- Top of page
The effects of the neurotransmitter 5-hydroxytryptamine (5-HT) are mediated through at least fourteen distinct receptors. The known 5-HT receptors comprise a ligand gated ion channel (the 5-HT3 receptor) and 13 G-protein coupled receptors (Hoyer et al., 1994; Boess & Martin, 1994).
The 5-HT6 receptor was first isolated from rat striatal mRNA by reverse transcription and polymerase chain reaction, with degenerate oligonucleotide primers derived from conserved regions of known G-protein coupled receptors (Monsma et al., 1993) or by low stringency screening with probes derived from the rat histamine H2 receptor (Ruat et al., 1993). Subsequently, the human 5-HT6 receptor has been isolated (Kohen et al., 1994). In rats the highest levels of 5-HT6 receptor mRNA are present in olfactory tubercle, nucleus accumbens, striatum and hippocampus (Monsma et al., 1993; Ruat et al., 1993; Ward et al., 1995; Gerard et al., 1996). In addition to these regions, 5-HT6 receptor-like immunoreactivity was also identified in frontal and entorhinal cortex and the molecular layer of the cerebellum (Gerard et al., 1997).
The 5-HT6 receptor is positively coupled to adenylyl cyclase and can be radiolabelled with [125I]-lysergic acid diethylamide (LSD), [3H]-LSD and [3H]-5-HT (Monsma et al., 1993; Boess et al., 1997). Many non-selective compounds such as tricyclic antidepressants, antipsychotic agents, tryptamine and ergoline derivatives bind to the 5-HT6 receptor with high affinity (Monsma et al., 1993; Roth et al., 1994; Boess et al., 1997) but to date no selective ligands for the 5-HT6 receptor have been described. Therefore, the identification of functional 5-HT6 receptors in physiological preparations could only be tentative based on the rank order of potency non-selective compounds (Sleight et al., 1997).
The only study exploring the functional significance of the receptor in vivo used antisense-oligonucleotides which should abolish or reduce the expression of the 5-HT6 receptor protein. Intracerebroventricular treatment of rats with 5-HT6 specific antisense-oligonucleotides produced a behavioural syndrome consisting of yawning, stretching and chewing which could be antagonized by atropine but not by haloperidol (Bourson et al., 1995; Sleight et al., 1996).
The further study of 5-HT6 receptors and their physiological function requires potent and selective ligands for the receptor. Here we characterize 2,6-dimethylamino-4-sulphanilamidopyrimidin (Ro 04–6790) and 2,6-dimethylamino-4-sulphanilamido-pyridine (Ro 63–0563) as potent and selective antagonists at the 5-HT6 receptor (Figure 1). These compounds were also used to determine whether 5-HT6 antagonists could produce the same behavioural syndrome as that induced by treatment with antisense oligonucleotides in vivo.
- Top of page
[3H]-LSD labelled recombinant rat 5-HT6 receptors with a dissociation constant (Kd) of 1.9 ± 0.1 nM and recombinant human 5-HT6 receptors with a Kd of 1.6 ± 0.1 nM. As can be seen in Figure 2, both Ro 04–6790 and Ro 63–0563 displaced [3H]-LSD from both the human and the rat 5-HT6 receptor with high affinity. The affinity values (mean ± s.e.mean pKi) of Ro 04–6790 for the rat and human 5-HT6 receptor were 7.35 ± 0.04 and 7.26 ± 0.06, respectively (Table 2). Ro 63–0563 had higher affinity than Ro 04–6790 for the rat and human 5-HT6 receptors with pKi values of 7.83 ± 0.01 and 7.91 ± 0.02, respectively (Table 2). The curves shown in Figure 2 represent the means of 3 experiments for both compounds.
Figure 2. Displacement of [3H]-LSD from rat recombinant 5-HT6 receptors by Ro 04–6790 and Ro 63–0563 and from human recombinant 5-HT6 receptors by Ro 04–6790 and Ro 63–0563. Each data point represents the mean of 3 separate competition experiments; vertical lines show s.e.mean.
