5) since it prior to 1993 was considered a selective agonist for the 5-HT_1A receptor. When comparing pK_i values obtained from human cloned receptor with those from rat cloned receptor some remarkable differences in affinity can be noted. In most cases, pK_i values from rat cloned receptor are higher than those from human cloned receptors. Although these differences are not larger than 1 log unit, as in the case of methiothepin, they point to possible discrepancies between binding data obtained from human and rat cloned receptors. In competition assays [³H]5-HT, [³H]5-CT, and [³H]LSD have been used as radioligands to determine binding affinities as shown in Table 6.

Table 5.

pK_i values of selected ligands at recombinat 5-HT₇ receptor from human, rat, mouse, and guinea-pig expressed in Cos-7 cells using [³H]5-HT as radioligand

Ligand	Humana	Ratb	Mousec	Guinea pigd
5-CT	9.0	9.8	9.0	9.4
Methiothepin	8.4	9.4	8.2	8.0
5-MeOT	8.3	9.2	8.2	9.0
Metergoline	8.2	8.2	7.5
5-HT	8.1	8.8	8.3	9.0
Mesulergine	7.7	7.7	7.6
2-Br-LSD	7.5		8.0
Methysergide	7.1	7.9	7.9	7.4
Spiperone	7.0		7.2	7.0
Cyproheptadine	6.9	7.5
Tryptamine	6.8	7.8
8-OH-DPAT	6.3	7.5	6.6	7.3
Sumatriptan	6.0	6.6	5.7	6.6
Ketanserin	5.9	6.6	6.4

^aBard et al. (1993)

^bShen et al. (1993)

^cPlassat et al. (1993)

^dTsou et al. (1994)

Table 6.

Ki values [nM] of selected ligands at human cloned 5-HT₇ receptor obtained from competition assays with different radioligands.

Ligand	[3H]5-HT	[3H]5-CT	[3H]LSD
5-CT	0.93a	0.79b
5-HT	8.1a	6.3b
8-OH-DPAT	466a	251b
Clozapine	48c	63.1b	17.9d
Ketanserin	1134a	794b
Mesulergine	18a	31.6b
Risperidone	4.3c		6.6d
Amisulpride		135.5e	11.5e

^aBard et al. (1993)

^bThomas et al. (1998)

^cFernández et al. (2005)

^dKroeze et al. (2003)

^eAbbas et al. (2009)

1.2. Expression and distribution of the 5-HT₇ receptor

1.2.1. Messenger RNA

The distribution of mRNA encoding the 5-HT₇ receptor protein has been studied by several groups using a variety of techniques. The results of these studies are summarized in Tables 6 and 8. In all species studies, 5-HT₇ receptor mRNA has been identified in the central nervous system. Particularly high levels have been detected in the hypothalamus (notably within the suprachiasmatic nucleus), thalamus and hippocampus. In peripheral tissues, 5-HT₇ receptor mRNA has been described in the ileum, spleen, endocrine glands, and arteries. In blood vessels and the gastrointestinal tract the expression has generally been localized to smooth muscle cells. Some discrepancies have been noted, as in the study by Plassat et al., (1993) where Northern blot analysis was unable to detect 5-HT₇ mRNA in mouse brain, whereas RT PCR demonstrated high levels in the brainstem. Thus, differences observed, both between species and even within the same species, may depend on the sensitivity of the technique used. Also, low detection of 5-HT₇ receptor mRNA may arise when it is restricted to a small population of cells of a given organ. The distribution of the mRNA encoding the various isoforms of the receptor in the rat has been determined by using either RT-PCR or in situ hybridization (Heidmann et al., 1998). The authors ruled out large tissue specific differences in splicing of the 5-HT₇ mRNA.

Table 8.

Localization and abundance of 5-HT₇ receptor mRNA in guinea-pig, human, mouse, and pig

Species	Technique	Localization (Relative Abundance)
Guinea piga	Northern Blot	parietal cortex (+++), hippocampus (++), frontal cortex (++), cerebellum (+), ileum (+), spleen (+)
Guinea piga	In situ hybridization	hippocampus (+++), periventricular thalamus (+++), superficial cortex (+++), cerebellar granule cells (+++)
Guinea pigb	In situ hybridization	medial thalamic nucleus (+++), hippocampal formation (+++), superficial layer cortex (++), medial geniculate nucleus (++), amygdala (++), hypothalamus (++), midbrain (+), hindbrain (+)
Guinea pigc	In situ hybridization	outer layer of the cortex, thalamus, hippocampus
Guinea pigd	Northern Blotb	thalamus (+++), brainstem (+++), hypothalamus (+++), substantia nigra (+++), olfactory bulb (+++), olfactory tubercle (+++) (not detected in peripheral organs)
Humane	RT-PCR	artery smooth cells (+++), pulmonay artery smooth cells (++) (not detected in coronary artery, pulmonary artery, aortic endothelial)
Humanf	RT-PCR	dorsal root ganglia
Humang	RT-PCR	granulosa-lutein cells
Humang	Northern Blot	granulosa-lutein cells
Humanh	RT-PCR	brain (+++), kidney (+), liver (+), pancreas (+), spleen (+), coronary artery (++), stomach (++), descending colon (++), ileum (++)
Mousei	Northern Blot	not detected in brain, heart, kidney, lung, liver
Mousei	RT-PCR	forebrain (+++), brain stem (+++), cerebellum (++), colliculi (++), intestine (+), heart (+), not detected in spleen, kidney, lung, liver
Pige	RT-PCR	pulmonary artery (++), coronary artery (++), cerebral artery (++), cerebral vein (++)
Pigj	RT-PCR	myometrium
Pigk	RT-PCR	brain cortex, trigeminal ganglion, cerebellum, pulmonary artery, coronary artery, superior vena cava, saphenous vein, not detected in heart

^aTsou et al. (1994);

^bTo et al. (1994);

^cMengod et al. (1996);

^dRuat et al. (1993);

^eUlmer et al. (1995);

^fPierce et al. (1997);

^gGraveleau et al. (2000);

^hBard et al. (1993);

ⁱPlassat et al. (1993);

^jKitazawa et al. (2000);

^kBhalla et al. (2002)

1.2.2. Protein

1.2.2.1. Radioligand binding studies

An early study attempted to define 5-HT₇ receptor binding sites in rat hypothalamic membranes by labeling with [³H]5-HT in the presence of 100 nM pindolol. Although the level of specific binding detected was low (it represented approximately 70% of the total binding), the pharmacology of the identified binding sites correlated well with that of rat recombinant 5-HT₇ receptors (Sleight et al., 1995). These results were challenged by a subsequent study (Gobbi et al., 1996) which found that 100 nM pindolol does not completely mask 5-HT_1A and 5-HT_1B receptors and that the population of pindolol-insensitive [³H]5-HT receptors in rat hypothalamus appeared to be heterogeneous. Also [³H]5-CT failed to define a homogeneous population of 5-HT binding sites in rat hypothalamus homogenates despite the use of a ‘cocktail’ of drugs ((±)-pindolol, sumatriptan, DOI) (Stowe & Barnes 1998).

To et al. (1995) reported that a single population of receptors isolated from guinea pig cerebral cortex membranes could be labeled by [³H]5-CT in the presence of cyanopindolol and sumatriptan as masking agents. The pharmacology of these binding sites was well correlated to that of the cloned guinea pig 5-HT₇ receptor. Because affinities were consistently lower in the tissue binding assay, the authors hypothesized that cyanopindolol and sumatriptan might occupy at least in part the 5-HT₇ receptors.

[³H]Mesulergine has been used to label 5-HT₇ sites in guinea-pig ileal longitudinal muscle. Also in this case, due to the lack of selectivity of mesulergine over a range of receptors, several masking drugs (cinanserin, RS 102221, raclopride, prazosin, and yohimbine) were used. It was also noted that no binding could be detected in the rat jejunum under the same experimental conditions (Hemedah et al., 1999).

1.2.2.2. Autoradiographic studies

Initial attempts to visualize the distribution of 5-HT₇ receptor binding sites used [³H]5-CT in the presence of various masking agents (To et al., 1995; Mengod et al., 1996; Waeber & Moskowitz, 1995; Gustafson et al., 1996). In guinea pig and rat brain, the distribution of the 5-HT₇ sites was found to be largely consistent with that reported for 5-HT₇ receptor mRNA. The highest densities were in the medial thalamic nuclei and related limbic and cortical regions. However, as densities were very low, the authors concluded that the masking compounds added in these autoradiographic experiments might have sufficient 5-HT₇ affinity to occupy part of the 5-HT₇ binding sites at the concentration used, thus preventing full visualization of the receptor.

In an effort to circumvent these limitations Bonaventure et al. (2002) performed an autoradiographic study using 5-HT_1A receptor knockout and 5-HT_1A/B receptor double knockout mice. They found that, despite its high affinity for the 5-HT₇ receptor in tissue homogenates, [³H]5-CT was not a good tracer for measuring 5-HT₇ receptor binding sites autoradiographically. Instead they concluded that the lower affinity ligand [³H]8-OH-DPAT was much more suited for autoradiographic studies of 5-HT₇ receptor binding sites. The anatomical distribution of the [³H]8-OH-DPAT binding sites observed in knockout mice matched the distribution of 5-HT₇ receptor mRNA and 5-HT₇ receptor immunoreactivity reported in the literature. Within the hippocampal formation, strong labeling was found in the CA3 region, whereas the densities in CA1 were low. 5-HT₇ receptor binding sites were also found within the dorsal raphe. High densities of 5-HT₇ receptor binding sites were observed throughout the hypothalamus (including the suprachiasmatic nucleus).

By using [³H]8-OH-DPAT in combination with the antagonists pindolol or ritanserin at concentrations selectively blocking either 5-HT_1A or 5-HT₇ receptors respectively, Duncan et al. (1999) were able to demonstrate that both of these receptor subtypes are present in important components of the circadian timing system (intergeniculate leaflet, lateral geniculate nucleus, suprachiasmatic nucleus) and dorsal and median raphe nucleus of the Syrian hamster.

Varnäs et al. (2004) have used the selective and high-affinity radioligand [³H]SB-269970 to visualize the distribution of 5-HT₇ receptors in human whole hemisphere brain sections. The data were in good agreement with the autoradiographic studies by Gustafson et al. (1996) and To et al. (1995). [³H]SB-269970 binding was mainly found in the thalamus, hypothalamus and hippocampal formation of the human brain. Some important differences were found compared to results obtained with [³H]mesulergine (Martin-Cora & Pazos, 2004). The authors found that using the selective radioligand [³H]SB-269970 very low densities were found in the caudate nucleus and the putamen, contrary to the high levels found in the investigation with the non-selective radioligand [³H]mesulergine. It was suggested that these differences could be due to different brain levels studied, or to the possibility that the concentration of non-labeled masking compounds used in the [³H]mesulergine autoradiographic study was not sufficiently high to abolish binding to other receptor subtypes. It should also be noted that SB-269970 has affinity for 5-HT_5A receptors (Varnäs et al., 2004).

1.2.2.3. Immunocytochemical studies

Immunocytochemistry has been used to localize the distribution of 5-HT₇ receptors in rat forebrain. 5-HT₇ receptors were detected within the cerebral cortex, hippocampal formation, tenia tecta, thalamus, and hypothalamus. These results were in accordance with the reported localization of the 5-HT₇ receptors mRNA. At a microscopic level, both cell bodies and proximal fibers were strongly stained in the suprachiasmatic nucleus, suggesting a somatodentric subcellular distribution (Neumaier et al., 2001). Similar results were obtained by Belenky and Pickard (2001). By using electron microscopic immunocytochemical procedures, they determined the presence of 5-HT₇ receptors in both pre- and postsynaptic GABA, vasoactive intestinal polypeptide, and vasopressin processes in the suprachiasmatic nucleus in mouse. An immunohistochemical study has demonstrated the presence of 5-HT₇ receptors in Purkinje cells of rat cerebellum (Geurts et al., 2002).

An immunocytochemical study of 5-HT₇ receptor distribution at the lumbar level has been performed by Doly et al. (2005). 5-HT₇ immunolabeling was localized mainly in the two superficial laminae of the dorsal horn and in small and medium-sized dorsal root ganglion cells, which is consistent with a predominant role in nociception. In addition, moderate labeling was found in the lumbar dorsolateral nucleus (Onuf’s nucleus), suggesting involvement in the control of pelvic floor muscles (Doly et al., 2005). Electron microscopic examination of the dorsal horn revealed a postsynaptic localization on peptidergic cell bodies in laminae I–III and in numerous dendrites, a presynaptic localization on unmyelinated and thin myelinated peptidergic fibers (Doly et al., 2005).

2. Medicinal chemistry of 5-HT₇ receptor agents

2.1. Nonselective 5-HT₇ receptor agents

As already discussed, immediately after the cloning of the 5-HT₇ receptor, a number of high-affinity nonselective agents were identified (Tables 3–6). From the analysis of their chemical structures (Figure 1) it is apparent that the compounds belong to different classes (tryptamines, ergolines, tricyclic neuroleptics, and other compounds). Because all of these compounds display multi-receptor affinity, none of them has been used as lead compound in the discovery of selective 5-HT₇ agents. Thus, no structure-activity relationship studies for those chemical classes have been reported.

Figure 1.

Chemical structures of nonselective 5-HT₇ agents

This section will focus on studies that have allowed elucidation of stucture-activity relationships for the chemical classes of compounds that bind at 5-HT₇ receptors. Comprehensive overviews on patent literature dealing with 5-HT₇ receptor agents can be found elsewhere (Leopoldo 2004, Pittalà et al. 2008, Leopoldo et al. 2010).

2.2. “Long-chain” arylpiperazine derivatives

1-Aryl-4-alkylpiperazines, bearing in the terminal position of the alkyl chain a bulky substituent (the so-called “long-chain” arylpiperazines) are a class of compounds extensively studied for the identification of 5-HT_1A ligands. Due to the structural similarity between the 5-HT₇ and 5-HT_1A receptors, several research groups have modified “long-chain” arylpiperazine templates, in order to identify selective 5-HT₇ receptor ligands.

Researchers at Meiji Seika Kaisha Ltd. have reported several studies on the development of selective 5-HT₇ antagonists which culminated in the identification of the selective antagonists DR-4004 and DR-4365 (see below). From the screening of a corporate compound library against human cloned 5-HT₇ receptors, Kikuchi et al. (1999) identified 1-arylpiperazine derivatives 1–3 (Table 9). The basic nitrogen-containing substructure was structurally modified because this part of the molecule was responsible for high 5-HT₇ affinity and selectivity over other receptor systems (Kikuchi et al. 2002a). A few simple modifications on the framework of compounds 1–3 gave some key information on structure-activity relationships: optimization of the alkyl chain length indicated that a butyl linker (compound 3) was preferred; replacement of the phenyl with cyclohexyl removed all affinity (3 vs 4); affinities of isomers bearing a OCH₃ group on the phenyl ring (5–7) were ranked 3 > 2 > 4, being the 2-OCH₃ substituted compound 5 the most selective over 5-HT_2A receptor. On such basis, a variety of 2-substituted derivatives with a butyl spacer were evaluated (compounds 8–14). The nature of the substituent had significant effect on affinity with pK_i values ranging from 7.13 to 8.82. The most potent compounds were the cyano (9) and acetyl (11) derivatives, which showed pK_i of 8.42 and 8.10, respectively.

Table 9.

Binding affinities at 5-HT₇ and 5-HT_2A receptors of “long-chain” arylpiperazine derivatives with a tetrahydrobenzindole nucleus in the terminal position.

			pKi
Compound	Ar	n	5-HT7a	5-HT2Ab
1c	Ph	2	6.99	8.27
2c	Ph	3	8.29	7.79
3c	Ph	4	8.48	7.37
4c	cyclohexyl	4	<6	<6
5c	2-MeO-Ph	4	8.29	6.95
6d	3-MeO-Ph	4	8.63	7.19
7d	4-MeO-Ph	4	7.76	6.69
8d	2-Cl-Ph	4	7.91	7.01
9c	2-CN-Ph	4	8.42	6.98
10d	2-CONH2-Ph	4	7.76	6.05
11d	2-COCH3-Ph	4	8.10	6.45
12d	2-CF3-Ph	4	7.13	5.86
13d	2-NO2-Ph	4	7.62	7.34
14d	2-CH3-Ph	4	7.98	7.35
15d	2,6-diCH3-Ph	4	6.83	5.85

^aDetermined at human 5-HT₇ receptors in COS-7 cells using [³H]5-CT.

^bDetermined at 5-HT_2A receptors in rat cerebral cortex membranes using [³H]ketanserin.

