Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The function of serotoninergic systems is highly dependent on the membrane serotonin transporters that actively clear serotonin from the synaptic space. In the brain, these transporters have an exclusively neuronal localization on serotoninergic neurons of the raphe nuclei and their projections in rats (Blakely et al., 1991) and humans (Ramamoorthy et al., 1993). The study of serotonin transporters can therefore provide relevant information on the density of serotoninergic neurons as well as on its functioning. Several reports have shown that both these factors can be changed in several disease situations. It has been shown that the serotonin transporter can decline with age (Pirker et al., 2000; Van Dyck et al., 2000) and in several neurodegenerative disorders such as Parkinson's disease (Chinaglia et al., 1993) and Alzheimer's disease (Tejani-Butt et al., 1995). Decreases in serotonin uptake and/or radioligand binding have also been observed postmortem (Stanley et al., 1982; Leake et al., 1991) as well as in vivo (Malison et al., 1998) in human brains following depression or suicide. Moreover, serotonin transporters are major targets of antidepressant drugs including fluoxetine, sertraline, paroxetine, fluvoxamine, and citalopram. The serotonin transporter is therefore involved in major physiological processes as well as in various neurological and psychiatric disorders. This is the basis for exploration of this target protein to provide better understanding of the mechanisms linked to central nervous system diseases and to help in the diagnosis and treatment of these disorders.

Several compounds labeled with + or  emitters have been developed for in vivo exploration of serotonin transporters by positron emission tomography (PET) and single photon emission tomography (SPET). The PET tracers [¹¹C]McN 5652 and [¹¹C]DASB seem currently the most appropriate ligands allowing in vivo quantification of serotonin transporters in the human brain (Parsey et al., 2000; Ginovart et al., 2001). The cocaine derivatives -CIT (Brücke et al., 1993; Malison et al., 1998) and nor--CIT (Hiltunen et al., 1998) have been used for SPET studies, but neither is selective as they also bind to dopamine transporters. Several derivatives from another chemical family such as 5-chloro-2-((2-((dimethylamino)methyl)phenyl)thio)benzyl alcohol or 403U76 (Ferris et al., 1995), iodo-2-((2-((dimethylamino)methyl)phenyl)thio)benzyl alcohol (Oya et al., 1999), and 2-((2-((dimethylamino)methyl)phenyl)thio)-5-iodophenylamine benzyl alcohol or ADAM (Oya et al., 2000), have recently been described as potent ligands of serotonin transporters. ADAM seems to be the best tracer among these compounds in terms of affinity, selectivity, and in vivo properties for SPET exploration of the serotonin transporter (Choi et al., 2000). We recently described ADAM labeled with ¹¹C for PET exploration of serotonin transporters for this purpose (Vercouillie et al., 2001). The in vivo kinetics of [¹¹C]ADAM, however, are not appropriate for the half-life of ¹¹C (20 min), as the equilibrium of specific binding to serotonin transporters is obtained in the primate brain at 2 h postinjection (Halldin et al., 2001a). To obtain a more efficient PET tracer, we recently synthesized a derivative of ADAM, N,N-dimethyl-2-(2-amino-4-methylphenylthio)benzylamine (MADAM), which has a high affinity for the serotonin transporter compared with the dopamine and noradrenaline transporters (Emond et al., 2002). In addition, preliminary in vivo scintigraphy in the nonhuman primate showed early and specific binding of [¹¹C]MADAM to the serotonin transporter (Halldin et al., 2001b). In view of these initial findings, we hypothesized that MADAM could be an efficient ligand both for in vitro and in vivo study of the serotonin transporter. To assess this, we labeled MADAM with ³H and undertook the extensive in vitro pharmacological characterization of [³H]MADAM in rat and human brains.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Preparation of Stable and Radiolabeled MADAM

