Melatonin exerts a variety of physiologic activities that are mainly relayed through the melatonin receptors MT1 and MT2 Low expressions of these receptors in tissues have led to widespread experimental use of the agonist 2-[125I]-iodomelatonin as a substitute for melatonin. We describe three iodinated ligands: 2-(2-[(2-iodo-4,5-dimethoxyphenyl)methyl]-4,5-dimethoxy phenyl) (DIV880) and (2-iodo-N-2-[5-methoxy-2-(naphthalen-1-yl)-1H-pyrrolo[3,2-b]pyridine-3-yl])acetamide (S70254), which are specific ligands at MT2 receptors, and N-[2-(5-methoxy-1H-indol-3-yl)ethyl]iodoacetamide (SD6), an analog of 2-[125I]-iodomelatonin with slightly different characteristics. Here, we further characterized these new ligands with regards to their molecular pharmacology. We performed binding experiments, saturation assays, association/dissociation rate measurements, and autoradiography using sheep and rat tissues and recombinant cell lines. Our results showed that [125I]-S70254 is receptor, and can be used with both cells and tissue. This radioligand can be used in autoradiography. Similarly, DIV880, a partial agonist [43% of melatonin on guanosine 5′-3-O-(thio)triphosphate binding assay], selective for MT2, can be used as a tool to selectively describe the pharmacology of this receptor in tissue samples. The molecular pharmacology of both human melatonin receptors MT1 and MT2, using a series of 24 ligands at these receptors and the new radioligands, did not lead to noticeable variations in the profiles. For the first time, we described radiolabeled tools that are specific for one of the melatonin receptors (MT2). These tools are amenable to binding experiments and to autoradiography using sheep or rat tissues. These specific tools will permit better understanding of the role and implication in physiopathologic processes of the melatonin receptors.
The neurohormone melatonin (MLT) is produced by the pineal gland during periods of darkness, particularly at night (Arendt and Broadway, 1987; Zawilska et al., 2009). Melatonin pleiotropic effects include exerting control over biologic rhythms, such as circadian rhythms through the suprachiasmatic nuclei (Coomans et al., 2015), and circannual rhythms, e.g., seasonal reproduction and molting (Lincoln et al., 2006). The melatonergic system has been implicated in depression, leading to development of the MLT receptor–targeting antidepressive drug agomelatin (Valdoxan®) (Millan et al., 2003, 2005; de Bodinat et al., 2010).
Melatonin actions are exerted through its interactions with three known targets: two of these targets are the G-protein–coupled melatonin receptors MT1 and MT2 (Jockers et al., 2008), which are mainly coupled to Gi protein, leading to adenylate cyclase inhibition (Brydon et al., 1999; Barrett and Bolborea, 2012); and the third melatonin target is quinone reductase 2 (formerly known as MT3) (Dubocovich, 1988b; Paul et al., 1999; Nosjean et al., 2000), which has raised several questions regarding melatonin mechanisms of action in some models, particularly at higher concentrations (Vella et al., 2005; Jockers et al., 2008; Boutin, 2015). An interaction between the RAR-related orphan receptors and melatonin has been reported (Carlberg and Wiesenberg, 1995), but remains elusive to confirmatory studies.
Prior to the cloning of these receptors in the late 1990s (Reppert et al., 1994, 1995), it was very difficult to study MLT binding on biologic membranes, since these receptors are naturally expressed at very low levels and the available radioligand [3H]-MLT had a poor specific activity. After the description of the strongly labeled super agonist 2-[125I]-iodomelatonin (2-[125I]-MLT) (Vakkuri et al., 1984, 1985), all experiments were conducted with it.
