During translational studies to develop 4-phenyl-2-dimethylaminotetralin (PAT) compounds for neuropsychiatric disorders, the (2R,4S)-trans-(+)- and (2S,4R)-trans-(−)-enantiomers of the analog 6-hydroxy-7-chloro-PAT (6-OH-7-Cl-PAT) demonstrated unusual pharmacology at serotonin (5-HT) 5-HT2 G protein–coupled receptors (GPCRs). The enantiomers had similar affinities (Ki) at human (h) 5-HT2A receptors (∼70 nM). In an in vivo mouse model of 5-HT2A receptor activation [(±)-(2,5)-dimethoxy-4-iodoamphetamine (DOI)–elicited head twitch], however, (−)-6-OH-7-Cl-PAT was about 5-fold more potent than the (+)-enantiomer at attenuating the DOI-elicited response. It was discovered that (+)-6-OH-7-Cl-PAT (only) had ∼40-fold-lower affinity at mouse (m) compared with h5-HT2A receptors. Molecular modeling and computational ligand docking studies indicated that the 6-OH moiety of (+)- but not (−)-6-OH-7-Cl-PAT could form a hydrogen bond with serine residue 5.46 of the h5-HT2A receptor. The m5-HT2A as well as m5-HT2B, h5-HT2B, m5-HT2C, and h5-HT2C receptors have alanine at position 5.46, obviating this interaction; (+)-6-OH-7-Cl-PAT also showed ∼50-fold lower affinity than (−)-6-OH-7-Cl-PAT at m5-HT2C and h5-HT2C receptors. Mutagenesis studies confirmed that 5-HT2A S5.46 is critical for (+)- but not (−)-6-OH-7-Cl-PAT binding, as well as function. The (+)-6-OH-7-Cl-PAT enantiomer showed partial agonist effects at h5-HT2A wild-type (WT) and m5-HT2A A5.46S point-mutated receptors but did not activate m5-HT2A WT and h5-HT2A S5.46A point-mutated receptors, or h5-HT2B, h5-HT2C, and m5-HT2C receptors; (−)-6-OH-7-Cl-PAT did not activate any of the 5-HT2 receptors. Experiments also included the (2R,4S)-trans-(+)- and (2S,4R)-trans-(−)-enantiomers of 6-methoxy-7-chloro-PAT to validate hydrogen bonding interactions proposed for the corresponding 6-OH analogs. Results indicate that PAT ligand three-dimensional structure impacts target receptor binding and translational outcomes, supporting the hypothesis that GPCR ligand structure governs orthosteric binding pocket molecular determinants and resulting pharmacology.
Serotonin (5-HT) 5-HT2A, 5-HT2B, and 5-HT2C G protein–coupled receptors (GPCRs) are pharmacotherapeutic targets of interest due in part to their involvement in several psychiatric disorders, as inferred from human genetic studies (Dracheva et al., 2008; Sun et al., 2008; Nichols, 2009; Thanseem et al., 2012). Drug discovery programs targeting 5-HT2 receptors are focused on treating psychosis (5-HT2A antagonists), sleep disorders (5-HT2A antagonists), cluster headaches (5-HT2 agonists), anxiety (5-HT2B/2C antagonists), and obesity (5-HT2C agonists) (Sewell et al., 2006; Abbas and Roth, 2008; Smith et al., 2010). In addition, compounds with combined 5-HT2C agonist plus 5-HT2A/2B antagonist activity, similar to certain trans-4-phenyl-2-dimethylaminotetralins (PATs), may be useful for treating substance abuse and other compulsive behavioral disorders (Booth et al., 2009; Canal et al., 2013; Cunningham et al., 2013; Higgins et al., 2013; Morgan et al., 2013).
The (±)-(2,5)-dimethoxy-4-iodoamphetamine (DOI)-elicited head-twitch response (HTR) in mice is an animal model that is dependent on 5-HT2A receptor activation, and the HTR is produced reliably by most hallucinogenic 5-HT2 agonists (Glennon, 1992; Canal and Morgan, 2012; Halberstadt and Geyer, 2013). Thus, use of the DOI-elicited HTR model is expedient for translating drug discovery targeting 5-HT2 receptors. Although the C57Bl/6 mouse 5-HT2A (m5-HT2A) receptor shares ∼99% sequence identity with the human 5-HT2A (h5-HT2A) ortholog in the transmembrane domains (TMDs) where orthosteric ligands bind, there is significant divergence regarding binding affinities (Ki or KD) of 5-HT2A-targeting ligands at the two receptors (Dougherty and Aloyo, 2011). Similar observations are reported for the rat versus human 5-HT2A receptors (Nelson et al., 1993; Wainscott et al., 1996).
The only two residues within the TMDs that differ between the m5-HT2A and h5-HT2A receptors are at positions (Ballesteros and Weinstein, 1995) 1.42 (alanine in human and tyrosine in mouse) and 5.46 (serine in human and alanine in mouse). Position 1.42 is far from the orthosteric ligand binding pocket (Córdova-Sintjago et al., 2012a), and the functional and/or structural role of amino acids at position 1.42 is unknown. Amino acid residues in TMD5, however, have received attention due to their importance in ligand binding and function of GPCRs (Almaula et al., 1996a; van Rhee and Jacobson, 1996; Wang et al., 2013). Pointedly, residue 5.46 of aminergic neurotransmitter GPCRs (serotonin, dopamine, norepinephrine, and histamine) is an important component for binding of certain ligands in the orthosteric binding pocket (van Rhee and Jacobson, 1996; Shapiro et al., 2000; Armbruster et al., 2007; Cummings et al., 2010; Istyastono et al., 2011).
