In Vitro Comparison of Ulotaront (SEP-363856) and Ralmitaront (RO6889450): Two TAAR1 Agonist Candidate Antipsychotics

Abstract

Background

Trace amine-associated receptor-1 (TAAR1) agonists have been proposed as potential antipsychotics, with ulotaront and ralmitaront having reached clinical trials. While ulotaront demonstrated efficacy in a recent Phase II trial, a corresponding study studies of ralmitaront failed to show efficacy as a monotherapy or as an adjunct to atypical antipsychotics. In addition to TAAR1 agonism, ulotaront is a partial agonist at the serotonin 1A receptor (5-HT_1AR). However, little is known about ralmitaront.

Methods

We compared ulotaront and ralmitaront at TAAR1, 5-HT_1AR, and dopamine D₂ using luciferase complementation-based G protein recruitment, cAMP accumulation, and G protein–coupled inward rectifier potassium channel activation assays.

Results

Ralmitaront showed lower efficacy at TAAR1 in G protein recruitment, cAMP accumulation, and GIRK activation assays. Moreover, ralmitaront lacked detectable activity at 5-HT_1AR and dopamine D₂.

Conclusions

Compared with ulotaront, ralmitaront shows lower efficacy and slower kinetics at TAAR1 and lacks efficacy at 5-HT_1AR. These data may be relevant to understanding differences in clinical profiles of these 2 compounds.

Significance Statement.

Chemical compounds that activate trace amine-associated receptor-1 (TAAR1) have gained interest as a novel class of putative antipsychotics and are expected to have pro-cognitive effects, potentially capable of addressing presently undertreated symptoms of schizophrenia. Ulotaront and ralmitaront are the 2 most clinically advanced TAAR1 agonists to date. However, whereas the in vitro pharmacology of ulotaront has recently been described in several publications, little has been made publicly known about the corresponding properties of ralmitaront. Here, we compared the activities of ulotaront and ralmitaront at TAAR1, dopamine D₂ (D₂R), and serotonin 5-HT_1A (5-HT_1AR) receptors using potassium channel activation, G protein recruitment, and cyclic AMP modulation assays. We found that ralmitaront is a less efficacious agonist with slower binding kinetics than ralmitaront at TAAR1. Moreover, ralmitaront lacks activity at 5-HT_1AR and D₂R, where ulotaront is a low-efficacy agonist. These findings highlight the in vitro pharmacological differences between ralmitaront and ulotaront.

INTRODUCTION

Trace amine-associated receptor-1 (TAAR1) is a G protein–coupled receptor whose endogenous agonists include trace amines such as p-tyramine, β-phenethylamine, octopamine, and tryptamine, as well as the thyroid hormone metabolite 3-iodothyronamine (Zucchi et al., 2006; Gainetdinov et al., 2018). Although TAAR1 expression is low throughout tissues, this receptor has been described in the pancreas, gastrointestinal tract, and central nervous system, notably including dopaminergic neurons (Berry et al., 2017; Mantas et al., 2021). TAAR1 couples to G_s/olf proteins and stimulates adenylate cyclase activity but also opens G protein–coupled inward rectifier potassium (GIRK) channels to reduce dopaminergic neuron excitability (Bradaia et al., 2009; Gainetdinov et al., 2018). In addition, TAAR1 activation suppresses the activity of serotonergic neurons in the raphe nuclei and elicits antipsychotic-like and pro-cognitive effects in animal models (Revel et al., 2011; Dedic et al., 2019; Kokkinou et al., 2021; Saarinen et al., 2022).

Based on their regulatory effects on dopaminergic neurotransmission and efficacy in animal models, TAAR1 agonists have attracted attention as potential antipsychotics. Ulotaront (SEP-363856; developed by Sunovion Pharmaceuticals and PsychoGenics) and ralmitaront (RO6889450; developed by Roche) are arguably the 2 most clinically advanced TAAR1 agonists to date (Kane, 2022; Figure 1A). A recent trial of ulotaront, a dual TAAR1 and serotonin 1A receptor (5-HT_1AR) agonist (Dedic et al., 2019), yielded encouraging results in treating acute psychosis in schizophrenia (Koblan et al., 2020; Correll et al., 2021).

Figure 1.