^cKikuchi et al. (1999)

^dKikuchi et al. (2002a)

Replacement of the 1-phenylpiperazine group of 3 was accomplished with various basic moieties (Table 10). 4-Phenylpiperidine, 1,2,3,4-tetrahydroisoquinoline, and 4-tetrahydropyridine appeared good substitutes which delivered compounds 16, 17, and 18 that showed affinities comparable to 3. Isosteric replacement on the aromatic ring of tetrahydroisoquinoline 18 was detrimental for 5-HT₇ affinity (compounds 19–22) (Kikuchi et al., 2002b). It should be noted that the 4-tetrahydropyridine derivative 17 (DR-4004) was the most potent compound identified by this group. The compound 17 behaved as an antagonist because it inhibited 5-HT-induced stimulation if cAMP accumulation. This antagonist was also found to be at least 50-fold selective over serotonergic 5-HT_1A, 5-HT₂, 5-HT₄, 5-HT₆ and dopamine D₂ receptors. The structural modification of this class of compounds evolved further toward the identification of a basic moiety alternative to 1-arylpiperazine (Kikuchi et al., 2002a). From the comparison of the affinity values of phenyl and 2,6-diCH₃-phenyl derivatives 3 and 15 (Table 9), the authors hypothesized that the difference in affinity could be due to the conformational difference generated by rotation around the bond between the phenyl ring and the piperazine ring. To rationalize this hypothesis, the energy minimization of 1-phenylpiperazine and 1-(2,6-dimethylphenyl)piperazine was carried out, then lowest energy conformations of the compounds were superimposed. It was found that 1-(2,6-dimethylphenyl)piperazine showed no planar conformation. The authors hypothesized that such a difference could account for the difference in affinity of compounds 3 and 15. On such basis, the basic moiety of a tetrahydropyridoindole nucleus was tested to replace the arylpiperazine (Table 11). Experimental data confirmed the hypothesis because the phenyl derivative 3 was nearly equipotent to the corresponding tetrahydropyridoindole 23 (pK_i 8.48 vs 8.24). The authors also evaluated various substituents on the nitrogen of the indole ring and found that various substituents were capable of increasing affinity. In particular, compound 30 (DR-4365) had high selectivity over 5-HT_1A, 5-HT_1B, 5-HT_2C, 5-HT₂, 5-HT₃, 5-HT₄, 5-HT₆ receptors and behaved as an antagonist at the 5-HT₇ receptor.

Table 10.

Structure-activity relationships of tetrahydrobenzindole derivatives

		pKi
Compound	X	5-HT7a	5-HT2Ab
16c		8.40	6.84
17 (DR4004)d		8.67	7.01
18e		8.35	6.21
19e		8.19	6.08
20e		7.80	<6
21e		6.25	<6
22e		6.09	<6

^aDetermined at human 5-HT₇ receptors in COS-7 cells using [³H]5-CT.

^bDetermined at 5-HT_2A receptors in rat cerebral cortex membranes using [³H]ketanserin.

^cKikuchi et al. (1999)

^dKikuchi et al. (2002a)

^eKikuchi et al. (2002b)

Table 11.

Binding affinities at 5-HT₇ and 5-HT_2A receptors of 2a-[4-(tetrahydropyridoindol-2-yl)butyl]tetrahydrobenzindole derivatives^a

		pKi
Compound	R	5-HT7b	5-HT2Ac
23	H-	8.24	6.62
24	CH3-	7.72	6.98
25	CH3OCH2-	7.77	6.24
26	CH3CO-	8.19	6.27
27	CH2=CHCH2-	8.22	6.55
28	H2NCOCH2-	8.06	6.09
29	(CH3)2NCOCH2-	7.72	<6
30 (DR-4365)	(CH3)HNCOCH2-	8.45	<6

^aKikuchi et al. (2002a)

^bDetermined at human 5-HT₇receptors in COS-7 cells using [³H]5-CT.

^cDetermined at 5-HT_2A receptors in rat cerebral cortex membranes using [³H]ketanserin.

In a series of papers, Perrone and coworkers at Bari University, Italy, have studied the structure-activity relationships of a number of ligands characterized by an 1-arylpiperazine as the basic moiety. The main issue addressed with this framework was selectivity over 5-HT_1A receptors. That was indeed achieved in various ligands such as LP-44, LP-12, and LP-211 (see below for structures and details). In the first report, a series of 1-[ω-(4-aryl-1-piperazinyl)alkyl]-1-aryl ketones was disclosed (Perrone et al., 2003). Initially the position of an OH or OCH₃ on the aryl ketone moiety was evaluated together with various intermediate alkyl chains. These modifications led to 31 (R= 2-OH, n = 5) that demonstrated high 5-HT₇ receptor affinity (K_i = 5.8 nM). Modification of 31 on the aromatic ring linked to the piperazine ring gave interesting information (Table 12). Affinity data for compounds 32 and 33 revealed that the absence of an aryl group linked to the piperazine (R = CH₃, cyclohexyl) had a dramatic effect (K_i > 1000 nM), whereas aryl groups other than 2-methoxyphenyl (i.e. 2-Py, Ph, 3-CF₃-Ph, 2-benzoxazolyl, 2-benzimidazolyl) significantly reduced affinity (compounds 34–38). On the other hand, the 1,2-benzisoxazol-3-yl derivative 40 displayed high 5-HT₇ affinity (K_i = 2.93 nM) and low 5-HT_1A affinity (K_i = 189 nM). It was also highlighted that the simple introduction of a methoxy group on the 1,2-benzisoxazolyl moiety (40 vs 39) gave a significant loss in 5-HT₇ affinity (K_i = 462 nM), revealing how sensitive to structural manipulation that part of the molecule was. On the other hand, methylation of the hydroxy group of 40 gave derivative 41 which showed subnanomolar 5-HT₇ affinity. Both 40 and 41 showed good selectivity over the 5-HT_1A receptor, but they were poorly selective over the 5-HT_2A receptor. Derivatives 40 and 41 displayed agonist properties like 5-CT when tested for 5-HT₇ receptor mediated relaxation of substance P-induced guinea-pig ileum contraction.

Table 12.

Structure-activity relationships of 1-[ω-(4-aryl-1-piperazinyl)alkyl]-1-arylketone derivatives^a

			Ki(nM)
Compound	R1	R2	5-HT7	5-HT1A
31	2-OH	2-CH3O-Ph	5.8	5.8
32	2-OH	CH3	>1000	>850
33	2-OH	cyclohexyl	>800	>850
34	2-OH	2-Py	105	56
35	2-OH	Ph	43	137
36	2-OH	3-CF3-Ph	384	282
37	2-OH		148	1459
38	2-OH		682	>850
39	2-OH		462	389
40	2-OH		2.93	189
41	2-OCH3		0.90	175

^aPerrone et al. (2003)

^bDetermined at rat 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

^cDetermined at 5-HT_1A receptors in rat cerebral hippocampus membranes using [³H]8-OH-DPAT.

Leopoldo et al. (2004b) have also evaluated a series of 4-(2-phenylsubstitued-ethyl)-1-(2-methoxyphenyl)piperazines (Table 13). Substitution of the 2-phenylethyl group with one or more methoxy substituents was evaluated and moderated differences in affinity at 5-HT₇ receptor were found (compounds 42–50). On the other hand, replacement of the 2-methoxyphenyl group linked to the piperazine ring in compound 42 by 1,2-benzisoxazol-3-yl (51) or 2-hydroxyphenyl (52), or 4-methoxyphenyl (53) ring resulted in dramatic effects on affinity, suggesting this part of the molecule was relevant for interaction with the 5-HT₇ receptor. All the compounds studied lacked selectivity over 5-HT_1A receptors. Compound 51 behaved as an agonist when tested for 5-HT₇ receptor mediated relaxation of substance P-induced guinea-pig ileum contraction. This result replicated the finding obtained for compound 40 discussed above.

Table 13.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of 4-arylethyl-1-arylpiperazine derivatives^a

			Ki (nM)
Compound	Ar	Ar’	5-HT7b	5-HT1Ac
42	3-OCH3-Ph	2-OCH3-Ph	24.5	2.37
43	3-OH-Ph	2-OCH3-Ph	35.4	8.12
44	2-OCH3-Ph	2-OCH3-Ph	160	not tested
45	4-OCH3-Ph	2-OCH3-Ph	126	not tested
46	2,6-di-OCH3-Ph	2-OCH3-Ph	747	not tested
47	2,5-di-OCH3-Ph	2-OCH3-Ph	31.9	21.9
48	3,4-di-OCH3-Ph	2-OCH3-Ph	48.3	8.36
49	2,3,4,-tri-OCH3-Ph	2-OCH3-Ph	71	4.50
50	2,4,6-tri-OCH3-Ph	2-OCH3-Ph	300	not tested
51	3-OCH3-Ph	3-(1,2-benzisoxazolyl)-	8.2	3.63
52	3-OCH3-Ph	2-OH-Ph	51.5	3.50
53d	3-OCH3-Ph	4-OCH3-Ph	> 1000	not tested

^aLeopoldo et al. (2004b)

^bDetermined at rat 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

^cDetermined at 5-HT_1A receptors in rat cerebral hippocampus membranes using [³H]8-OH-DPAT.

^dLeopoldo et al. (unpublished)

Leopoldo et al. (2004a, 2007) have also reported on a series of N-(1,2,3,4-tetrahydronaphthalen-1-yl)-4-aryl-1-piperazinealkylamides as 5-HT₇ receptor agents. Initial exploration of structure-affinity relationships indicated that the length of the intermediate alkyl chain had a relevant role, with a five methylene chain being preferred. Also, the presence and phe position of one methoxy group on the 1,2,3,4-tetrahydronaphthalen-1-yl ring had a marginal role on affinity. Thus, derivative 54 (Table 14) was selected for further structural modifications. Exploration of various aryl rings linked to the piperazine ring evidenced the critical role played by this portion of the molecule. In particular, the possible isomers of OCH₃, CN, and COCH₃ substituted arylpiperazines were evaluated (compounds 54–62, Table 14). Affinity data revealed a key role of substitution pattern, being the 4-substituted derivatives nearly devoid of affinity and the 2-substitued derivative the most potent among the isomers. However, none of the above described structural modifications afforded significant selectivity over 5-HT_1A receptors. The study of this class of compounds proceeded further by evaluating substituents of different natures in the 2-position of the aryl ring linked to the piperazine ring (Table 15). It was found that the nature of the substituent markedly influenced the affinity: polar substituents were detrimental for affinity, whereas bulky apolar groups gave high affinity ligands. Again, 5-HT_1A receptor affinities of these compounds generally paralleled those for the 5-HT₇ receptor. Relevant exceptions were compounds 63 (R= 2-SCH₃), 74 (R= 2-isopropyl), 76 (2-phenyl), and 77 [2-N(CH₃)₂]: these compounds showed good selectivity over 5-HT_1A receptors. Thus, for the first time in this series of compounds a structural feature was identified that was well tolerated by the 5-HT₇ receptor but not by the 5-HT_1A receptor. The intrinsic activities at 5-HT₇ receptors of selected compounds were assessed in an isolated guinea-pig ileum assay by measuring 5-HT₇ agonist mediated relaxation of substance P-induced contraction. The intrinsic activity of the phenyl derivative 66 (40% Maximal Activity, EC₅₀= 8 µM) indicated that the N-(1,2,3,4-tetrahydronaphthalen-1-yl)-4-phenyl-1-piperazinealkylamide framework was able to activate the 5-HT₇ receptor independently from the presence of a substituent in 2-position. The presence of the 2-nitro- or 2-chloro substituent (compounds 67 and 68, respectively) had little effect on intrinsic activity, both compounds being weak partial agonists. The impact on the ability to activate the 5-HT₇ receptor was more marked in the case of bulky lipophilic groups. In fact, compounds 74 (83% Maximal Activity, EC₅₀= 0.90 µM) and 76 (74% Maximal Activity, EC₅₀= 1.77 µM) acted as partial agonists. Derivatives 63 and 77 were full agonists (100% Maximal Activity, EC₅₀= 1.17 and 2.56 µM, respectively). By contrast, 2-OH and 2-NHCH₃ substituents switched intrinsic activity toward antagonism: compound 78 antagonized the 5-HT₇ receptor (0% Maximal Activity, pA₂= 7.7) to a similar extent as the 2-hydroxy derivative 65.

Table 14.

K_i values at 5-HT₇ and 5-HT_1A receptors of N-(1,2,3,4-tetrahydronaphthalen-1-yl)-4-aryl-1-piperazinealkylamides^a

		Ki (nM)
Compound	R	5-HT7b	5-HT1Ac
54	2-OCH3	6.64	8.6
55	3-OCH3	119	105
56	4-OCH3	2100	not tested
57	2-CN	48.7	16.6
58	3-CN	97.8	291
59	4-CN	1400	not tested
60	2-COCH3	4.14	3.8
61	3-COCH3	496	676
62	4-COCH3	2639	not tested

^aLeopoldo et al. (2004a)

^bDetermined at rat cloned 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

^cDetermined at human cloned 5-HT_1A receptors in HEK-293 using [³H]8-OH-DPAT.

Table 15.

K_i values at 5-HT₇ and 5-HT_1A receptors of 1-arylpiperazine derivatives^a

		Ki (nM)
Compound	R	5-HT7b	5-HT1Ac
63 (LP-44)	SCH3	0.22	52.7
64	CH3	15.2	279
65	OH	11.4	24.0
66	H	65.6	128
67	NO2	63.3	183
68	Cl	40.1	96.0
69	CONH2	229	494
70	SO2CH3	298	3124
71	CH2CH3	7.10	79.2
72	(CH2)2CH3	49.6	168
73	(CH2)3CH3	2810	60.0
74	CH(CH3)2	1.10	167
75	C(CH3)3	538	1196
76 (LP-12)	Ph	0.13	60.9
77	N(CH3)2	0.90	112
78	NHCH3	25.4	133
79	NH2	8178	415
80	NHCOCH3	338	2500
81	NHSO2CH3	4253	not tested
82	F	131	29.2

^aLeopoldo et al. (2004a, 2007)

^bDetermined at rat cloned 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

^cDetermined at human cloned 5-HT_1A receptors in HEK-293 using [³H]8-OH-DPAT.

In a subsequent development, structural simplification of the N-(1,2,3,4-tetrahydronaphthalen-1-yl)-4-aryl-1-piperazinealkylamide was performed in order to obtain compounds endowed with suitable physicochemical properties for rapid and extensive penetration into the brain (Leopoldo et al., 2008). In particular, the structural modifications were targeted on the 1,2,3,4-tetrahydronaphthalene nucleus as suggested by the previous structure-affinity relationship studies. Three structurally related moieties (isoindoline, 1,2,3,4-tetrahydroisoquinoline, benzyl) that would have less impact on the lipophilicity of the target molecules were considered (Tables 16 and 17). Comparing the isoindolines 83–86, isoquinolines 87–90, and benzyl derivatives 91–94, with the corresponding tetrahydronaphthalenyl derivatives 63, 76, 74 and 54 a reduction in 5-HT₇ affinity was observed. It was suggested that the apolar nature of the terminal fragment was not the only feature that modulates the interaction of this part of the ligand with the 5-HT₇ receptor, because plotting the pK_is against the values of ClogP (logarithm to the base 10 of calculated partition coefficent 1-octanol/water) of compounds 83–94 revealed a low correlation coefficient. Replacement of the benzyl group attached to the amide nitrogen of compounds 91–94 (Table 17) by more polar groups had a limited impact on 5-HT₇ affinity, except in the case of compounds 97 and 101, which were 42- and 25-fold more potent than 93. It is interesting to note that the presence of 4-pyridinylmethyl, 4-cyanophenylmethyl, and 4-methanesulfonylphenylmethyl as terminal fragment was also well tolerated in the case of 2-phenyl derivatives 96, 100, and 104 that showed high 5-HT₇ affinity (5.7 nM < K_i < 0.58 nM). Affinity data for the 5-HT_1A receptor of the simplified compounds were quite similar to those of the reference 1,2,3,4-tetrahydronaphthalenyl derivatives 63, 76, 74 and 54. Nonetheless, the 2-phenyl derivatives 88, 96, and 100 still retained nanomolar affinity at 5-HT₇ receptors, showed lipophilicity within the target range, and showed good selectivity over 5-HT_1A receptors (25- to 324-fold). The authors further characterized the ligands 88, 96, and 100 for their intrinsic activity at 5-HT₇ receptor, by measuring the 5-HT₇ agonist mediated relaxation of substance P-induced contraction in an isolated guinea-pig ileum assay. The compounds showed full agonism at the 5-HT₇ receptor. Compounds 96 and 100 demonstrated EC₅₀ values comparable to that of the standard 5-HT₇ agonist 5-CT, and higher potency than the tetrahydronaphthalenyl derivative 76. It was observed that the modification of the terminal fragment of 76 left unchanged the intrinsic activity of 88, 96, and 100, confirming the predominant role of the arylpiperazine moiety on the intrinsic activity in this class of compounds.