Mutagenesis studies have revealed that in vitro pharmacological properties of some 5-HT2A ligands are impacted by the difference in residue 5.46 at rodent versus human 5-HT2A receptors (Kao et al., 1992; Johnson et al., 1994). Few studies, however, have examined species differences in receptor protein structure-function with regard to in vivo behavioral pharmacology (Miller et al., 2009; Suratman et al., 2011), with most reports just including a comparison of ligand-receptor in vitro pharmacology. Even among in vitro studies, there are no reports that delineate molecular determinants for binding of different enantiomers of the same compound with 5-HT2A residue 5.46.
Using a structure-based drug design approach, novel PATs targeting 5-HT2 receptors are being developed to treat compulsive behavioral and other neuropsychiatric disorders (Rowland et al., 2008; Booth et al., 2009; Canal et al., 2013; Morgan et al., 2013). Reported here are results from in silico m5-HT2A and h5-HT2A molecular modeling/ligand docking studies, in vitro molecular pharmacology studies, and in vivo neurobehavioral studies using the DOI-elicited HTR model for the trans-(−)- and trans-(+)-enantiomers of two PAT-type drug candidates, 6-OH-7-Cl-PAT and 6-OMe-7-Cl-PAT (Fig. 1). In preliminary studies, it was observed that for 6-OMe-7-Cl-PAT, both enantiomers have similar affinity across h5-HT2 subtypes, and there was no difference between the enantiomers regarding potency and efficacy for attenuating the DOI-elicited HTR in mice.
For 6-OH-7-Cl-PAT, the (−)-enantiomer had significantly higher affinity than the (+)-enantiomer at h5-HT2B and h5-HT2C receptors, but there was no difference in affinity for the two enantiomers at the h5-HT2A receptor. Nevertheless, the (−)-enantiomer showed significantly greater potency and efficacy for attenuating the 5-HT2A-dependent DOI-elicited HTR in mice. Docking studies of the 6-OH-7-Cl-PAT enantiomers at molecular models of 5-HT2A receptors (Córdova-Sintjago et al., 2012a) led to an examination of the putative role of 5-HT2A residue 5.46 in mediating (+)-6-OH-7-Cl-PAT ligand affinity and function. The results pinpoint a single amino acid difference at position 5.46 in the m5-HT2A and h5-HT2A receptor that impacts translation of novel PAT drug candidates for neuropsychiatric disorders.
Materials and Methods
Structures of (+)- and (−)-6-OH-7-Cl-PAT and 6-OMe-7-Cl-PATs used in the studies are shown in Fig. 1. The PATs were synthesized in our laboratory as racemic mixtures. Single enantiomers were resolved by chiral stationary-phase high-performance liquid chromatography and converted to hydrochloride salts as previously described (Bucholtz et al., 1999; Booth et al., 2009; Vincek and Booth, 2009). DOI and mianserin hydrochloride were purchased from Sigma-Aldrich (St. Louis, MO). 5-HT hydrochloride was purchased from Alfa Aesar (Ward Hill, MA). [3H]Ketanserin, [3H]mesulergine, and [3H]myoinositol were obtained from PerkinElmer Life and Analytical Sciences (Waltham, MA). All solid compounds were weighed on an XP26 microanalytical balance (Mettler-Toledo, Columbus, OH).
Human and Mouse Receptor Constructs and Point Mutations
A description of the wild-type (WT) m5-HT2 and h5-HT2 receptor cDNA constructs used was reported recently (Canal et al., 2013). Point mutations of residue 5.46 of both m5-HT2A and h5-HT2A receptors were made by polymerase chain reactions using the QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA) according to the manufacturer’s protocol. WT m5-HT2 receptor cDNAs, cloned in the pCMV6 vector, were purchased from OriGene (Rockville, MD). WT h5-HT2 receptor cDNAs, cloned in the pcDNA3.1+ vector, were purchased from Missouri S&T cDNA Resource Center (Rolla, MO; http://www.cdna.org/). Mutagenesis and sequencing primers were obtained from Life Technologies (Carlsbad, CA). Primers used to make the m5-HT2A A5.46S point mutation were 5′-cctcataggctcttttgtgtcatttttcatccccctaac-3′ (sense) and 5′-gttagggggatgaaaaatgacacaaaagagcctatgagg-3′ (antisense). Primers used to make the h5-HT2A S5.46A point mutation were 5′-cctgatcggctcttttgtggcatttttcattcccttaac-3′ (sense) and 5′-gttaagggaatgaaaaatgccacaaaagagccgatcagg-3′ (antisense). Polymerase chain reaction was performed as previously described, with optimization (Canal et al., 2011). Parental DNA in the reaction mixture was digested using DpnI at 37°C for 1 hour. Digestion mixture (2 μl) was transformed into 50 μl of XL1-Blue supercompetent cells by heat pulse at 42°C for 45 seconds. The transformed reaction was incubated in Super Optimal broth with Catabolite repression medium (Sigma-Aldrich), then plated onto LB agar plates containing 100 μg/ml kanamycin (mouse 5-HT2A WT and A5.46S mutant receptor) or 100 μg/ml ampicillin (human 5-HT2A WT and S5.46A mutant receptor). Single colonies were selected for sequencing, and point mutations were confirmed by Genewiz (South Plainfield, NJ).
