Targeting the 5-HT_2A Receptor for Developing Psychedelic Drugs and Beyond

Classical psychedelics such as psilocybin (4-phosphoryloxy-N,N-dimethyltryptamine), lysergic acid diethylamide (LSD), N,N-dimethyltryptamine (DMT), and mescaline (3,4,5-trimethoxyphenethylamine) hold considerable therapeutic potential for the treatment of anxiety, depression, substance use disorders, and other neuropsychiatric disorders (Figure 1A).¹ Recently, results of several phase II clinical trial studies have shown that psilocybin possesses a great potential for the treatment of depression and anxiety.² Psilocybin is currently under study in distinct phases of clinical trials and some are highlighted here (Figure 1A), including a phase II study to assess its therapeutic potential for the treatment of opioid use disorder (NCT06067737), depression in Parkinson’s disease (NCT06455293) and a phase III clinical trial to test safety, tolerability and efficacy of a 25 mg dose psilocybin relative to placebo in patients with major depressive disorder (NCT06308653). LSD, DMT and mescaline have also been suggested to enhance management of treatment-resistant depression, alcohol and drug abuse, and other neuropsychiatric conditions.^3-5 LSD is currently under study in phase II clinical trial for the treatment of cluster headache (NCT03781128) (Figure 1A). Despite the great potential of classical psychedelics as neurotherapeutics, clinical use of hallucinogens dictates costly, resource-intensive medical management in patients, limiting equitable clinical access due to florid hallucinations, abuse liability and safety risks (distress, anxiety, cardiovascular).⁶ Moreover, in clinical trials of psilocybin in depression, a majority of potential participants were excluded due to various risk factors (e.g., history of psychosis).^2,7 To date, the precise mechanisms of action of psychedelics and their analogs responsible for prospective therapeutic effects remain unclear. Despite complex polypharmacological profiles, it is well established that classical psychedelics bind as agonists to the 5-HT_2A receptor (5-HT_2AR), a G protein-coupled receptor (GPCR),⁸ and their psychedelic effects in humans are mediated by 5-HT_2AR agonist actions. In addition, the 5-HT_2AR antagonist ketanserin is reported to alleviate the effects of LSD and psilocybin in humans, validating that 5-HT_2AR stimulation triggers psychedelic-like effects.⁹ At present, there are two hypotheses on the molecular signaling mechanisms engaged by psychedelics. One hypothesis is that β-arrestin biased signaling at 5-HT_2AR triggers hallucinations.⁸ Another states that a threshold of 5-HT_2AR-G_q stimulation (E_max > 70%) is required to produce psychedelic-like effects.⁹ From the neuronal point of view, psychedelics may accelerate neuronal growth, enhancing brain capacity for neuroplasticity to drive therapeutic corrections of disrupted neuronal functions.¹

Figure 1.

Representative 5-HT_2AR ligands. (A) Chemical structures of classical psychedelics. (B) Chemical structures of analogs of LSD. (C) Chemical structures of β-arrestin-biased 5-HT_2AR agonists. (D) Chemical structures of 5-HT_2AR positive allosteric modulators.

Figure 2A illustrates the structure of 5-HT_2AR within the orthosteric binding pocket (OBP) and extended binding pocket (EBP) composed of a hydrophobic residue L229 of extracellular loop 2 (ECL2) and transmembrane helices TM3, TM6, and TM7, which lies on the top of OBP (Figure 2A).¹⁰ Very recently disclosed structures of tryptamines, such as 5-hydroxytryptamines (5-HT), psilocin (4-hydroxy-N,N-dimethyltryptamine), and DMT bound to the 5-HT_2AR solved by cryo-EM reveal that these molecules engage in an ionic interaction with residue D155 and the primate specific residue S242 consistent with previous studies (Figure 2B-D).^1,11,12

Figure 2.

(A) Orthosteric binding pocket (OBP) and extended binding pocket (EBP) on 5-HT_2AR. (B–K) Binding poses of 5-HT (PDB/EMDB-ID = 9ARX/EMD-43797), psilocin (PDB/EMDB-ID = 9AS7/EMD-43807), DMT (PDB/EMDB-ID = 9AS1/EMD-43801), LSD (PDB/EMDB-ID = 9AS3/EMD-43803), lisuride (PDB = 8UWL), BOL (PDB/EMDB-ID = 9ARZ/EMD-43799), mescaline (PDB/EMDB-ID = 9AS5/EMD-43805), 25CN-NBOH (PDB = 6WHA), IHCH-7086 (PDB = 7WC9), and RS130-180 (PDB/EMDB-ID = 9AS9/EMD-43809) bound 5-HT_2AR. (L) Overlay pose of 5-HT (red orange), psilocin (orange), DMT (green), LSD (violet), lisuride (gray), BOL (plum), mescaline (magenta), 25CN-NBOH (yellow), IHCH-7086 (red), and RS130-180 (pink) with 5-HT_2AR.