Table 16.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of some 1-arylpiperazine derivatives reported by Leopoldo et al. (2008)

			Ki (nM)
Compound	n	R	5-HT7a	5-HT1Ab
83	1	SCH3	42	1.07
84	1	Ph	8.3	124
85	1	CH(CH3)2	54	22
86	1	OCH3	179	3.2
87	2	SCH3	32.9	2.04
88	2	Ph	3.81	95
89	2	CH(CH3)2	6	3.7
90	2	OCH3	31	40

^aDetermined at rat 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

^cDetermined at human cloned 5-HT_1A receptors in HEK-293 using [³H]8-OH-DPAT.

Table 17.

K_i values at 5-HT₇ and 5-HT_1A receptors of 1-arylpiperazine derivatives^a

			Ki (nM)
Compound	Ar	R	5-HT7b	5-HT1Ac
91		SCH3	22	12
92		Ph	not determined	161
93		CH(CH3)2	215	139
94		OCH3	224	13
95		SCH3	34.8	8.2
96		Ph	0.98	70
97		CH(CH3)2	5.1	325
98		OCH3	389	19
99		SCH3	9.0	94
100 (LP-211)		Ph	0.58	188
101		CH(CH3)2	8.6	53
102		OCH3	296	43
103		SCH3	148	10.3
104		Ph	5.7	106
105		CH(CH3)2	76	57
106		OCH3	71	29

^aLeopoldo et al. (2008)

^bDetermined at rat cloned 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

^cDetermined at human cloned 5-HT_1A receptors in HEK-29 using [³H]8-OH-DPAT.

Bojarski et al. (2004) have investigated the role of intermediate alkyl chains of various “long-chain” arylpiperazines with respect to the affinity for 5-HT₇ receptor. Although none of the compounds were significantly selective over the 5-HT_1A receptor, the results gave interesting insight into the interaction of the molecules with the 5-HT₇ receptor. The authors developed some constrained analogues of the 5-HT_1A receptor antagonist 107 (NAN-190) (Table 18). It was observed an increase in 5-HT₇ affinity in relation to 107 for trans derivative 108. The cis isomer 109 and bismethylbenzene derivative 110 was about 10-fold less active than the corresponding trans analogue. The rigid compound 111, containing a cyclohexane moiety, was devoid of 5-HT₇ activity. Since such linearly constrained compound did not bind to the 5-HT₇ receptor, the authors suggested that the bent conformation of flexible “long-chain” arylpiperazines should be regarded as bioactive. Therefore, the partly constrained trans derivatives should be able to adopt a bent conformation during the interaction with this receptor. It was finally noted that the flexible butyl derivative could easily adopt a bent bioactive conformation that was close to that of a pharmacophore model for 5-HT₇ antagonism developed by López-Rodríguez et al. (2003).

Table 18.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of 1-(2-methoxyphenyl)piperazine derivatives^a

		Ki (nM)
Compound	Linker	5-HT7b	5-HT1Ac
107 (NAN-190)		87	0.6
108		36	5
109		353	77
110		367	405
111		2045	8

^aBojarski et al. 2004

^bDetermined at 5-HT₇ receptors in rat hypothalamus membranes using [³H]5-CT.

^cDetermined at 5-HT_1A receptors in rat cerebral hippocampus membranes using [³H]8-OH-DPAT.

Researchers at EGIS Pharmaceuticals Plc have developed a series of “long-chain” arylpiperazines bearing as terminal fragment an oxindole group (Table 19) (Volk et al., 2008). The authors formally deduced the framework from the tetrahydrobenzindole derivative previously studied by Kikuchi et al. (1999, 2002a,b). Actually, the ring opening of compounds 5 and 8 left the affinity for the 5-HT₇ receptor practically unchanged (compounds 112 and 115, Table 19). For methoxy substituted isomers 112–114 the affinities were ranked as follows: meta > ortho > para, whereas for chloro derivatives 115–117 the order was para > meta > ortho. The authors noted that the trend observed for compounds 112–117 was different from that found by Leopoldo et al. (2004a, 2007) in the case of 1-arylpiperazine dervatives 54–62 (Table 14). Because the methoxy derivatives 112–114 were less selective over the 5-HT_1A receptor than the chloro derivative 117, the authors evaluated derivatives 118 and 119, in order to assess optimal spacer length for high 5-HT₇ affinity and low 5-HT_1A affinity. The highest affinity for 5-HT₇ receptors was obtained with derivative 116 with a four methylene spacer. Inclusion of either Cl, F, or CH₃ in the 4-position of the aryl ring linked to the piperazine ring of 116 was well tolerated (derivatives 120–122). Removal of the ethyl group on the oxindole ring of 116 was also well tolerated (123) whereas replacement by isobutyl (124) caused a 4-fold loss in affinity. The authors also evaluated mono and dihalo substitution on the oxindole ring of 116: compounds 125–128 evidenced a 10-fold loss in affinity except the fluoro derivative 126.

Table 19.

Binding affinities at 5-HT₇ receptors of a series 1-arylpiperazine derivatives^a

Compound	R1	R2	n	R3	Ki (nM)b
112	H	Et	4	2-OCH3	5.38
113	H	Et	4	3-OCH3	2.55
114	H	Et	4	4-OCH3	20.40
115	H	Et	4	2-Cl	5.11
116	H	Et	4	3-Cl	0.41
117	H	Et	4	4-Cl	0.38
118	H	Et	3	3-Cl	21
119	H	Et	5	3-Cl	1.45
120	H	Et	4	3-Cl, 4-F	0.60
121	H	Et	4	3-Cl, 4-CH3	0.66
122	H	Et	4	3,4-diCl	0.63
123	H	H	4	3-Cl	0.49
124	H	i-But	4	3-Cl	1.80
125	5-Cl	Et	4	3-Cl	8.27
126	5-F	Et	4	3-Cl	0.67
127	5-F, 7-Cl	Et	4	3-Cl	4.75
128	5,7-diCl	Et	4	3-Cl	9.46

^aVolk et al. (2008)

^bDetermined at human cloned 5-HT₇ receptors in CHO cells using [³H]LSD.

Paillet-Loilier et al. (2005) have reported the results of a screening at 5-HT₇ receptors of a series of 1-(2-methoxyphenyl)piperazines bearing on the basic nitrogen a phenylpyrrolemethyl substituent. The compounds studied displayed poor affinity for the 5-HT₇ receptor (Table 20). Nonetheless, some information can be deduced. As an example, compounds 131 and 132 showed good affinities for the receptor (K_i = 4.7 and 5.4 nM, respectively) whereas the corresponding regioisomers 129 and 130 displayed 63% and 24% of inhibition at 10⁻⁶ M, respectively. Also, replacement of CN in 132 with OCH₃ (135) completely abolished 5-HT₇ affinity. It is noteworthy that 1,2-benzisoxazol-3-yl derivatives 133 and 134 were considerably less potent than the corresponding 2-methoxyphenyl derivatives 131 and 132, suggesting that the binding mode of these ligands was different from that of similar arylpiperazines described by Leopoldo et al. (2004b).

Table 20.

K_i values at 5-HT₇ receptor of phenylpyrrole derivatives reported by Paillet-Loilier et al. (2005)

Compound	Substitution	R	Ar	Ki (nM)a
129	2	CH3	2-MeO-Ph	(63% @ 10−6 M)
130	2	CN	2-MeO-Ph	(24% @ 10−6 M)
131	3	CH3	2-MeO-Ph	4.7
132	3	CN	2-MeO-Ph	5.4
133	3	CH3	1,2-benzisoxazol-3-yl	(42% @ 10−6 M)
134	3	CN	1,2-benzisoxazol-3-yl	(34% @ 10−6 M)
135	3	OCH3	2-MeO-Ph	(6% @ 10−6 M)

^aDetermined at human cloned 5-HT₇ receptors in sf9 cells using [³H]LSD.

Na et al. (2008) have reported a new class of “long-chain” arylpiperazine derivatives, bearing as terminal fragment a quinazolinone group (Table 21). The authors evaluated two groups of compounds characterized by a spacer with three or four carbon atoms and varied the nature and the position of the substituent on the aryl group linked to the piperazine ring. The affinity trend shown by this class of compounds generally paralleled that previously observed for other “long-chain” arylpiperazines. For example, compounds with a four methylene spacer showed binding affinities higher than those with a three methylene chain (136–140 vs 141–145). Also, the compounds with a substituent in ortho position of the phenyl ring linked to the piperazine ring displayed higher 5-HT₇ affinities, whereas para-substituted phenyl derivatives (137, 140, 142, 145) were significantly less potent or devoid of affinity for the target receptor. The nature of the substituent in ortho position was also relevant for 5-HT₇ affinity: OCH₃ (143) and OC₂H₅ (146) substituents were preferred over 2-F or Cl (147 and 148, respectively). The authors also investigated fluorosubstitued quinazolinone derivatives but found no substantial differences with the unsubstituted counterparts (143, 146, 147). The selectivity of selected compounds was assessed: the most potent compound 150 (IC₅₀ = 12 nM) was >42-fold selective over 5-HT_1A, 5-HT_2A, 5-HT_2C, and D₂ receptors (Na et al., 2008).

Table 21.

IC₅₀ values at 5-HT₇ receptor of quinazolinone derivatives^a

Compound	X	n	R	IC50 (nM)
136	H	3	3-Cl	510
137	H	3	4-Cl	370
138	H	3	2-OCH3	80
139	H	3	3-OCH3	1400
140	H	3	4-OCH3	710
141	H	4	3-Cl	110
142	H	4	4-Cl	450
143	H	4	2-OCH3	21
144	H	4	3-OCH3	>10000
145	H	4	4-OCH3	2000
146	H	4	2-OC2H5	26
147	H	4	2-F	400
148	H	4	2-Cl	130
149	F	4	2-OCH3	120
150	F	4	2-OC2H5	12
151	F	4	2-Cl	200

^aNa et al. (2008)

^bDetermined at human cloned 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

Another series of arylpiperazine derivatives has been reported by the same research group (Yoon et al., 2008). The authors studied several 4-methoxyphenyl- and 2-naphthalenyl sulfonamides (Table 22) that were designed on the basis of the pharmacophoric hypothesis for 5-HT₇ antagonism reported by López-Rodríguez et al. (2000). According to the hypothesis, the compounds presented an aromatic ring, a basic nitrogen atom, an H-bonding acceptor group and a hydrophobic region at 4.9–5.9 Å apart from the basic center. The modifications were for the most part focused on substitutions on the arylpiperazine moiety. Considering the 4-methoxysulfonamides, it was found that the rank order of 5-HT₇ receptor affinity among the methoxy-substituted derivatives was 2 > 3 > 4 (compounds 152–154), and 3 > 4 for the CF₃ derivatives 155 and 156. Presence of NO₂ and COCH₃ substituents in the 4-position (compounds 158 and 159) abolished the 5-HT₇ affinity. The same effect was observed for the phenyl derivative 157. A similar trend was shown by the corresponding 2-naphthalenesulfonamide derivatives. However, the most potent compounds 160 and 163 lacked selectivity over the 5-HT_1A receptor. In general, structure-affinity relationships highlighted in this study paralleled those found by other research groups (Kikuchi et al., 2002a; Leopoldo et al., 2004b; Leopoldo et al., 2007) for ligands belonging to the class of “long-chain” arylpiperazines.

Table 22.

Binding affinities at 5-HT₇ receptor of some 1-arylpiperazine derivatives reported by Yoon et al. (2008)

Compound	Ar1	Ar2	IC50 (nM)
152	4-OCH3-Ph	2-OCH3-Ph	37
153	4-OCH3-Ph	3-OCH3-Ph	339
154	4-OCH3-Ph	4-OCH3-Ph	9314
155	4-OCH3-Ph	3-CF3-Ph	89
156	4-OCH3-Ph	4-CF3-Ph	384
157	4-OCH3-Ph	Ph	>10000
158	4-OCH3-Ph	4-NO2-Ph	>10000
159	4-OCH3-Ph	4-COCH3-Ph	>10000
160	2-naphthalenyl	2-OCH3-Ph	20
161	2-naphthalenyl	3-OCH3-Ph	223
162	2-naphthalenyl	4-OCH3-Ph	3241
163	2-naphthalenyl	3-CF3-Ph	12
164	2-naphthalenyl	4-CF3-Ph	567
165	2-naphthalenyl	Ph	681

^aDetermined at human cloned 5-HT₇ receptors in CHO cells using [³H]LSD.

2.3. 4-Methylpiperidine compounds and their structural evolution

One of the first structure-activity relationship studies on 5-HT₇ receptor antagonists was published by researchers at SmithKline Beecham (later GlaxoSmithKline, GSK) in 1998 (Forbes et al., 1998). The studies started with the identification of the sulfonamide 166, after high throughput screening of the corporate Compound Bank against the human cloned 5 HT₇ receptor (Table 23). The lead compound 166 showed modest 5-HT₇ receptor affinity (pK_i= 7.2) and acceptable selectivity over a range of other receptors. Evaluation of all four possible enantiomers of 166 (compounds 167–170) indicated that the R chirality of the center closer to the sulfonamide function was essential for 5 HT₇ receptor affinity. On such basis, the chiral centre on the piperidine ring was removed by shifting the methyl group from the 3- to the 4- position. Thus, compound 167 originated 171 (Table 24) which showed slightly improved affinity (pK_i = 7.5). Subsequently, the naphthalene was replaced by other aromatic rings (compounds 172–175) but 5-HT₇ affinity remained basically unchanged. In particular, derivative 172 (SB-258719) showed good 5-HT₇ receptor affinity, 100-fold selectivity over a range of serotonergic receptors, including 5-HT_1A and 5-HT_2A, and competitive antagonism at human 5-HT₇ receptor.

Table 23.

Binding affinities at 5-HT₇ receptor of naphthalenesulfonamide derivatives^a

Compound	Stereochemistry 1	Stereochemistry 2	pKib
166	R,S	R,S	7.2
167	R	R	6.9
168	R	S	6.2
168	S	R	5.8
170	S	S	<5.0

^aForbes et al. (1998)

^bDetermined at human cloned 5-HT₇ receptors in HEK-293 cells using [³H]5-CT.

Table 24.

Binding affinities at 5-HT₇ receptors of a series of 4-methylpiperidines. ^a

Compound	Ar	pKib
171	1-naphthalenyl	7.5
172 (SB-258719)	3-methylphenyl	7.5
173	3,4-dichlorophenyl	7.5
174	3,4-dibromophenyl	7.7
175	3,4-dibromo-2-thienyl	7.8

^aForbes et al. (1998)

^bDetermined at human cloned 5-HT₇ receptors in HEK-293 cells using [³H]5-CT.

Starting from the observation that the chiral center in the flexible chain of 171 was critical for 5 HT₇ receptor affinity, the authors performed conformational analysis of this side chain. It was observed that all bonds were relatively free to rotate apart from the S–N and N(Me)–C(Me) bonds. An energy minimum was observed when the two methyl groups were oriented gauche with respect to each other, which could represent the binding conformation. This study suggested that analogues in which both methyl groups were tied together into a ring would possess 5-HT₇ receptor affinities (Lovell et al., 2000). Consequently, the authors designed cyclic analogues of the naphthalene derivative 171 by incorporating both 2-piperidinylethyl and 2-pyrrolidinylethyl side chains (compounds 176 and 177, respectively, Table 25). In both cases the (R)-isomer showed higher affinity than the (S)-isomer, and the pyrrolidine ring constrains the side chain in a more optimal conformation. Using the template of (R)-177, the authors evaluated aromatic rings different from 1-naphthalenyl. This modification was well tolerated (compounds 178–182, Table 26) and 3-hydroxyphenylsufonamide 182 showed very high 5-HT₇ receptor affinity (pK_i = 8.9). It should be noted that the hydroxyl group of 182 would favour additional interaction of the framework with the 5 HT₇ receptor, but it is not strictly necessary for 5-HT₇ affinity. Compound 182 (SB-269970) was found to be >100-fold selective against a total of 50 receptors, enzymes, or ion channels. It displayed a profile consistent with competitive antagonism when tested for the ability to inhibit agonist induced stimulation of adenylyl cyclase in HEK 293 cells stably expressing the 5 HT₇ receptor (Lovell et al., 2000; Hagan et al., 2000).