Radioligand Binding Assays
Radioligand saturation isotherm and competitive displacement binding assays were performed in 96-well plates using 5-HT2A, 5-HT2B, or 5-HT2C receptors transiently expressed in HEK293 cells as previously described (Canal et al., 2011, 2013); 10 μg of htr2A, htr2B, or htr2C cDNA and 20 μl of Lipofectamine 2000 (Life Technologies) or TurboFect Transfection Reagent (Thermo Scientific, Waltham, MA) was used for transfections. For saturation binding assays, 0.02–16 nM [3H]ketanserin was used to obtain KD values at the WT and mutant 5-HT2A receptors. The BCA protein assay kit (Thermo Scientific) was used to quantify protein concentration in receptor-expressing cell membranes, for later determination of m5-HT2A and h5-HT2A WT receptor binding site densities, Bmax values, calculated using GraphPad software (GraphPad Software, La Jolla, CA; http://graphpad.com/quickcalcs/radcalcform/). For competition binding assays, 1–3 nM [3H]ketanserin or [3H]mesulergine (∼KD concentration) was used to obtain the Ki values for unlabeled ligands at each of the 5-HT2 receptor subtypes. Nonspecific binding was determined in the presence of 10 μM mianserin. Each assay was performed a minimum of three times, using at least sample triplicates for each ligand concentration.
Phosphoinositide Hydrolysis Assay
Receptor-mediated phosphoinositide hydrolysis by phospholipase C (PLC) was measured as previously described (Canal et al., 2013). In brief, HEK293 cells transiently expressing a 5-HT2 receptor were incubated with 16 ml inositol-free Dulbecco’s modified Eagle’s medium (DMEM) containing 1 μCi/ml [3H]myoinositol. Cells were apportioned into 48-well CellBind plates (Corning, Lowell, MA), and [3H]myoinositol incorporation continued for 24 hours in an incubator. Cells were then acclimated to serum-free media, treated for 30 minutes with experimental drugs that had been prediluted in media to contain a final concentration of 50 mM LiCl per well. Media were discarded, and 50 mM formic acid was added to lyse cells. Plates were stored at −80°C for at least 24 hours. After thawing, solution from each well was added to individual anion-exchange columns. Bound [3H]inositol phosphates were eluted with 800 mM ammonium formate, and eluate was added to scintillation fluid (ScintiVerse Cocktail; Thermo Fisher Scientific, Pittsburgh, PA). 3H-Induced scintillations were counted with a scintillation analyzer. Experiments were performed with duplicates for 5-HT and triplicates for all other drugs, and each independent experiment was performed a minimum of three times.
IP-One HTRF Assay
5-HT2A receptor–mediated inositol phosphate 1 production was measured using the IP-One HTRF assay kit (Cisbio, Bedford, MA). These assays were performed during a transition from using the time-consuming method of the phosphoinositide hydrolysis assay, as described above. We have found consistent results between the two assays in the rank order of potency and efficacy of all ligands tested relative to 5-HT at 5-HT2A receptors, and others also report congruency between the IP-One HTRF and Ca2+ flux methods at 5-HT2A receptors (Jensen et al., 2013). The assay was performed according to the Cisbio protocol yet with the following optimization. HEK293 cells transiently transfected with 5-HT2A WT or 5.46 mutant receptors were incubated in serum-free DMEM for 1 hour. Cells were then harvested and pelleted by centrifugation at 200g at 37°C for 5 minutes, followed by resuspension in serum-free DMEM. Cell number was counted using a hemacytometer (Hausser Scientific, Horsham, PA), and 500 cells/μl DMEM were used. Five microliters of cell suspension and 5 μl of test compounds in stimulation buffer (included in kit and containing 50 mM LiCl) were added to each well of a 384-well plate (Greiner Bio-one, Monroe, NC). The plate was incubated at 37°C for 2 hours. After incubation, the reaction was terminated by adding 5 μl of the donor and acceptor fluorescent conjugates in lysis buffer. The plate was incubated at room temperature for 1 hour to reach equilibrium. Inositol phosphate 1 levels were assessed by the ratio of fluorescence emission at 665 nM and 620 nM (fluorescence resonance energy transfer, HTRF) using a Synergy H1 plate reader with an HTRF filter cube (BioTek, Winooski, VT). Experiments were performed with duplicates for 5-HT and triplicates for all other drugs, and each independent experiment was performed a minimum of three times.
Male C57Bl/6J mice, purchased at 60 days old from The Jackson Laboratory (Bar Harbor, ME), were used in all experiments. Mice were housed in standard cages with ad libitum access to food and water and were acclimated in the vivarium for at least 4 days before testing. Animal procedures were approved by the Institutional Animal Care and Use Committee and are in accordance with the principles in the National Institutes of Health Guide for Care and Use of Laboratory Animals (Publication No. 85-23, revised 1985).
All compounds were dissolved in sterile Milli-Q water (EMD Millipore, Billerica, MA), which served as the vehicle control. All compounds were administered subcutaneously at a volume of 10 ml/kg. A single dose of DOI (1 mg/kg) was used for all experiments, as it is known to reliably elicit a robust, consistent, and submaximal number of HTRs in C57Bl/6J mice (Canal et al., 2010; Canal and Morgan, 2012). Two or three doses were used to test the efficacy of novel PAT ligands for attenuating the DOI-elicited HTR. Doses for the compounds were as follows: (−)- and (+)-6-OMe-7-Cl-PAT, 3 and 5.6 mg/kg; (−)- and (+)-6-OH-7-Cl-PAT, 1.5, 3, and 5.6 mg/kg. Solutions of all compounds were made fresh on the day of behavioral testing. Drug or vehicle was injected 10 minutes prior to injection of DOI. Mice were placed in an activity chamber (43 × 43 cm; Med Associates, Inc., St. Albans, VT) 10 minutes later, and HTRs were counted for the next 10 minutes by an observer blind to the experimental treatment.