Download figure to PowerPoint
Table 2. Affinities of Ro 04–6790 and Ro 63–0563 for various binding sites
|Receptor||Ro 04–6790||Ro 63–0563|
|h5-HT6||7.26 ± 0.06||7.91μ 0.02|
|r5-HT6||7.35 ± 0.04||7.83 ± 0.01|
|h5-HT1A||IC50 > 10 μM||IC50 > 10 μM|
|5-HT1B||IC50 > 10 μM||IC50 > 10 μM|
|h5-HT1D||IC50 > 10 μM||IC50 > 10 μM|
|h5-HT2A||IC50 > 10 μM||5.32 ± 0.02|
|h5-HT2C||IC50 > 10 μM||5.69 ± 0.01|
|5-HT3||IC50 > 10 μM||IC50 > 10 μM|
|5-HT4||ND||IC50 > 10 μM|
|h5-HT7||IC50 > 10 μM||IC50 > 10 μM|
|hD12||IC50 > 10 μM||IC50 > 10 μM|
|hD2||IC50 > 10 μM||IC50 > 10 μM|
|hD3||IC50 > 10 μM||IC50 > 10 μM|
|hD4||IC50 > 10 μM||IC50 > 10 μM|
|hD5||IC50 > 10 μM||IC50 > 10 μM|
|α1 -Adrenoceptor||IC50 > 10 μM||IC50 > 10 μM|
|α2-Adrenoceptor||IC50 > 10 μM||IC50 > 10 μM|
|β-Adrenoceptor||IC50 > 10 μM||IC50 > 10 μM|
|hM1||IC50 > 10 μM||IC50 > 10 μM|
|hM2||IC50 > 10 μM||IC50 > 10 μM|
|hM3||IC50 > 10 μM||IC50 > 10 μM|
|hM4||IC50 > 10 μM||IC50 > 10 μM|
|hM5||IC50 > 10 μM||IC50 > 10 μM|
|r μ-Opioid||IC50 > 10 μM||IC50 > 10 μM|
|r k-Opioid||IC50 > 10 μM||IC50 > 10 μM|
|Benzodiazepine||IC50 > 10 μM||IC50 > 10 μM|
Ro 04–6790 has over 100 fold selectivity for the 5-HT6 receptor with respect to the other receptor binding sites examined. We were unable to measure the affinity of Ro 04–6790 for any of the other 23 receptor binding sites studied (Table 2), since it failed to displace more than 50% of the specific binding at a concentration of 10 μM. Indeed, for all of the other 5-HT receptors tested there was no displacement of specific binding by 10 μM Ro 04–6790.
Ro 63–0563 also had over 100 fold selectivity for the 5-HT6 receptor with respect to the other binding sites tested. The only other receptors for which Ro 63–0563 had any measurable affinity were the 5-HT2A and the 5-HT2C receptors with pKi values of 5.32 ± 0.02 and 5.69 ± 0.01, respectively (Table 2). In all other binding assays, Ro 63–0563 failed to displace more than 50% of the specific binding at a concentration of 10 μM. This was also true for all the binding sites tested at CEREP (see Methods section).
At the human 5-HT6 receptor, neither Ro 04–6790 nor Ro 63–0563 had any agonist activity (Figure 3), since they had no effect on the accumulation of cyclicAMP. In addition no inverse agonist activity was seen in that the compounds themselves did not significantly reduce basal accumulation of cyclicAMP (Figure 3). However, both Ro 04–6790 and Ro 63–0563 behaved as competitive antagonists. Ro 04–6790 had a pA2 value of 6.75 ± 0.07 calculated from Schild regressions with a slope of 1.16 ± 0.04. The pA2 value of Ro 63–0563 at the 5-HT6 receptor was 7.10 ± 0.09 calculated from Schild regressions with a slope of 1.14 ± 0.13. All of these values are means ± s.e.mean of 3 separate experiments each with 4 concentrations of each antagonist. Figure 4 shows a typical example of concentration-response curves to 5-HT in the presence of different concentrations of Ro 04–6790 and, likewise, Figure 5 shows a typical example of concentration-response curves to 5-HT in the presence of different concentrations of Ro 63–0563.
Figure 3. Effect of 5-HT, Ro 04–6790 and Ro 63–0563 on cyclicAMP accumulation in Hela cells stably expressing the recombinant human 5-HT6 receptor. Data points represent the means of at least 3 separate experiments.
Download figure to PowerPoint
Figure 4. Stimulation of cyclicAMP accumulation in Hela cells stably expressing the human 5-HT6 receptor by either 5-HT alone or 5-HT in the presence of Ro 04–6790 (1, 3, 10 or 30 μM). This figure shows a typical example of 3 separate experiments. Inset, Schild regression of the data.