Table 25.

pK_i values at 5-HT₇ and 5-HT_1A receptors of 4-methylpiperidine derivatives^a

	pKi
Compound	5-HT7b	5-HT1Ac
R-176	7.8	5.8
S-176	6.4	--
R-177	8.0	6.0
S-177	6.4	--

^aLovell et al. (2000)

^bDetermined at human cloned 5-HT₇ receptors in HEK-293 cells using [³H]5-CT.

^cDetermined at human cloned 5-HT_1A receptors in HEK-293 cells using [³H]8-OH-DPAT.

Table 26.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of 4-methylpiperidine derivatives

		pKi
Compound	Ar	5-HT7a	5-HT1Ab
178c	3,4-diCl-Ph	8.4	6.5
179c	3-Br-Ph	8.7	6.4
180c	3-CH3-Ph	8.5	6.0
181c	3-OCH3-Ph	8.0	-
182 (SB-269970)c	3-OH-Ph	8.9	<5
183d		8.62	-
184d		7.64	-
185d		8.19	-
186d		8.19	-
187d		7.59	-
188d		7.24	-

^aDetermined at human cloned 5-HT₇ receptors in HEK-293 cells using [³H]5-CT.

^bDetermined at human cloned 5-HT_1A receptors in HEK-293 cells using [³H]8-OH-DPAT.

^cLovell et al. (2000)

^dForbes et al. (2002)

Tests in vivo in the rat with 182 evidenced extremely high blood clearance, almost certainly due to the presence of the phenolic hydroxyl group. To overcome this limitation, 182 was structurally manipulated by replacing the phenolic moiety with metabolically more stable bioisosters (Forbes et al., 2002) In all cases the proposed changes caused a reduction in 5 HT₇ receptor affinity (compounds 183–188, Table 26). Only the 6-indolyl derivative 183 was nearly equipotent to 182, but it showed high blood clearance and zero bioavailability in the rat.

Then, the authors turned their attention on the 4-methylpiperidine moiety of 183, because metabolism studies on closely related structures indicated that hydroxylation could occur on such moiety. The authors replaced the methyl group because modeling studies indicated the presence of a large lipophilic pocket around this region. This may be considered a turning point of the studies at GSK because they moved from compounds with a single aromatic moiety to a class of compounds with an additional aromatic ring (Forbes et al., 2002). However, it was not commented whether the lipophilic or the aromatic nature of the group was relevant for affinity. This modification significantly enhanced the 5-HT₇ receptor affinity (Table 27). The indole 189, the benzimidazolone 191, and the 4-fluorophenoxy 193, showed the greatest increase in affinity (pK_i > 9.1) as compared to 183. On the other hand, it was found that this modification markedly increased unwanted affinity for the adrenergic α_1B receptor subtype. Due to its relatively low affinity for α_1B receptor, the 4 chlorophenoxy derivative 196 (SB-656104) progressed for further pharmacological evaluations. 196 showed ≥ 100 fold binding selectivity over most other 5-HT receptor subtypes, except 5-HT_2A (30 fold), 5-HT_2B (50 fold) and 5-HT_1D (12 fold). 196 inhibited 5-CT induced stimulation of cAMP accumulation (pA₂ = 8.1), thus behaving as 5-HT₇ antagonist. 196 showed greatly improved pharmacokinetic profile in the rat compared to 182, possessing a lower blood clearance and a significantly improved half life. The antagonist 196 passed the blood brain barrier in a similar extent of 182. The oral bioavailability of 196 was 16%, whereas 182 did not show any oral bioavailability. Evaluation of 196 in an in vivo functional assay (5-CT induced hypothermia in guinea pigs) resulted in a ED₅₀ profile similar to 182, but with administration being via the oral route.

Table 27.

Binding affinities at 5-HT₇ and α_1B receptors of 4-arylpiperidine derivatives^a

		pKi
Compound	R	5-HT7b	α1Bb
189	indol-3-yl	9.17	8.89
190	indol-2-yl	8.20	7.05
191	benzimidazol-2-on-3-yl	9.15	7.36
192	benzoxazol-2-on-3-yl	8.74	6.51
193	4-F-phenoxy	9.30	7.52
194	4-Cl-benzoyl	9.10	7.03
195	4-F-benzoyl	9.08	7.72
196 (SB-656104)	4-Cl-phenoxy	8.70	6.66

^aForbes et al. (2002)

^bDetermined at human cloned 5-HT₇ receptors in HEK-293 cells using [³H]5-CT.

^cExperimental details not specified.

2.4. 2-Aminotetralin, 3-aminochroman, and aporphine

The 2-aminotetralin, 3-aminochroman, and aporphine moieties have served as templates for the mapping of the ligand binding sites of certain 5-HT receptors. In 2000, a research group at Uppsala University (Sweden) initiated a research program aimed at the identification of selective 5-HT₇ receptor agents based on the above mentioned chemical classes (Linnanen et al., 2000). Initially, the novel (R)-1,11-methyleneaporphine ring system was designed (compound 198, Table 28). This new ring system showed interesting affinity at the 5-HT₇ receptor (K_i= 6.9 nM) as compared to the parent aporphine 197 (K_i = 88 nM). Subsequently, the authors found that selectivity over the 5-HT_1A receptor could be significantly modulated by introducing substituents at C12. The most relevant compound was the 12-amino substituted derivative 199 which showed high 5-HT₇ affinity (K_i = 18 nM) and approximately 20-fold selectivity over the 5-HT_1A receptor. Also the derivative disubstituted by methyl and methoxy at C12 showed selectivity over 5-HT_1A receptors (compounds 200 and 201). In a subsequent study (Linnanen et al., 2001) the aporphine core was modified further by introducing an aryl group in 11-position (Table 29). This structural modification offered significant advantages in terms of selectivity when the aryl ring was disubstituted in 2’ and 6’ position. Such substitution pattern generated stable atropisomers that displayed pharmacological differences. In particular, derivative 205 displayed affinity for the 5-HT₇ receptor in the nanomolar range (K_i= 3.79 nM) and 40-fold selectivity over the 5-HT_1A receptor. This compound behaved as a 5-HT₇ receptor antagonist. Three years later, the same authors reported on novel 2-aminotetralin and 3-aminochroman derivatives (Table 30) that can be considered formally derived from the above 11-arylaporphines (Holmberg et al., 2004). This structural simplification affected neither affinity nor selectivity for the 5-HT₇ receptor: in fact, compounds 208 and 215 showed similar properties as 202. Subtle changes of the 5-aryl substituent (8-aryl in the chromans) afforded selectivity toward the 5-HT_1A receptor. In particular, introduction of a single ortho methoxy substituent (209 and 216) reduced the affinity for the 5-HT_1A receptor 8–13 times, while the affinity for the 5-HT₇ receptor was unchanged. Further introduction of a second ortho methoxy substituent (211 and 218) substantially reduced the affinity for the 5-HT_1A receptor while it only marginally affected binding to the 5-HT₇ receptor. The authors suggested that the ortho substituents, by forcing the 5- (or 8-) aryl group out of plane of the tetralin (or chroman), reduced favorable interactions with the 5-HT_1A receptor. Decreasing the size of the N-alkyl substituents increased the selectivity for the 5-HT₇ receptor, due to a decrease in affinity at 5-HT_1A receptors (compounds 212 and 213 for the tetralin series). Comparison of binding affinities of 216 and its enantiomer (not shown in Table 30) revealed that interaction of the ligand with the 5-HT₇ receptor was stereospecific, favoring the (R)-chromans, with an eudismic ratio of approximately 130. The nature of the substituent on the basic nitrogen appeared to influence the signal transduction, because 211 and 218 were full agonists whereas the corresponding dimethylamino derivatives 213 and 219 were antagonists or partial agonists. However, it should be noted that the structurally related dimethylamino derivative 214 (AS-19, Table 30) is a potent and selective 5-HT₇ agonist (Brenchat et al., 2009).

Table 28.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of 1,11-methyleneaporphine derivatives^a

			Ki[nM]
Compound	R1	R2	5-HT7b	5-HT1Ac
197	--	--	88.0	80.0
198	H	H	6.9	40.7
199	NH2	H	18.0	355
200	CH3	OCH3	1.1	17.3
201	OCH3	CH3	1.1	16.9

^aLinnanen et al. (2000)

^bDetermined at rat cloned 5-HT₇ receptors in CHO cells using [³H]5-HT.

^cDetermined at human cloned 5-HT_1A receptors in CHO cells using [³H]8-OH-DPAT.

Table 29.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of 11-phenylaporphine derivatives^a

			Ki (nM)
Compound	R	R1	5-HT7b	5-HT1Ac
202	H	H	9.78	1.8
203	OCH3	OCH3	13.0	554
204	CN	CH3	20.8	778
205	CH3	CN	3.79	142
206	OH	CH3	42.1	319
207	CH3	OH	23.0	48.8

^aLinnanen et al. (2001)

^bDetermined at rat cloned 5-HT₇ receptors in CHO cells using [³H]5-HT.

^cDetermined at human cloned 5-HT_1A receptors in CHO cells using [³H]8-OH-DPAT.

Table 30.

K_i values at 5-HT₇ and 5-HT_1A receptors of 2-aminotetralin and 3-aminochroman derivatives

			Ki (nM)
Compound	Ar	R	5-HT7a	5-HT1Ab
208c	Ph	C3H7	3.38	0.87
209c	2-MeO-Ph	C3H7	1.73	11.7
210c	4-MeO-Ph	C3H7	12.4	1.49
211c	2,6-diMeO-Ph	C3H7	7.9	347
212c	2,6-diMeO-Ph	H	78.7	>1000
213c	2,6-diMeO-Ph	CH3	2.55	1420
214 (AS19)d	1,3,5-Trimethylpyrazol-4-yl	CH3	0.6	89.7
215c	Ph	C3H7	2.92	1.13
216c	2-MeO-Ph	C3H7	2.7	10.3
217c	4-MeO-Ph	C3H7	12.6	1.31
218c	2,6-diMeO-Ph	C3H7	6.44	174
219c	2,6-diMeO-Ph	CH3	5.29	>1000

^aDetermined at rat cloned 5-HT₇ receptors in CHO cells using [³H]5-HT.

^bDetermined at human cloned 5-HT_1A receptors in CHO cells using [³H]8-OH-DPAT.

^cHolmberg et al. (2004)

^dBrenchat et al. (2009)

2.5. 2-Arylethylamine derivatives and related analogues

Paillet-Loilier et al. (2007) have reported on a class of aminoethylbiphenyls that were designed by analysing the structural homologies between phenylpyrrole, 11-phenylaporphine, and phenyl-2-aminotetraline derivatives (see Tables 20, 29, and 30). However, as the phenylpyrroles 220 and 221 were found to be inactive (Table 31), the pyrrole ring was replaced by a phenyl ring to better align with both 11-phenylaporphine and 2- aminotetraline compounds. In this series, the 5-HT₇ affinity and selectivity depended on the substitution of the phenyl ring denoted with A (Table 31). Introduction of ortho substituents (Me or OMe) into the phenyl ring A increased the 5-HT₇ receptor affinity (222 and 223). These ortho substituted compounds showed a good selectivity toward 5-HT_1A receptor. The presence of a substituent in ortho position was essential for the 5-HT₇ receptor affinity. The authors hypothesized that the aryl groups A and B should not be placed in the same plane. Analysis from Cambridge Software Database data (Bruno et al., 2004) of the distributions of the dihedral angle between the two phenyl rings, unsubstituted and ortho substituted, showed a different distribution between the two fragments. The hypothesis was supported by the lack of affinity of the unsubstituted derivative 240. Introduction of a second substituent in ortho position in compounds 222 and 223 did not modify the 5-HT₇ receptor affinity but improved the selectivity over the 5-HT_1A receptor: 222 and 223 were about 60-fold selective over 5-HT_1A receptor, whereas 225 and 282 showed increased selectivity (360- and 560-fold, respectively). It was also observed that the unsubstituted amino group was tolerated (228 vs 222). Compounds 222 and 226 have been evaluated for their pharmacological profile by using a specific test of aldosterone secretion from perifused rat adrenal cortex stimulated by serotonin through 5-HT₇ receptor. Compound 222 behaved as 5-HT₇ antagonist (pK_b= 7.03), whereas 226 was found to be a partial agonist.

Table 31.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of aminoethylbiphenyl derivatives^a

			Ki (nM)
Compound	X	R	5-HT7b	5-HT1Ac
220	--	CH3	(5% @ 10−8 M)d	(0% @ 10−8 M)d
221	--	H	(16% @ 10−8 M)d	(14% @ 10−8 M)d
222	2-CH3	CH3	7.6	443
223	2-OCH3	CH3	8.4	531
224	H	CH3	(18% @ 10−8 M)d	(2% @ 10−8 M)d
225	2,6-di-CH3	CH3	6.2	2250
226	2,6-di-OCH3	CH3	8.6	4826
227	4-OCH3	CH3	(9% @ 10−8 M)d	(0% @ 10−8 M)d
228	2-CH3	H	7.5	1470

^aPaillet-Loilier et al. (2007)

^bDetermined at human cloned 5-HT₇ receptors in sf9 cells using [³H]LSD.

^cDetermined at 5-HT_1A receptors in rat hippocampus membranes using [³H]8-OH-DPAT.

^dPercentage of displacement at the concentration given.

Thomson et al. (2004) at Merck Sharp & Dohme laboratories have developed a series of 5-HT₇ receptor ligands starting from compound 229 (Table 32). Simple structural modifications of 229 gave interesting information: removal of 4-pyridine ring caused a loss in affinity (230), whereas replacement with phenyl or t-butyl was well tolerated (compounds 231 and 232, respectively). Variation of the core heterocycle was explored on derivative 231 (compounds 233, 234–237): when the thiazole of 229 was replaced by a pyridine with the 2,6-substitution pattern very high affinity value at 5-HT₇ receptors was achieved (237). The position of the pyridine nitrogen was important, as 2,4-substituted isomer 236 did not show subnanomolar binding affinity. Replacement of the phenyl ring of compound 237 by Br, cyclohexyl, t-butyl afforded high binding affinities at 5-HT₇ receptors. Compound 237, which was 26-fold selective over the 5-HT_1A receptor, was tested in a functional assay and found to behave as a partial agonist, giving 80% of the response of the full agonist 5-CT in 5-HT₇ receptors.

Table 32.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of thiazoles and thiopyridine derivatives^a

			Ki (nM)
Compound	R	Cy	5-HT7b	5-HT1Ac
229	4-Py		27	90
230	H		6800	--
231	Ph		12	210
232	t-Butyl		22	2200
233	Ph		3200	--
234	Ph		1200	--
235	Ph		>400	--
236	Ph		100	1200
237	Ph		0.6	16
238	Br		4.1	37
239	Cyclohexyl		0.7	8.7
240	t-Butyl		1.3	17

^aThomson et al. (2004)

^bDetermined at cloned 5-HT₇ receptors in CHO cells using [³H]5-HT (species not specified).

^cDetermined at cloned 5-HT_1A receptors in HeLa cells using [³H]5-HT (species not specified).

2.6. Aminotriazines and related analogues

Researchers at Bristol-Myers Squibb have identified 5-HT₇ receptor antagonists containing the aminotriazine core as basic moiety (Mattson et al., 2004). The lead compound 241 (K_i = 60 nM) was structurally modified to improve affinity and metabolic stability (Table 33). Replacement of the 3-phenylpropyl group with a 2-phenoxyethyl residue (242) resulted in an increase in affinity. Interstingly, the (S)-enantiomer 243 showed K_i = 3 nM whereas its enantiomer 244 was devoid of 5-HT₇ affinity (K_i > 1000 nM). Removal of the amino group linked to the triazine ring in the eutomer was well tolerated (compound 245, K_i = 3 nM), whereas substitution by F, CH₃ or N(CH₃)₂ gave a moderate loss in affinity (compounds 246–248). Introduction of an F substituent on both ends of 242 gave rise to compound 249 with no substantial changes in affinity (K_i = 5 nM). Also in the case of 249, the (S)-enantiomer was more potent than the (R)- (250 vs 251), but the eudismic ratio was considerably lower than that displayed by compounds 243 and 244. Compounds 243 and 250 displayed > 30-fold selectivity over 5-HT₆, α₁ and 5-HT_2C receptors and potent antagonistic activity at the 5-HT₇ receptor.

Table 33.

