Introduction

Summary Introduction Procedures Results Discussion References

5-HT exerts a wide variety of actions in the central and peripheral nervous systems by stimulating 14 different receptor subtypes (1). With the exception of 5-HT₃ receptors, which form an ion channel, all the known subtypes belong to the superfamily of receptors coupled to G proteins. The 5-HT₁ class (5-HT_1A, 5-HT_1B, 5-HT_1D, 5-HT_1E, and 5-HT_1F) is linked to adenylate cyclase inhibition. 5-HT₂ receptors (5-HT_2A, 5-HT_2B, and 5-HT_2C) increase phosphatidylinositide turnover. 5-HT₄, 5-HT₆, and 5-HT₇ receptors stimulate adenylate cyclase. The second messenger system used by 5-HT_5A and 5-HT_5B receptors is unknown. Although most of this diversity has been demonstrated in recent years by molecular cloning techniques, radioligand binding techniques (both on brain homogenates and tissue sections) have been instrumental in the discovery of the first receptor subtypes. Binding studies are easier and faster than physiological or biochemical (second messenger) experiments. They have, however, two major drawbacks: they are not useful to predict agonist efficacy and they require a high affinity radioligand for the target receptor.

The activation of G proteins by specific receptors has been assayed by measuring [³⁵S]GTPS binding in isolated membrane preparations (2). This nucleotide is an analogue of GTP, which is exchanged for GDP bound to the  subunit of the G protein after its activation by the agonist/receptor complex. Unlike GTP, [³⁵S]GTPS cannot be hydrolyzed by the intrinsic GTPase activity of the  subunit, and its incorporation into the membrane can be measured after filtration by liquid scintillation counting. This method addresses successfully the drawbacks mentioned above for radioligand binding studies. It has been used in membrane preparations for several receptors, including adenosine A₁ (3); acetylcholine muscarinic M₂ (4); µ-opioid (5); dopamine D₂ and D₃ (6); metabotropic glutamate₂, metabotropic glutamate₄, and metabotropic glutamate₆ (7, 8); and 5-HT_1A (9), 5-HT_1B, and 5-HT_1D (10) receptors. In all these reports, agonist-stimulated binding of [³⁵S]GTPS could be observed only in the presence of 1-10 µM GDP, which seemed to be required to keep G proteins in their GDP-liganded form. Recently, Sim et al. (11) adapted this technique (essentially by increasing the GDP concentration to 2 mM) to autoradiographically visualize [³⁵S]GTPS binding in brain sections after activation of µ-opioid, cannabinoid, and -aminobutyric acid_B receptors.

We now report the use of this autoradiographic approach to study the regional pattern of receptor-stimulated [³⁵S]GTPS binding in guinea pig and rat brain using drugs active at different 5-HT receptor subtypes (5-HT_1A, 5-HT_1B, 5-HT_1D, 5-HT_1F, 5-HT_2A, 5-HT_2C, 5-HT₄, and 5-HT₇). We also investigated the pharmacological profile of these responses and analyzed the distribution of [³⁵S]GTPS labeling at the light microscopic level. [Part of this work has been presented previously in abstract form (12)].

	Experimental Procedures

Summary Introduction Procedures Results Discussion References

Materials. [³H]5-CT and [³⁵S]GTPS were obtained from New England Nuclear Research Products (Boston, MA) (specific activity, 22.8 and 1000-1500 Ci/mmol, respectively). [³H]8-OH-DPAT was obtained from Amersham (Arlington Heights, IL) (specific activity, 205 Ci/mmol). GR-127,935 (2-methyl-4-(5-methyl-[1,2,4]oxadiazol-3-yl)-biphenyl-4-carboxylic acid-[4-methoxy-3-(4-methyl-piperazin-1-yl)-phenyl]amide), naratriptan (N-methyl-2-[3,1-methylpiperidin-4-yl)-1H-indol-5-yl]ethanesulfonamide), and sumatriptan were provided by Glaxo. CP-122,288 [5-methylaminosulfonylmethyl-3-(N-methylpyrrolidin-2R-ylmethyl)-1H-indole] was obtained from Pfizer (Groton, CT). GTI was obtained from Immunotech (Marseilles, France). GTPS and GDP were purchased from Sigma Chemical (St. Louis, MO). All other drugs were obtained from Research Biochemicals (Natick, MA).