Functional evaluation of ralmitaront at trace amine-associated receptor-1 (TAAR1). (A) Chemical structures of the TAAR1 agonists ulotaront and ralmitaront. (B) Comparison of the recruitment of LgBiT-miniGα_s to 3xHA-9β2-TAAR1-SmBiT by p-tyramine, ulotaront, and ralmitaront in the NanoBiT assay. Each data point represents mean ± SEM from 3 separate experiments, each using 2 replicate wells for each agonist concentration. Data shown are from 15 minutes after agonist addition. (C) Time dependence of pEC₅₀s in the TAAR1 NanoBiT assay. (D) Stimulation of cAMP production measured as light intensity change in HEK293T cells expressing 3xHA-9β2-TAAR1 and GloSensor 22F at different time points following agonist addition. Each data point represents mean ± SEM from 3 separate experiments, each using 8 replicate wells for each agonist concentration. (E) Time dependence of pEC₅₀s in the TAAR1 cAMP assay. (F) Activation of G protein–coupled inward-rectifying potassium channels upon ulotaront and ralmitaront application to Xenopus oocytes expressing G protein-coupled inward rectifier potassium (GIRK) 1/4 channels, 3xHA-9β2-TAAR1, and G_s. Current amplitudes at 1 minute following application of each agonist concentration were normalized, within cells, to the response to 100 µM p-tyramine. Each data point represents mean ± SEM from 4 to 8 separate oocytes. (G) Activation rates of TAAR1-mediated GIRK responses elicited by ralmitaront (slope 0.006 ± 0.001 s⁻¹µM⁻¹; y-intercept 0.019 ± 0.007 s⁻¹, n = 3–7 oocytes, concentrations tested: 1 µM, 3 µM, 6.5 µM, 10 µM) and ulotaront (slope 0.163 ± 0.022 s⁻¹µM⁻¹; y-intercept 0.035 ± 0.008 s⁻¹, n = 3–6 oocytes, concentrations tested: 100 nM, 300 nM, 650 nM). (H) Response decay time courses following washout of 3 µM ralmitaront (n = 4) or 300 nM of ulotaront (n = 5). The TAAR1 agonists were applied for 15 seconds, and the washout duration was 260 seconds. Data are presented as mean ± SEM.

On the other hand, 2 Phase II trials of ralmitaront as monotherapy for treating acute exacerbations of schizophrenia and schizoaffective disorder and as adjunctive therapy for negative symptoms, respectively, failed to show efficacy (NCT04512066) (NCT03669640). Although several in vitro characterizations of ulotaront have recently been published (Dedic et al., 2019; Saarinen et al., 2022), little has been made publicly known about ralmitaront. The present study compared the activities of ulotaront and ralmitaront at TAAR1, 5-HT_1AR, and dopamine D₂ receptor (D₂R).

METHODS

Molecular Biology and Compounds

cDNA encoding codon-optimized human trace amine-associated 1 receptor (hTAAR1) with N-terminal triple HA tags and 9 amino acids from the β2-adrenoceptor N terminus as well as a C-terminal SmBiT tag (3xHA-9β2-TAAR1-SmBiT) and SmBiT-tagged serotonin 1A (5-HT_1AR-SmBiT) and dopamine D₂ (D₂R-SmBiT) receptors were previously described (Barak et al., 2008; Saarinen et al., 2022). 3xHA-9β2-TAAR1 (without C-terminal SmBiT), untagged human 5-HT_1AR and D₂R (long isoform) constructs, the human G_αs subunit (GNAS), and G protein–coupled inwardly-rectifying potassium channel (GIRK) 1 and GIRK4 were synthesized by GenScript, Inc. (Piscataway, NJ, USA) and cloned into pXOOM (a gift from Dr. Søren-Peter Olesen, University of Copenhagen, Denmark). For cRNA synthesis, the restriction enzyme XbaI was used to linearize 5-HT_1AR and GNAS, whereas XhoI was used to linearize 3xHA-9β2-TAAR1, D₂R, and GIRK1/4. In vitro transcription was performed using the T7 mMessage mMachine kit (Ambion, Austin, TX, USA). cRNA concentration and purity were assessed by spectrophotometry. p-Tyramine, dopamine, quinpirole, and 5-HT were from Sigma-Aldrich (St. Louis, MO, USA), and ulotaront, ralmitaront, and RTI-7470-44 were from MedChemExpress (Monmouth Junction, NJ, USA).