Binding affinities at 5-HT₇ receptor of aminotriazine derivatives^a

Compound	Chirality	R1	R2	X	R3	Ki (nM)b
241	±	H	NH2	CH2	H	60
242	±	H	NH2	O	H	8
243	S	H	NH2	O	H	3
244	R	H	NH2	O	H	>1000
245	S	H	H	O	H	3
246	S	H	F	O	H	10
247	S	H	CH3	O	H	23
248	S	H	N(CH3)2	O	H	54
249	±	F	NH2	O	F	5
250	S	F	NH2	O	F	2
251	R	F	NH2	O	F	36

^aMattson et al. (2004a)

^bDetermined at human cloned 5-HT₇ receptors in CHO cells using [³H]5-CT.

In a subsequent paper the same authors explored the structure-affinity relationships of the aminotriazine ring (Denhart et al., 2004). Several diaminopyridine and diamino pyrimidine varieties were studied (Table 34). Considering the analogues of compound 245 the most active compounds were pyridines with the nitrogen in position denoted W or Y in the general structure (compounds 252 and 254, Table 34). A substituted pyridyl ring as in 255 or a phenyl ring (256) abolished 5-HT₇ affinity. The addition of a second nitrogen in the pyridyl ring of 255 produced the pyrimidine derivative 257 which showed again nanomolar 5-HT₇ affinity.

Table 34.

Binding affinities at 5-HT₇ receptor of ligands reported by Denhart et al. (2004)

Compound	W	X	Y	Z	Ki (nM)a
252	N	CH	CH	CH	0.4
253	CH	N	CH	CH	4
254	CH	CH	N	CH	0.4
255	CH	CH	CH	N	> 1000
256	CH	CH	CH	CH	>1000
257	CH	CH	N	N	2

^aDetermined at human cloned 5-HT₇ receptors in CHO cells using [³H]5-CT.

2.7. N-N’-Disubstituted guanidine derivatives

Researchers at Pfizer have reported on a series of (4,5-dihydroimidazol-2yl)biphenylamines as 5-HT₇ receptor agonists (Parikh et al., 2003). The basic fragment employed for the synthesis of the target compounds deserves interest because it is an alternative to the most frequently used piperazine, piperidine, or pyrrolidine scaffolds. However, the compounds displayed significant α₂ adrenergic activity that resulted in marked changes in blood pressure and heart rate after in vivo administration in rats, presumably due to the structural similarity of the compounds with the α₂ adrenergic clonidine. The most potent ligand of this series (compound 258) is shown in Figure 2.

Figure 2.

Examples of N-N’-Disubstituted guanidine derivatives

In the search of alternative basic moieties in the design of 5-HT₇ ligands, Zajdel et al. (2009) have studied a series of N-alkyl-N’-dialkylguanidines that showed moderate affinity for the 5-HT₇ receptor. Comparing the 5-HT₇ receptor affinities of the new derivative 259 (Figure 2) with that of its analogue without the guanidine moiety 160 (Table 22), it emerges that the introduction of delocalized positive charge in the ionic center of molecule attenuated affinity for the 5-HT₇ receptor.

2.8. Miscellaneous compounds

Raubo et al. (2006) have reported structure-affinity relationships on a series of alkylphenylsulfones. This fragment was attached to a variety of basic moieties. For this reason, the classification of the compounds reported in Tables 35 and 36 does not fit the criterion used in the present overview. Preliminary modeling studies suggested that a gem-disubstituted cyclobutyl phenyl sulfone could overlay with one of the low energy conformations of 182 (SB-269970, Table 26). Thus, this fragment originated the compounds listed in Table 35. Obtained affinity data suggested that an aromatic moiety near the basic nitrogen was essential for 5-HT₇ receptor affinity, and this points out that the compounds do not bind at 5-HT₇ receptors in the same fashion as 182. In fact, compound 263 displays 5-HT₇ affinities significantly lower than 182. The best orientation of the aromatic ring was shown by tetrahydroquinoline derivative 267, which however showed modest selectivity over 5-HT_1A and 5-HT_2A receptors (4- to 8-fold). This compound was also evaluated in a functional model of 5-HT₇ receptor activation and found to behave as an antagonist. Replacement of the tetrahydroquinoline ring of 267 with the full hydrogenated counterpart gave 269, endowed with high 5-HT₇ affinity and 35-fold selectivity over the 5-HT_1A receptor. This finding suggested that the hydrophobic nature of the basic moiety was relevant for affinity. Taking this information into account, the authors studied a set of compounds related to 269 (Table 36). Replacement of the spirocycle of 269 by a gem-dimethyl substituent gave 270, which showed an excellent selectivity profile over 5-HT_1A and 5-HT_2A receptors (>100-fold) and partial agonist properties. Evaluation of amine structure-activity relationships of compound 270 demonstrated that functional activity can be attenuated by installing a pendant substituent at the 3-position (Table 36, compounds 271–275). A variety of substituents were tolerated in the 3-position. In particular, sulfones 272 and 274 exhibited high affinity at the 5-HT₇ receptor with minimal or no functional activity and >30-fold selectivity over 5-HT_1A, 5-HT_2A, 5-HT_1B, and D₂ receptors. This study is an interesting example of how modifications of the same fragment can modulate intrinsic activity at the 5-HT₇ receptor.

Table 35.

K_i values at 5-HT₇ receptor of ligands reported by Raubo et al. (2006)

Compound	NR1R2	Ki (nM)a
260		>7000
261		2700
262		>7000
263		550
264		5800
265		22
266		26
267		8
268		9
269		6

^aDetermined at human cloned 5-HT₇ receptors in CHO cells using [³H]5-HT.

Table 36.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of phenylsulfone derivatives^a

		Ki (nM)
Compound	NR1R2	5-HT7a	5-HT1Ab
270		8	1500
271		54	>7000
272		24	2400
273		455	NT
274		15	>7000
275		7	400

^aRaubo et al. (2006)

^bDetermined at human cloned 5-HT₇ receptors in CHO cells using [³H]5-HT.

^cDetermined at human cloned 5-HT_1A receptors in HeLa cells using [³H]5-HT.

2.9. Pharmacophore models for 5-HT₇ receptor models

The first pharmacophore model for 5-HT₇ receptor antagonists was proposed by López-Rodríguez et al. (2000). The study was performed using the software package Catalyst to analyze a set of thirty structurally different 5-HT₇ receptor antagonists. This preliminary model was later optimized, with the incorporation in the training set of analogues of 182 (SB-269970, Table 26) and of 11-phenylaporphine derivatives (Table 29). The optimized model consisted of five features: a positive ionizable atom (PI), an H-bonding acceptor group (HBA), and three hydrophobic regions (HYD1–3). The proposed distances between these chemical features are reported in Figure 3. The proposed hypothesis has been supported by design and synthesis of a series of naphtholactam and naphthosultam derivatives. Considering the compounds in Table 37, the authors noted that the hydrophobic region HYD3 of the model must be aromatic since compounds 283, 284, 287 lacked affinity, contrary to the aromatic substituted counterparts 286, 288, 289. The only notable exception to this trend was 280 (R= isopropyl) which showed moderate affinity. Also, it was observed that replacement of the piperazine with a piperidine or tetrahydropyridine ring caused a dramatic loss in affinity (285 compared to 281 and 282). The optimum spacer length was four or five methylene units, because they defined the optimum distance between the HBA and the basic center. On the other hand, compounds 276–278 with a shorter spacer were inactive. One exception was 279 (n= 3) but this compound had less affinity than 286. An increase in the size of the alkyl chain to n= 6 caused a loss in affinity (compounds 290 and 291). Moreover, it was noted that naphtholactam derivatives were more potent than naphthosultam analogues (286 vs 292; 289 vs 293). However the authors did not give any explanation for this behaviour in terms of ligand geometry. Considering the length of the spacer the only difference with the naphtholactam series was that compound 295 (n= 6) showed moderate affinity. It was suggested that a 6-carbon chain may adopt a folded conformation giving the appropriate distances for interaction with the receptor. It should be noted that although the authors address the model as being for 5-HT₇ receptor antagonists, the intrinsic activity of these ligands was not investigated.

Figure 3.

Optimized pharmacophore model for 5-HT₇ antagonists proposed by López-Rodríguez et al. (2003)a

^aDistances between pharmacophoric features: PI-HYD3 = 5.4–6.4 Å; PI-HYD1 = 5.2–6.2 Å; PI-HBA = 5.6–6.6 Å; HYD1-HBA = 3.6–4.8 Å; HYD1–HYD3 = 9.3–10.3 Å

Table 37.

Binding affinities at 5-HT₇ receptor of naphtholactam derivatives^a

Compound	X	n	Y	R	pKib
276	CO	1	N	Ph	<5
277	CO	1	N	2-MeO-Ph	<5
278	CO	3	N	Ph	<5
279	CO	3	N	2-MeO-Ph	6.4
280	CO	4	CH	i-Pr	6.7
281	CO	4	CH	Ph	<6
282	CO	4	C	Ph	<6
283	CO	4	N	Me	<5
284	CO	4	N	cyclohexyl	<5
285	CO	4	N	Ph	6.2
286	CO	4	N	2-MeO-Ph	7.2
287	CO	5	N	cyclohexyl	<5
288	CO	5	N	Ph	7.1
289	CO	5	N	2-MeO-Ph	7.0
290	CO	6	N	Ph	<6
291	CO	6	N	2-MeO-Ph	<6
292	SO2	4	N	2-MeO-Ph	6.7
293	SO2	5	N	2-MeO-Ph	6.6
294	SO2	6	N	Ph	<6
295	SO2	6	N	2-MeO-Ph	6.7

^aLópez-Rodríguez et al. (2003)

^bDetermined at 5-HT₇ receptors in rat hypothalamus membranes using [³H]5-CT.

The first pharmacophore for 5-HT₇ receptor agonism was determined by Vermeulen et al. (2003). Full conformational analysis of a set of 20 diverse 5-HT₇ receptor agonists (including 5-HT, 5-CT, 8-OH-DPAT, 1-(1-naphalenyl)piperazine, 1-(2-methoxyphenyl)piperazine, LSD, 214 (AS-19, Table 30), 258) in their protonated form was performed with the MacroModel molecular modeling software package, and followed by a pharmacophore-identifying procedure through ligand overlap using the Automated PharmacOphore Location through Ligand Overlap (APOLLO) procedure. The obtained model defined the distances between four pharmacophoric elements: a basic nitrogen atom (PI), an H-bonding acceptor group (HBA) and two hydrophobic domains (HYD) (Figure 4). The results of a comparative molecular field analysis (CoMFA) were then used to map the agonist binding site of the model of the 5-HT₇ receptor. Important roles in ligand binding was attributed to Asp^3.32 (according to Ballesteros & Weinstein 1995) (interaction with a protonated nitrogen), and Thr^5.43 (interaction with a substituent at an aromatic moiety). Amino acid residues of the aromatic cluster of transmembrane region 6 (TM6) were hypothesized to play an important role in ligand binding as π-π stacking moieties. It was also proposed that agonists missing a hydrogen-bond-accepting moiety, but possessing an aromatic substituent instead, could bind to the receptor with high affinity as well by occupying a lipophilic pocket hosted by residues of TM5 and TM6.

Figure 4.

Pharmacophore model for 5-HT₇ agonists proposed by Vermeulen et al. (2003)a

^aDistances between pharmacophoric features: PI-HYD1 = 5.7 Å; PI-HBA = 6.2 Å; HYD1-HBA = 3.0 Å; HYD1–HYD2 = 4.2 Å

Vermeulen et al. (2004) have also developed a pharmacophore model for 5-HT₇ receptor inverse agonists. A computational study was carried out on a set of 22 inverse agonists, which included various arylsulfonamides, exemplified by 296 and 297 (Figure 5). The model shown in Figure 6 resembles the model proposed by López-Rodríguez for selective antagonists, except for the additional HYD4 region mapping, among others, an aromatic substitutent at the sulfonamide group. Slightly shorter distances between HYD3-HBA and HYD2-HYD3 were also hypothesized. Interestingly, the model of inverse agonists revealed a close similarity to that of agonists, the main difference between the models being the presence of both HYD2 and HYD3 regions.

Figure 5.

5-HT7 inverse agonists developed by Vermeulen et al. (2004)

Figure 6.

Pharmacophore model for 5-HT₇ inverse agonists proposed by Vermeulen et al. (2004)a

^aDistances between pharmacophoric features: PI-HYD1 = 4.4 Å; PI-HBA = 5.8 Å; HYD1-HBA = 3.3 Å; HYD1–HYD4 = 4.4 Å; HYD4-HBA = 3.8 Å; HY3-HBA = 8.7 Å;

The first receptor-based pharmacophore for the 5-HT₇ receptor was constructed by Kołaczkowski et al. (2006), which evaluated, through docking studies, the mode of interaction of selective and nonselective antagonists with the receptor binding site. It was hypothesized that selective and nonselective antagonists might have different binding modes with the receptor. Thus, the pharmacophore model for antagonists was divided into two distinct submodels: an “affinity” pharmacophore and a “selectivity” pharmacophore.

The “affinity” pharmacophore was characterized by six features representing specific interactions points in the ligand structure: a protonated nitrogen (PI), three hydrophobic/aromatic regions (HYD/AR1-3) and two H-bond acceptors (HB1,2) (Figure 7). For affinity, at least three of the features must be present in a specific spatial arrangement. Among them, PI and one of ARs (capable of specific CH–π or π-π interaction) are strictly necessary, while the third may be HBA or another HYD/AR region. Each of the features interacts specifically with the receptor structure: PI is involved in the salt bridge formation with Asp^3.32; AR1 interacts with Phe^3.28 (CH–π or π-π ) and/or Arg^7.36 (ion-π), AR2 and AR3 have CH–π contact with Phe^6.52 and Phe^6.51, while HBA1 and HBA2 form H-bonds with Tyr^7.43 and Ser^5.42, respectively.

Figure 7.

Receptor based pharmacophore model for 5-HT₇ antagonists proposed by Kołaczkowski et al. (2006)a

^aDistances between pharmacophoric groups: PI-HYD/AR1= 6.9–9.4 Å; PI-HBA1 = 3.7–5.1 Å; PI-HYD/AR3= 3.7–6.1 Å; PI-HYD/AR2= 4.3–7.8 Å; PI-HBA2= 5.1–6.9 Å

Also the “selectivity” pharmacophore was characterized by three crucial features. Two of them were common for all the selective antagonists: PI and AR1, which form strong specific interactions with the residues in the pocket between transmembrane helices (TMH) 7-3 (especially Phe^3.28 and Arg^7.36). The third feature necessary for selectivity could be either HBA1, an H-bond acceptor situated in the vicinity of Tyr^7.43 or HYD/AR2, a hydrophobic or an aromatic moiety penetrating the pocket between TMHs 4–6. This latter interaction should not dominate that of the AR1 feature. The authors underlined that the geometry of the terminal moiety containing AR1 (aromatic imide/amide/sulfonamide), which should enable the formation of π-π stacking with Phe^3.38, ion-π interaction and/or with Arg^7.36, was of crucial importance. Thus, nonselective antagonists such as ergolines, aporphines, tricyclic psychotropes, and aminotetralines were docked to the receptor model in a quite similar way, because they accommodated (one of) the benzene ring by the pocket between TMHs 4–6.

By contrast, the arylsulfonamide part of molecules such as 182 (SB-269970), was situated in the pocket between TMHs 7-3, and the sulfone oxygen formed an H-bond with Tyr^7.43, while the aromatic moiety was in an optimal position for the π-π stacking and ion-π interaction with Phe^3.28 and Arg^7.36, respectively.

On the other hand, complex ligands such as those listed in Tables 9–11 consist of two pharmacophoric groups: the main arylpiperidine/arylpiperazine/β-carboline moiety (containing the crucial basic nitrogen) and the “terminal amide/imide” fragment of different structure (frequently containing another aryl ring), linked together by a flexible alkyl spacer (2–5 methylene units). Compounds of that numerous and diversified class of 5-HT₇ receptor antagonists occupied both pockets of the binding site simultaneously (i.e. one between TMHs 4–6 and the other between TMHs 7-3), having different interactions because of their diversified structure. The aromatic ring of an arylpiperidine/arylpiperazine/β-carboline moiety penetrated the pocket between TMHs 4–6, forming a specific CH–π interaction or having van der Waal contacts with one or more residues from the aromatic cluster of TMH6 (Phe^6.52, Phe^6.51, Trp^6.48) (compounds 17, 19, 23, 30, 65). The terminal amide/imide moiety occupied the pocket between TMHs 7-3 (compounds 17, 19, 23, 30, 65), Aromatic rings from terminal groups formed specific interactions with Phe^3.28 and/or Arg^7.36 (compounds 17, 19, 23, 30, 65). In the case of tetrahydrobenzindole as a terminal imide (compounds 17, 19, 23, 30), development of π-π stacking with Phe^3.28 and an ion-π interaction with Arg^7.36 were possible. That was due to the geometry of the tetrahydrobenzindole group, which positioned its benzene ring in a way that enabled those favorable interactions.

