Biased Signaling in Psychedelic Action

Abstract

Psychedelics show tremendous promise for treating psychiatric disorders and other illnesses, including pain and migraine. Despite decades of research, there is uncertainty which signaling mechanisms are necessary for rapid-acting and durable therapeutic effects of psychedelics. Although activation of the serotonin 5-HT_2A receptor is critical for their psychopharmacological effects, the precise signaling pathways and receptor conformations responsible are still under investigation. This review summarizes progress in studying 5-HT_2A signaling mechanisms and recent developments in the discovery of biased agonist tool compounds to disentangle therapeutic from adverse effects. Moreover, we review insights from structural studies regarding the design of psychedelic-derived compounds with tailored pharmacology, and briefly discuss other 5-HT receptors that may be important for shaping therapeutic effects. Finally, by drawing parallels between 5-HT_2A biased signaling and the opioid field, we conclude with lessons learned and discuss the need for more rigor and reproducibility to facilitate the development of novel psychedelic-based pharmacotherapies.

INTRODUCTION

Psychedelics include (in)famous compounds such as the synthetic lysergic acid diethylamide (LSD)(1), or natural substances such as psilocybin from so-called magic mushrooms (2) and mescaline found in the peyote cactus (3). These mind- manifesting substances typically produce altered states of consciousness (ASCs) and psychosis-like states (4). Herein we focus on classic psychedelics defined as agonists of the serotonin (5-hydroxytryptamine; 5-HT) 5-HT_2A receptor, in contrast to other drugs such as MDMA and ketamine, which have also been described as psychedelics, but possess distinct pharmacological mechanisms (5; 6). While psychedelics have been used in naturalistic and spiritual settings for millennia (7; 8), they have recently moved into the focus of clinical studies due to their therapeutic potential. Psychedelics predominantly show considerable promise in the treatment of psychiatric disorders such as depression and substance use disorders (9; 10). In fact, psychedelics are currently in clinical trials for a wide spectrum of illnesses (11), including substance use disorders and various forms of anxiety and depression (12), but also posttraumatic stress disorder (13), cluster headaches (14) and neuropathic pain (15). Despite these efforts to harness the profound physiological and psychological effects of psychedelics for clinical use, it should also be noted that studies have reported on a multitude of potential adverse effects of psychedelic use (16–18). Depending on dose, adverse effects can range from short-lived acute effects such as nausea, paranoia, or ataxia, to more persistent effects. Most commonly, people have reported worsening of feelings of anxiety or rare side effects such as the development of hallucinogen persisting perception disorder, in which perceptual changes such as visual hallucinations persist well past the psychedelic experience. Moreover, studies reaching back to the original discovery of LSD already reported on the negative effects that psychedelics can have on individuals suffering from schizophrenia, bipolar disorders, or other illnesses with psychotic conditions (19).

Overall, these studies highlight the current challenges in developing psychedelic-based therapeutic approaches. One practical challenge is the need for medical supervision during treatment, as studies have shown that guidance during the psychedelic experience can mitigate some of the side effects such as bouts of heightened anxiety. Moreover, psychedelics are often used in conjunction with psychotherapy, as the psychedelic experience provides a unique level of access to interact with and work on the human psyche. It should be noted that different psychedelics may not be one-size-fits-all for all illnesses, and even differences in patient cohort and disease etiology may pose distinct risks depending on the psychedelic used. Currently, psilocybin appears to be the most used psychedelic in clinical trials (11), which is in part due to its favorable duration of action between 4 and 6 h, compared to longer-lasting effects of LSD, and perhaps the lower reported intensity of the psychedelic experience compared to other psychedelics such as N,N-dimethyltryptamine (DMT) and 5-methoxy-N,N-dimethyltryptamine (5-MeO-DMT).

The promising clinical potential of psychedelics in treating diverse mental and neurological illnesses as well as their well-documented risks and side effects, argues for a deeper understanding of how psychedelics act at the circuit, cellular, and molecular levels. Such mechanistic insight would not only enable elucidation of disease etiology but also guide better selection of the appropriate psychedelic based on disease and patient context. Beyond the immediate facilitation of clinical applications, researchers have explored the development of safer psychedelic-based pharmacotherapies. At the forefront of this work are efforts to retain clinical efficacy, while reducing acute and persistent adverse effects. One major approach to accomplish this goal has been to identify the key pathways and molecular targets involved in the complex physiological effects of psychedelics and then rationally engineer pathway-selective or so-called biased agonists tailored to specific therapeutic outcomes, while avoiding adverse events. Herein we review the current state of what is known about the targets and pathways linked with distinct psychedelic pharmacology with a particular focus on the 5-HT_2A receptor, and highlight some of the ongoing work on the design and activity of novel compounds. We conclude with a discussion of the findings in the context of other fields such as opioid drug discovery and highlight which lessons can be learned to facilitate the development of novel psychedelic-based pharmacotherapies (see the sidebar titled Psychedelic Psychopharmacological Effects).

PSYCHEDELIC PSYCHOPHARMACOLOGICAL EFFECTS.

Altered States of Consciousness

Psychedelics induce altered states of consciousness (ASCs), which were suggested to be responsible for their sustained therapeutic effects but also acute and sometimes persistent adverse perceptual and psychological changes. ASCs typically include five or more dimensions (oceanic boundlessness, anxious ego dissolution, visionary restructuralization, auditory alterations, and reduction of vigilance), which are assessed through a questionnaire.