Competition and saturation binding data were analyzed using nonlinear regression commands in GraphPad Prism 6.02 for Windows (GraphPad Software). Binding data were fit using the “one-site” models. Two-site curve-fitting did not result in an improved fit (data not shown). Approximate Ki values were determined by conversion of the IC50 data using the equation Ki = IC50/1 + L/KD, where L is the concentration of radioligand (Cheng and Prusoff, 1973). Calculations of unpaired t tests and one-way analyses of variance (ANOVAs) with Tukey’s multiple-comparison post hoc tests were performed to compare calculated Ki and EC50 values of individual compounds for receptors. Ligand efficacies (Tables 1 and 2) are presented as the mean (± S.E.M.) of the percentages of maximal 5-HT response from all assays; 5-HT served as a positive control, full agonist in all functional assays.
Separate ANOVAs were performed to analyze effects of experimental compounds on the DOI-elicited HTR. Dunnett’s post-test compared the number of HTRs elicited by DOI after pretreatment with each experimental compound to the number of HTRs elicited by DOI and vehicle administration. Two-way ANOVA and Student-Newman-Keuls comparisons were used to identify dose-dependent effects and differences in efficacy across test compounds.
Molecular and Computational Modeling
A β2-adrenergic receptor (β2AR)-based (Protein Data Bank entry: 2RH1) homology model of the h5-HT2A receptor was used in docking and molecular dynamics (MD) studies. Detailed homology modeling procedures are described elsewhere (Córdova-Sintjago et al., 2012b). In brief, the 5-HT2A native sequence (WT) was aligned to the β2AR sequence using ClustalW multiple sequence alignment (Thompson et al., 1994; Larkin et al., 2007). Other structures (e.g., the inverse agonist carazolol, T4-lysozyme, and cholesterol molecules) present in the β2AR/T4-lysozyme chimera crystal structure were deleted. To obtain the 5-HT2A receptor model, point mutations were performed as needed, followed by the appropriate sequence additions and deletions to match the WT amino acid sequence. TMDs and loops were built using the Biopolymer module of Sybyl-X 2.0 (Tripos International, St. Louis, MO), and the crude models of the unbound receptor were minimized using the Powell method with Tripos force field (Clark et al., 1989) and AMBER charges (Cornell et al., 1995) followed by equilibration in a 1-palmitoyl-2-oleyl-sn-glycero phosphatidyl choline bilayer (Heller et al., 1993). The system was relaxed using the Tripos force field to a gradient 0.05 kcal/Å.mol prior to MD simulations in the 1-palmitoyl-2-oleyl-sn-glycero phosphatidyl choline membrane. MD simulation conditions were time run for 5 microseconds, time step 1 femtosecond, with snapshots collected every 5 femtoseconds. Other parameters were the canonical ensemble (number of particles, volume, and temperature were constants), 300K temperature, Boltzmann initial velocities, and nonbonded cutoff set at 8 Å. Constraints for alpha carbons in the TMDs were employed. Subsequently, the constraints were removed for a 1000-picosecond MD simulation run. The final unbound WT 5-HT2A homology model was obtained from the median structure after clustering analysis of the frames from the last 10 picoseconds of the MD simulation, and optimized using the Tripos force field to a convergence of 0.05 kcal/Å.mol. The stability and validation of the model were tested as described previously (Córdova-Sintjago et al., 2012b).
While this article was in preparation, the crystal structure of h5-HT2B GPCR was reported (Wang et al., 2013). To assess the validity of our h5-HT2A model using β2AR as a template, we similarly built an h5-HT2B homology model and compared it to the crystal structure of the chimeric h5-HT2B–BRIL in complex with ergotamine, Protein Data Bank code 4IB4, by aligning the alpha carbon atoms. The root mean square deviation of 3.34 Å indicates that the h5-HT2B homology model is similar to the crystal structure and validates the strategy used to generate the h5-HT2A model (Fig. 6). A similar conclusion was reached when comparing the β2AR-based human histamine H1 homology model to its crystal structure (Cordova-Sintjago et al., 2012c).
The ligands (+)- and (−)-6-OH-7-Cl-PAT were built as monocations (protonated amines) using HyperChem 8.0 (Hybercube, Gainesville, FL) and optimized using PM3 model Hamiltonian to a gradient of 0.01 kcal/Å mol. The binding site was defined by assigning residue D3.32 as a definitive binding site interaction point with the PAT ligand’s protonated amine moiety and including residues within a 7-Å radius (Kristiansen et al., 2000; Canal et al., 2011). The initial ligand-receptor complex configuration was used for flexible ligand docking with Flexidock in Sybyl-X 2.0. Structure preparation was carried out prior to docking studies by assigning AMBER charges (Cornell et al., 1995) for the protein and Gesteiger-Marsili charges (Gasteiger and Marsili, 1980) for the ligand. Rotatable bonds in the ligand and the side chains of residues defining the receptor putative active site were screened for optimal positioning of the ligand and side chains in the conformational space; remaining residues were frozen during docking. Default FlexiDock parameters were set at 80,000-generation. The best docking solution, according to the highest FlexiDock score, was minimized using the Tripos force field to a gradient of 0.05 kcal/Å.mol, prior to MD simulation.
The selected high-score pose of the docked ligand was subjected to an MD simulation run for 500 picoseconds, with other parameters the same as above, to allow adjustment of the positions of side chains and helices. The final structure of the ligand docked at the receptor model was obtained from the average of the last 10 picoseconds of the MD simulation.
Affinity and Function of 6,7-Substituted PATs at Mouse and Human WT 5-HT2A Receptors.