The arylsulfonamidoalkylamines with a bulky aromatic substituent at the 4-position of piperidine (Table 27) shared with complex arylpiperidines-piperazine the binding mode with the receptor. Again, arylsulfonamide part of the molecule was situated in the pocket between TMHs 7-3, and the sulfone oxygen formed an H-bond with Tyr^7.43, while the aromatic moiety was in an optimal position for the π-π stacking and ion-π interaction with Phe^3.28 and Arg^7.36, respectively. p-Cl-Phenoxy or indole substituents at the piperidine ring interacted with the hydrophobic/aromatic residues in the pocket created by TMHs 4–6. It was postulated that compounds that fit well the TMHs 7-3 pocket and were threfore independent of the interaction with the TMHs 4–6 cavity, which is highly conserved in monoamine GPCRs, were more likely to show selectivity.

It should be noted that the binding mode for 182 (SB-269970) suggested by Kołaczkowski et al. (2006) differed from that proposed by Vermeulen et al. (2004). Apart from the electrostatic interaction between the protonated nitrogen atom with Asp^3.32, Vermeulen and coworkers suggested: C-H˙˙π interactions between Ph^7.38 and the piperidine ring; interaction between the sulfonamide oxygen and the hydroxyl group of Thr^6.45; interaction between the phenolic-OH and Thr^6.45.

Very recently, Medina et al. (2009) have designed a new class of 5-HT₇ ligands on the basis of the pharmacophore model for antagonists (Table 38) proposed by the same research group (López-Rodríguez et al., 2003). From the structure-affinity relationship standpoint the study confirmed previous findings about spacer characteristics and the nature of the aryl linked to the piperazine ring. Computational models of the 5-HT₇ receptor in complex with ligands 299, 300, and 309 were also reported. As an example, it was proposed that the lactam carbonyl group of 299 interacts with Ser^6.55, and the aromatic ring of this same moiety with Phe^5.67 and Phe^6.52 (i.e. all residues present in the TMHs 5 and 6). Most notably, the methoxy group on the aryl ring linked to the piperazine ring is reported to interact with Arg^7.36. The authors suggested this interaction to be a “hot spot” of the interaction of the molecule with the receptor, responsible for high selectivity. However, it was not explained why the corresponding derivative 301 which lacked of such feature was still selective. As can be noted, the interaction of these ligands with the 5-HT₇ receptor was quite different from that proposed by Kołaczkowski for 1-arylpiperazine derivatives. These findings confirm that the 5-HT₇ binding site can be divided into two pockets: one between TMHs 4–6 and the other between TMHs 7-3. Moreover both pockets have similar interaction points for aromatic moieties and H-bond acceptors, and this makes them fairly symmetrical and, therefore, equivalent upon ligand docking.

Table 38.

Binding affinities at 5-HT₇ and 5-HT_1A receptors of 1-arylpiperazine derivatives^a

			Ki (nM)
Compound	Spacer	Ar	5-HT7a	5-HT1Ab
298	(CH2)4	Ph	74	124
299	(CH2)4	2-OCH3-Ph	32	>1000
300	(CH2)4	1-naphthalenyl	47	22
301	(CH2)5	Ph	63	>1000
302	(CH2)5	2-OCH3-Ph	63	>1000
303	(CH2)5	1-naphthalenyl	62	16
304		1-naphthalenyl	250	11.5
305		1-naphthalenyl	69	39
306		1-naphthalenyl	>1000	26
307		1-naphthalenyl	>1000	>1000
308		1-naphthalenyl	>1000	181
309	(CH2)4	d	7	219
310	(CH2)3	d	105	>1000
311	(CH2)2	d	350	>1000

^aMedina et al. (2009)

^bDetermined at human cloned 5-HT₇ receptors in HEK-293 cells using [³H]LSD.

^cDetermined at human cloned 5-HT_1A receptors in CHO cells using [³H]8-OH-DPAT.

^dThe basic moiety is 1,2,3,4-tetrahydroisoquinoline.

3. 5-HT₇ receptor function in normal and pathological processes

3.1. Background

It has consistently been possible to correlate 5-HT₇ receptor distribution with function as there is a significant agreement between the localization of 5-HT₇ receptors and the functions they are implicated in (Hedlund & Sutcliffe, 2004, 2006; Hedlund, 2009). Within the brain, which has attracted the most interest, its presence in the hypothalamus correlates with involvement in circadian rhythm, thermoregulation, and endocrine regulation. Thalamic and cortical 5-HT₇ receptors may be of importance for sleep, mood regulation, and epilepsy. The distribution of 5-HT₇ receptors in the hippocampus is of relevance for its role in learning and memory. Finally, its localization in the spinal cord is in agreement with demonstrated functions in nociception and locomotion.

Interest in the possible involvement of the 5-HT₇ receptor in brain disorders was prompted by the early finding that several antipsychotics (Roth et al., 1994) and antidepressants (Monsma et al., 1993; Mullins et al., 1999) have high affinity for the 5-HT₇ receptor in combination with its presence in relevant regions of the brain. The result is several preclinical studies that have evaluated the possible involvement of the 5-HT₇ receptor in psychiatric disorders and other pathological processes of the nervous system (Hedlund, 2009). These studies have used pharmacological tools and/or constitutive knockout mice lacking 5-HT₇ receptors in animal behavioral models designed to mimic, at least in part, human disorders. The available pharmacological tools and even more under development to study 5-HT₇ receptors are discussed in detail above. Briefly, most studies have used antagonists that are considered relatively selective, the most widely used being SB-269970 (Hagan et al., 2000) and DR4004 (Kikuchi et al., 1999). Constitutive knockout mouse strains lacking the 5-HT₇ receptor have been created independently by several laboratories (Hedlund et al., 2003; Guscott et al., 2005; Sprouse et al., 2005; Witkin et al., 2007). One strain was developed at The Scripps Research Institute (Hedlund et al., 2003) and has been made available to other academic institutions (Landry et al., 2006; Liu et al., 2009). This strain has been bred on a C57BL/6J background. The other published strains have been made by the pharmaceutical industry. As seen below, most evidence has been collected to support a role for the 5-HT₇ receptor in depression. In fact, a recent study suggests that the clinically established antidepressant effect of amisulpride is due to its action at 5-HT₇ receptors (Abbas et al., 2009). Studies aimed at evaluating the role of 5-HT₇ receptors in disorders such as anxiety and schizophrenia have also generated interesting results.

3.2. Anxiety

Efforts to evaluate a possible link between the 5-HT₇ receptor and anxiety have generated mixed results. Two different strains of mice lacking the 5-HT₇ receptor (5-HT₇^−/−) have been evaluated in anxiety models with differing results when compared to 5-HT₇^+/+ mice. Mice bred on a C57BL/6J background were evaluated in a light-dark transfer test. It was found that both genotypes spent the same amount of time in the light compartment and that they had an equal number of transitions between the light and dark compartments (Roberts et al., 2004). The same strain of mice was also evaluated in the marble burying test. Here 5-HT₇^−/− mice buried significantly fewer marbles than the 5-HT₇^+/+ mice, a response considered anxiolytic (Hedlund & Sutcliffe, 2007). A different strain of 5-HT₇^−/− mice bred on a mixed 129SvEv/C57BL/6J background was tested in an elevated plus maze. There was no difference in the time spent exploring the open arms or in the number of entries onto the open arms of the maze between the two genotypes (Guscott et al., 2005). The selective 5-HT₇ receptor antagonist SB-269970 has been evaluated in both rats and mice for potential anxiolytic effect. In rats the drug induced an anxiolytic effect in the Vogel conflict drinking test and the elevated plus maze models of anxiety after both systemic (Wesolowska et al., 2006a) and intra-hippocampal (Wesolowska et al., 2006b) administration. Thus, SB-269970 increased the number of shocks tolerated in the Vogel test. In the elevated plus maze SB-269970 increased both the number of entries onto the open arms and the amount of time spent on those arms. It should be noted that SB-269970 generally had an inverted U-shaped dose-response effect and its anxiolytic effect was not as pronounced as that of diazepam, a reference anxiolytic of the benzodiazepine class (Wesolowska et al., 2006a, 2006b). SB-269970 also had a similar limited anxiolytic effect in the conflict-based four-plate test in mice as the drug increased the number of punished crossings (Wesolowska et al., 2006a). Again the effect was smaller than seen for diazepam. SB-269970 was also effective in the marble burying model (Hedlund & Sutcliffe, 2007).

The clinical relevance of these models for anxiety in humans has been questioned. Traditional benzodiazepine anxiolytics are effective in all the animal models, but only in certain types of human anxiety, whereas chronically administered selective serotonin reuptake inhibitors (SSRI) are effective in most, if not all, forms of human anxiety disorders (Borsini et al., 2002). It appears that inactivation or blockade of the 5-HT₇ receptor more consistently produce an anxiolytic response in models responding to SSRI than those responding only to benzodiazepines. The latter include light-dark transfer and elevated plus maze (Borsini et al., 2002). Particularly interesting is the observation that the 5-HT₇ receptor appears to modulate the response in the marble burying test, as this procedure has been suggested as a model for the specific anxiety disorder obsessive-compulsive disorder (Njung'e & Handley, 1991). The pharmacological treatment of choice for obsessive-compulsive disorder is antidepressants, specifically SSRIs (Heyman et al., 2006). Taken together with the findings using depression models described below the combined data suggest that the 5-HT₇ receptor might be an alternative or complementary target for disorders currently treated with SSRI. It has also been suggested that modulating the 5-HT₇ receptor might be a new approach to design anxiolytic treatment, based on effects within the bed nucleus of the stria terminalis (Guo et al., 2009).

3.3 CNS Development

The developing cortex is densely innervated by serotonergic neurons with 5-HT acting on several different receptor subtypes. It has been shown that the 5-HT₇ receptor is involved in the postnatal formation of synaptic connectivity in the prefrontal cortex, where early postnatally there is a high expression of 5-HT₇ receptors that later declines and is replaced by an increase in 5-HT_1A receptor expression (Beique et al., 2004a,b). The early postnatal appearance of 5-HT₇ receptors in the hypothalamus has also been demonstrated underlining its importance in this region (Russo et al., 2005). Furthermore, serotonergic projections from the dorsal raphe have been shown to modulate cholinergic and noncholinergic neurons within the ventral pallidum through mechanisms that at least partly appear to involve the 5-HT₇ receptor (Bengtson et al., 2004). These studies were made in slices from immature Wistar rats and the findings might be of importance for the etiology of psychiatric disorders (Bengtson et al., 2004). In the developing rat brain 5-HT₇ receptor immunoreactivity has been observed in a cytoplasmic inclusion termed stigmoid body (Muneoka & Takigawa, 2003). In neonatal animals these immuno-positive inclusions were most prominent within the hypothalamus and they have been linked to the development of sexual dimorphism. As discussed further below, the 5-HT₇ receptor is important in regulating spinal locomotion, something that is evident already in neonatal mice (Madriaga et al., 2004).

3.4. Circadian rhythm

Circadian rhythms, sleep and mood are closely linked physiological phenomena that are all regulated by the 5-HT₇ receptor. Soon after the 5-HT₇ receptor was discovered, it was found that 8-OH-DPAT-induced phase resetting within the suprachiasmatic nucleus (SCN) was mediated by this receptor, and not the 5-HT_1A receptor as previously thought (Lovenberg et al., 1993). This reevaluation was due to 8-OH-DPAT previously being considered a selective agonist for 5-HT_1A receptors (Hjorth et al., 1982). An extensive body of work has confirmed the importance of the 5-HT₇ receptor in SCN function (Antle et al., 2003; Glass et al., 2003; Duncan et al., 2004; Sprouse et al., 2004, 2005). That the phase shifting induced by 8-OH-DPAT is mediated by the 5-HT₇ receptor is further supported by the observation that the shift can be inhibited by the selective antagonists SB-269970 and DR-4004, but not by selective 5-HT_1A receptor antagonists (Ehlen et al., 2001; Sprouse et al., 2004). However, in one study the putative 5-HT₇ receptor agonist AS-19 failed to mimic the effect of 8-OH-DPAT after systemic administration (Cuesta et al., 2009). Possible explanations for the inability of AS-19 to induce a phase shift include that only one dose was tested, and/or its lack of selectivity (Bosker et al., 2009; Brenchat et al., 2009). It may also be speculated that AS-19 is metabolically unstable even though a dose-dependent biological effect has been achieved after systemic administration in a study on nociception (Brenchat et al., 2009). Shifting the SCN pacemaker neurons with 8-OH-DPAT is a non-photic stimulus involving serotonergic input from the dorsal and median raphe nuclei. In these nuclei 5-HT₇ receptors have been shown to modulate SCN phase resetting (Glass et al., 2003). There is also evidence to support a role for the 5-HT₇ receptor in photic regulation of the SCN. It has been demonstrated using pharmacological profiling with unselective drugs (Ying & Rusak, 1997) and with DR-4004 (Ehlen et al., 2001) that 5-HT-mediated reduction of photic stimulation of SCN neurons is most likely mediated by the 5-HT₇ receptor. The inhibitory effect of 8-OH-DPAT on spontaneous SCN activity is also most likely mediated by the 5-HT₇ receptor, as it could be blocked by 5-HT₇ receptor antagonists but not 5-HT_1A receptor antagonists (Yu et al., 2001). The above studies were performed in either rats or hamsters. In contrast, one study in the mouse concluded that the effects of 8-OH-DPAT on SCN function was not as pronounced (Antle et al., 2003). Nevertheless, the involvement of the 5-HT₇ receptor in both photic and non-photic phase resetting has been verified in mice lacking the receptor (Gardani & Biello, 2008).

3.5. Depression

It has been suggested that the action of antidepressants, at least in part, might be mediated directly by the 5-HT₇ receptor (Monsma et al., 1993; Mullins et al., 1999). Thus, several antidepressants, both tricyclics and SSRI, induced c-fos expression in a way consistent with 5-HT₇ receptor activation within the SCN (Mullins et al., 1999). The effect on c-fos expression was attenuated after chronic treatment with antidepressants. Additionally, chronic antidepressant drug treatment led to a downregulation of 5-HT₇ receptor binding (Mullins et al., 1999).

The forced swim test and the tail suspension test are two of the most common behavioral models for evaluating the antidepressant potential of a drug or to evaluate the phenotype of transgenic animals with respect to depression (Cryan et al., 2002; Cryan & Holmes, 2005). In both of these models, pharmacological blockade of the 5-HT₇ receptor or inactivation of the receptor gene leads to an antidepressant-like behavioral profile; that is, reduced immobility (Guscott et al., 2005; Hedlund et al., 2005; Wesolowska et al., 2006a, 2006b, 2007; Bonaventure et al., 2007; Sarkisyan et al., 2010). More interestingly, it has been shown that there is a synergistic interaction between an individually ineffective dose of the selective antagonist SB-269970 and an individually ineffective dose of one of several antidepressants, leading to reduced immobility in both the forced swim test and the tail suspension test (Wesolowska et al., 2007; Bonaventure et al., 2007; Sarkisyan et al., 2010). Thus, concurrent administration of citalopram, an antidepressant of the SSRI type, and SB-269970 has been shown to reduce immobility in the tail suspension test in C57BL/6J mice (Bonaventure et al., 2007; Sarkisyan et al., 2010). At a higher dose, SB-269970 alone reduces immobility in the mouse tail suspension test (Hedlund et al., 2005; Bonaventure et al., 2007). The higher dose of SB-269970 alone did not alter 5-HT concentration in the rat frontal cortex, but the combination of a low dose of SB-269970 and a low dose of citalopram increased the level of 5-HT in the frontal cortex (Bonaventure et al., 2007). A similar synergistic interaction between individually ineffective doses of SB-269970 and citalopram has also been demonstrated in the mouse forced swim test (Wesolowska et al., 2007; Sarkisyan et al., 2010). These studies further reported that interactions also occur between SB-269970 and other classes of antidepressants. Thus, ineffective doses of imipramine, desipramine, and moclobemide all reduced immobility in the mouse forced swim test when given in combination with an ineffective dose of SB-269970 (Wesolowska et al., 2007; Sarkisyan et al., 2010). The interaction between SB-269970 and imipramine has also been shown in Wistar rats using the forced swim test in which the effect on immobility was accompanied with an increase in 5-HT levels in the prefrontal cortex (Wesolowska & Kowalska, 2008). Besides the prefrontal cortex, the hippocampus has also been implicated in the effects of SB-269970 and imipramine in the rat forced swim test, as direct intra-hippocampal administration reduced immobility (Wesolowska et al., 2006b).