Download figure to PowerPoint
Figure 5. Stimulation of cyclicAMP accumulation in Hela cells stably expressing the human 5-HT6 receptor by either 5-HT alone or 5-HT in the presence of Ro 63–0563 (0.1, 0.3, 1 or 3 μm). This figure shows a typical example of 3 separate experiments. Inset, Schild analysis of the data.
Download figure to PowerPoint
The measurement of Ro 04–6790 and Ro 63–0563 in both plasma and cerebrospinal fluid (CSF) showed that a small but significant fraction of the plasma concentration of Ro 04–6790 could be detected in the CSF (Figure 6). After treatment with Ro 04–6790 (30 mg kg−1, i.p.), its concentration in the CSF was sufficient to occupy more than 70% of the 5-HT6 receptors. Ro 63–0563 was detected in plasma at similar concentrations to Ro 04–6790 but could not be detected in the CSF at any time point (results not shown). Therefore, Ro 04–6790 alone was used for behavioural studies, to determine whether the same behavioural syndrome could be produced as in rats treated with antisense oligonucleotides designed to prevent the translation of 5-HT6 mRNA. Animals were habituated to the observation cages for 4 days. On the 5th day, rats treated with Ro 04–6790 exhibited a behavioural syndrome of yawning, stretching and chewing identical to that produced by antisense oligonucleotides. As can be seen from Figure 7, Ro 04–6790 produced a dose-related and statistically significant increase in the number of stretches observed over a 1 h period immediately after treatment. A maximal effect was observed at a dose of 30 mg kg−1, i.p., which is in agreement with the concentration of Ro 04–6790 in the CSF being sufficient to occupy 5-HT6 receptors. Similar results were observed for yawning. However, this failed to reach significance (P = 0.17, data not shown).
Figure 6. Mean concentrations of Ro 04–6790 in rat plasma and CSF at various time points following administration of Ro 04–6790 (30 mg kg−1, i.p.). Separate rats were used for measurements at each time point. Data points at 0.5 h are the mean ± s.e.mean of determinations from 3 rats, at the 1 h time point, 2 rats were used and at all other time points 1 rat was used.
Download figure to PowerPoint
- Top of page
The present study describes two compounds, Ro 04–6790 and Ro 63–0563, which are potent and highly selective antagonists for the cloned 5-HT6 receptor (Monsma et al., 1993; Ruat et al., 1993; Kohen et al., 1994; Boess et al., 1997). Both compounds had similar binding affinities at the rat and human 5-HT6 receptor showing that there are no species differences between the human and the rat receptors for these compounds. Furthermore, both compounds were highly selective for the 5-HT6 receptor, with Ro 04–6790 having over 100 fold selectivity with respect to 23 other binding sites including 8 5-HT receptors. The same was true for Ro 63–0563 which was also tested in a further 45 binding assays by CEREP. In total, Ro 63–0563 was tested in 69 binding assays and was found to have over 100 fold selectivity for the 5-HT6 receptor. When studied in functional assays, neither compound had any significant effect on basal cyclicAMP accumulation, suggesting that they are not either agonists or inverse agonists. However, both compounds behaved as competitive antagonists at the human 5-HT6 receptor with pA2 values of 6.75 ± 0.07 and 7.10 ± 0.09, respectively. These values are in agreement with the binding affinities of the compounds at the human 5-HT6 receptor (7.26 ± 0.06 and 7.91± 0.02, respectively).