Four PAT compounds were tested for affinity and function at each of the m5-HT2 and h5-HT2 receptor subtypes, excluding m5-HT2B (which was not procured). 5-HT was used as a positive control, full agonist for assessing the functional efficacies and potencies of PAT compounds. Representative binding curves for (−)- and (+)-6-OH-7-Cl-PAT at m5-HT2A and h5-HT2A are shown in Fig. 2. A summary of all Ki results (mean ± S.E.M.), excluding h5-HT2B data, is shown in Table 1. The mean ± S.E.M. Ki of (−)-6-OH-7-Cl-PAT at h5-HT2B receptors was 102.4 ± 2.6 nM, whereas (+)-6-OH-7-Cl-PAT had a Ki of 1018 ± 144.8 nM. The mean ± S.E.M. Ki of (−)-6-OMe-7-Cl-PAT at h5-HT2B receptors was 29.3 ± 4.8 nM, whereas (+)-6-OMe-7-Cl-PAT had a Ki of 72.7 ± 10.3 nM.
ANOVA of Ki values of (−)-6-OH-7-Cl-PAT at each receptor revealed a significant main effect of receptor subtype (F4,16 = 18.71; P < 0.0001). The only post-test comparisons that did not reach a significant difference were the affinities at m5-HT2A versus h5-HT2A receptors, m5-HT2C versus h5-HT2A receptors, and m5-HT2A versus m5-HT2C receptors; similar results were obtained for (+)-6-OH-7-Cl-PAT (F4,12 = 51.46; P < 0.0001). Only h5-HT2B versus h5-HT2C and m5-HT2A versus m5-HT2C Ki values did not reach a significant difference. Unpaired t tests comparing Ki values of the (−)- versus (+)-enantiomers at each of the 5-HT2 subtypes revealed that the h5-HT2A receptor was the only receptor where the Ki values were not significantly different. Regarding functional effects, the (+)-6-OH-7-Cl-PAT compound was a partial agonist at h5-HT2A receptors only (Table 1). Neither 6-OH-7-Cl-PAT enantiomer showed activity up to 10 μM at all other 5-HT2 receptors screened, including m5-HT2A.
Separate ANOVAs of Ki values of (−)-6-OMe-7-Cl-PAT and (+)-6-OMe-7-Cl-PAT at each receptor revealed a significant main effect of receptor subtype (F4,15 = 20.40; P < 0.0001 and F4,11 = 17.43; P < 0.0001, respectively). Post-test comparisons that did not reach a significant difference were the same across the two enantiomers: m5-HT2A versus h5-HT2A receptors, h5-HT2B versus h5-HT2C receptors, h5-HT2A versus m5-HT2C receptors, and m5-HT2A versus m5-HT2C. Unpaired t tests comparing Ki values of the (−)- versus (+)-enantiomers at each of the 5-HT2 subtypes revealed that the Ki values of the enantiomers were statistically different only at h5-HT2B receptors and m5-HT2C receptors (P < 0.05). Regarding functional effects, (−)-6-OMe-7-Cl-PAT was a partial agonist at both m5-HT2C and h5-HT2C receptors yet was inactive up to 10 μM at all other 5-HT2 receptors (Table 1). (+)-6-OMe-7-Cl-PAT was a nonpotent, partial agonist at m5-HT2C receptors and was inactive up to 10 μM at all other 5-HT2 receptors (Table 1).
Bmax values for [3H]ketanserin-labeled h5-HT2A and m5-HT2A receptors were obtained from HEK293 cells transfected separately from cells used for competition binding and functional assays, yet transfection procedures for all assays were the same. m5-HT2A and h5-HT2A receptor Bmax values varied across transfections; the mean ± S.E.M. values were 1311 ± 137.9 and 1046 ± 595.8 fmol/mg, respectively, which were not statistically different (P > 0.05).
Affinity and Function of (+)- and (−)-6-OH-7-Cl-PATs at Mouse and Human 5.46 Point-Mutated 5-HT2A Receptors.
The KD value of [3H]ketanserin at mutant m5-HT2A A5.46S and h5-HT2A S5.46A receptors was not statistically different. The mean ± S.E.M. KD values of [3H]ketanserin at m5-HT2A A5.46S and h5-HT2A S5.46A receptors were 1.2 ± 0.3 and 1.6 ± 0.4 nM, respectively, closely matching many reported literature values (Almaula et al., 1996a; Braden and Nichols, 2007) and not statistically different from 5-HT2A WT receptors (P > 0.05). The KD value of [3H]ketanserin at m5-HT2A WT and h5-HT2A WT receptors was 1.3 ± 0.2 and 1.6 ± 0.5, respectively. Figure 2 shows representative competition binding curves for (−)- and (+)-6-OH-7-Cl-PAT at each of the [3H]ketanserin-labeled mutant receptors—only these two enantiomers were screened at the mutant receptors, because only they showed unusual differences in pharmacology between m5-HT2A and h5-HT2A WT receptors. Ki values are shown in Table 2. The Ki value of (+)-6-OH-7-Cl-PAT at h5-HT2A S5.46A receptors increased 30-fold relative to 5-HT2A WT; this value, 2813 ± 266 nM, was not statistically different from m5-HT2A WT, 3005 ± 321 nM (P > 0.05). Similarly, the Ki value of (+)-6-OH-7-Cl-PAT at m5-HT2A A5.46S receptors decreased 45-fold relative to the WT receptor, and this value, 66 ± 10 nM, was not statistically different from h5-HT2A WT, 80 ± 4.4 nM (P > 0.05). Consistent with the results from WT receptors, the point mutations did not greatly affect the Ki values of (−)-6-OH-7-Cl-PAT; only the affinity at m5-HT2A A5.46S receptors, 35 ± 1.6 nM, was statistically different relative to WT, 61 ± 0.8 nM (P < 0.05).