The validity of the tail suspension test and forced swim test as models of depression has been questioned since they are acute, whereas antidepressant treatment in human usually require several days to weeks before clinical effect is seen (Borsini & Meli, 1988; Cryan & Mombereau, 2004). In order to overcome this, models attempting to induce a depression-like state in laboratory animals have been developed. One such model is chronic unpredictable mild stress (Willner et al., 1992). In a recent study it was shown that 5-HT₇ receptor mRNA was upregulated in the hippocampus and hypothalamus, but not cortex, in rats after exposure to such stress (Li et al., 2009). The change in mRNA levels could be inhibited by treatment with fluoxetine and curcumin, an active ingredient in turmeric extracts (Li et al., 2009).

Indirect evidence that targeting the 5-HT₇ receptor for the treatment of depression might be successful was recently presented in a study showing that the antidepressant effect of amisulpride is most likely mediated by the 5-HT₇ receptor. Amisulpride is an atypical antipsychotic that is also a proven antidepressant (Lecrubier et al., 1997; Smeraldi, 1998). Previously it has generally been thought that the antidepressant effect of amisulpride somehow relied on its properties as a dopamine D₂/D₃ receptor antagonist, although the mechanism had never been satisfactorily explained. It has now been demonstrated that amisulpride has high affinity for the 5-HT₇ receptor and that amisulpride reduces immobility in both the tail suspension test and the forced swim test in 5-HT₇^+/+ mice but not in 5-HT₇^−/− mice (Abbas et al., 2009). As such, these findings provide at least indirect evidence that the antidepressant effect of amisulpride is mediated by the 5-HT₇ receptor. It should also be noted that aripiprazole, another atypical antipsychotic that is successfully used to augment the effect of traditional antidepressants (Berman et al., 2009), has high affinity for the 5-HT₇ receptor (Lawler et al., 1999; Shapiro et al., 2003). In fact, aripiprazole has recently been shown to reduce immobility in the tail suspension and forced swim tests in 5-HT₇^+/+ mice but not in 5-HT₇^−/− mice (Sarkisyan et al., 2010). Interestingly, aripiprazole has, like SB-269970, been shown to enhance the effect of an antidepressant, in this case fluoxetine (Kamei et al., 2008).

Hippocampal neurogenesis is a physiological phenomenon that has attracted considerable interest for its potential importance in depression (Sahay & Hen, 2007), especially since depressed patients have been found to have reduced hippocampal volume (Colla et al., 2007) and since neurogenesis is increased by chronic antidepressant treatment (Malberg et al., 2000; Perera et al., 2007). Thus, it has been speculated that the 5-HT₇ receptor might modulate hippocampal neurogenesis (Nandam et al., 2007). This notion is supported by the finding that the 5-HT₇ receptor modulates neurite growth in hippocampal cell cultures (Kvachnina et al., 2005). However, one study that looked directly at hippocampal cell proliferation did not find any difference between 5-HT₇^+/+ and 5-HT₇^−/− mice (Sarkisyan & Hedlund, 2009). This lack of change could possibly be due to compensatory mechanisms in constitutive knockout mice. The lack of selective 5-HT₇ receptor antagonists that are metabolically stable enough to permit long-term treatment has to date made it impossible to study any possible effect of pharmacological inhibition on hippocampal cell proliferation.

Taken together with findings in sleep studies (see below) there is nevertheless a virtually consistent body of evidence using both pharmacological and genetic tools that implicates the 5-HT₇ receptor in depression, which has led to the hypothesis that new antidepressants could be developed that target the 5-HT₇ receptor (Hedlund, 2009; Mnie-Filali et al., 2007, 2009; Nandam et al., 2007)

3.6. Epilepsy

The first study implicating the 5-HT₇ receptor as a modulator of seizure activity was performed before the availability of selective antagonists (Bourson et al., 1997). Instead an affinity correlation analysis was done showing that the relative affinity of seven compounds for the 5-HT₇ receptor correlated with their ability to protect against audiogenic seizures in DBA/2J mice. The compounds tested had in common that they inhibited 5-HT-stimulated cAMP formation in cells expressing the 5-HT₇ receptor and were thus antagonists at the receptor (Bourson et al., 1997). As outlined above, autoradiographic mapping of 5-HT₇ receptor binding sites has revealed that the highest density of these receptors is found in the thalamus (Bonaventure et al., 2004). It has been hypothesized that the 5-HT₇ receptors in this region might be of importance for epilepsy, especially in the WAG/Rij rat model of absence epilepsy (Graf et al., 2004). These particular rats exhibit spontaneously occurring spike-wave discharges that can be exacerbated by 8-OH-DPAT. With the realization that 8-OH-DPAT is an agonist not only for 5-HT_1A receptors, but also for 5-HT₇ receptors, it is likely that at least part of the potentiating effect of 8-OH-DPAT on epileptic activity is mediated by the 5-HT₇ receptor. It was further shown that the selective 5-HT₇ receptor antagonist SB-269970 reduced spontaneous epileptic activity in WAG/Rij rats (Graf et al., 2004).

In a study investigating both electrically- and chemically-induced tonic-clonic seizures it was found that the seizure threshold was lower in 5-HT₇ receptor knockout mice bred on a CD-1 background (Witkin et al., 2007). Using transcorneal electrical stimulation it was found that the 5-HT₇^−/− mice had lower tonic but not clonic seizure threshold. For chemically induced seizures it was found that 5-HT₇^−/− mice exhibited lower thresholds for seizures induced by the GABA antagonist pentylenetetrazole, as well as for seizures induced by cocaine. Another study found that 5-CT, a potent 5-HT_1/7 receptor agonist, increased the threshold for seizures chemically induced with picrotoxin (Pericic & Svob Strac, 2007). The effect was present in both previously stressed and unstressed animals. The 5-CT-induced increase of the seizure threshold for picrotoxin in stressed mice could be counteracted with SB-269970, but not with the selective 5-HT_1A receptor antagonist WAY-100635. It was thus concluded that the 5-CT effect was mediated by the 5-HT₇ receptor (Pericic & Svob Strac, 2007). Finally, it has been suggested that the 5-HT₇ receptor is involved in regulating tonic-clonic seizure-induced antinociception (Freitas et al., 2009).

In summary, it is evident that further studies are needed to sort out the role of the 5-HT₇ receptor as being either proconvulsant or anticonvulsant depending on the type of epilepsy, species, or 5-HT₇ receptor manipulation studied.

3.7. Learning and memory

An early attempt to assess the role of the 5-HT₇ receptor in various behavioral paradigms used antisense oligonucleotides to inhibit receptor synthesis (Clemett et al., 1998). This report found no effects of the treatment in feeding, locomotor activity or anxiety-like behavior using an elevated plus maze. A more comprehensive study used 5-HT₇ receptor knockout mice to evaluate the role of this receptor in various behavioral and learning tasks (Roberts et al., 2004). It was found that mice lacking the 5-HT₇ receptor had a specific impairment in contextual fear conditioning. Contextual fear conditioning is generally believed to be a hippocampus-dependent learning, as are other types of place learning. It is therefore of interest that a first Barnes maze test included in this study showed no difference between wild-type and knockout mice (Roberts et al., 2004). However, in a later more extensive variant of the Barnes maze task a spatial learning deficit was revealed in the 5-HT₇^−/− mice (Sarkisyan & Hedlund, 2009). This later study also found novel location recognition in the knockout mice, another measurement of hippocampus-dependent learning. The conclusion drawn was that the 5-HT₇^−/− mice had a specific impairment that reduced their ability to respond to spatial changes in their environment. The two studies did not find any differences in hippocampus-independent learning tasks: cued fear conditioning, operant food conditioning, motor learning (rotarod), and novel object recognition (Roberts et al., 2004; Sarkisyan & Hedlund, 2009). The impairments seen were determined not to be due to alterations in motor skills, visual acuity or anxiety level. The finding that impaired contextual fear conditioning is an early phenomenon in models of Alzheimer’s disease opens up the intriguing possibility for an involvement of the 5-HT₇ receptor in this disorder (Corcoran et al., 2002; Dineley et al., 2002). A possible electrophysiological correlate to the hippocampus-dependent deficits has been established, as there was a reduced ability to induce LTP in the CA1 region of the hippocampus in 5-HT₇^−/− mice (Roberts et al., 2004). The 5-HT₇ receptor has also been implicated in memory formation in a Pavlovian learning test as the selective antagonists SB-269970 and DR4004 could reverse amnesia induced by scopolamine and dizocilpine (Meneses, 2004). Also in support of a pro-cognitive effect of the 5-HT₇ receptor it has been shown that the putative agonist AS-19 enhances memory formation and that this effect can be reversed by SB-269970 (Perez-García & Meneses, 2005). Attempts to reconcile the somewhat conflicting results on the 5-HT₇ receptor in learning and memory are ongoing with more work still needed (Perez-Garcia & Meneses, 2009).

3.8. Locomotion

As mentioned above, the 5-HT₇ receptor has been found to be involved in spinal locomotion already in early development (Madriaga et al., 2004). A series of studies have attempted to characterize this involvement in more detail, again using both pharmacological tools and knockout mice. It has been demonstrated that 5-HT has an important role in the central pattern generator for locomotion, parts of which are located within the spinal cord (Jordan et al., 2008). Serotonergic neurons located in the parapyramidal region (PPR) produce locomotion when stimulated. These neurons constitute the first anatomically discrete group of spinally-projecting neurons demonstrated to be involved in the initiation of locomotion in mammals. It has been shown that locomotion evoked from the PPR is mediated by 5-HT₇ and 5-HT_2A receptors (Jordan et al., 2008). Furthermore, it has been demonstrated that 5-HT₇ receptor antagonists block locomotion in cat, rat and mouse preparations, but that they have little effect in mice lacking 5-HT₇. In 5-HT₇ receptor knockout mice 5-HT induced rhythmic activity, but coordination among flexor and extensor motor nuclei and left and right sides of the spinal cord is disrupted. In adult 5-HT₇^+/+ mice, 5-HT₇ receptor antagonists impair locomotion, producing patterns of activity resembling those induced by 5-HT in 5-HT₇^−/− mice (Jordan et al., 2008; Liu et al., 2009). Moreover, 5-HT₇^−/− mice displayed greater maximal extension at the hip and ankle joints than 5-HT₇^+/+ mice during voluntary locomotion (Liu et al., 2009). Further evidence that the 5-HT₇ receptor is involved in spinal central pattern generator was obtained in a study examining electrically stimulated locomotion (Dunbar et al., 2010).

Knowledge of the mechanisms involved in spinal locomotion is of extreme importance when trying to find ways to restore function after spinal cord injury. In mice with the spinal cord transected at the thoracic level it has been shown that 8-OH-DPAT administered intraperitoneally acutely induced, within 15 min, hindlimb movements that share some characteristics with normal locomotion (Landry et al., 2006). Paraplegic mice pretreated with the selective 5-HT_1A receptor antagonists WAY-100135 or WAY-100635 displayed significantly less 8-OH-DPAT-induced movement. A similar reduction of 8-OH-DPAT-induced movements was found in animals pretreated with SB269970. Moreover, a near complete blockade of 8-OH-DPAT-induced movement was observed in 5-HT₇^+/+ mice pretreated with 5-HT_1A and 5-HT₇ receptor antagonists, and in 5-HT₇^−/− mice pretreated with 5-HT_1A receptor antagonists (Landry et al., 2006). The overall conclusion of these studies was that 8-OH-DPAT potently induces locomotor-like movement in the previously paralyzed hindlimbs of low-thoracic-transected mice and that both 5-HT_1A and 5-HT₇ receptors are involved in spinal locomotor rhythmogenesis in vivo. In a related field it has been found that the 5-HT₇ receptor is also involved in modulating phrenic nerve motor output, a finding that might also be of relevance for spinal cord injury (Macfarlane & Mitchell, 2009).

3.9. Migraine

The ability of 5-HT to induce cranial vasodilation under certain conditions is thought to be one of the mechanisms involved in migraine (Saxena & Ferrari, 1989). Although the effects of 5-HT are most likely not mediated by a single receptor subtype (Ullmer et al., 1995; Schoeffter et al., 1996; Verheggen et al., 2004), available evidence clearly suggest a role for the 5-HT₇ receptor. Initial interest in the 5-HT₇ receptor was driven by the observation that several migraine prophylactic drugs showed moderate to high affinity for the 5-HT₇ receptor (Bard et al., 1993; Ruat et al., 1993; Shen et al., 1993). Thus, it has been proposed that the 5-HT₇ receptor mediates 5-HT-induced dilation of the carotid artery following blockade of 5-HT_1B/1D receptors in combination with low sympathetic tone (Villalon et al., 1997). It is, however, not clear how relevant such a mechanism might be for migraine. Before the availability of selective 5-HT₇ receptor antagonists, pharmacological profiling was used to implicate the 5-HT₇ receptor as the most likely mediator of 5-HT induced vasodilation in preparations of the basilar and middle cerebral arteries (Terrón & Falcón-Neri, 1999). Later involvement of the 5-HT₇ receptor was confirmed since such vasodilation could be blocked in vivo by SB-269970 (Terrón & Martínez-García, 2007). This has been shown to also be true after 5-HT depletion, suggesting that the 5-HT₇ receptors involved are not sensitized by reduced availability of 5-HT (Martínez-García et al., 2009). The 5-HT₇ receptor has also been shown to mediate relaxation in a pial vein preparation (Ishine et al., 2000). As it is very difficult to model migraine in laboratory animals (Hess, 1996), further studies are needed to determine if the 5-HT₇ receptor can be targeted for the prophylaxis or treatment of migraine.

3.10. Pain

Several studies have investigated a possible role for the 5-HT₇ receptor in pain. Pain mediation and processing occur at many levels of the nervous system. Several of these, including peripheral, spinal, and thalamic mechanisms, have been evaluated with regard to the 5-HT₇ receptor. For peripheral pain, the 5-HT₇ receptor appears to mediate a pronociceptive effect that can be inhibited by SB-269970 (Rocha-González et al., 2005). This study used injections of formalin in the paw to induce flinching behavior. It was found that SB-269970 had a direct local peripheral antinociceptive effect when the drug was injected into the paw prior to formalin. Furthermore, SB-269970 could locally counteract the pronociceptive effect of 5-CT. When given intrathecally, SB-269970 had no direct effect on flinching, but could reverse the pronociceptive effect of 5-CT.

When additional studies performed on spinal and thalamic nociception are considered, the interpretation of 5-HT₇ receptor function becomes more complex. One study using mice lacking 5-HT₇ receptors did not see a difference in spinal pain mediation using the tail-flick test (Roberts et al., 2004). In another study also using the tail-flick model in mice, it was found that intrathecally administered SB-269970 blocked the analgesic effect of morphine given subcutaneously, thus suggesting that the antinociceptive activity of descending serotonergic pathways is mediated by 5-HT₇ receptors (Dogrul & Seyrek, 2009). Similar conclusions were drawn in a study assessing the effects of spinal microinjections of morphine and SB-269970 using a paw-flick model (Dogrul et al., 2009). It should be noted, that the anatomical localization of 5-HT₇ receptors within the spinal cord supports such an interpretation (Doly et al., 2005). Another study also showed that the antinociceptive properties of spinal serotonergic pathways were mediated by the 5-HT₇ receptor (Brenchat et al., 2009). Using both the putative agonist AS-19 and SB-269970, it was found that the 5-HT₇ receptor modulates capsaicin-induced mechanical hypersensitivity in mice. Thus, the agonist had a dose-dependent antinociceptive effect that could be counteracted by the antagonist (Brenchat et al., 2009). Capsaicin was injected in the plantar surface of a paw, which was then mechanically stimulated. The serotonergic compounds evaluated were administered systemically by subcutaneous injections. In a different study, rats were exposed to an electric shock to the tail. Here, it was found that 8-OH-DPAT had an antinociceptive effect when injected into the medial thalamus, specifically the nucleus parafascicularis and the central lateral thalamic nucleus (Harte et al., 2005). This action could be counteracted by selective antagonists for both the 5-HT_1A and the 5-HT₇ receptor. The effect was observed for pain behaviors organized at the medullary and forebrain levels, but not at the spinal level, of the neuraxis (Harte et al., 2005). As noted above, in a possible link between 5-HT₇ receptors, epilepsy, and pain modulation it has been suggested that 5-HT₇ receptors are involved in regulating tonic-clonic seizure-induced antinociception (Freitas et al., 2009). Taken together, available data appear to suggest differing roles for the 5-HT₇ receptor in peripheral versus central pain mediation. Future studies will in all likelihood help to clarify fully the relationship between the 5-HT₇ receptor and pain.

