Skip to main content
NCBI home page
ACS Medicinal Chemistry Letters logoLink to ACS Medicinal Chemistry Letters
. 2022 Mar 18;13(4):648–657. doi: 10.1021/acsmedchemlett.1c00694

Discovery of Highly Potent Serotonin 5-HT2 Receptor Agonists Inspired by Heteroyohimbine Natural Products

Meghan J Orr †, Andrew B Cao ‡, Charles Tiancheng Wang †, Arsen Gaisin †, Adam Csakai †, Alec P Friswold †, Herbert Y Meltzer §,∥, John D McCorvy ‡,*, Karl A Scheidt †,§,*
PMCID: PMC9014500  PMID: 35450369

Abstract

graphic file with name ml1c00694_0015.jpg

The serotonin 5-HT2 receptors are important pharmaceutical targets involved in signaling pathways underlying various neurological, psychiatric, and cardiac functions and dysfunctions. As such, numerous ligands for the investigation of these receptors’ activity and downstream effects have been developed synthetically or discovered in nature. For example, the heteroyohimbine natural product alstonine exhibits antispychotic activity mediated by 5-HT2A/2C agonism. In this work, we identified a heteroyohimbine metabolite containing a serotonin pharmacophore and truncated the scaffold, leading to the discovery of potent agonist activity of substituted tetrahydro-β-carbolines across the 5-HT2 receptor family. Extensive SAR development resulted in compound 106 with EC50 values of 1.7, 0.58, and 0.50 nM at 5-HT2A, 5-HT2B, and 5-HT2C, respectively. Docking studies suggest a π-stacking interaction between the tetrahydro-β-carboline core and conserved residue Trp6.48 as the structural basis for this activity. This work lays a foundation for future investigation of these compounds in neurological and psychiatric disorders.

Keywords: Serotonin 5-HT2 receptors, tetrahydro-β-carbolines, small molecule agonists, heteroyohimbine metabolism

From ancient use of the hallucinogen psilocybin to contemporary use of the antihypertensive reserpine, the indole alkaloid class of natural products has long been a source of medicinally relevant bioactive molecules. The degree to which compounds of this family have been studied is highly variable, however, and so there remains untapped potential for the discovery of new and unexpected pharmacological activity among these molecules. One such example is the heteroyohimbine natural product alstonine (1, Figure 1). Identified as the active component in a traditional Igbo remedy for psychosis,1 alstonine was subsequently found to mediate its antipsychotic effects via the serotonin-2A (5-HT2A) and 2C (5-HT2C) receptors, rather than the dopamine receptor modulation ubiquitous among antipsychotics’ pharmacological mechanisms.2 However, further investigation of alstonine’s unique properties was hindered by the limited natural supply of the compound, until our total synthesis of alstonine furnished us with 1, its diastereomer serpentine (2), and their saturated counterparts tetrahydroalstonine (3) and ajmalicine (4) in sufficient quantities for further evaluation.3 Though structurally similar, these compounds possess diverse, if underexplored, bioactivity: serpentine has antimalarial4 and anticancer effects;5,6 ajmalicine is an alpha-1 adrenergic antagonist;7 and the pharmacology of tetrahydroalstonine is unknown.

Figure 1.

Figure 1

Open in a new tab

Heteroyohimbine compounds.

The exposure of ajmalicine to mouse liver microsomes over 1 h produced a metabolite with mass [M + 16], with its concentration surpassing that of the parent compound after 30 min (Figure 2). Spectroscopic characterization of the metabolite revealed a 3:1 mixture of 11-hydroxyajmalicine (5) and 10-hydroxyajmalicine (6), consistent with previous reports of indole alkaloid metabolism.8,9 Primarily interested in profiling the 5-HT2A/2C activity of these compounds, we hypothesized that 10-hydroxyajmalicine was likely the more active metabolite given the serotonin pharmacophore present in its scaffold.

Figure 2.

Figure 2

Open in a new tab

Metabolism of ajmalicine.

As the metabolites were obtained in quantities only sufficient for analytical studies, we elected to truncate the scaffold, retaining just the 6-hydroxytetrahydro-β-carboline core for further exploration. These simplified analogues are accessible in a single synthetic step (Scheme 1), rendering them more amenable to rapid derivatization than the full natural products requiring complex multistep syntheses. The tetrahydro-β-carboline is a privileged scaffold,10,11 with even its simplest natural derivatives displaying diverse bioactivity.12 Perhaps the most well-known synthetic tetrahydro-β-carboline is tadalafil, the phosphodiesterase-5 inhibitor marketed as Cialis.13 Across the 5-HT2 family, this chemotype is known to display potent yet tunable biased agonist14 or antagonist activity, as demonstrated in the development of serotonin-2B (5-HT2B) antagonist LY266097.15

Scheme 1. Pictet–Spengler Reaction of 6-Substituted Tryptamines with Aldehydes to Afford Tetrahydro-β-Carboline Analogues.

Scheme 1

Open in a new tab

Conditions (a): dimethylformamide, 70 °C, 16–96 h or glacial acetic acid, microwave 110 °C, 5–90 min.