Few diverse tools have been developed for MLT molecular pharmacology studies (Markl et al., 2011), such as subtype-specific agonists or antagonists, or biased agonists (Devavry et al., 2012b; Legros et al., 2014). We further screened our compounds (Yan et al., 2008) to find various types of ligands that are amenable to chemical conversion to iodinated and radioactive derivatives. Through this study, we identified an MT2-specific very partial agonist, 2-(2-[(2-iodo-4,5-dimethoxyphenyl)methyl]-4,5-dimethoxy phenyl) (DIV880), with a Ki value that is 2 logs less potent at MT1 than at MT2. We synthesized the precursors of each ligand, make them iodinated, and assessed their characteristics as binders at the recombinant human MT1 and MT2 receptors. This approach has been briefly described in our recent publication (Legros et al., 2013).
In the present work, we described the behaviors of these new ligands as compared with the standard ligands [3H]-MLT and 2-[125I]-MLT. In particular, we investigated the ligand behaviors at their respective cloned receptors, finding that two were specific to the receptor isoform MT2. Furthermore, we established the molecular pharmacology of MT2 using those specific ligands, and compared it with the profiles obtained with other, nonspecific ligands (i.e., 2-[125I]-MLT and [3H]-MLT). This study contributes to the knowledge in this field, to help develop as many tools as possible to determine the characteristics of the purified receptors (Logez et al., 2014) as well as specific ligand(s) amenable for autoradiography in native tissues.
Materials and Methods
Reagents and Ligands
We purchased [125I]-SD6 (N-[2-(5-methoxy-1H-indol-3-yl)ethyl]iodoacetamide; specific activity 2175 Ci·mmol−1), [125I]-S70254 [(2-iodo-N-2-[5-methoxy-2-(naphthalen-1-yl)-1H-pyrrolo[3,2-b]pyridine-3-yl])acetamide; specific activity 2175 Ci·mmol−1], and [125I]-DIV880 (specific activity 2130 Ci·mmol−1) from ANAWA Trading SA (Wangen Zürich, Switzerland). We purchased 2-[125I]-MLT (specific activity 2200 Ci·mmol−1) from PerkinElmer (Boston, MA) and [3H]-MLT (specific activity 80–85 Ci·mmol−1) from American Radiolabeled Chemicals, Inc. (St. Louis, MO). We obtained MLT, 2-iodomelatonin, 6-chloromelatonin, serotonin, and D 600 (methoxy-verapamil) from Sigma-Aldrich (St. Louis, MO); 4P-PDOT (4-phenyl-2-propionamidotetraline) and luzindole (2-benzyl-N-acetyltryptamine) from Tocris (Bristol, UK); and 2-bromomelatonin from Toronto Research Chemicals, Inc. (North York, Canada). We also evaluated 15 analogs of MLT that stemmed from our product library (Legros et al., 2014). Compounds were dissolved in dimethylsulfoxide at a stock concentration of 10 mM, and stored at −20°C until use. All other reagents were obtained from Sigma-Aldrich.
Cell Membrane Preparation
CHO-K1 cell lines stably expressing the MT1 or MT2 receptor (human, rat, mouse, or sheep: hMT, rMT, mMT and oMT, respectively) were grown to confluence, harvested in phosphate-buffered saline buffer (GIBCO, Invitrogen, Cergy-Pontoise, France) containing 5 mM EDTA, and centrifuged for 20 minutes at 1000 × g (4°C). The resulting pellet was suspended in 5 mM Tris/HCl (pH 7.4) containing 2 mM EDTA, and homogenized using a Kinematica polytron (Kinematica AG, Luzern, Switzerland). The homogenate was then centrifuged for 30 minutes at 20,000 × g (4°C), and the resulting pellet was suspended in 75 mM Tris/HCl (pH 7.4) containing 2 mM EDTA and 12.5 mM MgCl2. Protein content was determined using the Bradford method (Bradford, 1976) with a Bio-Rad kit (Bio-Rad SA, Ivry-sur-Seine, France). Membrane preparation aliquots were stored at −80°C in resuspension buffer containing 75 mM Tris/HCl (pH 7.40), 2 mM EDTA, and 12.5 mM MgCl2.
Cell Membrane Binding Assays.