MADAM and its N-desmethylated precursor for labeling N-methyl-2-(2-amino-4-methylphenyl thio)benzylamine were synthesized as previously described (Tarkiainen et al., 2001). The labeling of [³H]MADAM was performed as follows; 700 µg of the precursor N-methyl-2-(2-amino-4-methylphenyl thio)benzylamine diluted in 300 µl of dimethyl formamide was mixed with 150 µl of [³H]methyl-iodide (specific activity, 85 Ci/mmol; Amersham Biosciences AB, Uppsala, Sweden) and heated for 15 min at 90°C. After cooling to room temperature, 300 µl of acetonitrile was added, and the tritiated product was purified by high-performance liquid chromatography using a C₁₈ reverse phase column (µ-Bondapack) and CH₃CN/NH₄CO₂H (45:55) as mobile phase. [³H]MADAM was separated from the precursor and methyl-iodide and obtained with a specific activity of 55 Ci/mmol.

Animals and Drugs

All experiments were conducted on male Wistar rats (Centre d'Elevage R. Janvier, Le Genest St. Isle, France), weighing 250-300 g, in accordance with French law regarding animal experiments. Stable ADAM and PE2I were synthesized as already described by Oya et al. (2000) and Emond et al. (1997), respectively. Natural cocaine was obtained from Coopération Pharmaceutique Française (Melun, France). GBR 12935, nisoxetine, desipramine, and fluoxetine were obtained from RBI Bioblock (Illkirch, France). Paroxetine was obtained from SmithKline Beecham (Nanterre, France), and citalopram was from Lundbeck (Copenhagen, Denmark).

In Vitro Binding Studies

Tissue Preparation. Rats were sacrificed by decapitation on the day of the assay, and the frontal cortex of each animal was removed on ice and weighed (two rats for each experiment). The tissue was homogenized in 10 volumes of 0.32 M sucrose using an Ultraturrax T25 (Bioblock, Illkirch, France). After centrifugation at 1000g for 10 min at 4°C, the supernatant was kept, and the pellet was treated as described above. Both supernatants were then pooled and centrifuged at 17,500g for 30 min at 4°C. Twenty volumes of the incubation buffer were added to the pellet, and the mixture was homogenized and centrifuged at 50,000g for 10 min at 4°C. The final pellet was suspended in a minimum volume of the assay buffer, and the protein concentration was measured according to Bradford (1976) using bovine serum albumin as standard.

Saturation Studies. [³H]MADAM was incubated at concentrations varying between 20 and 1500 pM with 60 of µg protein in a total volume of 1 ml in a Tris-HCl buffer, pH 7.4 (50 mM Tris-HCl, 120 mM NaCl, and 5 mM KCl) for 90 min at 22°C. Nonspecific binding was determined in the presence of 10⁶ M paroxetine. Samples were then rapidly filtered through Whatman GF/C fiber filters (Whatman, Clifton, NJ) soaked with 0.05% polyethylenimine (Sigma-Aldrich, St. Quentin-Fallavier, France). The filters were washed twice with 4 ml of cold buffer, and the residual radioactivity was measured in a  counter (LKB 1215 Rackbeta; Perkin Elmer, Courtaboeuf, France) in the presence of 6 ml of scintillator (LKB Optiphase). Results were analyzed with the EBDA RADLIG program (Biosoft, Cambridge, UK).

Competition Studies. For these studies, 50 pM of [³H]MADAM was incubated with 60 µg of protein in a total volume of 1 ml in the Tris-HCl buffer (50 mM Tris-HCl, 120 mM NaCl, and 5 mM KCl), pH 7.4, for 90 min at 22°C in the presence of different drugs, either MADAM and ADAM at concentrations of 10⁸ to 10¹³ M or paroxetine, citalopram, GBR 12935, PE2I, desipramine, and nisoxetine at concentrations of 10⁵ to 10¹⁰ M. Samples were then treated as described above. Total binding was determined in the absence of any drug, and nonspecific binding was measured in the presence of 10⁶ M paroxetine. The IC₅₀ values were determined graphically for each compound, and the K_i values were calculated according to Cheng and Prussoff (1973), as fully competitive inhibition was the assumed mechanism.