The 5-HT2 family comprises three G protein-coupled receptors (GPCRs) that share a high degree of sequence homology, and selective targeting of any of the three remains a challenge.16 5-HT2A is notable for its mediation of the hallucinogenic effects of psychedelics as well as for its antagonist role in the mechanism of action of second-generation or atypical antipsychotics, such as clozapine.17 5-HT2C is also a target of interest in the development of new antipsychotics18,19 and has been shown to play a role in obesity20 and substance use disorders.21−24 Recently, there has been renewed interest in 5-HT2A/2C mixed agonists, such as psilocin,25 which have shown remarkable long-lasting effects for treatment-resistant depression and other disorders.26 5-HT2B, however, is best known as an undesirable off-target whose activation results in valvular heart disease,27 underscoring the importance of achieving selectivity when targeting 5-HT2A or 5-HT2C. Given the ease of generating several tunable tetrahydro-β-carboline analogues and combined with high-throughput screening technologies, investigation of this scaffold appeared to be a fruitful area for generation of new drug-like compounds for 5-HT2 receptor subtypes.

We began by constructing a series of 6-hydroxytetrahydro-β-carbolines with variation at the 1 position and profiling their activity at 5-HT2A and 5-HT2C in Gq-mediated calcium flux assays (Table 1). The 1-phenyl analogue 7 was inactive at 5-HT2C and only weakly active at 5-HT2A, but, remarkably, introduction of a fluorine at the 2’ position (8) gave full agonism of both receptors with low nanomolar potencies. Switching to 2′-trifluoromethyl (9) attenuated potency about 6-fold at both receptors, and chloro (10) and methoxy (11) substitution at this position again reduced the compounds to inactivity at 5-HT2C and weak partial agonism at 5-HT2A. However, 2′-hydroxyl analogue 12 displayed full agonism at 5-HT2C with EC50 = 85.9 nM and partial agonism at 5-HT2A with EC50 = 106 nM. Shifting the hydroxyl substituent to the 3′ or 4’ position (13, 14) slightly attenuated activity at both receptors. While a 3′-methyl substituent (15) maintained similar activity, a 4′-methyl substituent (16) had deleterious effects on activity at both receptors.

Table 1. Functional Activity of 6-Hydroxytetrahydro-β-carbolines at 5-HT2A and 5-HT2C in the Gq-Mediated Calcium Flux Assaya.

graphic file with name ml1c00694_0007.jpg

graphic file with name ml1c00694_0008.jpg

Open in a new tab
a

All data were generated in stable cell lines expressing 5-HT2A or 5-HT2C receptors and represent at least three independent experiments performed in triplicate. Emax is calculated as percent 5-HT response performed for every experiment. Selectivity was calculated by the difference in log(Emax/EC50) between 5-HT2C and 5-HT2A.

As 12 was thus far the only fairly potent analogue to exhibit any degree of selectivity between 5-HT2A and 5-HT2C, we next synthesized a number of di- and trisubstituted analogues incorporating the 2′-hydroxyl substituent. Introducing a methoxy substituent at the 4’ position (17) led to complete loss of selectivity but demonstrated full agonism with subnanomolar potency at both receptors. This activity was not maintained with a switch to 4′-bromo (18), nor was it matched by any 2′-hydroxyl-5′-halo analogues (19–21). However, 2′,4′,5′-trimethoxy analogue 22 was a full agonist with EC50 values of 3.8 and 7.6 nM at 5-HT2A and 5-HT2C, respectively.

We then evaluated the effects of heterocyclic substituents at the 1 position. Unlike phenyl compound 7, the bioisosteric thiophene compound 23 was active at both 5-HT2A and 5-HT2C. Pyrimidine analogue 25 was a weaker agonist at both, but N-methyl imidazole compound 24 proved to be fairly potent with EC50 = 79 nM for 5-HT2A and 97 nM for 5-HT2C. Aliphatic substitution at the 1 position gave rise to active compounds as well (26–29), though only 1-ethyl compound 29 had full agonist activity at both receptors.

We next turned our attention to the 6 position of the tetrahydro-β-carboline scaffold to examine the feasibility of tuning the selectivity of our compounds. As halogenation at the analogous position on tranylcypromine derivatives previously demonstrated substantial impact on 5-HT2A versus 5-HT2C selectivity,28 we synthesized and evaluated parallel series of 6-chloro- and 6-fluorotetrahydro-β-carbolines (Table 2). In contrast to 7, 1-phenyl compounds 30 and 31 displayed robust activity at both 5-HT2A and 5-HT2C. Introduction of a 2′-fluoro substituent (32, 33) dramatically attenuated potency, perhaps rendering these analogues too electron-deficient. Other 2′- halo compounds (34–37) fared generally better, as did the 2′-methoxy analogues in both series (38, 39). As observed in the 6-hydroxyl series, the most favorable substituent 2’ substituent proved to be a hydroxyl group, affording potent agonists 40 and 41 with EC50 = 51.7 and 26 nM at 5-HT2A and EC50 = 15.3 and 4.3 nM at 5-HT2C. Shifting the hydroxyl to the 3′ or 4’ position reduced potency at both receptors (42–45), and the 3′- and 4′-methyl analogues were similarly less active (46–49).