Cell membrane binding assays were performed in 96-well plates at a final volume of 250 µl, with binding buffer containing 50 mM Tris/HCl (pH 7.4), 5 mM MgCl2, and 1 mM EDTA, plus 0.1% bovine serum albumin for [125I]-DIV880. For all radioactive compounds, the hMT1 and hMT2 membranes were used at a final concentration of 30 µg ml−1. In all protocols, the reaction was stopped by rapid filtration through GF/B unifilters (filters were presoaked in 0.1% polyethyleneimine for [125I]-DIV880), followed by three successive washes with ice-cold 50 mM Tris/HCl buffer (pH 7.4).
Kinetic parameters of [125I]-DIV880, [125I]-S70254, and [125I]-SD6 were measured with hMT1 and hMT2 at 37°C and at room temperature (Kon, Koff, and KDkinetics). Koff is the dissociation rate constant, Kon is the association rate constant, while Kd is the equilibrium binding constant, computed as Koff/Kon. For association studies, membranes were added to radioligands (0.1 nM [125I]-DIV880, 0.04 nM [125I]-S70254, and 0.08 nM [125I]-SD6), followed by incubation for increasing time periods ranging from 5 to 240 minutes. For dissociation studies, membranes were incubated with the same radioligand concentrations for 20 minutes, 1 hour, or 2 hours, followed by addition of 10 µM cold MLT to initiate dissociation, and then incubation for increasing time periods from 0 to 240 minutes. To account for membrane variability, kinetic measurements were repeated at least twice on each pool of membranes. For displacement tests, the membranes in binding buffer were first incubated with compounds diluted in 10% dimethylsulfoxide, followed by the same duration of incubation with the radioligands: [125I]-DIV880, 0.1 nM for hMT2, 1.5-hour incubations; [125I]-S70254, 0.04 nM for hMT2, 1-hour incubations; and [125I]-SD6, 0.08 nM for both hMT1 and hMT2, 1-hour incubations for hMT1 and 1.5-hour incubations for hMT2. The different incubation times were selected based on the kinetics results to be in equilibrium in accordance with the mass-action law. Ligand dilutions from 10−5 to 10−15 M were created using the MicroLab Starlet (Hamilton, Massy, France). Nonspecific binding was defined using 10 µM MLT.
Data were analyzed using the program PRISM (GraphPad Software Inc., San Diego, CA). For the saturation assay, the binding site density (Bmax) and the radioligand dissociation constant (KD) were calculated following the method of Scatchard (1949). Association kinetics data were analyzed by fitting specific binding data to the equation B = Bmax*[1 − exp(−k*t)], where B is binding at time t, and k is the observed association rate constant. Dissociation kinetics data were analyzed by fitting specific binding to the equation B = Bmax*exp (−k+t) + plateau, where k is the dissociation constant rate. For displacement experiments, inhibition constants (Ki) were calculated using the Cheng-Prussof equation (Cheng and Prusoff, 1973): Ki = IC50/[1+(L/KD)], where IC50 is the 50% inhibition concentration, and L is the [3H]-MLT concentration. The extra sum-of-squares F-test was used for preferred model analysis, at one or two sites, for the saturation and kinetics experiments. The Pearson product-moment correlation coefficient was used for correlation analysis of the inhibition constant values, as their log: pKi.
Tissue Membrane Preparation.
Lambs and young adult sheep were obtained from the SODEM slaughterhouse in February, and were killed during the morning between 7:00 and 12:00 (Le Vigeant, France). Retinas were collected, frozen in nitrogen liquid, and stored at −80°C, with a less than 10-minute interval between slaughter and freezing. For analysis, the retinas were thawed on ice, crushed, and homogenized with a Kinematica polytron (Kinematica AG) in 4°C grinding buffer comprising 5 mM Tris/HCl and 2 mM EDTA at pH 7.4. Homogenates were then centrifuged for 12 minutes at 1000 × g (4°C), and the supernatant was ultracentrifuged for 35 minutes at 20,000 × g (4°C). The resulting pellets were suspended in conservation buffer comprising 50 mM Tris/HCl, 6 mM ascorbic acid, 4 mM CaCl2, and 4% glycerol at pH 7.4. Protein content was determined following the Bradford method (Bradford, 1976) using the Bio-Rad kit. Aliquots of the membrane preparation were stored at −80°C until use.