Nanoluciferase Complementation Assay

TAAR1 and DRD2 kinetic measurements were performed using HEK293T cells (American Type Culture Collection, Manassas, VA, USA) transfected with a ratio of 1:1 or 1:10 LgBit tagged mini-G-protein to SmBiT tagged receptor (Saarinen et al, 2022), respectively, with a total plasmid dose of 1 µg mL⁻¹. Cells were seeded and transfected in a volume of 90 µL per well into white 96-well plates (Corning, Inc., Corning, NY, USA) in FluoroBrite Dulbecco’s modified eagle medium (ThermoFischer Scientific, Waltham, MA, USA) supplemented with penicillin-streptomycin (ThermoFischer Scientific) and 5% dialyzed fetal bovine serum (ThermoFischer Scientific). Twenty-four hours after transfection, 10 µM h-Coelenterazine (NanoLight Technologies, Pinetop, AZ, USA) was added to the cells. Baseline luminescence was read (SPARK 10M, Tecan, Männedorf, Switzerland) for 5 minutes to ensure signal stability before the addition of ligands. 10 µL of diluted ligands was added per well followed by luminescence measurements at a read time of 0.1 second per well. For 5-HT_1ARsignaling, the assay was performed as has been previously described (Saarinen et al, 2022). Expi293F cells were maintained according to manufacturer’s instructions in shaking culture flasks (NEST Scientific USA, Woodbridge, NJ, USA) at 8% CO₂ and 37°C.2.5 mL of cells at a density of 3e⁶ cells mL⁻¹ were transfected with Expifectamine at a gene dose of 1 µg mL⁻¹ with a 1:10 ratio of LgBit tagged mini-G-protein to SmBiT tagged receptor. Transfection enhancers were added the following day, and assay was performed 48 hours posttransfection. On the day of the assay, cells were resuspended in dPBS with 10 µM h-Coelenterazine (NanoLight Technologies) and seeded in a volume of 90 µL per well into white 96-well plates (Corning). . Luminescence signal counts were normalized as fold change relative to vehicle-treated control wells.

cAMP Assay

HEK293T cells were grown in 5% CO₂ atmosphere at 37°C on 10-cm dishes (VWR part of Avantor, Radnor, PA, USA) containing Dulbecco’s modified eagle medium with GlutaMAX (ThermoFischer Scientific) supplemented with 0.01% penicillin-streptomycin (ThermoFisher Scientific) and 10% fetal bovine serum (ThermoFisher Scientific). Cells were transfected using linear polyethylenimine (Polysciences Inc., Warrington, PA, USA) with 6 µg GloSensor 22F plasmid (Promega, Madison, WI, USA) and 4 µg (TAAR1) or 1 µg (5-HT_1AR and D₂R) of receptor together with empty pcDNA3.1 as needed to reach 20 µg plasmid DNA per dish. Twenty-four hours after transfection, cells were lifted off, pelleted, and resuspended in Hank’s buffered salt solution supplemented with 300 µM 3-isobutyl-1-methylxanthine (Sigma-Aldrich) and GloSensor reagent (Promega). Cells were dispensed into white 96-well plates (ThermoFischer Scientific) at a density of 50 000 cells per well and left to equilibrate at room temperature for 1 hour, followed by agonist addition. Luminescence intensity was read in a Berthold Tristar 2S plate reader (Berthold Technologies, Bad Wildbad, Germany) with a 1-second integration time at 10-minute intervals following agonist addition. Experiments where inhibition of cAMP production was observed (i.e., with the G_i/o-coupled 5-HT_1AR and D₂R) were performed with basal cAMP levels in the absence of exogenous adenylate cyclase stimulation, as described (Gilissen et al., 2015).

Oocyte Preparation

Xenopus oocytes were surgically isolated under tricaine (MS-222; Sigma-Aldrich) anesthesia from Xenopus laevis females (Nasco, Fort Atkinson, WI, USA). The procedure was approved by the Swedish National Board for Laboratory Animals and the Animal Welfare Ethical Committee in Stockholm (approval number 686–2021). Oocytes were incubated at 12°C in modified Barth’s solution containing 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 15 mM HEPES, 0.33 mM Ca(NO₃)₂, 0.41 mM CaCl₂, 0.92 mM MgSO₄, 2.5 mM sodium pyruvate, 25 U mL⁻¹ penicillin, and 25 µg mL⁻¹ streptomycin, adjusted to pH 7.6 using NaOH and injected individually on the day after extraction with 50 nL water containing cRNA encoding the proteins of interest. Oocytes were injected with either 0.2 ng (for concentration response experiments) or 1 ng (for response kinetics measurements) TAAR1 cRNA or 0.2 ng 5-HT_1AR or 0.2 ng D₂R cRNA, along with 40 pg each of cRNA encoding the GIRK1 and GIRK4 subunits using a Nanoject III microinjector (Drummond Scientific, Broomall, PA, USA). For TAAR1 experiments, 1 ng of GNAS cRNA was also injected. Oocytes were incubated at 12°C in modified Barth’s solution for an additional 6 days before experiments.

