5-HT2 receptor binding, functional activity and selectivity in N-benzyltryptamines

The last fifteen years have seen the emergence and overflow into the drug scene of “superpotent” N-benzylated phenethylamines belonging to the “NBOMe” series, accompanied by numerous research articles. Although N-benzyl substitution of 5-methoxytryptamine is known to increase its affinity and potency at 5-HT2 receptors associated with psychedelic activity, N-benzylated tryptamines have been studied much less than their phenethylamine analogs. To further our knowledge of the activity of N-benzyltryptamines, we have synthesized a family of tryptamine derivatives and, for comparison, a few 5-methoxytryptamine analogs with many different substitution patterns on the benzyl moiety, and subjected them to in vitro affinity and functional activity assays vs. the human 5-HT2 receptor subtypes. In the binding (radioligand displacement) studies some of these compounds exhibited only modest selectivity for either 5-HT2A or 5-HT2C receptors suggesting that a few of them, with affinities in the 10–100 nanomolar range for 5-HT2A receptors, might presumably be psychedelic. Unexpectedly, their functional (calcium mobilization) assays reflected very different trends. All of these compounds proved to be 5-HT2C receptor full agonists while most of them showed low efficacy at the 5-HT2A subtype. Furthermore, several showed moderate-to-strong preferences for activation of the 5-HT2C subtype at nanomolar concentrations. Thus, although some N-benzyltryptamines might be abuse-liable, others might represent new leads for the development of therapeutics for weight loss, erectile dysfunction, drug abuse, or schizophrenia.


Introduction
Serotonin or 5-hydroxytryptamine (5-HT) is a bioactive compound present in a large variety of plants and animals. In mammals it is an autacoid or mediator of important functions in the gut and in blood platelets where it is most abundant, but in spite of its relative scarcity in the central nervous system its most widely known functions are as a neurotransmitter. The discovery of increasingly selective 5-HT receptor inhibitors has shown that serotonin is not only involved in important peripheral functions, but is also implicated in cognition, memory, emotion, the regulation of mood, the sleep-wake cycle, food intake, sexual activity, and in migraine, obsessive-compulsive disorder, schizophrenia and hallucinations [1]. Serotonin activates a a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 The literature records only three N-benzylated tryptamine derivatives lacking the 5-methoxy substituent, comparing them with the corresponding 5-methoxytryptamines in a rat tail artery assay [12,14]. These compounds were partial agonists at the 5-HT 2A receptor, and were 2-4 times less potent than the 5-methoxy analogs, results that might be reasonably attributed to the absence of a hydrogen bond accepting methoxyl group on the indole moiety. It should be pointed out that the orthosteric binding site of 5-HT 2 receptors contains serine, threonine and tyrosine residues that form hydrogen bonds with agonist and antagonist ligands [22,23]. It could be further conjectured that N-benzylated compounds with less interactions in the 5-HT 2A receptor's orthosteric site might more clearly reveal effects due to unmapped interactions in the extended binding site described for the highly homologous 5-HT 2B and 5-HT 2C receptor crystal structures [22][23][24]. In fact Halberstadt [25], citing Braden et al. [13], points out that "compounds having low-to-moderate affinity tend to be the most sensitive to the (Nbenzyl) substitution". We therefore decided to use tryptamine instead of its 5-methoxy derivative as the starting point for the synthesis and evaluation of a more extensive series of N-benzyl compounds. Nevertheless, we also prepared and assayed a small number of 5-methoxytryptamine derivatives for comparison of our data with the literature, and to see if this substitution on the indole ring is responsible for any consistent changes in affinity or potency (Fig 1).
For specific substitution patterns, see Tables 1 and 2.