Finally, in congruence with the similarities in 6-OH-7-Cl-PAT pharmacology between the h5-HT2A WT and m5-HT2A A5.46S receptors, (+)-6-OH-7-Cl-PAT was a partial agonist at m5-HT2A A5.46S receptors (Fig. 3; Table 2). Furthermore, the EC50 (nM) and Emax (% of maximum 5-HT response) values of this compound at h5-HT2A WT and m5-HT2A A5.46S receptors were not statistically different (P > 0.05). Finally, (−)-6-OH-7-Cl-PAT showed no activity up to 10 μM at both of the 5-HT2A point-mutated receptors.
Molecular Modeling and Ligand Docking.
Selected low-energy poses of (−)- and (+)-6-OH-7-Cl-PAT docked at the h5-HT2A receptor model, as viewed from the extracellular domain, are shown in Figs. 4 and 5. Only these two enantiomers were docked at the h5-HT2A receptor, because (−)- and (+)-6-OMe-7-Cl-PAT did not show unusual differences in pharmacology between mouse and h5-HT2A WT receptors. Alignment of h5-HT2A and m5-HT2A receptors using ClustalW 2.1 showed that only two residues in the TMD helices 1 and 5 were different, 1.42 and 5.46. Residue 1.42 is far from the orthosteric pocket; therefore, the analysis of the drug-receptor interactions after docking and MD does not include residue 1.42.
For both 6-OH-7-Cl-PAT enantiomers, the protonated (at physiologic pH) dimethylamine moiety is capable of ionic interaction with the D3.32 carboxylate (Kristiansen et al., 2000; Ballesteros et al., 2001; Canal et al., 2011). Overall, however, the two enantiomers docked differently. For example, the 6-OH and 7-Cl substituents of (−)-6-OH-7-Cl-PAT are close to TMD7 (Fig. 4, left panel), whereas for the (+)-enantiomer the 6-OH and 7-Cl substituents are close to TMD5 (Fig. 4, right panel). Further examination of the docking results(Fig. 5) indicated that the 6-OH group of the (+)-enantiomer is able to form a hydrogen bond with the side chain of S5.46, whereas this interaction is not possible for the (−)-enantiomer. Both the 6-OH and 7-Cl substituents in the (+)-enantiomer (right panel) can interact with the hydroxyl group of S5.46. Conversely, the 6-OH group of the (−)-enantiomer (left panel) is close to Y7.43 in TMD7, far from S5.46. In addition, such a hydrogen bonding interaction is precluded at the m5-HT2A receptor model (not shown) where residue 5.46 is alanine, and similarly is precluded at the mouse and human 5-HT2B and 5-HT2C receptors that also have an alanine at position 5.46. After MD simulations for each 6-OH-7-Cl-PAT enantiomer docked at the h5-HT2B receptor, important binding site residues change orientation; for example, W6.48 and F6.51 are close to the pendant phenyl of the (−)-enantiomer, which is not observed for the (+)-enantiomer (Fig. 6).
When vehicle (“Veh” in Fig. 7) preceded DOI (1.0 mg/kg) administration, there were 35–40 HTRs observed in a 10-minute period (Fig. 7). Pretreatment with (−)-6-OH-7-Cl-PAT resulted in a dose-dependent and complete attenuation of the DOI-elicited HTR, with the lowest dose tested (1.5 mg/kg), producing an ∼50% effect. In contrast, 1.5 mg/kg (+)-6-OH-7-Cl-PAT failed to alter the effects of DOI, while a 50% attenuation was achieved with the highest dose tested, 5.6 mg/kg. Comparison of the two dose-effect curves revealed a significant main effect of enantiomer (F1,28 = 91.3; P < 0.001) and a significant main effect of dose (F2,28 = 31.66; P < 0.001), with all doses being different from each other. Relative to vehicle administration, all doses of the (−)- and the two higher doses of the (+)-enantiomer resulted in statistically significant decreases in HTRs (F6,41 = 37.51; P < 0.001). Both enantiomers of 6-OMe-7-Cl-PAT produced similar effects, with significant decreases at both doses relative to DOI preceded by vehicle administration. Comparison of the two dose-response curves revealed a significant effect of dose (F1,16 = 8.9; P = 0.009) but no interaction or difference between the enantiomers.
While developing 5-HT2-targeting PAT compounds for neuropsychiatric disorders, it was discovered that 6-OH-7-Cl-PAT demonstrates unexpected 5-HT2 pharmacology. For other PATs, the enantiomer that demonstrates highest affinity at h5-HT2A reliably translates as the more potent and efficacious enantiomer in mouse neurobehavioral models of 5-HT2 receptor activation, including the DOI-elicited HTR (Canal et al., 2013; Morgan et al., 2013). In addition, PAT enantiomers generally retain their affinity (Ki) differences across 5-HT2 subtypes, e.g., if one enantiomer is more potent at 5-HT2A, it is predictably more potent at 5-HT2B and 5-HT2C, probably because of relatively high sequence homology (∼75%) in 5-HT2 family TMDs.
Unusual was the observation that both (+)- and (−)-6-OH-7-Cl-PAT had indistinguishable affinity (Ki) at h5-HT2A, i.e., about 70 nM. Their affinities, however, at h5-HT2B and h5-HT2C diverged substantially. Thus, at h5-HT2B and h5-HT2C, affinity of (+)-6-OH-7-Cl-PAT was about 1 μM, whereas affinity of (−)-6-OH-7-Cl-PAT was ~100 and 20 nM, respectively. Furthermore, (+)-6-OH-7-Cl-PAT was a partial agonist at h5-HT2A, whereas (−)-6-OH-7-Cl-PAT did not activate h5-HT2A; neither enantiomer activated h5-HT2B nor h5-HT2C.