Electrophysiology

Electrophysiological recordings were performed at 22°C using 2-electrode voltage-clamp (OpusXpress 6000A; Molecular Devices, San Jose, CA, USA) (Ågren et al., 2021). Data were acquired at −80 mV and sampled at 156 Hz using the OpusXpress 1.10.42 (Molecular Devices) software. To increase the inward rectifier potassium channel current at −80 mV, a high-potassium extracellular perfusion buffer was used (64 mM NaCl, 25 mM KCl, 0.8 mM MgCl₂, 0.4 mM CaCl₂, 15 mM HEPES, adjusted to pH 7.4 using NaOH). For experiments with dopamine, 1 mM ascorbic acid was added to prevent oxidation. p-Tyramine, dopamine, and 5-HT were prepared fresh on the day of experiments and dissolved directly in assay buffer. Ulotaront and ralmitaront were prepared in DMSO. The maximum final concentration of DMSO used in any experiment was 0.1%. During concentration-response experiments, continuous perfusion was maintained at 1 mL min⁻¹. Subsequent ligand applications were performed, with exception for higher concentrations (>10 nM) of ralmitaront due to suspected slow wash-out (see Figure 1). For kinetic experiments, a flow rate of 4.5 mL min⁻¹ was used.

Data and Statistical Analysis

For NanoBiT and cAMP assay data, luminescence counts from drug-treated wells were normalized to the values of vehicle-treated control wells, resulting in “fold change over baseline” values. For electrophysiology TAAR1 data, the baseline current immediately before agonist application was subtracted from each individual recording, and the current amplitude at the end of each agonist application interval was normalized to the maximal response to p-tyramine recorded in the same cell.

Concentration-response relationships were analyzed by fitting a sigmoidal function using nonlinear regression in GraphPad Prism 8 or 9 (GraphPad Software, San Diego, CA, USA).

The following equation was used for fitting:

$${\displaystyle \begin{array}{l}{\displaystyle Y=Bottom+(Top-Bottom)/(1+{10}^{(\log \mathrm{E}{\mathrm{C}}_{50}-X)})}\end{array}}$$ (1)

where Top is the maximal response to ligand (either as a fraction of the maximal response to reference agonist or, in cAMP inhibition experiments, as a fraction of the baseline luminescence signal) and X is the logarithm of dopamine concentration. For NanoBiT, cAMP stimulation, and cAMP inhibition data, Bottom was set to 1, whereas for GIRK activation data, Bottom was set to 0. For cAMP inhibition data, Top would assume a value less than 1, indicating the maximal level of inhibition of cAMP production. Data points are presented as mean ± SEM throughout.

For fitting of pEC₅₀ time course data, the following equation was used:

$${\displaystyle \begin{array}{l}{\displaystyle Y=Y0+(Plateau-Y0)\ast (1-exp(-K\ast x))}\end{array}}$$ (2)

where Y is the pEC₅₀ value as a function of time, Y0 is the pEC₅₀ value at time 0, Plateau is the presumed steady state pEC₅₀ value, K is the rate constant (per unit time), and x is time.

Student’s 1-sample t test was used to determine whether drug-treated conditions significantly differed from control conditions.