A comparison of poses for 5-HT, psilocin, and DMT shows the modest shift within the indole ring. The most notable change is observed in the 5′-OH position of 5-HT and 4′-OH of psilocin. The hydroxy functionality of 5-HT resides in the proximity of residue N343 within a range favorable to hydrogen-bonding. However, the hydroxy group of psilocin is projected toward the amino tail.¹ Mutagenesis studies indicate that N343 mutation impacts the potency of 5-HT, suggesting that electrostatic and/or potential H₂O-facilitated interaction between 5-HT and the N343 mutation.¹ Notably, this interaction has not been observed for psilocin. A comparison of these structure (Figure 2A-C) with 5-HT complexed with 5-HT_1AR/5-HT_1DR, 5-carboxamidotryptamine (5-CT) bound to the 5-HT_5AR, and psilocin bound to the 5-HT_2CR shows that 5-HT, psilocin, and DMT occupy OBP instead of previously proposed EBP.¹⁰ LSD is reported to be the most potent psychedelics in the ergoline class but demonstrates hallucinogenic effects. In contrast, lisuride and 2-bromo-LSD (BOL), the analogs of LSD, are nonhallucinogenic (Figure 1B). BOL was initially identified in the 1950s as lacking hallucinogenic effects in humans^13,14 consistent with its inability to evoke the head twitch response (HTR) in rodents, a proxy for human hallucinations.¹⁵ BOL blocked the psychological effects of LSD in a small human study.¹⁶ A recent study validated that BOL lacks efficacy to evoke HTRs and exhibits an antidepressant-like profile in mice and evokes 5-HT_2AR-dependent neuronal plasticity.¹⁷

Recent studies have revealed that BOL is a potent G-protein biased 5-HT_2AR partial agonist (EC₅₀ = 0.81 nM, E_max = 60%), which does not induce HTRs in male C57BL/6J mice at doses (0.1–10 mg/kg) administered intraperitoneally and enhances spine growth and dendritic arborization in primary cultures of rat cortical pyramidal neurons.¹⁷ In a case study analysis, three single dose of BOL administered within 10 days reduced the frequency and intensity of cluster headaches in patients.¹⁴ The binding poses resolved by cryo-EM for LSD, lisuride and BOL demonstrate that LSD, lisuride, and BOL engage in hydrogen bonding interactions with residues D155 and S242 in OBP (Figure 2E-G).^1,18 The only difference between LSD and BOL in binding is the bromine atom of BOL, which forms van der Waals contact with residue I163 positioned deep into OBP and residue F340.¹ I163 is isoleucine associated with a canonical PIF (proline-isoleucine-phenylalanine) motif with evidence that the activated receptor triggers conformational changes in this motif. Therefore, interaction with I163 may serve to prevent 5-HT_2AR activation, leading to weak partial agonism.¹ Notably, F340 is conserved across all 5-HT₂Rs, which is crucial in ligand recognition. These structural insights of BOL may accelerate the development of novel ergoline-based 5-HT_2AR selective agonists devoid of hallucinogenic effects. Mescaline bound 5-HT_2AR structure solved by cryo-EM reveals that mescaline interacts with residues D155 and S242 similar to tryptamines and ergolines (Figure 2H).¹ In addition, 3′-methoxy of mescaline forms hydrophobic contact with the residue L229 of ECL2.¹ This interaction was earlier observed in cocrystal structures of LSD-bound to the 5-HT_2AR and 5-HT_2BR that creates a lid on the top of OBP.¹⁹ This ECL2 lid traps LSD in OBP, which increases the time of action of LSD. Residue L229 also forms a strong hydrophobic interaction with residue F234, which might be responsible for signal transmission via TM5 when the receptor is in the activated state.¹ Mutagenesis studies illustrate that mutation at L229A converts mescaline from a 5-HT_2AR agonist to an inverse agonist, which binds to the receptor to reduce 5-HT_2AR constitutive (basal) activity.¹ In addition, mutagenesis analyses illustrate that residue F234A plays a pivotal role in ligand-dependent activity evoked by 5-HT and mescaline. Thus, the stabilization of residue F234 is crucial to transmit signals through TM5 of the 5-HT_2AR.¹

The N-benzylated phenethylamine molecule 25CN-NBOH is a potent and selective full agonist of 5-HT_2AR with a binding affinity (K_i = 0.81 nM) and selectivity of 100-fold and 46-fold over 5-HT_2CR and 5-HT_2BR, respectively (Figure 1C).²⁰ The structure of 25CN-NBOH bound to the 5-HT_2AR and solved by cryo-EM revealed that residue D155 forms a salt bridge with the positively charged nitrogen, which is a conserved interaction across serotonin (5-HT) and other monoamine receptors (Figure 2I).¹¹ In addition, 25CN-NBOH makes hydrophobic contacts with residues V156 and V235 and also engages in aromatic interactions with residues F339 and F340 (Figure 2I).¹¹ The 2-hydroxyphenyl part of 25CN-NBOH seats deep into the pocket, forming hydrophobic interactions with the indole moiety of residue W336 (Figure 2I).¹¹ Residue S159 concurrently forms two hydrogen bonds, one with the positively charged nitrogen and another with hydroxy functionality appended to the ortho position of the phenyl ring of 25CN-NBOH (Figure 2I).¹¹ Mutagenesis studies unveiled that mutation of residue I181 of ICL2 eliminates Gq activation by 25CN-NBOH and facilitates β-arrestin recruitment to 5-HT_2AR, indicating that 25CN-NBOH is a biased agonist.¹¹ Optimization of 25CN-NBOH has led to the identification of a β-arrestin-biased 5-HT_2AR agonist 9 with high potency (~100 nM) and efficacy (E_max = ~ 90%) relative to LSD (Figure 1C).²¹ IHCH-7086 is the first reported β-arrestin-biased 5-HT_2AR partial agonist (E_max = 13%) with a high binding affinity (K_i = 12.6 nM) compared to reference 5-HT (Figure 1C).¹⁰ IHCH-7086 does not trigger HTRs and exhibits efficacy in preclinical antidepressant-like models.¹⁰ The cocrystal structure of IHCH-7086 complexed with 5-HT_2AR provides structural insights responsible for β-arrestin-biased signaling (Figure 2J).¹⁰ Like other psychedelics, a conserved interaction is observed between residue D155 and the positively charged nitrogen of IHCH-7086. Notably, no interaction has been observed between the residues S239 and S242 of TM5, which lends insight into why IHCH-7086 does not exhibit detectable G-protein activity at 5-HT_2AR. The tetracyclic moiety of IHCH-7086 forms major interactions with the EBP, which is hypothesized to mediate β-arrestin-bias signaling at 5-HT_2AR. RS130-180 is another recently reported β-arrestin-biased 5-HT_2AR agonist (Figure 1C). The structure of RS130-180 bound to the 5-HT_2AR solved by cryo-EM has shown that the compound interacts with residue W336 similar to 25CN-NBOH (Figure 2K).¹ The toggle switch adopts a completely downward projecting position that has not been observed in any previously reported 5-HT_2AR structures possibly due to the steric hindrance and the additional bulk of RS130-180. The displacement of residue W336 triggers an inward rotation of residue F332 of the PIF motif, which subsequently pushes the Y380 outward compared to the 5-HT structure and all other structures. This outward shift at the bottom of TM7 is found in the inactive state crystal,²² which generates a less favorable conformation for G-protein activation due to the presence of TM5 and TM6 in active state conformations, while TM7 in an inactive conformation.¹ This noncanonical state stabilized by RS130-180 may explain β-arrestin-biased 5-HT_2AR agonist signaling. The overlay poses for all the ligands described herein demonstrate that, except for IHCH-7086, all ligands occupy the OBP (Figure 2L). Interestingly, IHCH-7086 occupies both OBP and EBP (Figure 2L).

Allosteric modulation of GPCRs is an emerging approach in therapeutic drug discovery and development.²³ Targeting topographically distinct (allosteric) sites on GPCRs by small molecules can trigger conformational changes in GPCRs to modulate the binding affinity and/or efficacy of the orthosteric ligands.^24,25 Allosteric modulators that increase or decrease the functional response to the orthosteric ligand are referred to positive allosteric modulators (PAMs) or negative allosteric modulators (NAMs), respectively. One of the core challenges in targeting 5-HT_2AR is to achieve selectivity over all three 5-HT₂Rs (5-HT_2AR, 5-HT_2BR, 5-HT_2CR) due to overall 80% homology across the transmembrane domains. One potential strategy to selectively modulate 5-HT_2AR is to develop PAMs as allosteric sites are postulated to have higher sequence divergence across the 5-HT₂ family relative to classical orthosteric site where 5-HT binds.²⁴ Although allosteric modulation for 5-HT₂R is in its infancy, several 5-HT_2CR PAMs with in vitro and in vivo profiles different from agonists have been reported.^26-29 To date, the literature reveals only two 5-HT_2AR PAMs, CTW0404 and CTW0419 (Figure 1D),³⁰ offering new avenues in drug discovery targeting the receptor to achieve high selectivity devoid of potential adverse effects.

In conclusion, recently disclosed cocrystal structures of classical psychedelics and their analogs bound to the 5-HT_2AR provide key molecular interactions between ligand and receptor. These structural insights will be helpful facilitating structure-based drug design to discover novel psychedelic drugs with and/or without hallucinogenic effects. As it is well-known that classical psychedelics possess promiscuous polypharmacological profiles, developing 5-HT_2AR PAMs offers a promising strategy to achieve selectivity over highly homologous subtypes 5-HT_2BR and 5-HT_2CR. We anticipate that novel analogs of classical psychedelics as 5-HT_2AR partial agonists as well as β-arrestin-biased 5-HT_2AR partial agonists and 5-HT_2AR PAMs without psychedelic-like side effects will be developed toward novel neurotherapeutics in the near future.