How Mescaline Derivatives and Related Amphetamines Interact with Brain Receptors

Full names and abbreviations of the test compounds are provided in Supplementary Table S1. The 3,5-dimethoxy-4-substituted phenethylamines (mescaline [5], MDFM [11], DFM [12], TFM [13], FE [16], DFE [17], TFE [18], IP [22], DFIP [23], FP [25], TFP [26], CP [29], MAL [32], BZ [33]) and the 3,5-dimethoxy-4-substituted amphetamines (TMA [6], 3C-DFM [14], 3C-DFE [19], 3C-FE [20], 3C-E [21], 3C-FP [27], 3C-P [28], 3C-AL [31]) were synthesized as racemates as previously described (Shulgin and Shulgin 1991; Trachsel 2002; Trachsel 2003; Trachsel et al., 2013), and provided as hydrochloride salts for pharmacological testing by ReseaChem (Burgdorf, Switzerland). Purity of all substances was >98%. [³H]serotonin (80.0 Ci/mmol) was purchased from Anawa (Zurich, Switzerland). [³H]dopamine (30.0 Ci/mmol) and [³H]norepinephrine (13.1 Ci/mmol) were obtained from Perkin-Elmer (Schwerzenbach, Switzerland).

Radioligand binding affinity (K_i) for monoamine receptors and transporters was assessed according to previously described methods (Luethi et al., 2018b). In short, different cell line derived membrane preparations overexpressing respective monoamine receptors (human genes with the exception of rat and mouse TAAR1) or transporter were briefly incubated with corresponding radiolabeled selective ligands at a concentration equal to the dissociation constant K_d. The cell membrane preparations were obtained from Chinese hamster ovary cells (for hα_1A adrenergic receptor), Chinese hamster lung cells (for hα_2A adrenergic receptor) and HEK 293 cells (for h5-HT_1A, h5-HT_2A, h5-HT_2C, and hD₂ receptors, TAAR1, and hNET, hDAT, and hSERT). The specific binding of radioligand to the target site was defined by measuring the difference between total binding and nonspecific binding (calculated in the presence of the respective receptor competitor in excess). This was used to measure the ligand displacement by the substances under investigation.

The following radioligands and their respective competitors were used: 0.90 nM [³H]8-hydroxy-2-(dipropylamine)tetralin (8-OH-DPAT) and 10 μM pindolol (h5-HT_1A receptor), 0.40 nM [³H]ketanserin and 10 μM spiperone (h5-HT_2A receptor), 1.4 nM [³H]mesulergine and 10 μM mianserin (h5-HT_2C receptor), 3.5 nM or 2.4 nM (rat or mouse isoform, respectively) [³H]RO5166017 and 10 μM RO5166017 (TAAR1), 0.11 nM [³H]prazosin and 10 μM chlorpromazine (hα_1A adrenergic receptor), 2 nM [³H]rauwolscine and 10 μM phentolamine (hα₂ adrenergic receptor), 1.2 nM [³H]spiperone and 10 μM spiperone (dopaminergic hD₂ receptor), 2.9 nM N-methyl-[³H]nisoxetine and 10 μM indatraline (hNET), 1.5 nM [³H]citalopram and 10 μM indatraline (hSERT), 3.3 nM [³H]WIN35,428 and 10 μM indatraline (hDAT).

To assess the functional activity at the serotonin 5-HT_2A receptor, mouse embryonic fibroblasts (NIH-3T3 cells) expressing human 5-HT_2A receptor were seeded at a density of 70,000 cells per 100 μl in poly-D-lysine-coated 96-well plates according to methods previously described by Luethi, Trachsel et al. (2018). In brief, the NIH-3T3 cells were incubated in HEPES-Hank’s Balanced Salt Solution (HBSS) buffer (Gibco) for 1 h at 37°C. Subsequently, the plates were incubated with dye solution (100 μl/well) for 1 h at 37°C (fluorescence imaging plate reader [FLIPR] calcium 5 assay kit; Molecular Devices, Sunnyvale, CA, United States). Twenty-five microliter of test drugs diluted in HEPES-HBSS buffer composed of 250 mM probenecid were added to the plate online. Using nonlinear regression, the rise in fluorescence was measured and EC₅₀ values were calculated from the concentration-response curves. The efficacy was calculated relative to 5-HT activity, which was defined as 100%.