Subjective Effects

Psychedelic-induced ASCs are often felt as some of the most meaningful life experiences, not unlike the birth of a child or death of a loved one. When corrected for intensity, such effects are highly subjective and vary in their content and experience between users. There is considerable ambiguity as to whether these effects are necessary for therapeutic efficacy, but their study is highly complicated by the inability to model such effects in rodents and other animals.

Hallucinations

Psychedelics cause perceptual changes including visual or auditory distortions, synesthesia, and other effects. Such changes not only acutely impair the judgment of users but also can develop into hallucinogen persisting perception disorder, characterized by frequent re-experiencing of these changes. Because of their harmful potential, hallucinogenic effects are often considered as detrimental side effects that are to be limited in the therapeutic use of psychedelics.

COMPLEXITY OF PSYCHEDELIC PHARMACOLOGY

The effects of psychedelics in humans are complex, and often alter both psychological and physiological homeostasis. Generally, psychedelics are defined by being able to induce ASCs, which have often been divided into the five dimensions oceanic boundlessness, dread of ego dissolution, visionary restructuralization, auditory alterations, and reduction of vigilance (20). Contained within this set of descriptors are perceptual changes often simply termed hallucinations, but also mystical experiences with likely profound psychological impacts on the feeling of self (21). In addition, psychedelics also exhibit a myriad of other physiological effects, including altered core body temperature, raised heart rate, and increased wakefulness (22). These effects are not associated with either therapeutic or severe adverse outcomes, but further highlight the pharmacological complexity of attributing key physiological or psychological effects to distinct circuitry, sets of molecular targets, or cellular signaling pathways.

THE 5-HT_2A RECEPTOR IS NECESSARY FOR PSYCHEDELIC DRUG ACTION

A plethora of studies have shown that psychedelics are agonists of the G protein-coupled 5-HT_2A serotonin receptor, which has been shown to mediate many key aspects of psychedelic psychopharmacology. For instance, seminal studies in humans showed that the psychosis-like, or psychotomimetic, effects of psilocybin could be blocked by the 5-HT_2A-selective antagonist ketanserin (23), following prior preclinical studies in which ketanserin blocked discrimination between psychedelics and other serotonergic agonists in rats (24). Similarly, studies showed that LSD’s subjective effects, i.e. the meaningfulness of the psychedelic experience and its impact on one’s behavior and perception of the world, can also be blocked by ketanserin (25; 26).

While these subjective effects are challenging, if not impossible, to model in most animals, hallucination-like perceptual effects can somewhat reliably be measured via proxy assays in rodents. The perhaps most important preclinical assay is the head-twitch response (HTR) in which administration of psychedelics produces rapid side-to-side rotational head movements in rats and mice (27). It should be noted that although HTR lacks face validity (i.e., the response does not directly represent measures of subjective or perceptual effects as observed in humans), the assay has excellent predictive validity to model and predict the actions of known psychedelics in humans (28). Other assays include prepulse-inhibition (PPI) (29), drug discrimination (30; 31), and several behavioral paradigms including exploratory behavior patterns (32). Among these, HTR is largely considered the gold standard for determining hallucinogenic effects in rodent models. PPI, which is also observed in humans (33), is more commonly associated with the psychomimetic effects of psychedelics rather than acute perceptual changes, and PPI has been shown to also be induced by other drugs such as phencyclidine (PCP) (34) and amphetamines (35). Drug discrimination and alterations in exploratory behavior are both complex and could conceivably be driven by effects other than the hallucinogenic attributes of the psychedelic experience. Although not without flaw, these assays are currently the best paradigms to assess 5-HT_2A receptor target engagement and hallucinogenic-like effects in preclinical models (36). This is founded on a plethora of studies that show that 5-HT_2A receptor knockout or inhibitions block psychedelic-mediated HTR and PPI (37–40).

One important question that remains unresolved is why psychedelics cause hallucinations and other perceptual changes, when serotonin and other serotonergic drugs do not. A central hypothesis in the field is that psychedelics produce unique signaling patterns via 5-HT_2A receptors, leading to unique psychopharmacological effects. G protein–coupled receptors (GPCRs) such as the 5-HT_2A receptor have been known to engage a variety of different cellular signaling pathways, and receptor agonists are thought to stabilize different subsets of receptor conformations that lead to the engagement of distinct signaling pathways. This phenomenon termed functional selectivity or biased signaling, has been widely cited as the likely cause for why psychedelics such as LSD and psilocybin cause hallucinations, while other 5-HT_2A agonists such as lisuride (41) and 5-HT (42) do not. Empowered by this hypothesis, considerable efforts have been dedicated to developing biased 5-HT_2A receptor agonists that selectively engage some pathways while avoiding others (43–48). The goal, of course, is to generate novel and safer psychedelic-like drugs that retain their therapeutic effects, while no longer producing acute hallucinations or more persistent adverse reactions.

The current underlying hypothesis for the rapid-acting and durable therapeutic effects by psychedelics is increased neuroplasticity. Psychedelics such as psilocybin, 5-MeO-DMT and many others have shown neuroplastic effects in cortical neurons in culture and in vivo (49–51). According to some studies, psychedelic-mediated increases in neuroplasticity may be dissociated from hallucinogenic-like effects in animals (52) and humans (53). But other studies have argued that the profound and mystical experiences stemming from psychedelic use in clinical settings are essential for rapid-acting and durable therapeutic efficacy in humans (25). It thus remains to be seen whether enduring therapeutic effects can be separated from acute and persistent adverse effects, as well as whether all psychedelics can produce rapid-acting and durable therapeutic responses. Another important question is whether 5-HT_2A receptor agonism alone is sufficient for such rapid-acting and therapeutic responses clinically or other receptors are needed to augment such effects. At the very least, selective 5-HT_2A receptor biased agonists are desperately needed chemical tools to decipher which pathways mediate which aspects of the overwhelmingly complex psychopharmacology of psychedelics.