Table 2. Functional Activity of 1-Phenyl-6-halotetrahydro-β-carbolines at 5-HT2A and 5-HT2C in the Gq-Mediated Calcium Flux Assaya.

graphic file with name ml1c00694_0009.jpg

      5-HT2A
 5-HT2C
  
  R X EC50, nM (pEC50) Emax(% 5-HT) EC50 nM (pEC50) Emax(% 5-HT) selectivity, 2C over 2A
5-HT     0.35 (9.46 ± 0.03) 100 0.32 (9.49 ± 0.03) 100  
30 H Cl 248 (6.61 ± 0.03) 83.9 ± 0.9 68.9 (7.16 ± 0.06) 98.0 ± 2.0 4.18
31   F 88.5 (7.05 ± 0.03) 86.5 ± 0.9 22.4 (7.65 ± 0.05) 101.6 ± 1.5 4.82
32 2-F Cl 8004 (5.10 ± 0.12) 23.2 ± 0.8 1164 (5.93 ± 0.08) 13.9 ± 1.2 4.17
33   F 302 (6.52 ± 0.08) 48.9 ± 1.5 1006 (6.00 ± 0.06) 81.0 ± 2.3 0.53
34 2-Cl Cl 12.8 (7.89 ± 0.07) 95.3 ± 0.8 12.6 (7.90 ± 0.03) 109.1 ± 0.2 1.16
35   F 278 (6.56 ± 0.10) 28.1 ± 0.8 872 (6.06 ± 0.05) 88.1 ± 0.8 1.00
36 2-Br Cl 403 (6.39 ± 0.05) 91.4 ± 0.5 95.2 (7.02 ± 0.05) 98.4 ± 0.8 4.60
37   F 98 (7.01 ± 0.25) 10.5 ± 1.1 2099 (5.68 ± 0.02) 75.1 ± 1.0 0.33
38 2-OMe Cl 207 (6.68 ± 0.05) 41.6 ± 0.9 284 (6.55 ± 0.06) 82.9 ± 1.9 1.47
39   F 126 (6.90 ± 0.05) 43.1 ± 0.7 75.4 (7.12 ± 0.07) 94.3 ± 2.2 3.65
40 2-OH Cl 51.7 (7.29 ± 0.03) 92.7 ± 0.9 15.3 (7.81 ± 0.05) 103.4 ± 1.5 3.85
41   F 26 (7.58 ± 0.07) 96.2 ± 0.8 4.3 (8.36 ± 0.05) 100.3 ± 0.5 6.30
42 3-OH Cl 3304 (5.48 ± 0.07) 49.1 ± 0.7 688 (6.16 ± 0.03) 67.4 ± 0.8 6.54
43   F 208 (6.68 ± 0.04) 83.2 ± 1.3 130 (6.89 ± 0.04) 122.9 ± 2.1 2.40
44 4-OH Cl 93.8 (7.03 ± 0.08) 103.1 ± 1.3 105 (6.98 ± 0.07) 99.4 ± 0.9 0.86
45   F 58.2 (7.24 ± 0.05) 99.9 ± 2.1 29.8 (7.52 ± 0.04) 104.3 ± 2.4 2.03
46 3-Me Cl 481 (6.32 ± 0.08) 63.1 ± 0.5 256 (6.59 ± 0.09) 89.1 ± 0.5 2.67
47   F 160 (6.80 ± 0.07) 39.9 ± 1.2 306 (6.51 ± 0.04) 114.2 ± 2.1 1.55
48 4-Me Cl 303 (6.52 ± 0.08) 94.1 ± 0.9 62.5 (7.20 ± 0.06) 97.2 ± 0.5 4.98
49   F 208 (6.68 ± 0.07) 39.9 ± 1.2 2058 (5.69 ± 0.04) 85.4 ± 2.0 0.22
Open in a new tab
a

All data were generated in stable cell lines expressing 5-HT2A or 5-HT2C receptors and represent at least three independent experiments performed in triplicate. Emax is calculated as percent 5-HT response performed for every experiment. Selectivity was calculated by the difference in log(Emax/EC50) between 5-HT2C and 5-HT2A.

Fluoropyridine derivatives 50–52 were partial agonists at 5-HT2A and 5-HT2C, and they were 2–14-fold selective for the former over the latter (Table 3). Though pyrimidine compound 53 was a partial agonist at 5-HT2A and 5-HT2C, its 6-fluoro counterpart 54 was a full agonist at both receptors with EC50 = 54.9 and 10.1 nM at 5-HT2A and 5-HT2C, respectively. As for five-membered heterocycles, isoxazole analogues 55 and 56 were fairly potent at both receptors, while thiazole compound 57 was a micromolar partial agonist at 5-HT2C and inactive at 5-HT2A. Thiophene compounds 58–62 demonstrated markedly greater efficacy at 5-HT2C than 5-HT2A, a trend likewise observed in benzthiophene analogues 63 and 64. Other compounds with larger substituents at the 1 position generally had diminished activity (65–68), with the exception of naphthalene derivative 65, a full agonist at 5-HT2A and 5-HT2C with EC50 = 15.1 and 24.5 nM, respectively.

Table 3. Functional Activity of 6-Halotetrahydro-β-carbolines with Heteroaromatic and Bicyclic 1-Substituents at 5-HT2A and 5-HT2C in the Gq-Mediated Calcium Flux assaya.

graphic file with name ml1c00694_0010.jpg

graphic file with name ml1c00694_0011.jpg

Open in a new tab
a

All data were generated in stable cell lines expressing 5-HT2A or 5-HT2C receptors and represent at least three independent experiments performed in triplicate. Emax is calculated as percent 5-HT response performed for every experiment. Selectivity was calculated by the difference in log(Emax/EC50) between 5-HT2C and 5-HT2A.

As in the 6-hydroxyl series, aliphatic substituents at the 1 position gave rise to some active compounds as well (Table 4). Hydrocinnamyl compound 75 was particularly potent at both receptors with EC50 = 64.0 nM at 5-HT2A and 15.8 nM at 5-HT2C. Notably, deletion of a methylene unit gave compound 77, which was inactive at 5-HT2C and a weak partial agonist at 5-HT2A, as was its 6-fluoro counterpart 78. This finding is perhaps unsurprising given the structural similarity of these compounds to 5-HT2B antagonist LY266097.