Given that the DOI-elicited HTR in mice is principally a 5-HT2A receptor–mediated effect, the greater potency of (−)-6-OH-7-Cl-PAT over (+)-6-OH-7-Cl-PAT in this assay was unexpected in view of their similar affinities at h5-HT2A. These results initially suggested that other receptors, including 5-HT2C, could have played a modulating role in the behavior (Canal et al., 2010, 2013) and/or that the 5-HT2A partial agonist activity of (+)-6-OH-7-Cl-PAT negatively impacted its efficacy to attenuate the DOI-elicited HTR. Results from binding and functional assays using the m5-HT2A and m5-HT2C, however, were parsimonious with another hypothesis that accounted for the behavioral data of 6-OH-7-Cl-PAT, i.e., the (+), but not the (−)-, enantiomer has different pharmacology at m5-HT2A compared with h5-HT2A (see Table 1). Affinity (Ki) of (+)-6-OH-7-Cl-PAT was nearly 40 times lower at m5-HT2A compared with h5-HT2A, likely accounting for its lower potency versus (−)-6-OH-7-Cl-PAT for attenuating the DOI-elicited HTR. In contrast, potencies of (−)- and (+)-6-OMe-7-Cl-PAT for attenuating the DOI-elicited HTR were not different, probably caused by their indistinguishable affinities and lack of activity (e.g., neutral antagonism) at m5-HT2A receptors.
With the exception of unrestricted rotation of the C(4) phenyl group and limited boat/chair conformational flexibility of the cyclohexyl moiety, the PAT molecular scaffold is relatively rigid. Moreover, PATs are chiral and can serve as templates to define the three-dimensional arrangement of 5-HT2 receptor amino acids involved in ligand binding, leading to inferences about receptor structure and conformational changes required for activation. Thus, the PATs are especially well suited for ligand docking studies at 5-HT2 receptor molecular models (Córdova-Sintjago et al., 2012a,b).
The m5-HT2A and h5-HT2A vary at only one amino acid, residue 5.46, in their putative orthosteric binding pockets, where there is serine in the human receptor and alanine in the mouse receptor. Results from molecular docking studies indicated that the 6-OH substituent of (+)-6-OH-7-Cl-PAT could form a hydrogen bond with the side chain of S5.46 when docked at the h5-HT2A receptor, similar to the hydrogen bond interaction observed between the catechol para-hydroxyl group of norepinephrine and residue S5.46 of βAR (Strader et al., 1989). An equivalent hydrogen bonding interaction is not possible for (+)-6-OH-7-Cl-PAT at m5-HT2A (A5.46), likely accounting for its nearly 40-fold lower affinity and lack of partial agonist functional activity at the mouse compared with the human receptor, as well as its relatively low potency in the mouse DOI-elicited HTR assay. In contrast, (−)-6-OH-7-Cl-PAT docked differently than the (+)-enantiomer at m5-HT2A and h5-HT2A, having little interaction with TMD5, instead docking close to TMD7, with the 6-OH and 7-Cl substituents close to residue Y7.43. Thus, a single amino acid at position 5.46 likely accounts for the differences in (+)-6-OH-7-Cl-PAT pharmacology at m5-HT2A versus h5-HT2A and for the relatively low potency of (+)-6-OH-7-Cl-PAT in the mouse DOI-elicited HTR assay. Additionally, the low affinity of (+)-6-OH-7-Cl-PAT at 5-HT2B and 5-HT2C may be related to the presence of alanine at residue 5.46 in these receptors.
The role of residue 5.46 in the observed differences in PAT 5-HT2 pharmacology was confirmed by assessing affinity and function at the m5-HT2A A5.46S and h5-HT2A S5.46A point-mutated receptors. The mutations impacted the pharmacology of (+)-6-OH-7-Cl-PAT but did not alter the pharmacology of (−)-6-OH-7-Cl-PAT. For (+)-6-OH-7-Cl-PAT, the h5-HT2A S5.46A mutation essentially rendered a receptor that was indistinguishable from the m5-HT2A WT receptor; affinity (Ki) of (+)-6-OH-7-Cl-PAT was about 3 μM at the h5-HT2A S5.46A receptor, similar to its affinity at the m5-HT2A WT receptor. The converse was also true. Affinity (Ki) of (+)-6-OH-7-Cl-PAT was about 70 nM at the m5-HT2A A5.46S receptor, similar to its affinity at the h5-HT2A WT receptor. Thus, the interaction of the (+)-6-OH-7-Cl-PAT 6-OH moiety with S5.46 was crucial for its binding, whereas this amino acid hardly impacted binding of (−)-6-OH-7-Cl-PAT. This observation is unusual compared with most other 5-HT2A-targeting compounds reported (Almaula et al., 1996a; Braden and Nichols, 2007; Shan et al., 2012), including DOI. For example, of more than 20 compounds tested, only the ergot derivatives, e.g., ergonovine, showed major losses in affinity at the human S5.46A 5-HT2A receptor (Almaula et al., 1996a; Braden and Nichols, 2007).
In addition to binding, the interaction of the (+)-6-OH-7-Cl-PAT 6-OH moiety with S5.46 also impacted its function at the 5-HT2A receptor. The (+)-6-OH-7-Cl-PAT enantiomer had nil agonist activity at the m5-HT2A WT receptor (A5.46); however, it demonstrated partial agonist activity at the m5-HT2A A5.46S point-mutated receptor, similar to its partial agonist activity at the h5-HT2A receptor (S5.46). Similarly, activation of βARs by catecholamines involves hydrogen bonding between the catechol OH− group in the para position and the OH− in S5.46 (Strader et al., 1989; Warne and Tate, 2013). The similar Bmax values for m5-HT2A and h5-HT2A WT receptors suggested that the partial agonist activity of (+)-6-OH-7-Cl-PAT at h5-HT2A but not m5-HT2A WT receptors was not due to receptor reserve.