5-HT_2A RECEPTOR SIGNALING FEATURES

Although stabilization of distinct receptor conformations is at the heart of functional selectivity or signaling bias, biased agonism can manifest on multiple levels of interrogation, including transducer-specific, temporal and spatial contexts. For instance, GPCRs have been observed in different compartments such as a cholesterol-rich rafts (54), neuronal cilia (55), and even intracellular membranes including endosomes or golgi (56), thus contributing toward the complex nature of expression of GPCR biased agonist properties. In aggregate, GPCR bias can accordingly be described as a biased ligand’s ability to (a) activate distinct pathways through the same receptor, (b) stimulate receptors in a different cellular compartment, and/or (c) only activate cells with compatible signaling networks. With respect to psychedelic drug action at the 5-HT_2A receptor, several studies have put forward hypotheses with novel approaches and tool compounds to illuminate signaling pathways relevant for therapeutic versus hallucinogenic effects of psychedelics.

GPCRs typically engage only a limited number of signal transducers directly, the most prominent of which are heterotrimeric G proteins and β-arrestins (see the sidebar titled Principles of Serotonin GPCR Pharmacology). Historically, GPCR signaling bias has been described as the ability to direct signaling to either G protein– or β-arrestin-mediated cellular events, but even preferential engagement of distinct G protein subtypes plays an important role in the pluridimensionality of GPCR signaling. Ultimately, researchers hope to disentangle the complexity of GPCR signaling and identify the transducers and pathways responsible for select physiological effects mediated by individual receptor systems. For instance, for many years researchers have tried to generate safer opioids based on the now-disputed rationale that pain-relieving therapeutic effects and lethal respiratory depression are mediated via G protein and β-arrestin pathways, respectively (57–61). Other examples include using biased compounds as safer antipsychotics targeting the D2 dopamine receptor (62; 63) or biased angiotensin II receptor ligands to increase cardiac performance (64). Not surprisingly, several studies on 5-HT_2A receptor signaling have examined the ability of different ligands to stimulate G protein activation or β-arrestin recruitment, and studies have further investigated the link between distinct transducers and the diverse physiological effects of psychedelics.

PRINCIPLES OF SEROTONIN GPCR PHARMACOLOGY.

5-HT Receptor

5-Hydroxytryptamine or serotonin receptors are membrane proteins that mediate much of the pharmacological effects of psychedelics. The 5-HT_2A receptor is a G protein–coupled serotonin receptor and the primary site of action for psychedelics.

Partial Agonism

Partial agonists produce a lower stimulus compared to the endogenous ligand serotonin, which by definition exhibits full receptor efficacy. Depending on their efficacy, partial agonists produce a stimulatory response but can also antagonize full agonists in vivo.

Selectivity

G protein–coupled receptor (GPCR) ligands typically act on more than one receptor due to the similarity of receptor binding pockets. For instance, serotonin activates 13 different serotonin receptors. Selectivity is a measurement of a ligand’s affinity for one receptor over that of another.

β-Arrestins

β-Arrestins directly engage GPCRs and are named after their ability to terminate GPCR signaling. Over the past 20–30 years, studies showed that β-arrestins can elicit G protein–independent signaling and sequester GPCRs to different cellular compartments.

Biased Signaling/Functional Selectivity

5-HT_2A receptor activation stimulates several distinct signaling cascades, and different drugs can bias the receptor to selectively activate some pathways over others, most notably G proteins versus β-arrestins. Activation of select pathways in cellular compartments or tissues can cause distinct drug actions via the same receptor subtype.

The 5-HT_2A receptor has been well-established to couple to the G_q/11 family of heterotrimeric G proteins, thereby activating phospholipase C (PLC) and producing inositol-1,4,5-trisphosphate (IP₃) and diacylglycerol (DAG) from membrane lipids (Figure 1) (65; 66). Subsequently, the second messenger DAG activates protein kinase C, and liberated IP₃ opens IP₃ receptors, which are ligand-gated calcium-selective channels, on the endoplasmic reticulum. Studies have also described additional pathways, such as activation of phospholipase A₂ (PLA₂), which produces the secondary messenger arachidonic acid (AA), or the release of the endocannabinoid 2-arachidolylglycerol (67; 68). As seen for other GPCRs, the 5-HT_2A receptor also can recruit β-arrestins, which are known to desensitize and internalize the 5-HT_2A receptor (Figure 1) (45; 69). Because GPCRs can couple promiscuously to other transducers such as different G protein subtypes (70), the 5-HT_2A receptor has been investigated for promiscuous coupling to G proteins other than G_q/11. Evidence for alternative G protein coupling initially came from studies examining PLA₂ AA release assays that were dependent on either pertussis toxin–sensitive (G_i/o) or pertussis toxin–insensitive (e.g., G12/13 or others) transducers (68; 71), further demonstrating the complexity of 5-HT_2A receptor signaling. Since then, the 5-HT_2A receptor has been shown to couple to G_i/o/z subtypes (Figure 1), but the strength and degree of G_i/o/z coupling differ depending on the biosensor platform employed (45; 72; 73). For example, 5-HT_2A receptor coupling to G_i1 has been posited as important for blocking psychosis by the US Food and Drug Administration–approved 5-HT_2A receptor–selective inverse agonist, pimavanserin (74).