Tissue Membrane Saturation Assays.
Tissue membrane saturation assays were performed in 96-well plates with a 250-µl final volume in binding buffer comprising 50 mM Tris/HCl, 6 mM ascorbic acid, and 4 mM CaCl2 at pH 7.4 (with 0.1% bovine serum albumin for [125I]-DIV880). For all radioactive compounds, the final membrane concentrations were 600 µg·ml−1. Membranes were incubated for 2 hours at 37°C in the presence of 2-[125I]-MLT, [125I]-SD6, [125I]-S70254, or [125I]-DIV880 (0.1–2.5 nM). Nonspecific binding was defined using 10 µM MLT. Reactions were stopped by rapid filtration through GF/B unifilters, followed by three successive washes with ice-cold 50 mM Tris/HCl buffer (pH 7.4).
Autoradiography of Rat and Sheep Brains.
Autoradiography was performed using brains from three rats and two sheep. The animals were sacrificed by decapitation, after which their brains were quickly removed and frozen over liquid nitrogen. We used a microtome-cryostat (Leitz Kryostat 1720) at −20°C to generate frozen frontal 15-µm sections throughout the rat brains, and from four regions of interest (suprachiasmatic nuclei, pars tuberalis, hypothalamus and hippocampus) from the sheep brains. These sections were collected on TESPA (3-amino-propyl-ethoxy-silane; Sigma-Aldrich, Saint-Quentin Fallavier, France) gel-coated slides (Leica Microsystemes, Nanterre, France), and stored at −80°C until incubation.
Sections were first washed with 50 mM Tris buffer at 4°C, then incubated for 1 hour at room temperature with 100 µl of Tris buffer containing 2-[125I]-MLT 280 pM (specific activity, 2200 Ci·mmol−1) for melatonergic binding sites (MT1/MT2) or [125I]-S70254 140 pM for MT2 binding sites. After this incubation, the sections were rinsed twice (for 2 and 3 minutes) at 4°C with Tris buffer, fixed with 4% paraformaldehyde for 10 minutes at 4°C, and dipped in water for 10 minutes. To assess nonspecific 2-[125I]-MLT binding and [125I]-S70254 binding, respectively, adjacent sections were incubated with 20 and 10 nM MLT (Fluka; Sigma-Aldrich) (Malpaux et al., 1998).
Autoradiograms were generated by placing air-dried sections in X-ray cassettes with Biomax MR hyperfilm (Amersham, Courtaboeuf, France). [125I] microscale standards were generated using seven concentrations of 2-[125I]-MLT in solution with pure ethanol, dispersed on thin-layer chromatography silicate gel (Macherey-Nagel, Hoerdt, France). The exposure time was 28 days for 2-[125I]-MLT binding and 7 days for [125I]-S70254 binding at room temperature. To identify histologic structure, autoradiography sections were stained using the Klüver and Barrera methods (Kluver and Barrera, 1953) and compared with a rat brain atlas (Paxinos and Watson, 1997). Binding intensity was assessed using an image analysis system (Biocom Histo 500, Les Ulys, France). Nonspecific binding levels were measured on the control section incubated with cold ligands. After removing these levels, the mean gray density was transformed to counts per minute using the microscale standards, and data were converted to fmol·mgprotein−1 as previously described (Nazarali et al., 1989).
Autoradiography of Rat and Sheep Retinas.