These results add to the growing literature that crucial points of interaction between a ligand and the 5-HT2A receptor are ligand-dependent, even within a restricted orthosteric binding pocket, and these unique interactions could contribute to different physiologic outcomes, e.g., hallucinogenic or not (Almaula et al., 1996b; Weinstein, 2005). For example, results here with (+)-6-OH-7-Cl-PAT are the opposite of results reported recently by another group testing the h5-HT2C-preferring ligand (R)-9-ethyl-1,3,4,10b-tetrahydro-7-trifluoromethylpyrazino[2,1-a]-isoindol-6(2H)-one (Miller et al., 2009). This ligand exhibits low affinity and efficacy at h5-HT2A (S5.46) yet is a potent and efficacious agonist at rat 5-HT2A (A5.46) and shows 5-HT2A agonist activity in vivo in rats. The in vitro WT receptor pharmacology of the ligand was reversed at the h5-HT2A S5.46A and rat 5-HT2A A5.46S point-mutated receptors. Thus, the interaction of this compound with S5.46 is detrimental to binding affinity and function, putatively due to a reduction in binding pocket size and negative steric interaction between ligand and receptor when serine exists instead of alanine at position 5.46 (Miller et al., 2009), similar to results from certain indole 5-HT2A agonists (Ebersole et al., 2003).
The complete set of data here converge on evidence that the difference in residue 5.46 between m5-HT2A and h5-HT2A accounts for the difference in behavioral potencies of the 6-OH-7-Cl-PAT enantiomers to attenuate the DOI-elicited HTR. The behavioral data, however, are not completely accounted for by ligand interaction with 5-HT2A, as (+)-6-OH-7-Cl-PAT, with essentially nil affinity at the m5-HT2A and m5-HT2C receptor subtypes, also attenuated the DOI-elicited HTR at higher but behaviorally relevant doses, comparable to the 6-OMe-7-Cl-PAT enantiomers. Thus, other mechanisms that modulate the effects of DOI may have come into play. For example, α-adrenergic receptors may modulate activity of the 5-HT2 circuitry or be direct targets of DOI (Darmani, 1993; Ray, 2010). Relevant here is that an early (1993) NovaScreen of racemic 6-OH-7-Cl-PAT reported appreciable (Ki ∼70 nM) at rat brain α1-adrenergic receptors yet negligible affinity for all other aminergic neurotransmitter receptors and two dozen other central nervous system receptor sites, with the known exceptions of histamine H1 (Bucholtz et al., 1999) and 5-HT2 receptors. Pharmacological studies are under way to assess the role of specific α1-adrenergic receptor subtypes in regulating the DOI-elicited HTR, and molecular modeling studies are in progress to assess interactions between 6-OH-7-Cl-PAT enantiomers and α1-adrenergic receptor TMD amino acids. Regarding known affinity of PAT ligands for H1 receptors, it is also acknowledged that H1 receptor actions do not seem to affect the DOI-elicited HTR (Canal and Morgan, 2012).
Finally, we are unaware of any compound, other than (+)-6-OH-7-Cl-PAT reported here, that possesses appreciable potency to bind and activate h5-HT2A together with negligible affinity and no ability to activate h5-HT2B or h5-HT2C. Given that (+)-6-OH-7-Cl-PAT penetrates mammalian brain (i.e., modulates the DOI-elicited HTR), it could be a useful tool to study the neurobehavioral effects of specific 5-HT2A receptor partial agonism in humans without the confound of 5-HT2B and/or 5-HT2C receptor binding and activation. Notably, partial agonist activity at 5-HT2A is possessed by classic hallucinogens, including LSD (Marek and Aghajanian, 1996; Egan et al., 1998; Ebersole et al., 2003), yet these drugs also bind and activate 5-HT2B and 5-HT2C (Porter et al., 1999), posing a challenge to distinguish specific receptor activation effects without using selective antagonists that may possess behavioral effects of their own.
Ki determinations of racemic 6-OH-7-Cl-PAT were provided by NovaScreen Biosciences, Calipur Discovery Alliances and Services.
Participated in research design: Canal, Cordova-Sintjago, Kim, Booth.
Conducted experiments: Canal, Cordova-Sintjago, Kim, Liu, Morgan.
Contributed new reagents or analytic tools: Kim.
Performed data analysis: Canal, Cordova-Sintjago, Liu, Morgan.
Wrote or contributed to the writing of the manuscript: Canal, Cordova-Sintjago, Liu, Morgan, Booth.
- Received August 7, 2013.
- Accepted September 30, 2013.
This work was supported by the National Institutes of Health National Institute of Mental Health [Grant R01-MH081193]; and National Institutes of Health National Institute on Drug Abuse [Grants R01-DA023928 and R01-DA030989] (to R.G.B. and D.M.).
- analysis of variance
- β2-adrenergic receptor
- Dulbecco’s modified Eagle's medium
- G protein–coupled receptor
- human 5-HT2A receptor
- head-twitch response
- homogeneous time-resolved fluorescence
- mouse 5-HT2A receptor
- molecular dynamics
- 4-phenyl-2-dimethylaminotetralin (4-phenyl-N,N-dimethyl-1,2,3,4-tetrahydronaphthalene-2-amine)
- phospholipase C
- transmembrane domain
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics