Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The lack of ₂-adrenoceptor (AR) subtype-selective antagonists has precluded clarification of the role of each subtype in various physiological and behavioral studies (MacDonald et al., 1997). Homologous recombination strategies that lead either to gene knockouts or a substitution of a mutant receptor at the wild-type locus in the mouse genome have been developed. Using genetically engineered mouse strains it has been suggested that the _2C-AR subtype has a distinct inhibitory role in the processing of sensory information and in the control of motor and emotion-related activities in the central nervous system (Scheinin et al., 2001). It is therefore possible that _2C-AR-selective drugs may have therapeutic value in the treatment of various neuropsychiatric disorders. In addition to the central nervous system effects associated with _2C-AR, a remarkable role of _2C-AR in vascular dysfunction has recently been discovered. The cold-induced cutaneous arterial blood vessel constriction observed in Raynaud's disease may be linked to increased reactivity of smooth muscle _2C-AR. _2C-ARs are not functionally responsive at 37°C, i.e., silent in normal regulation of the vascular function, but are activated by catecholamines at 28°C (Chotani et al., 2000). Thus, selective blockade of these receptors by using _2C-AR-selective antagonists may provide a new therapeutic intervention for the treatment of Raynaud's disease. Findings from cardiovascular studies using _2A-, _2B-, and _2C-AR knockout mice indicated that although the peripheral _2A- and _2B-AR modulate blood pressure, the _2C-AR does not play a measurable role in the control of systemic blood pressure (MacDonald et al., 1997). This finding is consistent with the notion that an _2C-AR-selective drug may offer beneficial clinical uses with a minimal potential for systemic blood pressure effects (Rizzo et al., 2001). _2C-AR subtype-selective antagonists will help to further elucidate the physiological roles of this receptor and may provide leads for a variety of therapeutic disorders.

The present article illustrates the identification of potent and selective human _2C-AR antagonists using the bivalent ligand approach. The term bivalent ligand (or pharmacophore dimer) is defined as a molecule that contains two pharmacophores linked through a spacer. This approach to receptor ligand design would be predicted to generate ligands that exhibit a potency that is greater than that derived from the sum of its two monovalent counterparts. The design relies on the concept that a bivalent ligand should first undergo univalent binding, followed by the binding of the second pharmacophore to a recognition site on a neighboring receptor (Portoghese, 2001). Here, we selected the bivalent pharmacophore of yohimbine because this compound is a selective and potent ₂-AR antagonist (Bylund, 1985). The yohimbine dimers used were evaluated on human _2A-, _2B-, and _2C-AR subtypes expressed stably in Chinese hamster ovary (CHO) cells using a radioligand binding assay. Selective yohimbine dimers were also evaluated for binding affinities on human _1A-, _1B-, and _1D-AR subtypes expressed stably in human embryonic kidney (HEK) cells. Finally, cell systems containing cDNA clones of the ₂-ARs, were used to assay agonist and antagonist agents using a CRE reporter gene functional assay. Our approach was to evaluate changes in luciferase activities, as an index of cAMP accumulations, for highly selective yohimbine dimers on human ₂-AR subtypes expressed in CHO cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Sources of Materials. All cell culture reagents were obtained from Invitrogen (Carlsbad, CA). Human _2A-, _2B-, and _2C-AR subtypes expressed in CHO cells were obtained from Drs. Marc Caron and Robert Lefkowitz (Duke University Medical Center, Durham, NC) and Dr. Stephen Liggett (College of Medicine, University of Cincinnati, Cincinnati, OH). Human _1A-, _1B-, and _1D-AR subtypes expressed in HEK cells were obtained from Dr. Kenneth Minneman (Department of Pharmacology, Emory University School of Medicine, Atlanta, GA). MK-912 was obtained from Merck (Rahway, NJ). Yohimbine dimers (Fig. 1) were provided by Dr. Duane D. Miller (Department of Pharmaceutical Sciences, University of Tennessee, Memphis, TN). The yohimbine dimers (n = 2, 3, 4, 5, 6, and 9) were prepared as diamides by coupling the commercially available yohimbinic acid with aliphatic ,-diamines under standard peptide coupling condition (Zheng et al., 2000). The procedures for the synthesis of the glycinamide derivatives (n = 18 and n = 24) are as described by Zheng (1999). All the yohimbine dimers were hygroscopic and light-sensitive, and dimers, with the exception of n = 18 and n = 24, were provided as dihydrochloride salts and were soluble in water. The dimers n = 18 and n = 24, prepared as glycinamide derivatives and supplied as the free base, were dissolved in a 1:5 mixture of dimethyl sulfoxide/water. Stock solutions of 10 mM were prepared and diluted in water to appropriate concentrations for these studies. Samples and drug solutions were protected from light. [³H]Rauwolscine and [³H]prazosin were obtained from PerkinElmer Life Sciences (Boston, MA). All other chemicals were obtained from Sigma-Aldrich (St. Louis, MO).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Structures of yohimbine dimers of varying spacer atom lengths (n = 2, 3, 4, 5, 6, 7, 9, 18, and 24).

Cell Culture. CHO cells stably expressing human _2A-, _2B-, and _2C-AR subtypes were grown in 150-cm² Corning flasks with Ham's F-12 medium supplemented with 10% fetal bovine serum, 2 mM glutamine, penicillin (100 units/ml), and streptomycin 100 µg/ml). The flasks were incubated at 37°C (5% CO₂). Media were changed every 48 h until the cells were confluent. Upon confluence, the cells were detached by addition of 0.05% trypsin EDTA for 3 to 5 min.

Radioligand Binding Assays. Radioligand binding studies were performed in CHO cells expressing human _2A-, _2B-, and _2C-AR subtypes and HEK cells expressing human _1A-, _1B-, and _1D-AR subtypes. The detached cells were washed and centrifuged with Tris-EDTA buffer, pH 7.4, in which they were finally suspended. The competition binding assays were performed in duplicate by incubating 50,000 cells with [³H]rauwolscine and [³H]prazosin for the ₂- and ₁-ARs, respectively, and varying concentrations of the compounds under investigation for 1 h in a water bath at 37°C. Nonspecific binding was defined by addition of 10 µM yohimbine and 10 µM phentolamine for the ₂- and ₁-ARs, respectively. Cell suspensions were filtered using GF/B glass fiber filters (Whatman, Maidstone, UK) using a cell harvester (model 12-R; Brandel Inc., Gaithersburg, MD). The filter discs were washed three times with 5 ml of Tris-EDTA buffer, pH 7.4, at 4°C. The radioactivity was determined using a TriCarb 2900TR liquid scintillation counter (Packard Instrument Company, Inc., Downers Grove, IL). The displacement curves were generated using GraphPad Prism software (GraphPad Software, San Diego, CA). The displacement curves were plotted using a standard slope factor of 1.0; and the K_i values of the competing ligands were determined using the equation of Cheng and Prusoff (1973). The percentage of specific binding was determined by dividing the difference between total bound (dpm) and nonspecific bound (dpm) by the total bound (dpm).

Cyclic AMP Response Element-Luciferase Reporter Gene Assay. The functional responses of the yohimbine analogs in these ₂-AR subtypes were determined with the use of six copies of a cAMP response element-luciferase reporter gene construct (6 CRE-LUC, pADneo2-C6-BGL) provided by Dr. A. Himmler (Boehringer Ingelheim Research and Demonstration, Vienna, Austria) by the same transfection procedures described previously in CHO cells (Vansal and Feller, 1999). The antagonist effects of yohimbine and the two highly selective bivalent analogs (n = 3 and n = 24) were examined for their concentration-dependent reversal of medetomidine on forskolin-induced cAMP levels, as assessed by luciferase activity changes in CHO cells expressing the human _2A- and _2C-AR subtypes. The compounds were tested under the following conditions: the 6 CRE-LUC plasmid (5 µg/100 µl of cell suspension) was transiently transfected into CHO cells expressing the _2A- and _2C-AR subtypes using electroporation, at 150 V, 70 ms, single pulse. The transfected cells were plated at a density of 50,000 cells/200 µl/well in a 96-well microtiter plate and allowed to grow for 20 h. The compounds under investigation for agonist activity were added directly to the medium and allowed to incubate for 4 h. When tested for antagonist activity, the compounds were added at varying concentrations 20 min before addition of a fixed concentration of medetomidine (10 nM). Subsequently, the media were aspirated, the cells lysed, and the luciferase activity determined using a TopCount (Packard BioScience, Meriden, CT) luminometer after adding luciferin. Data were analyzed using GraphPad Prism software and expressed as a mean ± S.E.M. The response produced by 5 µM forskolin in each experiment was used as 100%.