Autoradiography was performed using retinas from two rats and two sheep. Sheep were sacrificed by decapitation, the eyes were quickly removed, and then the retinas were collected by scraping the back of the eyes. The collected retinas were frozen together over liquid nitrogen and then embedded in Tissue-Tek (akura Finetek Europe B.V., AV Alphen aan den Rijn, The Netherlands) and used to generate 15-µm retina sections for analysis. Rats were sacrificed by decapitation, their eyes were collected, and the crystalline was removed. The eyes were then embedded together on the lateral side in Tissue-Tek, frozen over liquid nitrogen, and used to generate 15-µm sagittal sections. Binding in the rat and sheep retina tissue sections was assayed using the previously described conditions.
Characteristics of Human MT1 and MT2 Receptors with Classic Ligands.
Membrane preparations were first characterized using the classic radioligands 2-[125I]-MLT and [3H]-MLT. As described earlier (Legros et al., 2013), average affinities and Bmax values are as follows. With 2-[125I]-MLT and hMT1, the average pKd value for the used preparations was 10.69 ± 0.07 nM, and Bmax was 688 ± 153 fmol·mgprotein−1. With 2-[125I]-MLT and hMT2, pKd was 10.16 ± 0.03 nM, and Bmax was 1998 ± 318 fmol·mgprotein−1. The radioligands [125I]-S70254 and [125I]-DIV880 showed specific binding only to hMT2, with pKd values of 9.61 ± 0.14 and 9.65 ± 0.07, respectively, and Bmax values of 1778 ± 87 and 2308 ± 539 fmol·mgprotein−1. [125I]-SD6 showed high affinity for both receptors, with a pKd of 10.85 ± 0.01 and Bmax of 276 ± 50 fmol·mgprotein−1 for hMT1, and a pKd of 10.18 ± 0.11 and Bmax of 1929 ± 308 fmol·mgprotein−1 for hMT2. Membranes were also characterized using [3H]-MLT, which showed biphasic curves for both receptors, with pKd values of 10.23 ± 0.07 and 9.46 ± 0.01 and Bmax values of 575 ± 77 and 96 ± 12 fmol·mgprotein−1 for hMT1, and pKd values of 9.87 ± 0.05 and 9.26 ± 0.05 and Bmax values of 2220 ± 178 and 463 ± 68 fmol·mgprotein−1 for hMT2.
Association and Dissociation Kinetics.
Kinetics parameters were determined for each radioligand on hMT1 and hMT2. Table 1 presents all kinetics parameters (kon, koff, KDkinetics, and half-life) as the mean ± S.E.M. from experiments with n of at least 2. The kinetics curves are presented in Fig. 1. Table 1 recapitulates the results. It seems to us outstanding that all the ligands presented low nanomolar affinities at one or the other receptor, if not both. For the two reference compounds, 2-[125I]-MLT and [3H]-MLT, no major differences were observed, as far as the affinities were concerned, as already reported previously. To the contrary, there are huge differences between 2-[125I]-MLT and [3H]-MLT association half-life: 3 minutes for [3H]-MLT and 20–36 minutes for 2-[125I]-MLT, at both receptor subtypes. [3H]-MLT is the only tested compound to show such short association half-lives (ca. 3 minutes) among the five radiolabeled compounds tested, whereas their dissociation half-lives were not dramatically different. Only [125I]-SD6 was studied on hMT1 since the other radioligands did not bind to this receptor. Whatever the experimental temperature, no major differences were recorded with the ligands, as far as their affinities were concerned. As stated previously, it is almost impossible to measure an affinity at hMT1 for [125I]-S70254 and [125I]-DIV880. Of note is the fact that only data recorded with [3H]-MLT showed a second inflection point, leading to the possibility of measuring a second pKd (9.91 ± 0.02 at MT1) whatever the experimental temperature. This feature is shared only by the DIV880 at MT2 with a pKd of 9.23 ± 0.29. The meaning of this second binding site remains elusive.