3.11. Schizophrenia

Soon after the 5-HT₇ receptor was discovered a screen of several drugs of different classes was made to determine their affinity for the new receptor. Among the findings was the discovery that several antipsychotics had high affinity for the 5-HT₇ receptor (Roth et al., 1994), thus opening up the possibility that some of their effects may be mediated by this receptor. Even though certain typical antipsychotics were included in this group, particularly high affinities were seen for atypical antipsychotics such as clozapine and risperidone. It was thus speculated that some of the unique properties (e.g. the lack of extrapyramidal side effects) of these drugs could be attributed to an action at the 5-HT₇ receptor (Roth et al., 1994). More recently another drug, amisulpride, with both antipsychotic and antidepressant properties has also been shown to have high affinity for the 5-HT₇ receptor (Abbas et al., 2009). As discussed above, action at the 5-HT₇ receptor is most likely mediating the antidepressant effect of amisulpride, but might also play a role in its antipsychotic action. Several studies have attempted to test the possible relevance of 5-HT₇ receptors in schizophrenia using prepulse inhibition (PPI). This is a particularly attractive model as it can be used in both humans and laboratory animals. Even though PPI does not reflect all modalities of schizophrenia, it provides a model of the sensorimotor gating deficits seen in schizophrenia patients (Geyer et al., 2001). Of particular interest for the evaluation of antipsychotics, possible new treatments, and new treatment mechanisms is the ability to pharmacologically disrupt PPI. Thus, it has been possible to identify dopaminergic, serotonergic, and glutamatergic components of PPI. Of these it seems that the dopaminergic mechanisms are most relevant for the action of typical antipsychotics and the glutamatergic mechanisms are most relevant for atypical antipsychotics (Geyer et al., 2001). Using the selective 5-HT₇ receptor antagonist SB-258741 (Lovell et al., 2000), it has been shown that the 5-HT₇ receptor can modulate behavior in the PPI model in Wistar rats (Pouzet et al., 2002). Specifically, it was found that SB-258741 did not affect amphetamine-induced disruption of PPI, but that phencyclidine (PCP)-induced disruption was counteracted by SB-258741. Thus, the conclusion was that the 5-HT₇ receptor could influence the glutamatergic, but not the dopaminergic, component of PPI (Pouzet et al., 2002). Results in support of that notion were for the most part obtained in a study using 5-HT₇^−/− mice (Semenova et al., 2008). Thus, the ability of the dopaminergic drugs apomorphine and amphetamine to disrupt PPI was unaltered in 5-HT₇^−/− mice compared to 5-HT₇^+/+ mice, whereas the ability of PCP to disrupt PPI was diminished in the 5-HT₇^−/− mice. However, this study did not find any influence of SB-269970 on PCP-disrupted PPI in either C57BL/6J mice (the background strain of the 5-HT₇^−/− mice) or Wistar rats (Semenova et al., 2008). This and another study using a different strain of 5-HT₇^−/− mice did not observe any inherent differences in PPI in mice lacking the 5-HT₇ receptor (Guscott et al., 2005; Semenova et al., 2008). A recently published study made somewhat conflicting observations using C57BL/6J mice (Galici et al., 2008). This report described an effect also on the dopaminergic component of PPI. Thus, SB-269970 was able to reverse amphetamine-induced disruption of PPI in C57BL/6J mice. Again, no effect of SB-269970 on the glutamatergic component of PPI was observed as evaluated using ketamine-induced disruption (Galici et al., 2008).

In a study of humans, it has been shown that 5-HT₇ receptor mRNA is downregulated in the dorsolateral prefrontal cortex, but not the hippocampus, of individuals with schizophrenia (East et al., 2002). A study analyzing a possible correlation between single nucleotide polymorphisms in the 5-HT₇ receptor gene and schizophrenia found a positive association for two of four evaluated polymorphisms within cohorts of 383 Japanese schizophrenia patients and 351 controls (Ikeda et al., 2006). It is unlikely, however, that the polymorphisms in question cause any differences in 5-HT₇ receptor function. In another population study no correlation was found between single nucleotide polymorphisms in the 5-HT₇ receptor gene and the clinical efficacy of risperidone (Wei et al., 2009), an antipsychotic with high affinity for the 5-HT₇ receptor (Roth et al., 1994).

3.12. Sleep

Among the clinical manifestations of depression is often a dysregulated circadian rhythm and sleep disturbances (Belmaker & Agam, 2008). Specifically, a decreased latency to and increased amount of rapid eye movement (REM) sleep can often be detected in EEG recordings (Yadid et al., 2000). Direct involvement of the 5-HT₇ receptor in sleep regulation has been shown using selective antagonists and knockout mice. Both SB-269970 and SB-656104, a structurally related selective antagonist (Thomas et al., 2003), when administered to Sprague-Dawley rats at the beginning of the light period, increased the latency to REM sleep and decreased the amount of time spent in REM sleep (Hagan et al., 2000; Thomas et al., 2003). Other sleep parameters were not affected. It has further been shown that mice lacking the 5-HT₇ receptor exhibit a similar reduction in time spent in REM sleep during the light period, again without affecting other sleep phases (Hedlund et al., 2005). The 5-HT₇^−/− mice also had less frequent and longer REM episodes than the 5-HT₇^+/+ mice. In the 5-HT₇^−/− mice there was no change in latency to REM and citalopram was equally effective in increasing the latency to REM in both genotypes (Hedlund et al., 2005). Of considerable interest is that, as in the depression models discussed above, an interaction has been observed between individually ineffective doses of SB-269970 and citalopram also for sleep where such a combination of drugs resulted in an increase in latency to REM and a decrease in the amount of REM sleep in rats (Bonaventure et al., 2007). A more detailed analysis revealed that the decrease in the amount of REM sleep was due to a reduced number of REM episodes. Citalopram alone caused an increased fragmentation of sleep as seen from an increase in the number of microarousals. This fragmentation could be reversed by SB-269970 (Bonaventure et al., 2007). Furthermore, three antidepressants of the SSRI class (citalopram, fluoxetine, and paroxetine), but not a tricyclic antidepressant (desipramine), have been demonstrated to augment the effects on latency to REM sleep and REM sleep duration seen in 5-HT₇^−/− mice (Shelton et al., 2009). Again the reduction in the amount of REM sleep was caused by a reduced number of REM episodes. A recent study found that microinjections of either the putative 5-HT₇ receptor agonist LP-44 or SB-269970 into the dorsal raphe nucleus both reduced REM sleep and the number of REM episodes when injected during the light cycle (Monti et al., 2008). Interestingly, the LP-44 effect could be counteracted by pretreatment with SB-269970. Further studies are clearly needed to fully clarify the interplay between 5-HT₇ receptor agonism and antagonism on sleep. In view of the importance of the 5-HT₇ receptor in the hippocampus and its effect on REM sleep it is of interest to note that the levels of 5-HT are influenced by the different sleep phases in this region of the brain (Peñalva et al., 2003).

Sleep is also often disturbed in schizophrenia. It is interesting to note that atypical antipsychotics with high affinity for the 5-HT₇ receptor, such as clozapine and risperidone, show the greatest improvements on sleep parameters (Cohrs, 2008; Roth et al., 1994). This is another area that warrants further study.

3.13. Substance abuse

A possible link between the 5-HT₇ receptor and substance abuse is an area that remains to be explored. Nevertheless, a phenomenon that is closely linked with drug addiction is novelty-seeking behavior and it has recently been hypothesized that the 5-HT₇ receptor influences such behavior (Ballaz et al., 2007a, 2007b). In a first study Sprague-Dawley rats were classified as high or low responders when measuring the amount of locomotor activity after the animals were placed in an enclosed open arena (Ballaz et al., 2007a). There were differences in 5-HT₇ receptor mRNA expression in several brain regions between low and high responders. Notably, there was higher expression in the hippocampus in the low responding rats. This finding was interpreted to indicate that low levels of 5-HT₇ mRNA expression correlated with decreased aversion to forced exposure to novelty (Ballaz et al., 2007a). In a follow-up study it was found that the low responding rats showed increased exploration of a new object in a novel object discrimination task and that this increase could be diminished by SB-269970 (Ballaz et al., 2007b). It should, however, be noted that mice lacking the 5-HT₇ receptor did not differ from 5-HT₇^+/+ mice in novel object recognition, but instead exhibited reduced novel location recognition which is a more hippocampus-dependent behavior (Sarkisyan & Hedlund, 2009).

In a more direct model of impulsivity, delayed reinforcement, with possible relevance for drug abuse and attention deficit hyperactivity disorder, it was found that rats treated with methylphenidate during adolescence showed reduced impulse behavior as adults (Leo et al., 2009). This behavior could be counteracted by SB-269970 administered at the time of testing. The observed behavioral changes are possibly related to changes in gene expression since methylphenidate has been shown to upregulate 5-HT₇ mRNA expression in the striatum (Adriani et al., 2006) and nucleus accumbens (Leo et al., 2009). For now it is unclear if and how these results relate to each other. Thus, any possible relevance for the 5-HT₇ receptor in novelty seeking, impulsivity, and substance abuse remains to be fully determined.

3.14. Thermoregulation

That 5-HT is involved in the regulation of body temperature is a well-known phenomenon. Injection of 5-CT or 8-OH-DPAT will induce hypothermia. It has been established that the effect is similar regardless of whether the drug is administered systemically or directly into the brain. Thus, there is in all likelihood a central mechanism of action. As 5-CT and 8-OH-DPAT are both 5-HT_1A receptor agonists, this receptor was generally considered the main mediator of the hypothermia, although some reports had suggested involvement of other receptor subtypes. The first evidence that the 5-HT₇ receptor is of importance for 5-HT-induced hypothermia came when it was reported that the effect of 5-CT on body temperature in guinea pigs could be blocked by SB-269970 (Hagan et al., 2000). Such an antagonistic effect has also been observed for SB-656104, another selective 5-HT₇ receptor antagonist (Thomas et al., 2003). It was not possible to block the hypothermic effect of 5-CT in mice with the 5-HT_1A receptor antagonist WAY-100635 or the 5-HT_1B/D antagonist GR127935 (Guscott et al., 2003). In contrast, the hypothermic effect of 8-OH-DPAT in rats could be inhibited by a 5-HT_1A antagonist (WAY-100135), but only partially by SB-269970 (Hedlund et al., 2004). The importance of the 5-HT₇ receptor was further emphasized when it was observed that 5-HT and 5-CT failed to induce hypothermia in 5-HT₇ receptor knockout mice (Guscott et al., 2003; Hedlund et al., 2003). When including 8-OH-DPAT in the analysis in combination with selective antagonists and knockout mice to discriminate between 5-HT_1A and 5-HT₇ receptors it was revealed that both receptor subtypes are involved in 5-HT-mediated hypothermia (Hedlund et al., 2004). These findings provide an explanation for the results obtained with antagonists and also why 8-OH-DPAT fails to induce hypothermia in 5-HT_1A receptor knockout mice (Heisler et al., 1998). LP-211 has been found to induce hypothermia in in mice. The hypothermic effect could be inhibited by SB-269970, but not WAY-100135, and was absent in 5-HT₇ receptor knockout animals (Hedlund et al., 2010). Interestingly, the 5-HT₇ receptor seems to be most important at low agonist concentrations, thus contributing to the fine-tuning of temperature homeostasis, whereas the 5-HT_1A receptor comes into play at higher agonist concentrations possibly providing a defense against hyperthermia (Hedlund et al., 2004).

4. Conclusion

Through the combined use of modern molecular biology, knockout animal models, and other more traditional research methods such as medicinal chemistry and classical pharmacology, a clearer picture of the pathophysiological role of the 5-HT₇ receptor is emerging. The availability of selective antagonists and knockout mice strains has significantly increased our knowledge about this receptor. Together with nonselective agonists, these tools have helped to reveal the 5-HT₇ receptor distribution and function in more detail. Important functional roles for the 5-HT₇ receptor in thermoregulation, circadian rhythm, hippocampal signaling and sleep have been established. Hypotheses driving current research indicate that 5-HT₇ receptor might be involved in mood regulation, suggesting that the 5-HT₇ receptor is a target in the treatment of depression. The recent availability of more selective 5-HT₇ receptor agonists is opening new perspectives about the involvement of 5-HT₇ receptors in learning and memory, and in the control of nociception. Hopefully, these early results will be substantiated by future studies, especially in vivo, that will use newer selective antagonists and agonists endowed with optimal druglike properties. While the identification of selective antagonists would take advantage from a large body of medicinal chemistry literature about the design of selective antagonists, the development of selective agonists will require more extensive studies because the structural features that differentiate a selective agonist versus a nonselective one are yet to be identified.

Table 7.

Localization and abundance of 5-HT receptor mRNA in the rat

Technique	Localization (Relative Abundance)
Northern Blota	hypothalamus (+++), hippocampus (++), mesencephalon (++), cortex (++), olfactory bulb (+), olfactory tubercle (+), spleen (+) (not detected in retina, pituitary, testis, stomach, prostate, ovary, skeletal muscle, lung, liver, kidney, gut)
Northern Blotb	hypothalamus (+++), brainstem (+++), hippocampus (+++), stomach (+),ileum (+)
In situ hybridizationb	retrosplenial cortex (+++), hippocampus (+++), tenia tecta (+++), indusium griseus (+++), posterior hypothalamus (+++), medial amygdala nucleus (+++), thalamus (+++), cerebellum-Purkinije cell layer- (+++), pontine nuclei (+++), superior colliculus (+), raphe nucleus (+)
Northern Blotc	hypothalamus (+++), thalamus (+++), hippocampus (+), cortex (+) (not detected in urinary bladder, testis, liver, spleen, adrenal gland, kidney, lung, heart, pituitary)
In situ hybridizationc	hippocampus (+++), thalamus (+), hypothalamus (+)
Northern Blotd	hypothalamus (+++), thalamus (+++), hippocampus (++), cortex (++), medulla (++) (not detected in cerebellum, striatum, heart, liver, kidney, adrenal glands, testis, ovaries, spleen)
In situ hybridizatione	hippocampus, thalamus, enthorinal and piriform cortices, tenia tecta
In situ hybridizationd	thalamus (+++), hippocampus (++), retrosplenial cortex (++), neocortex (++), hypothalamus (+) (not detected in suprachiasmatic nucleus)
RT-PCRf	lumbar dorsal root ganglia, superior cervical ganglia, lumbar synpathetic ganglia
RT-PCRg	vena cava (+++), femoral vein (++), aorta (+), renal artery (+), portal vein (+) (not detected in jugular vein)
In situ hybridizationh	outer layer of the cortex, thalamus, hippocampus
In situ hybridizationi	thalamus (+++), hippocampus (+++), retrosplenial cortex (+++), mammillary region (+++), posterior thalamic nucleus (+), suprachiasmatic nucleus (+)
RT-PCRj	frontocortical astrocytes, hypothalamus
RT-PCRk	adrenal gland
RT-PCRl	submandibular gland
RT-PCRm	thymus, peripheral blood lymphocytes, spleen, mitogen activate spleen cells
RT-PCRn	adrenal cortex
In situ hybridizationo	forebrain (+++), olfactory complex (+++), thalamus (+++), hippocampal formation (+++), hypothalamus (+++), suprachiasmatic nucleus (+++), septal region(++), amygdala (++), parvicellular (++)

^aShen et al. (1993);

^bRuat et al. (1993);

^cMeyerhof et al. (1993);

^dLovenberg et al. (1993);

^eKinsey et al. (2001);

^fPierce et al. (1997);

^gUlmer et al. (1995);

^hMengod et al. (1996);

ⁱVenero et al. (1997);

^jShimizu et al. (1996);

^kContesse et al. (1999);

^lBourdon et al. (2000);

^mStefulj et al. (2000);

ⁿLenglet et al. (2000);

^oNeumaier et al. (2001)

Abbreviations

AC

adenylyl cyclase

CoMFA

comparative molecular field analysis

GPCR

G protein-coupled receptor

PPI

prepulse inhibition

REM

rapid eye movement

SCN

suprachiasmatic nucleus

SSRI

selective serotonin reuptake inhibitor

TMH

transmembrane helix