Data Accumulation and Statistical Analyses. For ligand binding studies in cell lines, varying concentrations of each drug/ligand (ranging from 10⁴-10¹¹ M) were added in duplicate within each experiment, and the individual molar IC₅₀ values were determined using GraphPad Prism. The K_i values of each ligand were determined according to the equation described by Cheng and Prusoff (1973), and final data are presented as pK_i ± S.E.M. of n  4 experiments. The concentration-dependent reversal of medetomidine actions on luciferase activity changes in CHO cells by yohimbine and dimers (n = 3 and n = 24) were determined as molar EC₅₀ and negative log molar effective concentration-50 (pEC₅₀) values ± S.E.M. of n = 3 to 4 experiments. Differences of means of binding affinities and functional responses for individual ligands on the AR subtypes were done by analysis of variance and Tukey's post hoc analysis test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Radioligand Binding Assay. Radioligand binding analyses of yohimbine and its dimeric analogs were performed in CHO cells expressing human _2A-, _2B-, and _2C-AR subtypes. The binding affinities of these dimers are presented in Table 1. Yohimbine was about 4- and 15-fold more selective for the _2C-AR compared with the _2A- and _2B-AR, respectively (Table 1). MK-912 was also tested as a standard because it is known to have a higher affinity for the _2C-AR subtype (Uhlen et al., 1997). In our study MK-912 was about 14- and 3-fold more selective for the _2C-AR compared with the _2A- and _2B-AR, respectively (Table 1). Displacement curves demonstrating the _2C- versus _2A- and _2B-AR selectivity for yohimbine are presented in Fig. 2. The yohimbine analogs, with the exception of the dimer (n = 7) on the _2A-AR, displayed significantly lower binding affinities than yohimbine on the three ₂-AR subtypes. However, the dimeric analogs show much higher affinities for the _2A- and _2C-AR subtypes versus the _2B-subtype at all spacer atom lengths (n = 2 to n = 24), and there are relatively small changes in the K_i values for these analogs on the _2C-AR subtype. Comparison of the pK_i ± S.E.M. values for these bivalent yohimbine analogs on the three ₂-AR subtypes (Table 1) shows that there is at least a 6.5-fold higher binding affinity of the yohimbine analogs (n = 2, 3, 4, 18, and 24) for the _2C-subtype. Binding displacement curves for the dimers n = 3 and n = 24 are shown in Fig. 3. Compared with yohimbine the binding displacements curves for the selective dimers n = 3 and n = 24 are shifted to the right, especially for _2A- and _2B-AR subtypes. The yohimbine dimers (n = 2, n = 3, and n = 24) possess affinities for the _2C-AR, which are 23- and 891-fold, 32- and 355-fold, and 82- and 776-fold greater than their binding to the _2A- and _2B-AR subtypes, respectively (Table 1).

View this table: [in this window] [in a new window]   TABLE 1 Binding affinities (pKi values) of yohimbine and its bivalent analogs on human 2A-, 2B-, and 2C-adrenoceptors expressed in Chinese hamster ovary calls pK1 values are the negative log of the K1 value, determined using the Cheng and Prusoff equation: K1 = [IC50]/(1 + [RL]/KRL), where RL is radioligand used. See Materials and Methods. [3H]Rauwolscine was the radioligand used. Values are expressed as the mean ± S.E.M. of n = 4 to 6 experiments. Statistical analysis was carried out by one-way analysis of variance followed by Tukey's test.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Binding displacement curves of yohimbine, for 2A-, 2B-, and 2C-adrenoceptors expressed in Chinese hamster ovary cells. Data are expressed as mean ± S.E.M. (n = 4-6 experiments). 2A, ; 2B, ; and 2C, .

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Binding displacement curves of yohimbine dimers n = 3 (left) and n = 24 (right) for 2A-, 2B-, and 2C-adrenoceptors expressed in Chinese hamster ovary cells. [3H]Rauwolscine was used as the radioligand. Data are expressed as mean ± S.E.M. (n = 4-6 experiments). 2A, ; 2B, ; and 2C, .

View this table: [in this window] [in a new window]   TABLE 2 Binding affinities of yohimbine and selected bivalent analogs (n = 3 and n = 24) on human 1-adrenoceptors expressed in human embryonic kidney cells pK1 = log K1 (K1 value was determined according to the Cheng-Prusoff equation) and the data are the mean ± S.E.M. of n = 4 to 8. [3H]Rauwolscine and [3H]prazosin were used as the radloligands in the equilibrium competition radioligand binding assays for 2- and 1-ARs expressed in CHO and HEK cells; and nonspecific binding was measured in the presence of 10 µM of yohimbine and 10 µM phentolamine, respectively. Statistical analyses was carried out by one-way analysis of variance followed by Tukey's test.