Table 4. Functional Activity of 6-Halotetrahydro-β-carbolines with Aliphatic 1-Substituents at 5-HT2A and 5-HT2C in the Gq-Mediated Calcium Flux Assaya.

graphic file with name ml1c00694_0012.jpg

graphic file with name ml1c00694_0013.jpg

Open in a new tab
a

All data were generated in stable cell lines expressing 5-HT2A or 5-HT2C receptors and represent at least three independent experiments performed in triplicate. Emax is calculated as percent 5-HT response performed for every experiment. Selectivity was calculated by the difference in log(Emax/EC50) between 5-HT2C and 5-HT2A.

Docking studies performed with some of our most potent compounds and structures of both 5-HT2A29 and 5-HT2C30 indicate that in both receptors the tetrahydro-β-carboline core occupies the orthosteric site with its protonated amine forming a salt bridge with Asp3.32, the 6-substituent lying in the vicinity of transmembrane domain 5, and the indole N–H bond oriented upward toward the extracellular end of the receptor (representative pose with 5-HT2C shown in Figure 3; 5-HT2A not shown). This pose is in accordance with that observed in the crystal structure of LY266097 in complex with 5-HT2B.14 The favorable effects of incorporating a 2′-hydroxyl substituent, as demonstrated by analogue 41, are likely reflective of the additional hydrogen bond which can be formed with Asp3.32.

Figure 3.

Figure 3

Open in a new tab

Docked pose of 41 in 5-HT2C (6BQG). Yellow dashed lines indicate hydrogen bonding interactions, and pink dashed line indicates salt bridge formation.

We next expanded our library of compounds to include di- and trisubstituted phenyl substituents at the 1 position, holding the 2′-hydroxyl constant to retain that favorable hydrogen bonding interaction (Table 5). Though addition of halogens at the 3′ and 4’ positions did not improve activity (83–90), incorporating a second hydroxyl substituent at the 4’ position afforded compound 91, a full agonist at both receptors with EC50 = 1.5 and 1.9 nM at 5-HT2A and 5-HT2C, respectively. By contrast, 2′,4′-dimethoxy analogue 92 was around 100-fold less potent. However, 2′- hydroxy-4′-methoxy analogues 93 and 94 exhibited activity more akin to that of 91, possibly underscoring the impact of hydrogen bond donation from the 2’ position. As expected, dimethyl analogues 95 and 96 were far less active at both receptors.

Table 5. Functional Activity of 6-Halotetrahydro-β-carbolines at 5-HT2A, 5-HT2C, and 5-HT2B in the Gq-Mediated Calcium Flux Assaya.