In Vitro Autoradiographic Studies

Rat Brains. Two animals were sacrificed by decapitation, and their brains were removed and immediately frozen at 35°C in cooled isopentane. Twenty-micron coronal sections were cut with a cryostat microtome at different brain levels (Reichert-Jung Cryocut 1800; Leica, Rueil-Malmaison, France), thaw mounted on gelatin microscope slides, and kept at 80°C until use.

Human Postmortem Brains. The human brains used were obtained from clinical autopsy at the National Institute of Forensic Medicine (Karolinska Institutet, Stockholm, Sweden) and handled as previously described (Hall et al., 1998, 2001). The study was approved by the Ethics Committee at Karolinska Institutet and the Swedish Board of Social Welfare. Cryosectioning on whole hemisphere sections was performed as previously described (Hall et al., 1998, 2001). Experiments were performed on 100-µm horizontal sections.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

In Vitro Binding Studies. The affinity and density of specific [³H]MADAM binding sites were measured by saturation experiments on rat cortical membranes. In our experimental conditions, the nonspecific binding, determined in the presence of 10⁶ M paroxetine, was around 15%. The Scatchard transformation of the resulting data (Fig. 1) revealed a linear curve suggesting a one-site model (Hill coefficient = 1) with a K_d value of 60 ± 9 pM (mean ± S.D. of three independent determinations, each performed in triplicate) and a B_max value of 543 ± 181 fmol/mg of protein (mean ± S.D.).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Saturation experiments of [3H]MADAM binding in the rat frontal cortex. Membranes (60 µg of protein/assay) were incubated with [3H]MADAM at concentrations between 20 and 1500 pM in 1 ml of a Tris-HCl buffer, pH 7.4, for 90 min at 22°C. Nonspecific binding was determined in the presence of 106 M paroxetine. A, one representative saturation curve; B, Scatchard transformation of the data.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Displacement of [3H]MADAM binding in rat frontal cortex by various drugs. Membranes were incubated with 50 pM of [3H]MADAM together with increasing quantities of drugs (1013 to 105 M). Each curve is representative of three independent experiments, each performed in triplicate. The inhibitory constants of the drugs are given in Table 2.

View this table: [in this window] [in a new window]   TABLE 1 Inhibition of specific [3H]MADAM binding in rat cortical membranes Fifty picomoles of [3H]MADAM were incubated with 60 µg of protein in Tris-HCl buffer, pH 7.4, for 60 min at 22°C in the presence of different drugs at concentrations of 105 to 1013 M. Total binding was determined in the absence of drugs and nonspecific binding in the presence of 105 M paroxetine. IC50 values were determined graphically, and the Ki values were calculated according to the method of Cheng and Prussoff (1973): Ki = 1/[1 + (1/Kd)]. Values are the mean ± S.D. of three independent experiments, each performed in triplicate.

Autoradiographic Studies. As shown on Fig. 3 and Table 2, the binding of [³H]MADAM to rat brain sections was consistent with the distribution of serotonin transporters in rat brain regions such as the frontal cortex, latero-dorsal thalamus, superior colliculi, and raphe nuclei. Binding was totally abolished in the presence of 1 µM paroxetine (not shown).

View larger version (92K): [in this window] [in a new window]   Fig. 3.   Distribution of [3H]MADAM (500 pM) in the rat brain after in vitro binding experiment on coronal cerebral sections. Cerebral regions were defined according to Paxinos and Watson's atlas (1986). A, 1 = frontal cortex; B, 1 = cingulate cortex, 2 = caudate-putamen, 3 = accumbens nucleus; C, 1 = hippocampus, 2 = latero-dorsal thalamus, 3 = dorso-median thalamus, 4 = ventral thalamus, 5 = hypothalamus nuclei, 6 = amygdaloid nucleus; D, 1 = superior colliculus, 2 = dorsal raphe, 3 = median raphe.