It is important to evaluate the molecular pharmacology profiles of the three investigated new radioligands in comparison with those of 2-[125I]-MLT and [3H]-MLT. We assessed a set of 24 compounds that were previously described in the literature [including 4-phenyl-2-propionamidotetraline (Dubocovich et al., 1997), luzindole (Dubocovich, 1988a), and ramelteon (Uchikawa et al., 2002)] or that were issued from our own medicinal chemistry programs (Depreux et al., 1994; Audinot et al., 2003, 2008; Mailliet et al., 2004; Devavry et al., 2012a,b; Ettaoussi et al., 2013; Legros et al., 2013, 2014). These compounds were tested against [125I]-SD6 on both hMT1 and hMT2 membrane preparations, and against [125I]-S70254 and [125I]-DIV880 on hMT2 membrane preparations only. Under standard non-decoupling binding conditions, the binding data all consistently exhibited good correlations among the various data sets (Fig. 2; Tables 2 and 3). The molecular pharmacology profiles of [125I]-SD6, [125I]-S70254, and [125I]-DIV880 mostly replicated that obtained with 2-[125I]-MLT for both hMT1 and hMT2 (for all radioligands, r ≥ 0.95 and P < 0.0001, n ≥ 2).
Tissue Membrane Saturation Assays.
We next tested binding of the [125I]-SD6, [125I]-S70254, and [125I]-DIV880 radioligands on tissue membrane preparations. We chose to use sheep retina tissue because it is readily available in large amounts, and because MT1 and MT2 mRNAs have been detected in this tissue (Coge et al., 2009). As a positive control, membranes were also characterized using 2-[125I]-MLT.
The saturation curves for 2-[125I]-MLT showed biphasic curves and Scatchard regressions (Fig. 3) with a pKd1 (site 1) of 10.35 ± 0.03, a pKd2 (site 2) of 9.25 ± 0.13, a Bmax1 (site 1) of 1.46 ± 0.13 fmol·mgprotein−1, and a Bmax2 (site 2) of 2.32 ± 0.98 fmol·mgprotein−1 (n = 4). Similarly, [125I]-SD6 showed biphasic curves and Scatchard regressions (Fig. 3) with a pKd1 of 10.13 ± 0.18, a pKd2 of 6.79 ± 0.21, a Bmax1 of 0.60 ± 0.20 fmol·mgprotein−1, and a Bmax2 = 2.11 ± 1.13 fmol·mgprotein−1 (n = 3). Unfortunately, MT2 radioligands did not show any proper saturation curves. Moreover, the signals of [125I]-S70254 were very low and highly variable, and it was not possible to obtain a saturation curve better than that shown in Fig. 3. We observed no specific binding of [125I]-DIV880 on retina membranes.
Rat Brain Autoradiography.
Rat brains were randomly screened. We analyzed 150 sections from throughout the brain, processed in groups of six. Compared with 2-[125I]-MLT, [125I]-S70254 showed higher specific binding in the periacqueductal gray region (1.93 ± 2.00 vs. 0.25 ± 0.51 fmol·mgprotein−1, respectively), in the interpeduncular nucleus (4.14 ± 4.52 vs. 1.56 ± 0.31 fmol·mgprotein−1), and in olfactory bulbs (4.88 vs. 2.24 fmol·mgprotein−1). In the subiculum and pontine nuclei, we observed only [125I]-S70254 binding (3.05 ± 3.17 and 3.10 ± 2.33 fmol·mgprotein−1, respectively). In the suprachiasmatic nuclei and pars tuberalis, we observed only 2-[125I]-MLT binding (4.07 and 14.24 fmol·mgprotein−1, respectively). Representative images are shown in Fig. 4.
Rat and sheep retina autoradiography.
To complete our characterization of these new radioligands, we also assessed binding of 2-[125I]-MLT and [125I]-S70254 in rat and sheep retinas (Fig. 5). The binding density of [125I]-S70254 was greater than that of 2-[125I]-MLT, with values of 14.13 ± 6.53 fmol·mgprotein−1 for [125I]-S70254 and 4.11 ± 0.73 fmol·mgprotein−1 for 2-[125I]-MLT in the rat retina, and values of 4.00 ± 0.24 fmol·mgprotein−1 for [125I]-S70254 and 1.98 ± 2.64 fmol·mgprotein−1 for 2-[125I]-MLT in the sheep retina.