graphic file with name ml1c00694_0014.jpg

      5-HT2A
 5-HT2C
   5-HT2B
  R X EC50 nM (pEC50) Emax(% 5-HT) EC50 nM (pEC50) Emax(% 5-HT) sel EC50 nM (pEC50) Emax(% 5-HT)
5-HT     0.35 (9.46 ± 0.03) 100 0.32 (9.49 ± 0.03) 100   0.36 (9.44 ± 0.05) 100
83 2-OH-3-Cl Cl 336 (6.47 ± 0.03) 70.1 ± 2.3 176 (6.75 ± 0.04) 94.0 ± 2.7 2.56 106 (6.97 ± 0.03) 95.9 ± 1.2
84   F 273 (6.56 ± 0.04) 65.3 ± 2.8 78 (7.11 ± 0.04) 99.2 ± 2.3 5.33 88 (7.06 ± 0.02) 96.2 ± 1.6
85 2-OH-3-Br Cl 74.5 (7.13 ± 0.04) 97.9 ± 0.8 68.6 (7.16 ± 0.08) 107.2 ± 2.1 1.18 nt  
86   F 116 (6.94 ± 0.08) 93.1 ± 0.7 66.5 (7.18 ± 0.02) 109.4 ± 3.1 2.05 nt  
87 2-OH-4-Cl Cl 52 (7.29 ± 0.02) 98.5 ± 0.6 47 (7.33 ± 0.02) 102.4 ± 0.8 1.14 25 (7.61 ± 0.03) 108.9 ± 1.2
88   F 229 (6.64 ± 0.03) 73.2 ± 0.8 209 (6.68 ± 0.03) 106.6 ± 1.5 1.60 128 (6.90 ± 0.02) 105.2 ± 0.9
89 2-OH-4-Br Cl 537 (6.27 ± 0.02) 94.5 ± 1.0 794 (6.10 ± 0.02) 106.8 ± 1.4 0.76 204 (6.69 ± 0.02) 106.2 ± 0.9
90   F 147 (6.83 ± 0.02) 98.0 ± 0.7 95 (7.02 ± 0.02) 110.8 ± 0.8 1.76 59 (7.23 ± 0.02) 112.1 ± 0.7
91 2,4-OH Cl 1.5 (8.82 ± 0.02) 91.2 ± 0.2 1.9 (8.72 ± 0.04) 101.7 ± 1.1 0.89 nt  
92 2,4-OMe F 167 (6.78 ± 0.04) 45.6 ± 0.7 159 (6.80 ± 0.03) 100.5 ± 1.1 2.31 98 (7.01 ± 0.01) 91.3 ± 0.5
93 2-OH-4-OMe Cl 8.3 (8.08 ± 0.02) 100.0 ± 0.8 8.9 (8.05 ± 0.02) 97.2 ± 0.8 0.90 3.1 (8.50 ± 0.02) 99.4 ± 0.7
94   F 3.2 (8.50 ± 0.02) 102.3 ± 0.8 1.02 (8.99 ± 0.03) 105.2 ± 1.0 3.23 1.3 (8.88 ± 0.07) 101.9 ± 2.1
95 2,4-Me Cl 562 (6.25 ± 0.03) 69.8 ± 1.1 407 (6.39 ± 0.05) 99.2 ± 2.2 1.96 977 (6.01 ± 0.03) 89.4 ± 1.5
96   F 363 (6.44 ± 0.05) 47.8 ± 1.2 159 (6.80 ± 0.04) 104.3 ± 1.3 5.01 537 (6.27 ± 0.02) 107.4 ± 1.3
97 2-OH-5-Cl Cl >10 000   3560 (5.45 ± 0.05) 58.9 ± 1.9   1995 (5.70 ± 0.02) 110.9 ± 1.1
98 2-OH-5-Br Cl 812 (6.09 ± 0.06) 79.7 ± 2.4 1318 (5.88 ± 0.04) 95.2 ± 2.3 0.74 457 (6.34 ± 0.02) 110.7 ± 0.9
99   F 1112 (5.95 ± 0.02) 80.8 ± 0.9 376 (6.42 ± 0.02) 101.3 ± 0.9 3.71 338 (6.47 ± 0.03) 111.4 ± 1.4
100 2-OH-5-Me Cl inactive   inactive     4265 (5.37 ± 0.04) 106.2 ± 2.8
101   F 1230 (5.91 ± 0.07) 36.8 ± 1.5 446 (6.35 ± 0.03) 99.2 ± 1.2 7.43 316 (6.50 ± 0.02) 115.2 ± 1.1
102 2,5-OH Cl inactive   inactive     inactive  
103   F 72 (7.14 ± 0.02) 95.2 ± 0.9 42 (7.38 ± 0.03) 119.6 ± 1.1 2.02 25 (7.60 ± 0.01) 102.3 ± 0.4
104 2,5-OMe Cl 1820 (5.74 ± 0.15) 14.4 ± 1.3 2691 (5.57 ± 0.04) 42.4 ± 1.2 1.94 416 (6.38 ± 0.05) 32.0 ± 0.7
105   F >10000   1604 (5.80 ± 0.04) 55.3 ± 1.2   588 (6.23 ± 0.04) 52.1 ± 1.1
106 2,4,5-OMe Cl 1.7 (8.77 ± 0.04) 98.4 ± 1.1 0.50 (9.33 ± 0.07) 97.5 ± 1.9 3.58 0.58 (9.24 ± 0.06) 96.7 ± 1.7
107   F 39 (7.40 ± 0.01) 97.1 ± 0.5 20 (7.69 ± 0.02) 110.6 ± 1.1 2.19 10 (7.96 ± 0.02) 105.9 ± 0.6
108 3,4-OMe Cl 81 (7.09 ± 0.03) 96.0 ± 1.0 123 (6.91 ± 0.02) 99.3 ± 1.0 0.68 34 (7.47 ± 0.02) 101.6 ± 0.6
109   F 254 (6.60 ± 0.02) 83.2 ± 0.7 152 (6.82 ± 0.03) 102.7 ± 1.3 2.05 102 (6.99 ± 0.03) 89.8 ± 1.0
Open in a new tab
a

All data were generated in stable cell lines expressing 5-HT2A, 5-HT2C, or 5-HT2B receptors and represent at least three independent experiments performed in triplicate. Emax is calculated as percent 5-HT response performed for every experiment. “sel” = selectivity for 5-HT2C over 5-HT2A and was calculated by the difference in log(Emax/EC50) between 5-HT2C and 5-HT2A. “nt” = not tested.

Compounds with 2,5-substitution patterns (97–105) generally displayed decreased activity as compared to 40 and 41. Though 2′,5′-dimethoxy analogue 104 was a weak partial agonist at both receptors, incorporation of a third methoxy substituent at the 4’ position gave compound 106 with EC50 = 1.7 and 0.50 nM at 5-HT2A and 5-HT2C, one of the most potent compounds across all our series. Removal of the 2′-methoxy substituent resulted in 80- and 200-fold loss of potency at 5-HT2A and 5-HT2C, respectively (108).

The subnanomolar activity of 106 was particularly intriguing. The SAR thus far indicated that a hydrogen bond donor was preferred at the 2’ position and that activity at both receptors was diminished with replacement by a 2′-methoxy group, observations supported by the compounds’ docked poses. Additionally, the much weaker activity profiles of 2′,4′-dimethoxy analogue 92 and 2′,5′-dimethoxy analogues 104 and 105 did not suggest that a combination of their characteristics would lead to dramatic improvements in potency. However, the 2,4,5-trisubstituted phenyl motif has a well-documented history of giving rise to powerful psychedelics in the phenethylamine class.31,32 Docking studies with 106 and the structure of 5-HT2A29 (Figure 4) gave insight into its activity profile: the 2,4,5-substituted phenyl ring occupies the orthosteric site within the binding pocket, and the tetrahydro-β-carboline unit extends downward toward transmembrane domain 6, engaging in π interactions with Phe3406.52. Finally, an edge-to-face π-stacking interaction with Trp3366.48, the canonical “toggle switch” residue, may account for this compound’s robust agonist activity.

Figure 4.

Figure 4

Open in a new tab

Docked pose of 106 in 5-HT2A (6WHA). Pink dashed line indicates salt bridge formation, and blue dashed lines indicate π interactions.

We next evaluated a subset of our compounds at 5-HT2B. These results for our final series of compounds are summarized in Table 5 (see Table S1 for 5-HT2B data for other compounds). In general, the activity of analogues at 5-HT2B was similar to their activity at 5-HT2A and 5-HT2C with no distinct pattern of selectivity for any one receptor over the others (Figure 5). Another selection of compounds was submitted to the Psychoactive Drug Screening Program (PDSP) for screening across a wide array of GPCRs and other targets (Table 6).33 Outside of the 5-HT family, these analogues demonstrated little to weak affinity for off-target receptors. (See Table S2 for PDSP data for additional compounds.)

Figure 5.

Figure 5

Open in a new tab

5-HT2 Gq-mediated calcium flux activity of highly potent tetrahydro-β-carbolines. All compounds were assayed in stable cell lines measuring Gq-mediated calcium flux at (A) 5-HT2A, (B) 5-HT2B, and (C) 5-HT2C receptors. Data represent mean and standard error of the mean from at least three independent experiments performed in triplicate. All data were normalized to percent 5-HT response.

Table 6. Binding Affinity Data for Selected Compoundsa.

  Ki (nM)
  93 94 106 107
5-HT1A 79.9 94.4 22.4 1367
5-HT1B 164 70.8 52.1 nd
5-HT1D 31.5 110 38.8 660
5-HT1E 1750 1409 666 nd
5-HT2A 225 250 113 2445
5-HT2B 4.53 4.42 2.93 55.7
5-HT2C 35.9 49.8 25.2 671
5-HT3 2790 nd 2785 nd
5-HT5A 177 229 69.5 2839
5-HT6 47.7 68.8 29.0 530
5-HT7 7.98 25.4 2.50 249
α2C nd nd nd 964
β1 nd nd 1475 961
D3 1246 nd nd nd
D4 609 362 nd nd
SERT nd 1841 nd nd
σ1 1032 nd 1493 760
σ2 nd nd nd 2328
Open in a new tab
a

“nd” indicates Ki not determined as binding inhibition was <50%. Additionally, all four compounds were found to have binding inhibition < 50% at α1A, α1B, α1D, α2A, α2B, BZP, β2, β3, D1, D2, D5, H1, H2, H3, H4, M1, M2, M3, M4, M5, DOR, KOR, MOR, DAT, NET, BZP, PBR, and GABA-A.

Inspired by a newly identified ajmalicine metabolite and the motivation to prepare simplified analogues of neuroactive heteroyohimbine natural products, we synthesized three series of tetrahydro-β-carbolines and evaluated their activity at 5-HT2A and 5-HT2C. The 2′,4′,5′-trimethoxy compound 106 was found to be a highly potent agonist at both receptors, with EC50 = 1.7 and 0.50 nM for 5-HT2A and 5-HT2C, respectively. Docking studies suggest that a π-stacking interaction between the tetrahydro-β-carboline core and conserved residue Trp6.48 is the structural basis for this activity. Our work reveals determinants of 5-HT2 subtype activation and highlights 6-hydroxyl and 6-halo tetrahydro-β-carbolines as highly potent agonists of the 5-HT2 receptor family. Considering the importance of these receptors in various neurological and psychiatric diseases, this investigation lays a foundation for further development of this compound scaffold toward these targets.

Acknowledgments

We thank the National Institutes of Health (R21NS120521) and Northwestern University (K.A.S) and a Medical College of Wisconsin Research Affairs Counsel Pilot Grant (J.D.M.) for financial support. M.J.O. was supported by the NIH under award number F30DA050445. A.P.F. acknowledges support from Northwestern University’s Chemistry of Life Processes Institute.

Glossary

Abbreviations

GPCR

G protein-coupled receptor

5-HT2C

serotonin-2C receptor

5-HT2A

serotonin-2A receptor

5-HT2B

serotonin-2B receptor

PDSP

Psychoactive Drug Screening Program

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.1c00694.

  • Additional 5-HT2B and PDSP assay results, experimental details for functional assays and metabolism study, modeling protocol, compound synthesis and characterization (PDF)

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.

Supplementary Material

ml1c00694_si_001.pdf (8.8MB, pdf)