Figure 1 |. 5-HT_2A receptor signaling pathways.

Signaling components stemming from 5-HT_2A receptor activation include (a) G_q/11 signaling pathways that activate PLC, leading to IP₃ and DAG generation, followed by PKC activation; (b) β-arrestin recruitment, leading to internalization, which is implicated in neuritogenesis and neuroplastic potential and may be due to intracellular sustained signaling; and (c) alternative pathways such as AA generation, which may be due to a pertussis toxin–sensitive or –insensitive G_i/o/z pathway. Abbreviations: 5-HT, 5-hydroxytryptamine; AA, arachidonic acid; DAG, diacylglycerol; ER, endoplasmic reticulum; IP₃, inositol-1,4,5-trisphosphate; PKC, protein kinase C; PIP2, phosphatidylinositol 4,5-bisphosphate; PLA₂, phospholipase A₂; PLC, phospholipase C. Figure created in BioRender; McCorvy J. 2025. https://BioRender.com/zkdc0c4.

Although it has long been known that GPCRs localize to intracellular compartments and that signaling bias can be spatially encoded to engage different signaling pathways depending on their cellular location (75; 76), the spatial location of 5-HT_2A receptors as it applies to signaling has been less studied, especially in the context of psychedelic drug action. Curiously, the 5-HT_2A receptor in particular has primarily been found in intracellular compartments (77; 78), and 5-HT_2A receptor internalization and potential degradation can be induced by both agonists and antagonists (79). Other studies even proposed a connection of intracellular 5-HT_2A receptors and psychedelic-induced cortical plasticity (Figure 1) (80), similar to studies linking synaptic plasticity to intracellular populations of other receptors (81).

GENETIC APPROACHES FOR STUDYING 5-HT_2A SIGNALING PATHWAYS

Despite evidence for alternative G protein coupling at the 5-HT_2A receptor, G_q/11 and β-arrestins remain the canonical transducers for determining psychedelic action in vivo. Unfortunately, the use of genetic knockout approaches has only convoluted the relative contributions of G proteins versus β-arrestins on psychedelic drug action. For example, G_q gene deletion in mice reduces but does not fully abolish 2,5-dimethoxy-4-iodoamphetamine (DOI)-mediated HTR (82), but this effect could be due to contributions from G11 signaling. Studies with β-arrestin2 knockout mice revealed that effects of the 5-HT/serotonin precursor 5-hydroxytryptophan (5-HTP), but not DOI-induced HTR, are dependent on β-arrestin2 (83). Studies with LSD show that β-arrestin2, but not β-arrestin1 is critical for LSD-mediated effects in the HTR and PPI assays (40), but these studies were only performed with a single dose of LSD. Collectively, these findings argue that both G_q/11- and β-arrestin-mediated cellular events contribute to HTR and therefore psychedelic effects, but compensatory changes such as altered gene expression related to knockouts cannot be ruled out.

MODULATION OF 5-HT_2A RECEPTOR SIGNALING BY OTHER RECEPTOR SYSTEMS

Overwhelming evidence from pharmacological studies in humans and animal models shows that the 5-HT_2A receptor undoubtedly lies at the center of the psychotropic effects of psychedelics. However, it has long been known that psychedelics bind and activate many different receptors, and their contributions to the psychedelic experience are currently under investigation. Several independent studies have already suggested that 5-HT_2A receptor activation is insufficient to explain the complex therapeutic or hallucinogenic aspects of psychedelic physiology. For instance, the 5-HT_2A receptor antagonist ketanserin did not block psilocybin-mediated attenuation of anhedonia in mice (39) or psilocybin-induced structural plasticity (50), which has been associated with the therapeutic effects of psychedelics. In humans it was shown that the empathogenic effect of LSD was not blocked by ketanserin (84), and the 5-HT_1A receptor antagonist pindolol increased the subjective psychedelic effects of DMT (85). These data collectively highlight that receptors other than the 5-HT_2A receptor might play a major role in the therapeutic and/or subjective effects of psychedelics. This should not come as a surprise given the affinity of psychedelics for multiple targets. For instance, LSD activates 11 of the 12 human 5-HT G protein–coupled serotonin receptors (86; 87), as well as other receptors including dopamine, adrenergic, and trace amine receptors (53; 88–91). Recent studies suggest that 5-MeO-DMT, a psychedelic found in the secretions of the Sonoran desert toad, has higher efficacy and potency in activating the 5-HT_1A receptor compared to the 5-HT_2A receptor (92; 93). Moreover, these and several other studies suggest that the 5-HT_1A receptor likely contributes to the antidepressant effects of psychedelics (92; 94), and 5-HT_1A receptor activation has also been repeatedly shown to modulate HTR, thereby directly affecting hallucinogenic-like effects in animal models. For instance, the 5-HT_1A receptor antagonist WAY-100635 attenuated LSD-mediated locomotor activity in mice (95), and even eliminated 5-MeO-DMT-mediated PPI disruption (96), a sensorimotor-based proxy for hallucinogenic effects in animal models. Moreover, HTR can be masked by 5-HT_1A receptor stimulation, as studies show that WAY-100635 pretreatment can increase 5-MeO-DMT-mediated HTR (92; 93), and 5-MeO-DMT can attenuate DOI-induced HTR (97). These findings mirror observations that the 5-HT_1A receptor agonist 7-(dipropylamino)-5,6,7,8-tetrahydronaphthalen-1-ol (8-OH-DPAT) can inhibit HTR stimulated by 5-HTP and psilocybin (98). Perhaps even more importantly, the 5-HT_1A receptor agonist buspirone reduces psilocybin-induced visual hallucinations in humans (99), providing evidence that 5-HT_1A receptor–mediated modulation of hallucinogenic effects is not limited to animal models.