Radioligand Structure and Characteristics.
Figure 6 presents a comparison of ligand electrostatic potentials, as computed using Molecular Operating Environment 2012.10 (Chemical Computing Group, (www.chemcomp.com). Notably, the iodine atom (either on position 2 or at the tip of the side chain of MLT) does not affect the hydrogen bonding potential with the 5-methoxy oxygen atom and the amide group on the side chain, as previously described (Pala et al., 2013). The iodine and the naphthyl (S70254) on the indole C2 position have a similar hydrophobic shape in the vicinity of the indole ring. Furthermore, the second aromatic of the naphthyl group likely confers the MT2 selectivity to S70254. On the side chain of compounds S70254 and SD6, the iodine atom does not prevent the amide group from making hydrogen bonds to the receptor, suggesting a hydrophobic pocket in that region.
One main factor limiting studies of the melatonin receptors MT1 and MT2 is the lack of antagonist and discriminant radioligands. The two presently known radioligands, [3H]-MLT and 2-[125I]-MLT, are pure agonists, and exhibit the same affinity and almost identical pharmacological parameters for both receptors (Browning et al., 2000; Masana and Dubocovich, 2001; Legros et al., 2013, 2014). In the present work, we described three new radioligands that can be used for MT1 and MT2 receptor characterization. Our results indicated that each radioligand has specific properties that can be exploited for different kinds of pharmacological studies. In particular, we found that [125I]-S70254 is specific for the MT2 receptor, and can be used with cells and tissue.
The three newly investigated radioligands could each bind hMT1 and/or hMT2 receptors. Specifically, [125I]-SD6 showed affinities (pKd ± S.E.M.) of 10.85 ± 0.01 for hMT1 and 10.18 ± 0.11 for hMT2, and [125I]-DIV880 and [125I]-S70254 bound hMT2 with affinities of 9.61 ± 0.14 and 9.65 ± 0.07, respectively. Despite its slow dissociation from the hMT2 receptor, [125I]-SD6 showed affinities for MT receptors and kinetics parameters (including pKd and half-life) that were very similar to those of 2-[125I]-MLT. The set of compounds tested in our dose-response experiments also showed affinities for receptors that were similar to those of 2-[125I]-MLT. These findings suggest that the position of the iodine (which constitutes 50% of the size of MLT, i.e., 125 vs. 232 Da) does not influence the binding, at least when it is located at position 2 on the indolic moiety or at the extremity of the lateral chain.
Although both [125I]-S70254 and [125I]-DIV880 showed low affinities for hMT1 [pKi of 6.25 and 6.08, respectively (Legros et al., 2013)], neither showed specific binding for this receptor. Compared with 2-[125I]-MLT, [125I]-S70254 and [125I]-DIV880 each showed a slightly lower affinity (half-log lower) for the hMT2 receptor, although this difference disappeared when using the kinetics pKd. The lower affinity could be explained by the difficulty reaching a saturation plateau in the Scatchard assay. Apart from the lack of affinity for hMT1, [125I]-S70254 and [125I]-DIV880 mainly differ from 2-[125I]-MLT in that they show fast and total dissociation at 37°C and partial dissociation at room temperature, whereas 2-[125I]-MLT dissociates partially at 37°C and does not dissociate at room temperature.
Our kinetics experiments illustrated the ability of the radiocompound to associate and dissociate from the receptor. After 1 hour of incubation at room temperature, 2-[125I]-MLT did not reach an equilibrium plateau, and only a proportion was bound to a receptor (3/4 with hMT1 and 1/2 with hMT2). The 2-[125I]-MLT radioligand has been widely used to study melatonin receptors in the brain and peripheral organs by autoradiography (Morgan et al., 1994; Thomas et al., 2002; Sallinen et al., 2005). Usually an incubation time of about 1 hour is used, which allows a balance between radioligand binding and degradation of the tissue slice, which is not fixed in this kind of protocol. These data suggest that the real densities of receptors and binding sites are underestimated with the use of 2-[125I]-MLT. Some prior studies may have detected no melatonin receptors because there exists only a low density of binding sites, and because these sites are only partially detected due to the use of an incubation time that is too short—for example, in the finding of no 2-[125I]-MLT binding on the suprachiasmatic nuclei in sheep brain tissue (de Reviers et al., 1991).
The use of alternative radioligands could potentially help with this problem of partial binding site labeling. In the present work, we showed that such autoradiography experiments could be performed, and could lead to new specific information about MT2 localization. The ligand 2-[125I]-MLT shows similar affinities for both receptors as well as similar kinetics, making it a poor selection for autoradiography. On the other hand, [3H]-MLT has the advantage of fast association kinetics but also shows a very low specific activity that also makes it a poor candidate for autoradiography. We found that [125I]-S70254 labeled the MT2 receptor in autoradiography. However, as with 2-[125I]-MLT, we must consider the association time that results in partial labeling of the MT2 receptor (about 2/3 or 3/4). Our present results did not show [125I]-S70254 binding on the suprachiasmatic nuclei. This could have been due to the random screening without targeting of structures and the small size of the rat suprachiasmatic nuclei (2–4 mm2) (Paxinos and Watson, 1997), or it could have been because the rat MT2 expression was too low and the reinforcement of this problem by the partial labeling of the receptors by the radioligand with an incubation time of 1 hour.
In our assays using the retina membrane preparation, we observed specific binding with 2-[125I]-MLT and [125I]-SD6. Sheep retina tissue expresses both the ovine MT1 and MT2 receptors (Coge et al., 2009). In comparison with the Kd value on transfected cell membrane preparations, the first highest affinity binding site could be oMT1 or oMT2 in sheep retina tissue (Mailliet et al., 2004). The second binding site in the saturation experiment remains to be identified. The binding of [125I]-S70254 was highly variable, and the saturation curve was difficult to obtain. This may have been due to the low density of the oMT2 receptor, which could potentially be remedied by increasing the protein concentration only if we also changed the filtration protocol, as filter obstruction would occur with the presently used protocol. Another possible cause is the affinity of the radioligand for the ovine form of the receptor (Kiovine = 7.67 ± 0.07, unpublished data). This could be investigated by performing tests with tissue membrane preparations from different species.
It is unfortunately clear that, as of today, the use of [125I]-SD6 would not determine other information than that determined by 2-[125I]-MLT. It remains to be seen if further characterization of the agonistic nature of the compound, using as many different signaling pathways, will reveal a different view of these agonists (C. Legros and J.A. Boutin, manuscript in preparation).
Overall, our present data contribute information and potential new tools that will be useful in further studies of MLT molecular pharmacology. Greater availability of such tools would help pharmacologists to better understand the nature of involvement of MLT receptors in the many pathologic processes in which MLT is active. However, further structure-based modeling will be required to design new agonist compounds and to better understand the structure mechanism of the various action relationships (agonist, partial agonist, and antagonist) that can, incidentally, serve as radioligands after iodination. Continued investigations of this field and the development of new techniques could lead to novel therapies targeting the melatonergic system.
Participated in research design: Legros, Nosjean, Delagrange, Boutin.
Conducted experiments: Legros, Brasseur, Ducrot.
Wrote or contributed to the writing of the manuscript: Legros, Boutin, Ducrot.
- Received October 12, 2015.
- Accepted January 11, 2016.
- 2-(2-[(2-iodo-4,5-dimethoxyphenyl)methyl]-4,5-dimethoxy phenyl)
- human melatonin receptors
- melatonin receptors
- ovine melatonin receptors
- Copyright © 2016 by The American Society for Pharmacology and Experimental Therapeutics