RESULTS

TAAR1 Pharmacology

First, the activities of ralmitaront, ulotaront, and p-tyramine were studied in HEK293T cells using a nanoluciferase complementation assay (NanoBiT) measuring the recruitment of LgBiT-miniGα_s to 3xHA-9β2-TAAR1-SmBiT (Saarinen et al., 2022). The N-terminal additions to this construct help increase hTAAR1 surface expression and have been found not to modify the pharmacological properties of this receptor (Barak et al., 2008; Saarinen et al., 2022). Luminescence measurements were made at several time points between 1 and 15 minutes following agonist addition. At 15 minutes after agonist addition, p-tyramine and ulotaront showed efficacies of 1.30 ± 0.01 and 1.32 ± 0.01 relative to baseline (normalized to 1), whereas ralmitaront displayed a lower efficacy of 1.26 ± 0.01 (Figure 1B). The pEC₅₀ values for agonist-induced luminescence increase were observed to increase with prolonged incubation time (Figure 1C). Fitting of an exponential function to the pEC₅₀ time course suggested plateau (steady-state) pEC₅₀ values of 6.96 ± 0.02 and 6.97 ± 0.00 for p-tyramine and ulotaront, respectively, whereas ralmitaront tended toward a pEC₅₀ of 7.03 ± 0.05 (Figure 1C).

Next, adenylate cyclase modulation was measured in HEK293T cells transiently transfected with 3xHA-9β2-TAAR1 and GloSensor (Gilissen et al., 2015). Luminescence was measured for 1 second per well every 10 minutes following ligand application, initially revealing a time-dependent left-shift of the concentration-response relations for all 3 agonists (Figure 1D). Although ralmitaront was less potent than ulotaront and p-tyramine at the 10-, 20-, and 30-minute time points, the potency difference progressively decreased with longer incubation time. Measurements were made up until 70 minutes post-agonist application. Fitting of an exponential function to the pEC₅₀ time course suggested steady-state pEC₅₀s of 7.04 ± 0.03, 7.41 ± 0.03, and 7.75 ± 0.07 for p-tyramine, ulotaront, and ralmitaront, respectively (Figure 1E). At all time points, the efficacy of ralmitaront was lower compared with p-tyramine and ulotaront (Figure 1D). At 70 minutes post-agonist application, p-tyramine and ulotaront showed efficacies of 10.9 ± 1.3 and 10.8 ± 1.3 relative to baseline (normalized to 1), respectively, whereas ralmitaront displayed an efficacy of 6.9 ± 0.2.

Finally, we compared the potencies and kinetics of ulotaront and ralmitaront-mediated GIRK current responses in Xenopus oocytes coinjected with cRNA encoding 3xHA-9β2-TAAR1, GIRK1/4 channel subunits, and G_s, as recently described (Saarinen et al., 2022). GIRK channels open when bound to G_βγ subunits liberated upon G protein activation (Touhara and MacKinnon, 2018). Concentration-response curves were constructed by first applying the reference agonist p-tyramine, which was subsequently washed out and followed by increasing concentrations of ulotaront or ralmitaront (see Methods). Each agonist application lasted 1 minute, and the current response was measured at the end of the application interval.

GIRK response data were normalized within individual oocytes, and the relative efficacies of ulotaront and ralmitaront, compared with p-tyramine, were 0.91 ± 0.04 and 0.63 ± 0.08, respectively. Ulotaront, with a pEC₅₀ of 6.42 ± 0.10, was more potent than p-tyramine (pEC₅₀ 6.09 ± 0.05) and ralmitaront (pEC₅₀ 5.38 ± 0.24; Figure 1F). However, due to slow response equilibration (as noted in the NanoBiT and cAMP assays, above) and the limited stability of the oocyte preparation, we were unable to construct steady-state concentration-response curves for the 3 agonists.

Instead, we focused on the rates of GIRK response activation and termination to deduce information about agonist binding kinetics. The linear dependence of GIRK activation rate on agonist concentration, which can be used as a proxy measure of agonist association rate (k_on); (Ågren et al., 2021), was approximately 30-fold lower for ralmitaront than for ulotaront (Figure 1G). Furthermore, GIRK response termination rates can be used to estimate the (relative) rate of agonist dissociation from its receptor (k_off); (Bünemann et al., 2001; Benians et al., 2003; Ågren et al., 2021; Goldberger et al., 2022). While the ulotaront-induced GIRK response was readily reversible (97 ± 3%; n = 5) upon washout of the compound, ralmitaront-induced responses decreased by 28 ± 8% (n = 4) during the same washout interval of 260 seconds (Figure 1H). To test the possibility that the slow washout of ralmitaront resulted from lipophilic accumulation in the oocyte membrane followed by re-binding to TAAR1, we also conducted experiments where the novel, nM potency hTAAR1 antagonist RTI-7470-44 (Decker et al., 2022) was added in the continued presence of ulotaront or ralmitaront. Similar to what was observed under agonist washout conditions, ulotaront-evoked responses decreased by 95 ± 7 % (n = 6) upon addition of 1 µM RTI-7470-44, whereas ralmitaront-evoked responses decreased by 20 ± 11 % (n = 4; supplementary Figure 1).