Chemistry
All the compounds (thirty-six N-benzylated tryptamine derivatives and seven N-benzylated 5-methoxytryptamine derivatives, four of the latter described previously [11,19], were synthesized by treating appropriate benzaldehydes with tryptamine or 5-methoxytryptamine (free base) in methanol and reducing the unisolated imine intermediate with sodium borohydride, as described for phenethylamines and 5-methoxytryptamines and shown in Fig 2 [16,19,26]: The secondary amines thus obtained in good yields, and not requiring extensive purification, were used as their water-soluble salts (usually hydrochlorides) in receptor binding and functional pharmacological studies.
It is worth noting that the 1 H NMR spectra (in DMSO-d 6 ) of some salts with an orthohydroxyl group on the N-benzyl moiety indicate the presence of an intramolecular hydrogen bond between the protonated amine and the hydroxyl group (Fig 3).   It should be noted that the affinities at 5-HT 2A and 5-HT 2C receptors were determined by displacement of antagonist radioligands, and therefore reflect binding to both active and inactive receptor conformations, while an agonist, presumably binding to an active conformation, was used for the 5-HT 2B affinity determinations. Therefore, the 5-HT 2B affinities are not strictly comparable to the others, and were not considered in the selectivity estimates.
The vast majority of the our products revealed submicromolar affinities at all three receptor subtypes, and a small number gave K i values in the 10 nanomolar range, although these values did not seem to follow any general trend. N-benzyltryptamine (1) bound to the 5-HT 2A and 5-HT 2C receptors with rather similar affinities to those reported for N,N-dimethyltryptamine (237 and 424 nM, respectively), but significantly better than N,N-diisopropyltryptamine (1.2 and 6.5 μM, respectively) [27]. N-Benzyl-5-methoxytryptamine (37) bound to both receptors 5-15 times more strongly than N-isopropyl-N-methyl-5-methoxytryptamine (165 nM and 1.3 μM, respectively) [27]. Comparison of our tryptamine derivatives with the corresponding 5-methoxylated analogs seemed to indicate some degree of parallelism. Thus, compounds 38, 40, and 41, described recently [19], which in our hands gave similar binding results to those published, are the 5-methoxytryptamine analogs of the 2-methoxy-, 2-bromo-and 4-bromobenzyltryptamines 3, 6, and 17, respectively. While in our case the human receptors were expressed in CHO (Chinese hamster ovary) cells, and in the paper by Nichols et al. [19] in human embryonic kidney (HEK) cells, our results for compounds 40 and 41 agree with theirs within a factor of 2. In the case of 38 our results do not differ by more than 2.7 times, which seems reasonable for data from different laboratories and determined in different biological substrates.
Almost half of our compounds bear an ortho-oxygen substituent on the benzyl moiety, which was introduced as a test of Heim's "structure-activity concept" regarding the increased activity of 2-methoxybenzyl derivatives [12]. Comparing the N-benzyl- (1), N-2-hydroxybenzyl- (2) and N-2-methoxybenzyl-(3) tryptamines, 5-HT 2A (and also 5-HT 2B ) affinity appeared to increase slightly in that order, although at most by a factor of 2.7. At the 5-HT 2C receptor the affinities fluctuated in the same range. These changes can hardly be considered significant. Similarly, in several pairs of compounds in which binding could be compared, introduction of an ortho hydroxyl or methoxyl group had at most a very minor effect. The only exceptions were the N-3-chlorobenzyl- (10) and N-3-bromobenzyl-(11) tryptamines, where substitution to afford the N-5-chloro-2-methoxybenzyl (28) and N-5-bromo-2-methoxybenzyl (11) and to a lesser extent N-5-bromo-2-hydroxybenzyl (26) analogs, was markedly damaging, at least for 5-HT 2A receptor binding. Thus, the presence of an ortho-oxygen substituent on the benzyl ring does not seem to be generally beneficial and, with the exceptions of the 2-hydroxy-3-methoxybenzyl and the 2-hydroxy-3-bromobenzyl derivatives 19 and 21, the N-disubstituted benzyltryptamines showed rather low affinities. On the other hand, in our limited 5-methoxytryptamine series, the 2-hydroxy-5-fluoro derivative 43 showed surprisingly strong 5-HT 2A receptor binding, and it also bound fairly strongly to the 5-HT 2C receptor.

Functional assays
The functional activities of our compounds at human 5-HT 2A and 5-HT 2C receptors were determined fluorometrically as calcium mobilization in CHO h5-HT 2A and HeLa h5-HT 2C cells ( Table 2).
Comparison of our results for the N-2-methoxybenzyl-(38), N-2-bromobenzyl-(40), and N-4-bromobenzyl-(41) 5-methoxytryptamine derivatives with those of Nichols et al. [19] indicates complete agreement in functional potencies in one case, and differences of somewhat more than an order of magnitude in the others, although practically full agonism at the 5-HT 2C receptor seems to be the norm. The use of different cell lines is probably responsible in part for these discrepancies, but it must be kept in mind that 5-HT 2 receptors display differential functional selectivity [28,29] and β-arrestin signaling bias [24] which are other factors to be considered.
Comparing tryptamine with 5-methoxytryptamine derivatives, the N-2-methoxybenzyl (NBOMe) derivative 3 was several hundred times less potent than its 5-methoxytryptamine analog 38 in the h5-HT 2A receptor functional assay, and the N-2-chlorobenzyl 5 was almost 40 times less potent than its counterpart 39. A comparison of 17 with 41 was not possible because the functional activity of the former was too low for quantification. At the h5-HT 2C receptor, 3 and 5 were only 2-9 times less potent than the corresponding 5-methoxytryptamine derivatives. It seems likely that while a 5-methoxyl group on the indole moiety can result in considerably higher potency at the h5-HT 2A subtype, it may have a less significant effect at the h5-HT 2C receptor.
Almost two decades ago we found that the electrophysiologically determined rank potencies of several hallucinogenic and non-hallucinogenic phenylisopropylamines were consistent with the 5-HT 2A and 5-HT 2C affinities obtained in radioligand displacement assays, and the 2C/2A affinity ratios paralleled the potency ratios reported in that work [30]. Although both the voltage-clamp assay used then and the fluorescence assay used in the present work depend on intracellular calcium release, other mechanisms are active in cells and the final result of receptor activation may not show such correlations. As far back as 1997 a modified ternary receptor model was invoked to explain discrepancies [31], and it has recently been shown that the binding profile alone can reasonably predict strong hallucinogenic effects in vivo [18,32].
As seen with respect to the affinities of these compounds for the three receptor subtypes, apparently similar molecules sometimes behave quite differently, defying interpretation. Many of these substances seem uninteresting as 5-HT 2 agonists because of their low potencies. However, a few of them exhibit low nanomolar EC 50 values, at either subtype, without following any obvious rule. Interest in N-benzylated phenethylamines and indoleamines has focused mainly on their possible psychedelic activities related to full or partial agonism at 5-HT 2A receptors [4]. In this regard, it is worth noting that the reported 5-HT 2A affinities of N-benzyl-5-methoxytryptamines were generally several times lower than those of the correspondingly N-benzylated 2-(2,5-dimethoxy-4-iodophenyl)ethylamines, with the sole exception (30x) of the "superpotent" N-(2-methoxybenzyl) analog (25I-NBOMe), but their in vitro functional potencies did not follow this trend [19]. The rodent head twitch response is commonly believed to distinguish 5-HT 2A agonists that are psychedelic in humans from others that are not [25]. The EC 50 values determined for those N-benzyl-5-methoxytryptamines which elicited the response (not all did) indicated potencies at least 30 times lower than that of 25I-NBOMe [19]. One could therefore expect that very few (e. g. 38 and 43, Fig 3) of these compounds might be human hallucinogens in the low milligram dose range, significantly higher than the commonly abused NBOMe phenethylamines. Considering the binding affinities instead of the calcium mobilization data, a fair number of our compounds exhibit pK i values greater than 7 and might show psychedelic properties at doses of a few tens of milligrams (Fig 4).
It may be noted that compound 38 was described by Nichols [19] (as 5a) and found to be the most potent 5-HT 2A receptor agonist in his series of N-benzyl-5-methoxytryptamines, with K i = 16.6 nM, EC 50 = 1.9 nM and E max = 81% (our values are 44.7 nM, 2.0 nM and 84%, respectively, in quite good agreement). It also gave an ED 50 = 3.15 mg/kg in the mouse head twitch response (HTR) assay which is commonly viewed as a predictor of human psychedelic activity [25]. For the sake of comparison, the potent psychedelic 25I-NBOMe (4-iodo-2,5-dimethoxy-N-(2-methoxybenzyl)phenethylamine) exhibits a much lower K i = 0.52 nM, almost identical IC 50 and relative efficacy data, but its HTR result, ED 50 = 0.078 mg/kg, suggests a 40-fold higher in vivo potency. Assuming that the HTR is a trustworthy model, we again see that the binding affinity seems to be a better predictor of psychedelic activity than functional potency, at least when determined as calcium mobilization.
A result that appeared with striking regularity was that almost all the compounds were partial agonists at the h5-HT 2A and full agonists at the h5-HT 2C receptor (or possibly "super agonists" eliciting a stronger response than serotonin). Moreover, a small number of these showed significant 5-HT 2C selectivity, sometimes coupled with EC 50 values below 100 nM.
Lorcaserin is the first drug in its class to be approved by the FDA [34] as an anorexic. Like our N-benzylated tryptamine derivatives, lorcaserin is a full 5-HT 2C agonist, and a partial agonist (about 70% efficacy) at 5-HT 2A receptors [35]. Because of its modest 5-HT 2C /5-HT 2A selectivity, lorcaserin is recommended exclusively for patients meeting specific criteria and has been placed in Schedule IV (prescription only) due to its presumed ability "to produce hallucinations, euphoria, and positive subjective responses at supratherapeutic doses" [34]. While 21 is about 3 times less potent than lorcaserin, it is much more selective vs. 5-HT 2A receptors, at which it also exhibits low efficacy, and therefore might not be expected to produce the abovementioned side effects at any reasonable dose level. Unfortunately, its appreciable affinity for the 5-HT 2B receptor militates against its acceptance, at least for prolonged use.
To summarize, our large series of N-benzylated tryptamines revealed a very broad range of affinities for the serotonin 5-HT 2 receptor subtypes spread over three orders of magnitude and generally showing little selectivity (tenfold at most, but usually much less) between the 2A and 2C subtypes. The ability of these compounds to elicit calcium mobilization was also quite variable and with no obvious correlation with their affinity. Unlike the binding studies, the functional assays exhibited significant selectivity with an unexpected bias favoring the 5-HT 2C receptor. Besides, our compounds were generally full 2C agonists and only partial agonists, sometimes with rather low efficacy, at the 2A subtype. This fact, coupled with the selective activation of 5-HT 2C receptors by several of these substances points them out as possible leads for the development of a novel series of compounds of interest in the areas of appetite reduction and the treatment of drug abuse, schizophrenia and sexual dysfunctions.

Chemistry
The tryptamine used was a generous gift from Prof. Michel Leboeuf, Faculté de Pharmacie de Châtenay-Malabry, Université de Paris XI, France. 5-Methoxytryptamine and most of the aldehydes were purchased from AK Scientific Inc. (Palo Alto, CA). Some aldehydes were prepared by standard aromatic bromination or O-methylation from the above precursors. Solvents were of synthesis grade, purchased from Merck S.A. (Santiago, Chile), and were used without additional purification. Melting points were determined on a Stuart SMP 10 apparatus, and NMR spectra were recorded on a Bruker Avance 400 instrument using the residual solvent signal as the internal standard.
The appropriate aromatic aldehyde (approx. 5 mmol) and the equivalent amount of tryptamine or 5-methoxytryptamine, each dissolved in MeOH (5 mL each) were placed in a 100 mL flask and allowed to react at r.t. for at least 120 min. Then solid NaBH 4 (a small molar excess) was added in small portions with stirring, and the resulting solution was stirred at r.t. for another 12 h The MeOH was removed, the solid taken up in dilute HCl, the solution made alkaline and extracted with CH 2 Cl 2 . The organic extract was washed with water and dried over Na 2 SO 4 , the solvent was removed and the N-benzylated amine was purified by ball-to-ball distillation or column chromatography if necessary (Scheme 1).

Functional studies
Functional activities were assessed by measuring Ca 2+ release in CHO-5-HT 2A or HeLa-5-HT 2C cells. The day before the assay, 2000 (5-HT 2A ) or 10000 (5-HT 2C ) cells/well were seeded in 384 well black plates (Greiner 781091). The cells were incubated with 25 μL of Fura-2 QBTTM Calcium Kit (Molecular Devices), in buffer supplemented with 5 mM probenecid (Invitrogen) for 1 h at 37˚C. Changes in fluorescence due to intracellular Ca 2+ mobilization (λ ex = 340 nm, λ ex = 380 nm; λ em = 540 nm) were measured using a calcium imaging plate reader system (FDSS7000EX, Hamamatsu) every second after the establishment of a baseline. The agonist Ca 2+ peak in response to agonist addition occurred from 10 to 20 s following stimulation.

Statistics
Data were adjusted to non-linear fitting using Prism V2.1 software (Graph Pad Inc., Chicago, USA). K i values were calculated using the Cheng-Prusoff equation.