Cyclic AMP Response Element-Luciferase Reporter Gene Assay. Luciferase reporter gene assay experiments were conducted by adding a fixed concentration of medetomidine (10 nM) to block the cAMP-induced increases by 5 µM forskolin. A concentration of medetomidine was chosen that produced at least a 50% inhibition of the forskolin response in the _2C-AR subtype expressing CHO cell lines. Controls of yohimbine and medetomidine and basal responses (solvent control) were included for comparisons. The effects of yohimbine and two bivalent analogs (n = 3 and n = 24) are provided graphically in Fig. 4. As noted in the graphs, forskolin produced an increase in cAMP levels, as assessed by luciferase changes using a 6 CRE-LUC reporter gene bioassay. Concentrations of yohimbine and bivalent analogs (n = 3 and n = 24) were varied from 0.01 to 10 µM. The conditions of incubation were as follows: the addition of yohimbine antagonist for 20 min; addition of medetomidine for 20 min; and the addition and incubation with forskolin for 4 h. The EC₅₀ values of the yohimbine analogs for the reversal of medetomidine action against forskolin-induced cAMP are presented in Table 3. The results show that yohimbine and the two bivalent analogs were able to reverse the effect of medetomidine on the _2C-AR at lower concentrations than those required for reversal on the _2A-AR. Yohimbine, and its bivalent analogs, n = 3 and n = 24, were 10-, 42-, and 29-fold more potent at reversing the action of medetomidine on the _2C- versus the _2A-AR subtype, respectively. Although the fold differences between the _2A- and _2C-AR for these analogs are not equal in functional versus binding assays, the data of these functional studies show the same trend as binding affinities observed with these drugs on these two ₂-AR subtypes. Subtle differences between binding and functional results may arise and could be due to differences in receptor density for the two subtypes used for the study. Nevertheless the data indicate that these bivalent yohimbine dimers are highly selective _2C-AR antagonists.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Reversal of medetomidine inhibition of forskolin-induced cAMP elevations, as assessed by luciferase activity changes by yohimbine analogs n = 3 and n = 24 on human 2A- (A and C) versus 2C (B and D)-adrenoceptors. Concentrations used: forskolin, 5 µM; medetomidine, 10 nM; and yohimbine dimers, 0.01 to 10 µM. Data are expressed as mean ± S.E.M. of n = 3 to 4 experiments. F, forskolin (5 µM); M, medetomidine (10 nM); n = 3, yohimbine dimer, n = 3 (0.01, 0.1, and 1 µM) and n = 24, yohimbine dimer (0.01, 0.1, 1, and 10 µM).

View this table: [in this window] [in a new window]   TABLE 3 Concentration-dependent effects of yohimbine and yohimbine analogs for the reversal of medetomidine effects on forskolin-Induced cAMP elevations on the 2A- and 2C-ARs expressed in Chinese hamster ovary cells EC50 and pEC50 were calculated using GraphPad Prism and expressed as the mean ± S.E.M. of n = 3 to 4 experiments. Values were determined as the effective concentration (EC50) and negative log (pEC50) of each yohimbine analog that reversed the effect of medetomidine on the maximum cAMP response of forskolin.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

G protein-coupled receptors (GPCRs) known to exist as dimers are emerging at an accelerated rate (Bouvier, 2001). GCPRs are major pharmacological targets and therefore the existence of dimers could have important implications for the development and screening of new drugs. The bivalent ligand approach has been successfully used in developing highly potent and selective ligands in a diverse set of receptor systems, such as the opioid (Portoghese et al., 1982; Portoghese, 2001) and serotonergic (LeBoulluec et al., 1995) receptors, two members of the seven transmembrane G protein-coupled receptor superfamily, as well as the growth factor receptor system (e.g., the thrombopoietin receptor) (Cwirla et al., 1997). Several studies have also strengthened the possibility that _2A- and _2C-ARs can exist as dimers under normal physiological conditions (Maggio et al., 1993, 1996), which is supported by direct observations of _2A-AR clustering using immunomicroscopic techniques (Uhlen et al., 1995).

There is a high level of sequence and structure homology, especially in the seven transmembrane regions of the ₂-AR subtypes; however, the extracellular loops are very diverse in terms of the amino acid compositions among the three subtypes. To date three ₂-AR subtypes, _2A, _2B, and _2C, have been classified using functional and molecular techniques. However, available antagonist ligands such as idazoxan, rauwolscine, or prazosin have at best only marginal-subtype selectivity among the three ₂-AR subtypes (MacDonald et al., 1997). Therefore, in order to take advantage of receptor dimerization and clustering and the extracellular loop diversity among the ₂-AR subtypes, the bivalent ligand approach was used in our efforts to identify _2C-AR subtype-selective antagonists. For this purpose yohimbine was chosen as the antagonist pharmacophore and was found to be a potent and selective ₂-AR antagonist. Furthermore, our results show that yohimbine is about 4- and 15-fold more selective for the _2C-AR compared with the _2A- and _2B-AR, respectively. These findings agree with the previous work of Bylund et al. (1992) using these receptor subtypes from various sources. Also, the point of spacer attachment as an amide ester on the C-16 carboxyl of yohimbine was selected such that the ligand receptor binding interactions would not be impaired (Zheng et al., 2000). This site of modification was also chosen, based on the report that the C-16 yohimbine-agarose conjugate is an excellent affinity chromatography matrix for large-scale micropurification of multiple ₂-AR subtypes.

In the present study, a series of bivalent yohimbine analogs with atom linker lengths of n = 2 to n = 24 were prepared and evaluated on the human -AR subtypes. Each of the yohimbine dimers generally possessed a lower binding affinity than that of yohimbine for the _2A- and the _2C-AR subtypes. It is important to note that the dimeric yohimbine analogs show much higher affinities for the _2A- and _2C-AR subtypes versus the _2B-AR subtype at all linker lengths, and there was also a high degree of selectivity of yohimbine analogs for the _2C-AR subtype at spacers n = 2, 3, 4, 18, and 24. In particular, yohimbine dimers n = 3 and n = 24 bound with an affinity for the _2C-AR subtype that was 32- and 355-fold and 82- and 776-fold greater than their binding to the _2A- and _2B-ARs, respectively. Furthermore, the selectivity of binding affinities and functional antagonist potencies of these two yohimbine dimers for the _2C-AR subtype were at least 1000-fold greater than those on the ₁-AR subtypes; and they were >29-fold selective in reversing the action of medetomidine on cAMP accumulations on the _2C- versus the _2A-AR subtype, respectively. In these functional studies, yohimbine was more potent than the two yohimbine analogs, but was less selective (10-fold) as an antagonist of the _2C-AR (Table 3), and this correlates with the observed differences in the binding affinity selectivity for yohimbine versus these yohimbine analogs (Table 1).

Results of our studies using bivalent yohimbine pharmacophores have yielded highly selective and potent _2C-AR antagonists. To our knowledge, the two bivalent analogs (n = 3 and n = 24) represent the most selective antagonists identified so far for the human _2C-AR versus the other ₂-AR subtypes. There are other _2C-AR-selective ligands such as MK-912 (Uhlen et al., 1997). The yohimbine dimers n = 3 and n = 24 are 2.3- and 5.9-fold more selective for the _2C- versus the _2A-AR and 118- and 259-fold more selective for the _2C- versus the _2B-AR compared with MK-912 (Table 1).

It is unclear as to the exact mechanism underlying the _2C-AR selectivity observed for these yohimbine dimers. It is possible that one pharmacophore binds to the ligand receptor site, with the second pharmacophore is binding with an adjacent site of the ligand binding pocket (shorter linkers, n = 3), or with either one of the extracellular loops within the same receptor, or a ligand binding pocket on a second receptor. The latter possibility has been found in studies of bivalent pharmacophores on other GPCR systems (Portoghese, 2001). Interactions with extracellular loops or with adjacent receptors (molecular clustering of receptors) are most likely to occur with the longer linkers (n = 18 and n = 24) used in our study. However, we are unable to differentiate between these three possibilities with these bivalent yohimbine analogs at the present time.

Interestingly, varying the length of the spacer atoms linking the yohimbine molecules together had only modest effects on _2C-AR affinities, whereas the binding potencies were found to vary considerably in comparison with the _2A- and _2B-AR subtypes. These findings suggest that further changes in atom spacer length, and possibly of the linker composition, may lead to the development of more potent _2C-AR subtype-selective ligands.

Taken collectively, these results indicate the following: 1) all of the yohimbine dimers evaluated have higher affinities at the _2A- and _2C-AR subtypes compared with the _2B-AR subtype; 2) five of the dimeric compounds, i.e., n = 2, n = 3, n = 4, n = 18, and n = 24 have a high affinity and selectivity for the human _2C-AR subtype; 3) yohimbine dimers n = 3 and n = 24 represent potent and selective _2C-AR antagonists; and 4) the bivalent approach using yohimbine as a pharmacophore has been useful for developing _2C-AR-selective antagonists.