Tissues. Adult male rats (Sprague-Dawley, 200-250 g) and adult guinea-pigs (Hartley, 250-350 g) were sedated with inhaled chloroform and decapitated. The whole brain and upper cervical spinal cord was dissected and frozen in isopentane cooled at 40°. Frozen brains were sectioned using a cryostat-microtome (1720; Leica, Deerfield, IL). The sections were thaw-mounted onto gelatinized glass slides and stored at 80° for <1 month. Preliminary experiments using 8-, 14-, and 20-µm-thick sections showed that 10 µM 5-CT increased [³⁵S]GTPS binding by 1050 ± 185% (mean ± standard error), 680 ± 120%, and 450 ± 55%, respectively, in guinea pig hippocampus; therefore, 10-µm-thick sections were used in subsequent experiments.

Autoradiography. [³⁵S]GTPS binding was visualized using the method developed by Sim et al. (11), with minor modifications. Briefly, the tissue sections (from at least five different animals) were brought to room temperature 15 min before the experiment; incubated for 15 min at room temperature in 50 mM HEPES buffer, pH 7.5, containing 100 mM NaCl, 3 mM MgCl₂, 0.2 mM EGTA, and 0.2 mM dithiothreitol; and then incubated for an additional 15 min in the same buffer supplemented with 2 mM GDP. Agonist-stimulated binding was determined by incubating the sections for 60 min at 30° in buffer containing 2 mM GDP, 0.04 nM [³⁵S]GTPS, and the appropriate concentration of agonist and/or antagonist. Nonspecific binding was assessed by including 10 µM unlabeled GTPS in the incubation buffer. Slides were then washed twice for 3 min in ice-cold 50 mM HEPES buffer, pH 7.0, dipped briefly in ice cold distilled water, dried under a stream of cold air, and exposed to Hyperfilm max (Amersham) for 24 hr. All experiments were performed independently at least twice.

Light microscopic autoradiography. After exposure to Hyperfilm max films, selected slides were coated with Kodak NTB2 liquid emulsion (diluted 1:1 with water and maintained at 40°), allowed to dry in a humid chamber, and kept at 4° for 2 weeks in boxes containing Silicagel. They were developed in Kodak D-19 (diluted 1:1 with water) for 3 min at 16° and fixed with Kodak Polymax fixer (diluted 1:8 with water). The sections then were stained with 1% basic fuchsin for 10 sec and coverslips were affixed.

Image analysis. The absorbance of the autoradiograms was measured over selected brain regions using a computerized image analysis system (M4; Imaging Research, St. Catherines, Ontario, Canada). [³⁵S]GTPS autoradiograms were analyzed by comparing the absorbance of the films with the absorbance of a Kodak calibrated density step tablet. Because parallel experiments with ¹⁴C standards indicated that the maximal absorbances observed on [³⁵S]GTPS autoradiograms were still within the linear domain of the film exposure-response curve (except for WIN 55212-2-stimulated binding in globus pallidus and substantia nigra), radioactive standards were not routinely used. Agonist-induced [³⁵S]GTPS binding is expressed as percentage of basal binding.

Data analysis. Data points from autoradiographic measurements were fitted by nonlinear regression using Grafit (Erithacus Software, Staines, UK). The equation used was Stim = E_max/(1 + EC₅₀/Ago), where Stim is the stimulated binding (percent over basal), E_max is the maximal binding, EC₅₀ is the concentration of agonist resulting in half-maximal [³⁵S]GTPS binding, and Ago is the agonist concentration. The pK_B values for antagonists (NAN-190 at 5-HT_1A receptors) were calculated from the rightward shift of agonist concentration-response curves according to the formula pK_B = log (CR  1)  log (Anta), where CR is the ratio of agonist IC₅₀ values with or without antagonist, and Anta is the antagonist concentration.

	Results

Summary Introduction Procedures Results Discussion References

Unless otherwise stated, the following results refer to both guinea pig and rat brains.

Basal [³⁵S]GTPS labeling. In the absence of GDP in the incubation buffer, basal [³⁵S]GTPS labeling was homogeneously very high throughout the brain, and no increase in binding was detected in the presence of 10 µM 5-CT. As reported previously (11), optimal agonist stimulation was observed when 2 mM GDP was included. No improvement was found with 4 mM GDP, and the stimulation disappeared at 8 mM GDP (not shown). All subsequent experiments were thus carried out with 2 mM GDP. The basal [³⁵S]GTPS observed under these conditions was likely to be specific because the labeling was not different from film background when 10 µM unlabeled GTPS was included in the buffer. Basal binding (i.e., in the absence of added agonist) was regionally heterogeneous (Figs. 1A and 2A). The highest level was found in the substantia gelatinosa of the spinal cord and medulla, followed by the interpeduncular nucleus and substantia nigra; intermediate levels were also found in hippocampus, central gray, and superficial layer of the superior colliculus. The cortex and striatum bound only slightly more [³⁵S]GTPS than white matter areas (which contain very low but GTPS-displaceable labeling). The level of basal binding, in particular in the medulla, was not decreased when the animals were killed with pentobarbital overdose (no decapitation), when the sections were subjected to a longer preincubation (even in the presence of 1 mM GTP to accelerate the dissociation of endogenous ligands), or by increasing the washing time up to 15 min (results not shown). None of the antagonists used in the study (NAN-190, GR-127,935, or methiothepine) produced a detectable reduction of basal labeling.

View larger version (179K): [in this window] [in a new window]   Fig. 1.   Distribution of [35S]GTPS and [3H]5-CT binding sites in guinea pig brain at the level of posterior hippocampus (similar patterns were observed in four other animals in two independent experiments). A, Basal [35S]GTPS binding is low, but heterogeneous, with higher levels in substantia nigra (SN), hippocampus [dentate gyrus (DG); CA1, subfield of Ammon's horn], interpeduncular nucleus (IP), superior colliculus (SuG), and central gray (CG). B, 5-CT (10 µM) markedly increases [35S]GTPS binding in hippocampus, superior colliculus, and entorhinal cortex (Ent) but does so only weakly in substantia nigra. The increase in the latter region is blocked by the 5-HT1B/1D antagonist GR-127,935 (10 µM; D); in the former three regions, it is blocked by the 5-HT1A antagonist NAN-190 (10 µM; C). The labeling pattern obtained using [3H]5-CT as a ligand is shown (E); comparably high densities of binding sites are seen in hippocampus, superior colliculus, and substantia nigra. Intermediate densities of sites are also found in the superficial and deep cortical layers. The addition of blockers of the 5-HT1D (10 µM GR-127, 935) and 5-HT7 (1 µM spiperone) receptor subtypes does not affect 5-HT1A sites (F), concentrated in hippocampus, superior colliculus, and deep cortical layers. Note that 10 µM 5-CT increases [35S]GTPS binding in these deep layers but not in the superficial layers (B), which are known to contain 5-HT7 receptors. Scale bar, 5 mm.

View larger version (189K): [in this window] [in a new window]   Fig. 2.   Effect of several 5-HT1B/1D receptor agonists on [35S]GTPS binding in guinea pig brain at the level of posterior hippocampus (similar patterns were observed in four other animals in two independent experiments). A, Basal [35S]GTPS binding. Labeling patterns observed in the presence of 1 µM L-694,247 (B), 5-(nonyloxy)-tryptamine (C), naratriptan (D), and GTI (E). Enhanced binding is found not only in substantia nigra (containing 5-HT1D receptors) but also in hippocampus (containing 5-HT1A receptors). In the presence of 10 µM NAN-190 (F), GTI-enhanced [35S]GTPS binding disappears in hippocampus and superior colliculus. The binding remaining in substantia nigra is significantly higher than basal binding (t test, p < 0.01). Note that naratriptan (D), despite its high potency at 5-HT1F receptors, does not increase binding in the intermediate cortical layers (arrowheads), known to contain high densities of 5-HT1F receptors. Scale bar, 5 mm.

5-CT stimulated [³⁵S]GTPS labeling. The potent, but nonselective, 5-HT receptor agonist 5-CT (10 µM) increased [³⁵S]GTPS binding very strongly in the hippocampal formation and lateral septum (latter area not shown) but only weakly in the superficial gray layer of the superior colliculus, central gray, interpeduncular nucleus, substantia gelatinosa of the medulla (trigeminal nucleus caudalis; not shown), neocortex, substantia nigra, and globus pallidus (not shown) (Fig. 1). The selective 5-HT_1B/1D receptor antagonist GR-127,935 (10 µM) inhibited this effect only in substantia nigra and globus pallidus, whereas the selective 5-HT_1A receptor antagonist NAN-190 (10 µM) was effective in the other areas. A strikingly different labeling pattern was found when [³H]5-CT was used as a radioligand. It labeled equally high densities of sites in hippocampus, substantia nigra, and superior colliculus. Intermediate densities of [³H]5-CT binding sites [known to correspond to 5-HT₇ receptors (13)] were also observed in the superficial cortical layers. Fig. 1F shows the labeling pattern obtained after blockade of [³H]5-CT binding to 5-HT_1D (with 100 nM GR-127,935) and 5-HT₇ receptors (with 1 µM spiperone); it is mostly accounted for by 5-HT_1A receptors.

5-HT_1B/1D receptor stimulated [³⁵S]GTPS labeling. The effect of a series of agonists with high affinity for 5-HT_1B/1D receptors was examined in guinea pig brain (Fig. 2). At 1 µM, L-694,247 (Fig. 2B), 5-(nonyloxy)-tryptamine (Fig. 2C), naratriptan (Fig. 2D), GTI (Fig. 2E), and sumatriptan (not shown) all stimulated [³⁵S]GTPS binding in the substantia nigra. Varying degrees of stimulation were also observed in regions known to contain 5-HT_1A receptors (hippocampus, lateral septum, and superior colliculus). This effect was particularly strong with 10 µM L-694,247, which increased binding in the latter areas to a higher level than in the substantia nigra. The stimulation by 10 µM GTI in hippocampus, lateral septum, and superior colliculus (but not substantia nigra) was abolished in the presence of 10 µM NAN-190 (a 5-HT_1A receptor antagonist; Fig. 2F). In the absence of agonist, the 5-HT_1B/1D receptor antagonist GR-127,935 (10 µM) had no effect on [³⁵S]GTPS binding in guinea pig substantia nigra (not shown).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Effect of increasing concentrations of the 5-HT1B/1D receptor agonist 5-(nonyloxy)-tryptamine on [35S]GTPS binding in substantia nigra (open symbols) and hippocampus (closed symbols) of guinea pig (squares) and rat (circles). Data are expressed as percent increased [35S]GTPS binding over basal binding (error bars, mean ± standard error values observed between different animals in an individual experiment) and are representative of two independent experiments. The agonist shows comparable potencies in both species and is less potent in hippocampus ( and ) than in substantia nigra ( and ). Maximal stimulations of [35S]GTPS binding are 54%, 94%, and 120% in guinea pig substantia nigra, rat substantia nigra, and rat hippocampus, respectively. At 100 µM, 5-(nonyloxy)-tryptamine increases [35S]GTPS binding by 190% in guinea pig hippocampus.

5-HT_1A receptor stimulated [³⁵S]GTPS labeling in guinea pig brain. Fig. 4 shows the regional distribution of [³⁵S]GTPS binding stimulated by 10 µM (R)-8-OH-DPAT (Fig. 4, A-D) compared with the distribution of [³H]8-OH-DPAT binding sites at similar levels of the guinea pig brain (Fig. 4, A-D). Both techniques reveal a virtually identical distribution of recognition sites. This agreement is particularly noticeable in the hippocampal formation, in which both [³⁵S]GTPS- and [³H]8-OH-DPAT-labeled sites are concentrated in the molecular layer of the dentate gyrus and in strata oriens and radiatum of Ammon's horn. Much lower densities of bound tracers were seen in the polymorph and granule cell layers of the dentate gyrus and in the pyramidal cell layer of Ammon's horn, indicating that they were present mainly in the dendritic fields of pyramidal and granule cells (see discussion of light microscopy).

View larger version (100K): [in this window] [in a new window]   Fig. 4.   Distribution of [3H]8-OH-DPAT binding sites (A-D) and [35S]GTPS binding stimulated by 10 µM (+)-8-OH-DPAT (A-D) at several levels of guinea pig brain (similar patterns were observed in four other animals in two independent experiments). The labeling pattern of both ligands showed an excellent correlation in all brain regions investigated. Basal [35S]GTPS labeling (not shown on this figure) was similar to that displayed in Figs. 1A, 2A, and 7A. Nonspecific [3H]8-OH-DPAT labeling could not be distinguished from film background. High densities of binding sites for both radioligands were observed in the lateral septum [dorsal and ventral nuclei (LSD and LSV)], hippocampal formation [particularly in the CA1 subfield and dentate gyrus (DG)], superficial gray layer of the superior colliculus (SuG), dorsal raphe nucleus (DR), interpeduncular nucleus (IP), and entorhinal cortex (Ent), as well as in the deepest cortical layers (arrowheads). The laminar and subfield distributions of [35S]GTPS and [3H]8-OH-DPAT binding sites in hippocampus overlap completely. Scale bars, 5 mm.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Effect of increasing concentrations of three 5-HT1A receptor agonists known to possess various intrinsic activities on [35S]GTPS binding in guinea pig hippocampus. Data are expressed as percent increased [35S]GTPS binding over basal binding (error bars, mean ± standard error values observed in an individual experiment) and are representative of three independent experiments. , (R)-(+)-8-OH-DPAT; , (S)-()-8-OH-DPAT; , buspirone data points. , 5-HT1A receptor antagonist NAN-190 has no effect of [35S]GTPS binding (10 µM).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Effect of increasing concentrations of the 5-HT1A receptor antagonist NAN-190 on the dose-response of (R)-(+)-8-OH-DPAT on [35S]GTPS binding in guinea pig hippocampus. NAN-190 (, 10 nM; , 100 NM) produced a parallel rightward shift in the (R)-(+)-8-OH-DPAT dose-response curve (, absence of antagonist). Similar effects were observed in the lateral septum, raphe dorsalis, and superior colliculus (not shown).

Effect of 5-HT_1F, 5-HT₂, 5-HT₄, and 5-HT₇ receptor agonists on [³⁵S]GTPS labeling. Both 1 µM (see above) and 30 µM naratriptan (not shown) or 10 µM CP-122,288 (not shown) increased [³⁵S]GTPS labeling in the substantia nigra (5-HT_1B/1D receptors) and hippocampus (5-HT_1A receptors). However, in the presence of 10 µM concentration of both GR-127,935 and NAN-190, no enhanced [³⁵S]GTPS binding was observed with these compounds [in particular, not in the claustrum and neocortex, in which high densities of 5-HT_1F receptors have been reported (14)], despite the fact that both drugs possess nanomolar affinity for 5-HT_1F sites (10,14). [³H]5-CT has been shown to label high densities of 5-HT₇ sites in the superficial cortical layers and midline thalamic nuclei (13); however, no 5-CT-enhanced [³⁵S]GTPS labeling was detected in these areas (10 µM 5-CT). Finally, neither the 5-HT_2A/2C agonist (±)-1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane nor the 5-HT₄ agonist SC 53116 (4-amino-5-chloro-N-[(hexahydro-1H-pyrrolizin-1-yl)methyl]2-methoxybenzamide) was able to increase [³⁵S]GTPS binding in any brain area (both agonists were used at a concentration of 10 µM).

Species differences in agonist-induced [³⁵S]GTPS labeling. As mentioned above, 5-(nonyloxy)-tryptamine stimulated [³⁵S]GTPS binding more strongly in rat substantia nigra (as well as in mouse substantia nigra; not shown) than in guinea pig substantia nigra. The stimulation induced by agonists for other receptor classes (10 µM) was compared in rats and guinea pig (Fig. 7). In guinea pig, the [³⁵S]GTPS labeling in the presence of the µ-opiate agonist DAMGO or muscarinic M₂ agonist oxotremorine barely differed from the basal labeling (with the exception of the superficial gray layer of the superior colliculus and midline thalamic nuclei, in which binding was enhanced strongly by oxotremorine and the mammillary nuclei, densely labeled in the presence of DAMGO). In contrast, both agonists produced a marked and heterogeneous increase in [³⁵S]GTPS binding in various rat brain areas. DAMGO-enhanced labeling was observed in striatal patches (probably corresponding to striosomes), globus pallidus, several thalamic nuclei, and substantia nigra pars compacta. Oxotremorine increased binding in the rat striatum and superficial gray layer of the superior colliculus. The histamine H₃ receptor agonist imetit stimulated binding weakly in guinea pig striatum, globus pallidus, and substantia nigra; a similar pattern was found in rats but with higher densities of bound [³⁵S]GTPS. Only stimulation by the cannabinoid agonist WIN55,212-2 resulted in comparable [³⁵S]GTPS binding pattern in both species, with very dense labeling observed in the globus pallidus and substantia nigra, as well as, to a lesser extent, the hippocampus and cortex.

View larger version (97K): [in this window] [in a new window]   Fig. 7.   Effect of agonists for several Gi/o protein-coupled receptors on [35S]GTPS binding in guinea pig (A-E) and rat (F-J) brain (similar patterns were observed in four other animals of each species in two independent experiments). Basal [35S]GTPS labeling patterns are similar in both species, as are the labeling obtained in the presence of the cannabinoid agonist 10 µM WIN55,212-2 (B and G). However, stimulation of µ-opioid, histamine H3, and muscarinic M2 receptors with 10 µM DAMGO (C and H), imetit (D and I), or oxotremorine (E and J), respectively, results in only weak stimulation of [35S]GTPS binding in guinea pig brain, whereas marked and heterogeneous changes are observed in rat brain. DAMGO-enhanced labeling is prominent in rat globus pallidus (GP), striatal patches [probably corresponding to striosomes (arrowheads)], and a striatal subcallosal streak (arrow) typically observed with µ-opiate receptor radioligands. Imetit and oxotremorine increase [35S]GTPS binding in globus pallidus and caudate-putamen (CPu), respectively. Scale bars, 5 mm.

Light microscopic autoradiography. Fig. 8 shows the distribution of silver grains over different guinea pig brain areas labeled with [³⁵S]GTPS (dark field microscopy). In the absence of agonist (Fig. 8B), virtually no autoradiographic grains were found over the pyramidal cell layer of Ammon's horn (CA1) and the granular cell layer of dentate gyrus and only a low level of diffuse labeling was seen over the surrounding layers. In the presence of 1 µM 8-OH-DPAT (Fig. 8C), no increase in labeling was observed over the pyramidal and granular cell layers, and only a moderate increase was found in the polymorph layer of the dentate gyrus. In contrast, labeling was increased markedly in the strata oriens and radiatum of Ammon's horn and the molecular layer of dentate gyrus. In the superficial gray layer of the superior colliculus, 1 µM 8-OH-DPAT (Fig. 8D) seemed to increase binding homogeneously and mostly in the neuropil, whereas only few grains were observed directly over the cells. A similar distribution was observed in the globus pallidus (Fig. 8E) and substantia nigra reticulata (Fig. 8F) in the presence of a cannabinoid agonist, 1 µM WIN55,212-2.

View larger version (203K): [in this window] [in a new window]   Fig. 8.   High-resolution autoradiography of [35S]GTPS binding sites in guinea pig hippocampus (B and C), superficial gray layer of the superior colliculus (D), globus pallidus (E), and substantia nigra (F). [35S]GTPS-labeled slides have been coated with Kodak NTB2 liquid emulsion as described in Experimental Procedures. A, Basic fuchsin-counterstained cell nuclei in the ventral hippocampus. B, Autoradiographic silver grains (white dots under dark field illumination) over the same section, incubated with [35S]GTPS in the absence of agonist: labeling is low and homogeneous, except over the cell bodies, where only background labeling is observed. In the presence of 1 µM 8-OH-DPAT (C), binding is enhanced markedly in the strata oriens (Or) and radiatum (Rad) of Ammon's horn and the molecular layer of the dentate gyrus (MolDG) but virtually absent over the pyramidal and granular cell layers (CA1, pyr, and Gran). Only a small increase is seen in the polymorph layer of dentate gyrus (PoDG). At a higher magnification, 8-OH-DPAT-stimulated binding of [35S]GTPS in the superior colliculus (D), as well as WIN55,212-2-stimulated binding in the globus pallidus (E) and substantia nigra (F), produces homogeneously distributed silver grains in the neuropil of these areas, whereas only very few grains are observed over the cell bodies (arrowheads). Scale bars, 1 mm (A-C) and 0.2 mm (D-F).

	Discussion

Summary Introduction Procedures Results Discussion References

5-HT is known to interact with 14 different receptor subtypes, of which 13 are coupled to G proteins (1). Taken together, the drugs used in the current study have a high affinity for all these receptors, with the exception of 5-HT_1E, 5-HT_2B, and 5-HT₆ sites. G protein activation by only two (5-HT_1A and 5-HT_1B) of the remaining 10 receptors was detected using [³⁵S]GTPS autoradiography. None of the drugs used in this study discriminate between 5-HT_1B and 5-HT_1D receptors; however, in consideration of the predominance of 5-HT_1B receptors in the mammalian brain (15,16), it is likely that these sites account for the enhancement of [³⁵S]GTPS binding by 5-HT_1B/5-HT_1D agonists.

The failure to detect G protein activation by 5-HT_2A/C and 5-HT₄ receptors probably reflects the fact that these receptors are coupled to pertussis toxin-insensitive G proteins, and receptors coupled to these classes of G protein have not been reported to stimulate [³⁵S]GTPS binding in brain membrane preparations. In contrast, there are numerous accounts of enhanced [³⁵S]GTPS binding induced by G_i- or G_o-coupled receptors (see Introduction). Recently, this technique was adapted and used on brain sections to visualize the activation of cannabinoid, -aminobutyric acid _B, and µ-opioid (11), opioid receptor-like (17), -opioid (18), and somatostatin (19) receptors. The results of the current report add muscarinic M₂ and histamine H₃ receptors to the list of G_o/i-coupled receptors activating [³⁵S]GTPS binding in brain sections.

Several properties might account for the indirect labeling of receptors coupled to pertussis toxin-sensitive, but not other, G proteins (20). First, G_o and G_i are the major G proteins in the brain. Second, activation of the  subunit of G_s requires much higher (25-50 mM) Mg²⁺ concentrations than activation of G_i or G_q  subunits (3 mM Mg²⁺ was used in this and previous studies). Finally,  subunits of different classes show various rates of spontaneous GDP dissociation (2). At 30°, G_i proteins show a rapid spontaneous dissociation of GDP from their  subunits; excess GDP must be added to shift the equilibrium toward the  subunit/GDP complex and allow the detection of agonist-enhanced [³⁵S]GTPS binding. Alternatively, the reaction can be carried out at 4° but in the absence of added GDP and NaCl (2) (which probably uncouples G proteins from unoccupied receptors). The latter approach has not been used in brain sections, although it might be an interesting alternative to the approach used in this and previous studies. In addition to being relatively expensive, the large amounts of GDP added in the incubation medium might be responsible for the low potencies of agonists observed with this method. Indeed, differences in GDP concentrations might account for the lower EC₅₀ value of 8-OH-DPAT in our system (50-70 nM) compared with that obtained on membrane preparations in the presence of only 3 µM GDP (6 nM) (9).

At variance with G_i proteins, G_s proteins show a relatively slow spontaneous dissociation of GDP, and the addition of GDP or NaCl only decreases isoproterenol-induced [³⁵S]GTPS binding to turkey erythrocyte membranes (2). Several strategies have been proposed to reduce agonist-independent [³⁵S]GTPS binding on these membranes (2); their use on cryostat brain sections might permit detection of activated receptors coupled to G proteins other than G_i or G_o.

Because 5-HT_1F receptors have been reported to inhibit adenylate cyclase in transfected cells (21), the lack of effect of naratriptan and CP-122,288 on [³⁵S]GTPS labeling (in the presence of 5-HT_1A and 5-HT_1B/1D receptor blockers) is unexpected. The binding affinities (K_D) of naratriptan and CP-122,288 for 5-HT_1F binding sites are 4 nM (10) and 1.6 nM (14), respectively (i.e., comparable to the affinities of the 5-HT_1A and 5-HT_1B/1D agonists used in this study for their respective receptors). Furthermore, 5-HT_1D and 5-HT_1F recognition sites are present at comparable densities in guinea pig brain (14). Several explanations might account for the absence of effect of naratriptan and CP-122,288. Agonist binding to 5-HT_1F sites might be more sensitive to high concentrations of GDP, or 5-HT_1F receptor might be coupled to a different subtype of G_i/o protein than the other receptors visualized with [³⁵S]GTPS autoradiography. Adham et al. (21) reported that 5-HT_1F receptors can couple to multiple signal transduction pathways via pertussis toxin-sensitive G proteins, possibly via distinct subtypes of G proteins. It is possible that native brain 5-HT_1F receptors do not interact with the G protein subtype leading to inhibition of adenylate cyclase and/or with a subtype prone to detectable [³⁵S]GTPS binding.

Concerning the difference in maximal [³⁵S]GTPS binding with 5-HT_1A and 5-HT_1D agonists, it is worth noting that [³H]5-CT labels a comparable density of sites in hippocampus (5-HT_1A sites) and substantia nigra (5-HT_1D sites) (13). It is thus likely that 5-HT_1A receptors possess a larger amplification factor than 5-HT_1D receptors. This raises the issue of whether drugs assumed to be selective for 5-HT_1D receptors would activate more 5-HT_1A receptors in vivo than expected on the basis of their selectivity ratios. There are, however, very few systems in which this possibility can be explored because it is difficult to determine the contribution of the initial G protein activation step if distinct pathways lead from receptor stimulation to functional response. This question is nevertheless of interest because for most of the available drugs, the 5-HT_1A/5-HT_1D selectivity ratio (indicated in parentheses) is rather low: L-694,247 (3) (Ref. 22), naratriptan (106) (Ref. 10), sumatriptan (10-60) (Refs. 10 and 23), CP-122,288 (6) (Refs. 10 and 23), and GTI (21) (Ref. 24). 5-(Nonyloxy)-tryptamine, with a 5-HT_1A/5-HT_1D affinity ratio of 260 (Ref. 25), is one of the most selective agents, but at concentrations of >10 µM, it activates more 5-HT_1A than 5-HT_1D receptor-linked G proteins.

When compared with in vitro autoradiography with the use of radiolabeled receptor ligands, [³⁵S]GTPS autoradiography offers significant advantages, with only a few drawbacks. The major disadvantage is the fact that the current method is not applicable to receptors coupled to pertussis toxin-insensitive G proteins or even to some receptors coupled to G_i or G_o (e.g., 5-HT_1F receptors). Its quantification might also be less reliable because it is not known with certainty whether all subtypes of G proteins respond in the same manner in this system. The cause for the differences between rat and guinea pig brains is also unknown, and it cannot be ruled out that similar differences exist between brain regions and were unnoticed in this and previous reports. Minor shortcomings are the smaller resolution of ³⁵S versus ³H (used to label most receptor ligands) and the fact that film darkening with ³⁵S depends on the section thickness (a very reproducible cryostat-microtome should thus be used; only the first 4-5 µm of a ³H-labeled section are responsible for film exposure). The main advantage of [³⁵S]GTPS autoradiography is that one can determine agonist efficacies in various brain regions. This technique can be used even in the absence of a suitable receptor radioligand and requires short exposure times (1-2 days versus 2 weeks to 6 months for conventional autoradiography). The current report also shows that ³⁵S-GTPS-labeled sections can be directly coated with nuclear emulsion and potentially resolve receptor distribution at the cellular level. This can be achieved only because [³⁵S]GTPS binds to its target in a virtually irreversible manner (26). In contrast, receptor radioligand binding is usually reversible (even more so at the temperature required to coat the slides with nuclear emulsion), and cross-linking to the receptor can be performed only for selected ligands (in general peptides). The light microscopic distribution of [³⁵S]GTPS binding stimulated by 8-OH-DPAT is very similar to that reported previously for [³H]8-OH-DPAT binding sites (27). Interestingly, most of the [³⁵S]GTPS binding to activated G proteins occurs in the neuropil (in the superior colliculus, globus pallidus, and substantia nigra) or on the cell processes (hippocampus), which is in agreement with the expected distribution of G proteins (28, 29). Finally, [³⁵S]GTPS autoradiography, coupled with selective antibodies or peptides (30), can potentially be used to investigate which G protein subtypes are coupled to different receptors in the brain and regional differences in this coupling.