References

  1. Costa-Campos L.; Lara D. R.; Nunes D. S.; Elisabetsky E. Antipsychotic-Like Profile of Alstonine. Pharmacol., Biochem. Behav. 1998, 60, 133–141. 10.1016/S0091-3057(97)00594-7. [DOI] [PubMed] [Google Scholar]
  2. Linck V. M.; Bessa M. M.; Herrmann A. P.; Iwu M. M.; Okunji C. O.; Elisabetsky E. 5-HT2A/C receptors mediate the antipsychotic-like effects of alstonine. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 2012, 36, 29–33. 10.1016/j.pnpbp.2011.08.022. [DOI] [PubMed] [Google Scholar]
  3. Younai A.; Zeng B. S.; Meltzer H. Y.; Scheidt K. A. Enantioselective syntheses of heteroyohimbine natural products: a unified approach through cooperative catalysis. Angew. Chem., Int. Ed. 2015, 54, 6900–6904. 10.1002/anie.201502011. [DOI] [PubMed] [Google Scholar]
  4. Wright C. W.; Phillipson J. D.; Awe S. O.; Kirby G. C.; Warhurst D. C.; Quetin-Leclercq J.; Angenot L. Antimalarial Activity of Cryptolepine and Some Other Anhydronium Bases. Phytother. Res. 1996, 10, 361–363. . [DOI] [Google Scholar]
  5. Beljanski M.; Beljanski M. S. Selective Inhibition of in vitro Synthesis of Cancer DNA by Alkaloids of β-Carboline Class. Pathobiology 2004, 50, 79–87. 10.1159/000163131. [DOI] [PubMed] [Google Scholar]
  6. Beljanski M.; Beljanski M. S. Three Alkaloids as Selective Destroyers of Cancer Cells in Mice. Oncology 2004, 43, 198–203. 10.1159/000226363. [DOI] [PubMed] [Google Scholar]
  7. Roquebert J.; Demichel P. Inhibition of the α1-and α2-adrenoceptor-mediated pressor response in pithed rats by raubasine, tetrahydroalstonine and akuammigine. Eur. J. Pharmacol. 1984, 106, 203–205. 10.1016/0014-2999(84)90698-8. [DOI] [PubMed] [Google Scholar]
  8. Beck O.; Faull K. F.; Repke D. B. Rapid hydroxylation of methtryptoline (1-methyltetrahydro-β-carboline) in rat: Identification of metabolites by chiral gas chromatography-mass spectrometry. Naunyn Schmiedebergs Arch. Pharmacol. 1986, 333, 307–312. 10.1007/BF00512946. [DOI] [PubMed] [Google Scholar]
  9. Tsuchiya H.; Todoriki H.; Hayashi T. Metabolic hydroxylation of 1-methyl-1,2,3,4-tetrahydro-β-carboline in humans. Pharmacol., Biochem. Behav. 1995, 52, 677–682. 10.1016/0091-3057(95)00048-2. [DOI] [PubMed] [Google Scholar]
  10. Gryglewski R. J.; Misztal S. H.; Splawiński J. A.; Panczenko B. The Influence of Some 1-Substituted 6-Methoxy-1,2,3,4-tetrahydro-β-carbolines on the Metabolism and Activity of 5-Hydroxytryptamine. J. Med. Chem. 1966, 9, 471–474. 10.1021/jm00322a005. [DOI] [Google Scholar]
  11. Beato A.; Gori A.; Boucherle B.; Peuchmaur M.; Haudecoeur R. β-Carboline as a Privileged Scaffold for Multitarget Strategies in Alzheimer’s Disease Therapy. J. Med. Chem. 2021, 64, 1392–1422. 10.1021/acs.jmedchem.0c01887. [DOI] [PubMed] [Google Scholar]
  12. Maity P.; Adhikari D.; Jana A. K. An overview on synthetic entries to tetrahydro-β-carbolines. Tetrahedron 2019, 75, 965–1028. 10.1016/j.tet.2019.01.004. [DOI] [Google Scholar]
  13. Curran M. P.; Keating G. M. Tadalafil. Drugs 2003, 63, 2203–2212. 10.2165/00003495-200363200-00004. [DOI] [PubMed] [Google Scholar]
  14. McCorvy J. D.; Wacker D.; Wang S.; Agegnehu B.; Liu J.; Lansu K.; Tribo A. R.; Olsen R. H. J.; Che T.; Jin J.; Roth B. L. Structural determinants of 5-HT2B receptor activation and biased agonism. Nat. Struct. Mol. Biol. 2018, 25, 787–796. 10.1038/s41594-018-0116-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Audia J. E.; Evrard D. A.; Murdoch G. R.; Droste J. J.; Nissen J. S.; Schenck K. W.; Fludzinski P.; Lucaites V. L.; Nelson D. L.; Cohen M. L. Potent, selective tetrahydro-β-carboline antagonists of the serotonin 2B (5HT2B) contractile receptor in the rat stomach fundus. J. Med. Chem. 1996, 39, 2773–2780. 10.1021/jm960062t. [DOI] [PubMed] [Google Scholar]
  16. Roth B. L.; Willins D. L.; Kristiansen K.; Kroeze W. K. 5-Hydroxytryptamine2-family receptors (5-hydroxytryptamine2A, 5-hydroxytryptamine2B, 5-hydroxytryptamine2C): where structure meets function. Pharmacol. Ther. 1998, 79, 231–57. 10.1016/S0163-7258(98)00019-9. [DOI] [PubMed] [Google Scholar]
  17. Nichols D. E. Psychedelics. Pharmacol. Rev. 2016, 68, 264–355. 10.1124/pr.115.011478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Shen J. H. Q.; Zhao Y.; Rosenzweig-Lipson S.; Popp D.; Williams J. B. W.; Giller E.; Detke M. J.; Kane J. M. A 6-week randomized, double-blind, placebo-controlled, comparator referenced trial of vabicaserin in acute schizophrenia. J. Psychiatr. Res. 2014, 53, 14–22. 10.1016/j.jpsychires.2014.02.012. [DOI] [PubMed] [Google Scholar]
  19. Rosenzweig-Lipson S.; Comery T. A.; Marquis K. L.; Gross J.; Dunlop J.. 5-HT2C Agonists as Therapeutics for the Treatment of Schizophrenia. In Novel Antischizophrenia Treatments; Geyer M. A., Gross G., Eds.; Springer: Berlin, Heidelberg, 2012; pp 147–165. [DOI] [PubMed] [Google Scholar]
  20. Smith B. M.; Thomsen W. J.; Grottick A. J. The potential use of selective 5-HT2C agonists in treating obesity. Expert Opin. Investig. Drugs 2006, 15, 257–266. 10.1517/13543784.15.3.257. [DOI] [PubMed] [Google Scholar]
  21. Levin E. D.; Johnson J. E.; Slade S.; Wells C.; Cauley M.; Petro A.; Rose J. E. Lorcaserin, a 5-HT2C agonist, decreases nicotine self-administration in female rats. J. Pharmacol. Exp. Ther. 2011, 338, 890–6. 10.1124/jpet.111.183525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Rezvani A. H.; Cauley M. C.; Levin E. D. Lorcaserin, a selective 5-HT(2C) receptor agonist, decreases alcohol intake in female alcohol preferring rats. Pharmacol., Biochem. Behav. 2014, 125, 8–14. 10.1016/j.pbb.2014.07.017. [DOI] [PubMed] [Google Scholar]
  23. Gerak L. R.; Collins G. T.; France C. P. Effects of Lorcaserin on Cocaine and Methamphetamine Self-Administration and Reinstatement of Responding Previously Maintained by Cocaine in Rhesus Monkeys. J. Pharmacol. Exp. Ther. 2016, 359, 383–391. 10.1124/jpet.116.236307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Neelakantan H.; Holliday E. D.; Fox R. G.; Stutz S. J.; Comer S. D.; Haney M.; Anastasio N. C.; Moeller F. G.; Cunningham K. A. Lorcaserin Suppresses Oxycodone Self-Administration and Relapse Vulnerability in Rats. ACS Chem. Neurosci. 2017, 8, 1065–1073. 10.1021/acschemneuro.6b00413. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Klein A. K.; Chatha M.; Laskowski L. J.; Anderson E. I.; Brandt S. D.; Chapman S. J.; McCorvy J. D.; Halberstadt A. L. Investigation of the Structure–Activity Relationships of Psilocybin Analogues. ACS Pharmacol. Transl. Sci. 2021, 4, 533–542. 10.1021/acsptsci.0c00176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Agin-Liebes G.; Davis A. K.. Psilocybin for the Treatment of Depression: A Promising New Pharmacotherapy Approach; Springer: Berlin, Heidelberg; pp 1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Rothman R. B.; Baumann M. H.; Savage J. E.; Rauser L.; McBride A.; Hufeisen S. J.; Roth B. L. Evidence for possible involvement of 5-HT2B receptors in the cardiac valvulopathy associated with fenfluramine and other serotonergic medications. Circulation 2000, 102, 2836–2841. 10.1161/01.CIR.102.23.2836. [DOI] [PubMed] [Google Scholar]
  28. Cho S. J.; Jensen N. H.; Kurome T.; Kadari S.; Manzano M. L.; Malberg J. E.; Caldarone B.; Roth B. L.; Kozikowski A. P. Selective 5-Hydroxytryptamine 2C Receptor Agonists Derived from the Lead Compound Tranylcypromine: Identification of Drugs with Antidepressant-Like Action. J. Med. Chem. 2009, 52, 1885–1902. 10.1021/jm801354e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kim K.; Che T.; Panova O.; DiBerto J. F.; Lyu J.; Krumm B. E.; Wacker D.; Robertson M. J.; Seven A. B.; Nichols D. E.; Shoichet B. K.; Skiniotis G.; Roth B. L. Structure of a Hallucinogen-Activated Gq-Coupled 5-HT2A Serotonin Receptor. Cell 2020, 182, 1574–1588.e19. 10.1016/j.cell.2020.08.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Peng Y.; McCorvy J. D.; Harpsøe K.; Lansu K.; Yuan S.; Popov P.; Qu L.; Pu M.; Che T.; Nikolajsen L. F.; Huang X-P.; Wu Y.; Shen L.; Bjørn-Yoshimoto W. E.; Ding K.; Wacker D.; Han G. W.; Cheng J.; Katritch V.; Jensen A. A.; Hanson M. A.; Zhao S.; Gloriam D. E.; Roth B. L.; Stevens R. C.; Liu Z-J. 5-HT2C receptor structures reveal the structural basis of GPCR polypharmacology. Cell 2018, 172, 719–730. 10.1016/j.cell.2018.01.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Shulgin A. T.; Sargent T.; Naranjo C. Structure–Activity Relationships of One-Ring Psychotomimetics. Nature 1969, 221, 537–541. 10.1038/221537a0. [DOI] [PubMed] [Google Scholar]
  32. Nichols D. E. Structure–activity relationships of serotonin 5-HT2A agonists. Wiley Interdisciplinary Reviews: Membrane Transport and Signaling 2012, 1, 559–579. 10.1002/wmts.42. [DOI] [Google Scholar]
  33. Besnard J.; Ruda G. F.; Setola V.; Abecassis K.; Rodriguiz R. M.; Huang X.-P.; Norval S.; Sassano M. F.; Shin A. I.; Webster L. A.; Simeons F. R. C.; Stojanovski L.; Prat A.; Seidah N. G.; Constam D. B.; Bickerton G. R.; Read K. D.; Wetsel W. C.; Gilbert I. H.; Roth B. L.; Hopkins A. L. Automated design of ligands to polypharmacological profiles. Nature 2012, 492, 215–220. 10.1038/nature11691. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ml1c00694_si_001.pdf (8.8MB, pdf)

Articles from ACS Medicinal Chemistry Letters are provided here courtesy of American Chemical Society

RESOURCES