View this table: [in this window] [in a new window]   TABLE 2 Autoradiographic analysis of [3H]MADAM binding to several areas of coronal rat brain sections Sections were incubated with 500 pM of [3H]MADAM either alone (total binding) or in the presence of 1 mM paroxetine (nonspecific binding). Films were exposed for 4 weeks. Regional optical densities were measured on identified anatomical regions according to Paxinos and Watson's atlas (1986) and then converted to femtomoles per milligram of tissue according to 3H microscales. The results given are the mean ± S.D. for four rats.

View larger version (77K): [in this window] [in a new window]   Fig. 4.   Human whole hemisphere autoradiography of horizontal cryosections showing [3H]MADAM (1 nM) total binding (A) and binding in the presence of 10 µM fluoxetine (B). The frontal lobe faces right. Enlargement showing total binding (C) and binding in the presence of 10 µM fluoxetine (D) in the accumbens nucleus, hippocampus, and superior colliculus. Enlargement showing total binding (E) and binding in the presence of 10 µM fluoxetine (F) in the amygdala, raphe nucleus, and cerebellum. The numbers in lower right corner indicate brain number and the distance from the vertex in millimeters.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Numerous recent studies have disclosed relationships between disturbances in serotoninergic neurotransmission, cerebral sites of action of antidepressant drugs, and behavioral disorders. Imaging of the serotonin transporter, a major element of the function of serotonin systems, could be useful to explain the role of these systems in the pathophysiological processes of these disorders and to improve therapeutic strategies. To date, few radiotracers have had optimal properties to allow in vivo study of the serotonin transporter, due mainly to high in vivo nonspecific binding or inappropriate kinetics. Derivative compounds from the diphenyl sulfide structure (Ferris et al., 1995) have recently been proposed as potent tracers for scintigraphic exploration of the serotonin transporter, in particular iodo-2-((2-((dimethylamino)methyl)phenyl)thio)benzyl alcohol (Oya et al., 1999), ADAM (Oya et al., 2000; Acton et al., 2001), and DASB (Ginovart et al., 2001). Our team recently described the synthesis and evaluation of several new compounds from this chemical family (Emond et al., 2002); among these derivatives, MADAM appeared to be a good candidate because of its high selective binding to the serotonin transporter compared with dopamine and norepinephrine transporters. In addition, preliminary in vivo scintigraphy in monkeys showed that accumulation of [¹¹C]MADAM was rapid and highly specific in brain regions rich in serotonin transporters (Halldin et al., 2001b). Precise understanding of the pharmacological profile of such a radioactive probe, however, is essential for reliable interpretation of data obtained with it. We therefore characterized this profile using [³H]MADAM in vitro. This ligand bound to the serotonin transporters of cortical membranes with very high affinity, K_d around 60 pM, and a mean B_max value of 543 fmol/mg of protein. Our data fitted with the binding of [³H]MADAM to a one-site model, in agreement with results obtained in same experimental conditions with ADAM, a compound with a closely related structure (Choi et al., 2000). It cannot be excluded that [³H]MADAM may bind to multiple sites, as the buffer used (high sodium concentration and Tris) has already be shown to interfere with binding sites, as described for binding of cocaine derivatives to monoamines transporters (Calligaro and Eldefrawi, 1988; Laruelle et al., 1994). A single binding site, however, has already been determined in rat cortical membrane with different ligands of the serotonin transporter such as [³H]paroxetine (Habert et al., 1985), [³H]citalopram (D'Amato et al., 1987), and [¹²⁵I]-CIT (Boja et al., 1992). The affinity of [³H]MADAM for the serotonin transporter was at least 2 times greater than described for the most commonly used in vitro radioactive probes such as [³H]paroxetine (Habert et al., 1985), [³H]citalopram (D'Amato et al., 1987), and [¹²⁵I]-CIT (Boja et al., 1992), whereas the B_max was of the same order for [³H]MADAM and other tracers. Competition studies toward [³H]MADAM showed no inhibitory effect of desipramine (norepinephrine transporter) or GBR 12935 (dopamine transporter), a poor inhibitory effect of nisoxetine (norepinephrine transporter) and PE2I (dopamine transporter), and a strong effect of serotonin transporter ligands such as citalopram, paroxetine, and ADAM. This showed that MADAM had around 1000-fold selectivity for the serotonin transporter over the norepinephrine transporter and dopamine transporter, comparable to that of ADAM (Oya et al., 2000) and DASB (Wilson et al., 2000).

In vitro binding studies on cerebral rat slices followed by autoradiographic analysis provided characterization of the distribution of the tracer among brain areas. The results showed widespread [³H]MADAM binding throughout cerebral regions, fitting well with known serotonin transporter localization and serotoninergic innervation (Vergé and Calas, 2000). The highest levels of [³H]MADAM were found in the dorsal raphe, superior colliculi, frontal cortex, and latero-dorsal thalamic nuclei. The very high binding of [³H]MADAM in the dorsal raphe, corresponding to the serotonin transporters localized on serotoninergic cell bodies and in regions of nerve projections, such as the superior colliculi and latero-dorsal thalamic nuclei, is in agreement with previous studies using other serotonin transporter ligands such as [³H]paroxetine (De Souza and Kuyatt, 1987; Hrdina et al., 1990; Hrdina and Vu, 1993) and [³H]citalopram (D'Amato et al., 1987; Hébert et al., 2001). Some discrepancies, however, appeared between the results obtained with [³H]MADAM and these other ligands, e.g., a relatively high binding in the frontal cortex and a relatively low binding in the caudate-putamen. The high binding in the frontal cortex could be due to our choice of a more anterior cut level for quantification (Bregma 3.70 according to Paxinos and Watson, 1986) than in other studies, whereas the low intensity of striatal binding remains unclear. The distribution of [³H]MADAM binding in the human brain as determined by autoradiograms roughly paralleled that of the autoradiographic study in rats. The highest binding density in the brain was found in the superior colliculus and was totally abolished in the presence of fluoxetine. The strong labeling of the pineal gland indicates the presence of serotonin transporters in this structure. Serotonin is a precursor of the pineal gland hormone melatonin, and it has been shown that mammalian pinealocytes contain more 5-hydroxytryptamine than any other cells (Hayashi et al., 1999). To our knowledge, however, there has been no demonstration of serotonin transporters in the pineal gland. Ligands for the serotonin transporter have been shown to label the vesicular monoamine transporters in the pineal gland (Segonzac et al., 1985; Darchen et al., 1989), a property that may be shared with [³H]MADAM. Vesicular monoamine transporters, however, are located in most brain regions (Darchen et al., 1989), and it is therefore unlikely that the exceptional binding of [³H]MADAM in the pineal gland is to a vesicular monoamine transporters. Previous findings have shown high levels of norepinephrine transporters to the pineal gland (Madras and Kaufman, 1994), but this is not the cause of this strong [³H]MADAM binding. Taken together, autoradiographic studies in rat and human brains showed that [³H]MADAM had high specific binding to the serotonin transporter, with low nonspecific binding. In view of in vivo use of MADAM in humans, it would be now necessary to precisely characterize several of its properties such as the lipophilicity, association/dissociation rate, and plasma protein binding.

In conclusion, the high affinity and selectivity of [³H]MADAM for the serotonin transporter show that this ligand is very valuable for further in vitro studies of this transporter with homogenate assays as well as autoradiography, especially in animal models of serotoninergic system disorders. The high selectivity and low nonspecific binding of [³H]MADAM on postmortem human brains, together with preliminary in vivo results obtained with [¹¹C]MADAM in monkey brains (Halldin et al., 2001b), is a new argument for future use of this ligand for in vivo studies of the distribution, pharmacology, and pathophysiology of the serotonin transporter in the human brain with PET.