5-HT _1AR Pharmacology

When examining LgBiT-miniGα_i coupling to 5-HT_1AR-SmBiT in the NanoBiT assay, ulotaront behaved as a low-potency, low-efficacy agonist, with an almost 2 orders of magnitude lower potency (pEC₅₀ 4.79 ± 0.16) compared with 5-HT (pEC₅₀ 6.33 ± 0.17). The relative efficacy of ulotaront was 1.66 ± 0.05 times the baseline response, whereas the maximal response evoked by 5-HT was 2.87 ± 0.10 times the baseline response. There was no consistent increase in luminescence evoked by ralmitaront at any concentration tested, the highest being 300 µM (Figure 2A). In the GloSensor cAMP assay, 5-HT and ulotaront decreased luminescence output with pEC₅₀s of 8.24 ± 0.09 and 5.19 ± 0.16 at 40 min post-agonist application (Figure 2B). Ulotaront behaved as a partial agonist with a maximal decrease of 44 ± 6% relative to vehicle control, whereas 5-HT decreased the signal by 57 ± 2%. Ralmitaront potency could not be accurately quantified since no curve could be fit to the data (Figure 2B). The highest concentration tested (33.3 µM) evoked a 12 ± 3% (n = 3) reduction in luminescence output relative to vehicle control. However, this decrease was not statistically different from 0 (P = .054, Student’s 1-sample t test). In the GIRK channel activation assay, 30 µM ralmitaront evoked no appreciable response (Figure 2C; n = 6), whereas 30 µM ulotaront elicited 38.2 ± 3.4 % (n = 5) of the control response evoked by 100 nM 5-HT.

Figure 2.

Ralmitaront lacks detectable activity at the serotonin 5-HT_1A receptor (5-HT_1AR) and the dopamine D₂ receptor (D₂R). (A) Comparison of the recruitment of LgBiT-miniGα_i to 5-HT_1AR-SmBiT by 5-HT, ulotaront, and ralmitaront in the NanoBiT assay. Each data point represents mean ± SEM from 3 separate experiments, each using 2 replicate wells for each agonist concentration. (B) Inhibition of cAMP production, measured as light intensity change, in HEK293T cells expressing the 5-HT_1AR and GloSensor 22F at 40 minutes following agonist addition. Each data point represents mean ± SEM from 3 separate experiments, each using 8 replicate wells for each agonist concentration. (C) Average traces representing mean ± SEM showing responses to application of 30 µM ulotaront (n = 5), 30 µM ralmitaront (n = 6), and 100 nM 5-HT (n = 11) in Xenopus oocytes coexpressing the 5-HT_1AR with G protein-coupled inward rectifier potassium (GIRK) 1/4 channels. (D) Comparison of the recruitment of LgBiT-miniGα_i to D₂R-SmBiT by quinpirole, ulotaront, and ralmitaront in the NanoBiT assay. Each data point represents mean ± SEM from 3 separate experiments, each using 2 replicate wells for each agonist concentration. (E) Inhibition of cAMP production, measured as light intensity change, in HEK293T cells expressing the D₂R and GloSensor 22F at 40 minutes following agonist addition. Each data point represents mean ± SEM of data from 3 independent experiments, each using 8 replicate wells for each agonist concentration. (F) Average traces representing mean ± SEM showing responses to applications of 30 µM ulotaront (n = 4), 30 µM ralmitaront (n = 6), and 1 µM dopamine (n = 10), respectively, to Xenopus oocytes co-expressing the D₂R with GIRK1/4 channels.