Similarly, evidence has been provided for the involvement of 5-HT_2C receptors in modulating psychedelic response in animal models (100). For instance, in rats, the 5-HT_2C receptor antagonist SER-082 blocked some of the LSD-mediated behavioral patterns (101), and the 5-HT_2C receptor antagonist SB242084 increased psilocybin-mediated HTR in mice (102). These results are supported by the finding that the 5-HT_2C receptor antagonist RS-102221 enhanced HTR at lower doses but reduced it at higher doses (98), and other studies showed that 5-HT_2C receptor knockout or inhibition with the antagonists SB206553 or SB242084 attenuated DOI-induced HTR (103).

LEVERAGING 5-HT_2A RECEPTOR BIASED AGONISTS TO UNDERSTAND PSYCHEDELIC DRUG ACTION

Considering the limitations of genetic animal studies and lack of 5-HT_2A receptor selectivity with traditional psychedelics, an alternative approach for disentangling 5-HT_2A receptor signaling pathways associated with psychedelic effects is to design biased agonists for the 5-HT_2A receptor to determine the relative contributions of select signaling pathways. Some of the initial clues for diverging 5-HT_2A receptor signaling pathways came from determining second messenger IP₃ versus AA release, resulting in 5-HT_2A receptor agonists that preferentially stimulate IP₃ accumulation versus AA release (43; 104). Recently, mini-G_q recruitment versus β-arrestin recruitment assays have been used for the determination of 5-HT_2A receptor biased agonism of select psychedelic compounds (105), but it remains to be seen whether this approach can successfully identify biased agonists validated in vivo.

By leveraging 5-HT_2A receptor biased agonist discoveries in combination with animal studies such as HTR determination, there has been much progress using recently determined 5-HT receptor structures for discovering 5-HT_2A receptor biased agonists as tool compounds to explore the relative contributions of G protein versus β-arrestin recruitment associated with psychedelic effects (Table 1).

Table 1 |.

Reported 5-HT_2A receptor biased ligands

Compound names and type (reference)	Structure	5-HT2A activity	Selectivity measures	HTR response	Other behavioral notes
IHCH-7079 and IHCH-7086 Weakly β-arrestin biased (47)		IHCH-7079 Gq = 707.95 nM, 20.19% βarr2 = 174.78 nM, 39.00% IHCH-7086 Gq: not detectible βarr2 = 204.17 nM, 12.72%	IHCH-7079 5-HT2A Ki = 16.98 nM 5-HT2B Ki = 26.30 nM 5-HT2C Ki = 38.02 nM [3H]-LSD derivedb IHCH-7086 5-HT2A Ki = 12.59 nM 5-HT2B Ki = 36.31 nM 5-HT2C Ki = 39.81 nM [3H]-LSD derivedb	IHCH-7079 Inactive at 2 and 10 mg/kg Blocked LSD-induced HTR at 2 and 10 mg/kg IHCH-7086 Inactive at 2 and 10 mg/kg Blocked LSD-induced HTR at 2 and 10 mg/kg	IHCH-7079 and IHCH-7086 attenuated acute restraint-induced immobility and chronic corticosterone-induced immobility in TST and FST in C57BL/6J mice
(R)-69 and (R)-70 Gq biased, “unusual kinetics for G protein signalling versus arrestin recruitment” (48, p. 590)		(R)-69 Gq (calcium flux) = 41.3 nM, 90.1% βarr2: not reporteda (R)-70 Gq (calcium flux) = 109.3 nM, 73.3% βarr2: not reporteda	(R)-69 5-HT2B Gq (calcium flux) = 187 nM, 82.9% 5-HT2C Gq (calcium flux) = 2,061 nM, 75.0% (R)-70 5-HT2B Gq (calcium flux) = 701 nM, 50.5% 5-HT2C Gq (calcium flux) = 3,246 nM, 51.7%	(R)-69 Inactive at 1 mg/kg; slight activity at 3 mg/kg Partially blocked LSD-induced HTR at 1 and 3 mg/kg (R)-70 Inactive at 1 and 3 mg/kg Partially blocked LSD-induced HTR at 3 mg/kg	(R)-69 and (R)-70 did not disrupt PPI (R)-69 and (R)-70 reduced immobility in TST in VMAT2 Het mice (R)-70 reduced immobility in FST
25N-N1-Nap and 25N-NBPh β-arrestin biased (45)		25N-N1-Nap Gq = 1.45 nM, 22.5% βarr2 = 0.41 nM, 78.2% 25N-NBPh Gq = 5.50 nM, 25.7% βarr2 = 4.07 nM, 109.4%	25N-N1-Nap 5-HT2B Gq: no activity 5-HT2C Gq = 13.18 nM, 64.6% 25N-NBPh 5-HT2B Gq: no activity 5-HT2C Gq EC50 >1 μM, Emax not calculable	25N-N1-Nap Inactive at 0.3, 1, 3, 10, and 30 mg/kg Blocks DOI-induced HTR at 3, 10, 30 mg/kg 25N-NBPh Inactive at 3, 10, 30, and 100 mg/kg Blocks DOI-induced HTR at 30 and 100 mg/kg	25N-N1-Nap (20 mg/kg/day) induced tachyphylaxis for DOI HTR 25N-N1-Nap (3 mg/kg) reduced PCP-induced hyperlocomotion in C57BL/6J mice

^aValues not reported from Kaplan et al. (48) extended data figures 2 and 3.

^bNot tested functionally at 5-HT2B and 5-HT2C or others.

Abbreviations: 5-HT, 5-hydroxytryptamine; βarr2, β-arrestin 2; DOI, 2,5-dimethoxy-4-iodoamphetamine; HTR, head-twitch response; LSD, lysergic acid diethylamide; PCP, phencyclidine; PPI, prepulse inhibition; TST, tail suspension test; FST, forced swim test; VMAT2, vesicular monoamine transporter 2.

Using structure-based design, a β-arrestin-biased 5-HT_2A receptor compound (IHCH-7086) was developed based on the antipsychotic lumateperone-bound 5-HT_2A receptor crystal structure (47). IHCH-7086 does not appear to produce HTR in mice, but the level of 5-HT_2A receptor β-arrestin activity was low (<20%) compared to 5-HT maximal activity in this assay, suggesting this compound does not produce sufficient 5-HT_2A receptor agonism in vivo. Moreover, IHCH-7086 and analogs based on lumateperone display substantial affinity at other aminergic targets, including 5-HT_2C and 5-HT_2B receptors, and weaker affinity at 5-HT_1A and dopamine receptors, all of which are known to affect HTR response (98; 106), but not all potential targets were tested for function. However, these compounds did show antidepressant-like phenotypic behaviors by attenuating acute restraint-induced immobility in forced swim and tail suspension tests.

Virtual ligand docking using ultra-large chemical libraries can be leveraged to discover novel chemotypes, as has been done for a plethora of GPCRs (58; 107–109), including 5-HT_2A receptors (48). This recent virtual ligand docking study led to the identification of a novel chemotype series of azaindoles as 5-HT_2A receptor agonists that appear to be G_q biased, with considerable efficacy close to that of 5-HT/serotonin but devoid of HTR in mice. However, these compounds were only tested at two doses in HTR. Within this series, compounds (R)-69 and (R)-70 were weakly selective for 5-HT_2A over 5-HT_2B and 5-HT_2C receptors and did partially block the LSD-induced HTR, indicating brain penetration. Furthermore, these compounds demonstrated a lack of PPI disruption relative to LSD and induced antidepressant-like or anxiolytic activity in mice in several assays, including tail suspension and elevated plus maze. Intriguingly, the authors note that this compound series exhibits “unusual kinetics for G protein signalling versus β-arrestin recruitment” (48, p. 590), highlighting how binding and/or signaling kinetics might play a role in 5-HT_2A receptor stimulatory output in vivo. Specifically, it was recently shown that drug-binding kinetics at the 5-HT_2A receptor strongly influence signal bias patterns (45; 46; 72; 110), and, conceivably, engagement of different signaling pathways and circuits at different times may contribute to hallucinogenesis or other psychotropic properties.

Using a strategy to avoid influence from off-target receptors and engineer in 5-HT_2A receptor selectivity, researchers designed a series of N-benzyl analogs to engender 5-HT_2A receptor β-arrestin biased agonism (45). 5-HT_2A receptor selective β-arrestin-biased analogs, 25N-N1-Nap and 25N-NBPh, were designed to increase the bulk of the N-substituted ring system to interfere with the conserved W^6.48 (superscripts denote Ballesteros-Weinstein numbering) toggle switch involved in GPCR activation (72). These analogs not only showed substantially reduced 5-HT_2A receptor G_q/11 activity (22.5% and 25.7%, respectively) and preserved β-arrestin recruitment nearly to full agonism (78.2% and 109.4%, respectively), but also showed 5-HT_2A receptor selectivity across many tested targets. Importantly, both 25N-N1-Nap and 25N-NBPh were also inactive in HTR and able to block DOI-induced HTR, suggesting sufficient brain 5-HT_2A receptor occupancy. Interestingly, these compounds are reported to possess β-arrestin recruitment efficacy and to cause sufficient internalization in vitro despite lacking G_q activity, and they were further demonstrated to induce tachyphylaxis in the HTR in vivo. Finally, 25N-N1-Nap was able to block PCP-induced hyperlocomotor activity similar to the selective 5-HT_2A receptor antagonist M100907 (volinanserin), suggesting that 5-HT_2A receptor β-arrestin-biased agonism may have utility for treating psychotic symptoms. Importantly, this study further suggests that a G_q/11 efficacy threshold (>70%) is necessary for psychedelic potential, whereas partial agonists (<70%), such as 2-Br-LSD and lisuride, do not evoke sufficient intrinsic efficacy to cause hallucinations.

Further evidence for G_q efficacy leading toward psychedelic effects comes from measuring the efficacy of various psychedelics at 5-HT_2A receptors in vitro. The synthetic psychedelics LSD, DOI, and 25I-NBOMe as well as the natural substance mescaline are all full or near-full agonists in 5-HT_2A receptor–mediated G protein signaling, whereas psilocin, DMT, and 5-MeO-DMT appear to be partial agonists with sufficient efficacy to stimulate HTR (45; 92). These findings may also explain why tabernanthalog (TBG), an iboga alkaloid and 5-HT_2A receptor partial agonist with approximately 50% efficacy in G protein–mediated elevation of intracellular calcium levels, does not produce HTR in mice (52). Importantly, the psychedelic analogs lisuride (41), Ariadne (111), and 2-Br-LSD (53) have been known to be devoid of hallucinogenic effects in humans and animal models (44), and are all reported to have reduced 5-HT_2A receptor–mediated G_q/11 efficacy, thus likely not reaching the threshold levels in vivo required to produce hallucinations.

Studies in mice further suggested that responses to the hallucinogen LSD but not the nonhallucinogen lisuride also involve pertussis-sensitive G_i/o heterotrimers (38). However, it remains unclear whether the 5-HT_2A receptor indeed couples to inhibitory G proteins since studies using bioluminescence resonance energy transfer (BRET)-based biosensors, which measure signaling via G protein dissociation, transducer binding, or sequestration, have reported different results. For instance, one study suggested that only pertussis-insensitive G_z inhibitory proteins couple to 5-HT_2A receptors (72), while a separate study using indirect bystander BRET biosensors suggests that 5-HT_2A receptors can engage G_i1, G_i2, and G_i3 heterotrimers (112). However, compounds shown to preferentially target G_i over G_q signaling in this study were suggested to mediate HTR, but this is difficult to conclude since no off-target activity at other 5-HT receptors or other targets was examined.

5-HT_2A RECEPTOR STRUCTURE AND CONFORMATIONAL STATES

Signaling bias typically arises as different ligands stabilize divergent conformational ensembles of a receptor. Distinct conformations (and their lifetimes) encode how potently and efficaciously different sets of transducers are engaged and activated, thereby selectively activating receptor subpopulations across specific cellular compartments, linked to distinct signaling pathways, or expressed in distinct neuronal populations. Although structures have uncovered features of 5-HT_2A receptor active (46; 72) and inactive states (113), insights into the conformational differences that might link 5-HT_2A receptor activation to distinct signaling pathways remain limited. Recent cryo-EM structures of 5-HT_2A receptor–G_q signaling complexes revealed a noncanonical conformation of the 5-HT_2A receptor when bound by the β-arrestin-biased compound RS130–180 (46) (Figure 2a). Specifically, key 5-HT_2A receptor structural motifs, which translate agonist binding into an intracellular opening of the receptor for transducer activation, assume conformations that do not align with that of either the 5-HT_2A receptor inactive state or canonical 5-HT_2A receptor active state conformations. RS130–180 appears to push the conserved toggle switch residue W^6.48 closer toward the receptor core than observed for other 5-HT_2A receptor active states. This appears to induce a unique configuration of F^6.44 of the P^5.50-I^3.40-F^6.44 (PIF) motif, a key switch region in the receptor that links agonist binding to activation-related conformational transitions (86; 114). Changes in F^6.44 are typically thought to create the torque on transmembrane helix 6 (TM6) that leads to its outward movement and subsequent opening of the receptor on the intracellular site. Rather than moving inward to point toward the receptor core, in the RS130–180-bound state, F^6.44 appears to flip outward to point toward TM7, which can no longer move inward as observed in other 5-HT_2A receptor active states. Similarly, the structure of the 5-HT_2A receptor bound to 2-Br-LSD showed that the 2-Br substituent on LSD may impair the PIF trigger motif with the bromine atom specifically targeting I^3.40, thereby affecting receptor activation. However, it remains unknown how exactly these noncanonical states translate to preferred activation of β-arrestins over G proteins. One suggestion was that such noncanonical states are less optimal for G protein binding, as studies using 25N-N1-Nap ligands similarly suggested that unique interactions with W^6.48 can dampen G_q activation thus leading to β-arrestin-biased output (45). Together, these findings to some extent mirror structural studies of the related 5-HT_2B receptor that also suggested that the β-arrestin-bias of LSD and other ergolines may arise from conformations that are suboptimal for the engagement of G proteins (86). Additional structures of G_q- and β-arrestin-bound 5-HT_2B receptor signaling complexes showcase how larger TM6 movements may be required to engage β-arrestin (115), which provides additional clues as to which features drive engagement and activation of distinct signal transducers. This is also supported by studies of the 5-HT_1A receptor that revealed how ligand-stabilized conformational variability in TM6 leads to the differential activation of distinct inhibitory G proteins (116). Furthermore, molecular dynamics simulations of the 5-HT_2A receptor suggest that psychedelic and nonpsychedelic compounds stabilize distinct intracellular loop 2 (ICL2) conformations, which conceivably leads to engagement with different signal transducers, or potentially drive the formation of 5-HT_2A receptor homo- or heteromers (117). Along these lines, studies have previously reported on potential heterodimer formation of 5-HT_2A receptors and G_i/o/z-coupled metabotropic glutamate receptor 2 (mGluR2) (118). Specifically, postmortem human brain tissue of schizophrenic subjects was suggested to contain 5-HT_2A/mGluR2 receptor heterodimers that could be responsible for altered cortical processes. Since psychedelics induce hallucinations likened to psychosis-like states and are reported to exacerbate symptoms of schizophrenia, it was implied that the 5-HT_2A/mGluR2 receptor complex is the site of action for hallucinogenic substances such as LSD and other psychedelics (118).

Figure 2 |. Structural studies of the 5-HT_2A receptor illuminate mechanisms of psychedelic action and ligand bias.

(a) Structures of the 5-HT_2A receptor in active (light blue, PDB ID 6WHA; https://www.rcsb.org/structure/6WHA), inactive (red, PDB ID 6A93; https://www.rcsb.org/structure/6A93), and noncanonical (green, PDB ID 9ASA; https://www.rcsb.org/structure/9ASA) states. Structural models (top) and simplified schematics highlight differences in overall helical configurations and key motifs and residues. (b) Close-up of the LSD-bound 5-HT_2A receptor (purple) ligand-binding pocket (PDB ID 6WGT; https://www.rcsb.org/structure/6WGT), with lid residues (yellow) sealing LSD in the binding site (top). Schematic illustrating that mutating L229 increases LSD’s off rate and selectively diminishes β-arrestin activity (bottom). Abbreviations: 5-HT, 5-hydroxytryptamine; ECL2, extracellular loop 2; ICL2, intracellular loop 2; LSD, lysergic acid diethylamide; PDB, Protein Data Bank; TM, transmembrane helix; WT, wild type. Images created using the PyMOL Molecular Graphics System, Version 1.2r3pre, Schrödinger, LLC.

Studies of 5-HT_2A and related 5-HT receptors have further uncovered additional features that contribute to biased signaling of psychedelics and other serotonergic agents (86; 87; 110; 115; 119). For instance, several studies uncovered how LSD engages receptor residues near extracellular loop 2 (ECL2), which appears to be critical for LSD’s slow binding and receptor dissociation kinetics (Figure 2b). Curiously, it was also reported that accelerating LSD’s binding kinetics selectively diminishes β-arrestin recruitment, which further highlights the temporal dimension of biased signaling and links drug-binding kinetics with distinct cellular outcomes (72; 110).

SUMMARY AND FUTURE OUTLOOK

Psychedelics undoubtedly hold great potential in the treatment of a variety of psychiatric illnesses and other disorders, but their use, even under supervised clinical settings, is not without considerable risk. Their safe application in clinical settings and successful mounting of regulatory hurdles is, however, likely dependent on a better understanding of their pharmacological actions. This notably includes a comprehensive delineation of signaling mechanisms on (a) the atomic scale in potentially stabilizing distinct 5-HT receptor conformations, (b) the cellular scale in activating diverse signaling pathways, and (c) the circuitry scale in stimulating distinct neuronal/cell populations in various brain regions. As reviewed here, studies have begun elucidating aspects of psychedelic activity at varying levels of resolution and different layers of signal integration. That being said, it remains to be determined which receptor conformations, signaling pathways, and circuits are responsible for different aspects of psychedelic physiology. Such data could potentially facilitate development of safer pharmacotherapies that separate sustained therapeutic effects from adverse reactions such as persistent perceptual changes and exacerbation of anxiety or psychoses, as well as understudied physiological effects such as changes in core body temperature or gut motility. Several novel psychedelic-like compounds have already emerged from preclinical studies and represent powerful tools to study these mechanisms.

Despite these promising findings we would like to issue words of caution based on lessons learned from other receptor and drug systems. We specifically note that the current state of the field is all too reminiscent of efforts in designing safer opioid-like drugs based on novel signaling modalities. Decades were spent developing G protein–biased mu-opioid receptor compounds that demonstrated reduced side effects such as respiratory depression, the main reason for opioid-related deaths, in preclinical models. Ultimately, the first of these drugs, oliceridine (TRV-130), was approved, but respiratory depression was still an issue in some aspects of the clinical trials, especially compared to morphine (120). Several studies have since suggested that not the lack of β-arrestin-mediated effects, but low intrinsic G protein activity could be the key to safer opioid-like compounds (121). Other studies have called into question the reproducibility of the pre-clinical work altogether (122). While this road still lies ahead for novel psychedelic-based therapies, lessons from the opioid field present a unique opportunity to avoid such pitfalls, and highlight the importance of fundamental studies to better understand the mechanisms of classic psychedelics. An additional concern is the translational power of rodent models, which much of the current discoveries are based on. Opioid-related physiological processes, including pain regulation, respiration, and reward, are evolutionarily well-conserved (123–126), which arguably cannot be said for many cortical processes (127–130) that psychedelics alter. Proxy assays such as HTR correlate reasonably well with the acute perceptual effects of psychedelics and 5-HT_2A receptor engagement, but by no means model the subjective effects in humans, something that may be impossible in rodents. It should also be highlighted that brain structures high in 5-HT_2A receptor expression such as the claustrum, similarly, are poorly conserved between humans and most animal models (131). It thus remains to be seen how findings from pre-clinical work translate to effects in humans, and work in this space, therefore, likely demands exceptional rigor and reproducibility. This is by no means meant to diminish current accomplishments or novel probes, but rather caution against premature conclusions drawn from incomplete characterization of signaling modalities or rodent behavior studies which might not translate well going forward.

Overall, psychedelic research is at an exciting turning point. Once shunned as illicit drugs with no clinical use, their tremendous therapeutic potential has reinvigorated inquiries into their basic mechanisms. Empowered by recent chemical biology efforts, we are closer than ever to deciphering the fundamental signaling pathways and circuits involved in their complex physiological effects. Continued efforts, carried out with the appropriate scientific rigor, undoubtedly promise to facilitate the development of psychedelic and psychedelic-based therapies in the hopefully not too distant future.