The 5-HT6 receptor was identified by molecular biological methods and its pharmacology made the study of naturally occurring 5-HT6 receptors very difficult. All compounds previously identified as having high affinity for the 5-HT6 receptor were non-selective ligands. Potent agonists at the receptor are LSD, ω-N-methyl-5-HT, bufotenine, 5-methoxy-tryptamine, 5-HT, 2-methyl-5-HT, 5-benzyloxytryptamine and tryptamine all of which are non-selective, having affinities at other 5-HT receptors (Boess et al., 1997). LSD, lisuride, 2-methyl-5-HT, tryptamine and 5-benzloxytryptamine behaved as partial agonists compared to 5-HT. Potent antagonists at the receptor are also non-selective and include methiothepin, clozapine, mianserin and ritanserin (Boess et al., 1997). Therefore, the identification of physiological responses mediated by the 5-HT6 receptor has not been possible and responses with properties similar to that of the cloned receptor can only be classified as 5-HT6-like (Sleight et al., 1997). Indeed, with respect to receptor nomenclature the fact that no selective 5-HT6 ligands existed excluded the 5-HT6 receptor from being officially recognised as a functional, naturally occurring 5-HT receptor (Hoyer et al., 1994; Hoyer & Martin, 1997). We propose that responses can now be classified as mediated by the 5-HT6 receptor with the use of Ro 04–6790 and Ro 63–0563. For example, in vitro studies have suggested that naturally occurring 5-HT6-like receptors are present in mouse neuroblastoma-derived cell lines N18TG2 and NCB-20 (MacDermot et al., 1979; Berry-Kravis & Dawson, 1983; Conner & Mansour, 1990; Unsworth & Molinoff, 1994). However, this classification was only based on a comparison of the rank order of potency of compounds in the functional models with that at the recombinant receptor. In addition, 5-HT6-like receptors have also been demonstrated in pig caudate membranes and mouse striatal neurones in culture where they stimulate cyclicAMP accumulation (Schoeffter & Waeber, 1994; Sebben et al., 1994).
The lack of selective tools has also meant that the pharmacological evaluation of the function of 5-HT6 receptors in vivo has not been possible. The only study investigating the role of 5-HT6 receptors in vivo used i.c.v. injections of antisense oligonucleotides to reduce the expression of the 5-HT6 receptor protein. This treatment gave rise to a behavioural syndrome of yawning, stretching and chewing and caused a 30% reduction in the number of [3H]-LSD binding sites (measured in the presence of 300 nM spiperone). This behavioural syndrome was dose-dependently antagonized by atropine suggesting a modulatory role for 5-HT6 receptors on cholinergic neurones (Bourson et al., 1995; Sleight et al., 1996). In order to determine whether a 5-HT6 antagonist will produce the same behavioural syndrome, the ability of both Ro 04–6790 and Ro 63–0563 to cross the blood-brain barrier was determined by measuring the CSF and plasma concentrations of the two compounds. Ro 04–6790 but not Ro 63–0563 could be detected in the CSF at concentrations that would occupy more than 70% of the 5-HT6 receptors after administration of 30 mg kg−1, i.p. Interestingly Ro 04–6790 produced a similar behavioural syndrome to that produced by the 5-HT6 antisense consisting of yawning, stretching and chewing (Figure 7). The stretching behaviour was statistically significant while the yawning just failed to reach statistical significance. Since the behavioural syndrome is produced by both antisense oligonucleotide treatment and selective 5-HT6 antagonists, we conclude that it is indeed mediated by 5-HT6 receptor blockade. It should also be noted that the maximal effect of Ro 04–6790 at 30 mg kg−1, i.p. would also agree with its concentration in the CSF being sufficient to occupy 5-HT6 receptors. These data suggest that experiments with antisense oligonucleotides can be of use in the study of newly cloned receptors where no selective compounds are available, provided that correct controls and techniques are used.
Currently, there is no good binding assay to label 5-HT6 receptors in animal tissue although it has been suggested that [3H]-clozapine labels two populations of receptors in the rat brain, one of which may be the 5-HT6 receptor binding site (Glatt et al., 1995). However, 5-HT had no measurable affinity for these binding sites and [3H]-clozapine was not observed in the corpus striatum. The presence of 5-HT6 receptor binding sites in this brain region would have been predicted from in situ hybridization studies (Ward et al., 1995) and from the mapping of 5-HT6-like immunoreactivity with polyclonal antibodies (Gerard et al., 1996). Therefore, the lack of affinity for 5-HT and the localization of [3H]-clozapine binding raises questions as to the validity of this radioligand binding assay. Consequently, it would be interesting to radiolabel a compound such as Ro 63–0563 in order to determine whether there are indeed specific 5-HT6 receptor binding sites in the brain. This work has been started in our group and is now ongoing.
In summary, the present study describes and characterizes two potent and selective 5-HT6 receptor antagonists that can be used to study and characterize the functional role of naturally occurring 5-HT6 receptors. Perhaps, more importantly such compounds may now allow the identification of possible therapeutic uses of 5-HT6 receptor ligands.