D ₂R Pharmacology

In the NanoBiT assay, studying LgBiT-miniGα_i coupling to D₂R-SmBiT, ulotaront behaved as a low-potency, low-efficacy agonist, with an almost 2 orders of magnitude lower potency compared with the reference D₂R agonist quinpirole (pEC₅₀ 4.92 ± 0.08 vs 6.78 ± 0.03) and a relative efficacy of 1.51 ± 0.02 relative to baseline. Quinpirole displayed an efficacy of 2.94 ± 0.03 relative to baseline. There was no consistent increase in luminescence evoked by ralmitaront at any concentration tested, the highest being 300 µM (Figure 2D). Ulotaront similarly acted as a low potency, low efficacy agonist at the D₂R in the cAMP assay, decreasing luminescence output by an estimated maximal 29 ± 9% with a pEC₅₀ of 5.17 ± 0.35 at 40 min post-agonist application, while dopamine displayed 39 ± 2% maximal inhibition and a pEC₅₀ of 8.86 ± 0.12 (Figure 2E). Meanwhile, ralmitaront did not display any appreciable activity at D₂R in the cAMP assay (Figure 2E). Similarly, at 30 µM, ralmitaront was without effect on GIRK currents in oocytes coexpressing D₂R with GIRK1/4 subunits (Figure 2F), whereas 30 µM ulotaront evoked a minimal response which was 3.8 ± 3.7 % (n = 4) of the response to 1 µM dopamine.

DISCUSSION

Our results suggest that ralmitaront has lower efficacy than ulotaront at human TAAR1. Furthermore, the electrophysiology data indicate that ralmitaront binding kinetics (agonist association rate; k_on, and dissociation rate; k_off) at TAAR1 are much slower than ulotaront binding kinetics. This is also consistent with the slower gradual increase in ralmitaront potency in the NanoBiT and cAMP assays compared with ulotaront and p-tyramine. Notably, the estimated equilibrium potencies of ralmitaront in these assays were similar to or slightly greater than those of ulotaront.

5-HT_1AR agonism is considered relevant for pro-cognitive and mood-stabilizing effects of some antipsychotics, including clozapine and aripiprazole (Newman-Tancredi and Kleven, 2011). Ulotaront inhibits the firing of serotonergic neurons of the raphe nuclei through 5-HT_1AR, an action that was tentatively linked to its antipsychotic efficacy (Dedic et al., 2019). Here, we did not observe any effect of ralmitaront on 5-HT_1AR in either NanoBiT miniGα_i recruitment or the cAMP and GIRK activation assays. Although ulotaront is considered a TAAR1/5-HT_1AR agonist, showing negligible D₂R occupancy at behaviorally active concentrations, we did find D₂R activity of ulotaront. However, second messenger assays are particularly prone to signal amplification (Smith et al., 2018), which is likely to have made the ulotaront-induced D₂R agonism more readily apparent in our cAMP assay. Low-potency, low-efficacy D₂R activity in the NanoBiT miniGα_i assay and very weak ulotaront-induced GIRK activation via D₂R was also observed. Previous studies (Dedic et al., 2019; Saarinen et al., 2022) similarly reported weak agonism by ulotaront at D₂R. However, we were unable to pick up any D₂R activity of ralmitaront in the NanoBiT, cAMP, and GIRK assays.

In summary, ralmitaront displayed lower efficacy and slower binding kinetics than ulotaront at TAAR1, whereas at the 5-HT_1AR and the D₂R, ralmitaront lacked demonstrable activity. To extrapolate from these in vitro findings to the observed differences in clinical efficacy between ulotaront and ralmitaront, it would have been helpful to know the central nervous system receptor occupancies achieved with these agents in the respective trials. However, although well-established positron emission tomography tracers are available for both 5-HT_1AR and D₂R, a suitable positron emission tomography ligand for TAAR1 has yet to be developed (Sun et al., 2021). At present, we can only speculate as to whether the in vitro differences between ulotaront and ralmitaront may translate into differential clinical efficacy of these compounds.

Author Contributions

Richard Ågren (Conceptualization [Lead], Data curation [Lead], Formal analysis [Lead], Funding acquisition [Equal], Investigation [Lead], Writing—original draft [Equal]), Nibal Betari (Data curation [Equal], Formal analysis [Equal], Investigation [Equal], Writing—original draft [Equal]), Marcus Saarinen (Conceptualization [Equal], Data curation [Equal], Formal analysis [Equal], Funding acquisition [Equal], Investigation [Equal], Writing—original draft [Equal]), Hugo Zeberg (Formal analysis [Equal], Investigation [Equal], Writing—original draft [Equal]), Per Svenningsson (Conceptualization [Equal], Funding acquisition [Equal], Project administration [Equal], Resources [Equal], Supervision [Equal], Writing—original draft [Equal]), and Kristoffer Sahlholm (Conceptualization [Lead], Data curation [Equal], Formal analysis [Equal], Funding acquisition [Lead], Project administration [Lead], Resources [Equal], Supervision [Lead], Writing—original draft [Lead])

Interest Statement

None.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.