To assess the functional activity at the serotonin 5-HT_2B receptors, HEK 293 cells expressing the human 5-HT_2B receptor were seeded at a density of 50,000 cells per well in 96-well poly-D-lysine-coated plates overnight at 37°C, according to methods previously described by Luethi, Trachsel et al. (2018). In brief, the HEK 293 cells were incubated overnight at 37°C in high glucose Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Zug, Switzerland), 10% fetal calf serum (non-dialyzed, heat-inactivated), 250 mg/L Geneticin and 10 ml/L PenStrep (Gibco). Using snap inversion, the growth medium was removed and 100 μl of calcium indicator Fluo-4 solution (Molecular Probes, Eugene, OR, United States) was added to each well for an incubation time of 45 min at 31°C. Thereafter, the Fluo-4 solution was removed (snap inversion) and subsequently an additional 100 μl of the Fluo-4 solution was added (incubation of 45 min at 31°C). Next, using the EMBLA cell washer, the cells were washed just before testing with HBSS and 20 mM HEPES and exposed to 100 μl of assay buffer. The plate was placed inside the FLIPR and 25 μl of test drugs diluted in assay buffer were added to the plate online. Using nonlinear regression, the rise in fluorescence was measured and EC₅₀ values were calculated from the concentration-response curves. The efficacy was calculated relative to 5-HT activity, which was defined as 100%.

To assess the functional activity at the human TAAR1, HEK 293 cells expressing recombinant human TAAR1 were grown in 250 ml falcon culture flasks containing 30 ml of high glucose DMEM [10% heat inactivated fetal calf serum, 500 μg/ml Geneticin (Gibco, Zug, Switzerland) and 500 μg/ml hygromycin B] at 37°C and 5% CO₂/95% air, according to methods previously described by Luethi, Trachsel et al. (2018). Once 80–90% confluency was reached, the cells were collected by removing the medium, washing with PBS and then adding 5 ml of trypsin/EDTA solution for 5 min at 37°C. Next, 45 ml of medium was added and the entire mixture was transferred into a falcon tube. The tube was then centrifuged at room temperature for 3 min at 900 revolutions per minute (rpm). Next, the supernatant was removed in order to resuspend the remaining cell pellet in fresh medium to 5 × 10⁵ cells/ml. Hundred microliter of the cells was transferred into a 96-well plate (80,000 cells/well; BIOCOAT 6640, Becton Dickinson, Allschwil, Switzerland) and incubated for 20 h at 37°C.

For the cAMP assay, the aspirated medium was replaced with 50 μl PBS without Ca²⁺ and Mg²⁺ ions. Using snap inversion, the PBS was extracted and the plate was gently tapped against tissue. Next, 90 μl of Krebs-Ringer Bicarbonate buffer (KRB, Sigma-Aldrich) containing 1 mM IBMX was added and incubated for 60 min at 37°C and 5% CO₂/95% air. Each test compound was examined in duplicate in a concentration range between 300 pM and 30 μM. A standard curve with a range of cAMP concentrations (0.13 nM–10 μM) was created for each 96-well plate. Each experiment was accompanied with a reference plate that included RO5256390, β-phenylethylamine and p-tyramine. The cells were exposed to either 30 μl of compound solution, 30 μl of β-phenylethylamine (as maximal response), or a basal control in PBS (containing 1 mM IBMX) for 40 min at 37°C. Next, under forceful shaking using black lids, the cells were exposed to 50 μl of 3× detection mix solution (composed of Ru-cAMP Alexa700 anti-cAMP antibody and lysis buffer) for 120 min at room temperature. Using the NanoScan reader (Innovate Optische Messtechnik, Berlin, Germany; 456 nm excitation wavelength; 630 and 700 nm emission wavelengths), the fluorescence was examined and the FRET signal was determined using the following equation; FRET (700 nm)- P × FRET (630 nm), where P = Ru (700 nm)/Ru (630 nm).

To exclude activity of the scalines and 3C-scalines at monoamine transporters at pharmacologically relevant concentrations, a single high drug concentration was examined using HEK 293 cells stably transfected with the human serotonin, norepinephrine or dopamine transporters (hSERT, hNET, or hDAT, respectively) as previously described (Luethi et al., 2018a). In summary, the cells were cultured in DMEM (Gibco, Life Technologies, Zug, Switzerland) containing both 250 μg/ml Geneticin (Gibco) and 10% fetal bovine serum (Gibco). Once the cells were confluent (70–90%) they were detached and resuspended in KRB (Sigma-Aldrich, Buchs, Switzerland) at a density of 3 × 10⁶ cells/ml of buffer. For [³H]dopamine uptake experiments, the buffer additionally contained 0.2 mg/ml ascorbic acid. Hundred microliter of cell suspension per well was added to a round bottom 96-well plate. The cells were then incubated with 25 μl buffer containing the test drug (10 μM), vehicle control (0.1% dimethyl sulfoxide), or transporter-specific inhibitors [10 μM fluoxetine (SERT), 10 μM mazindol (DAT) or 10 μM nisoxetine (NET)] for 10 min by shaking on a rotary shaker (450 rpm) at room temperature. Uptake transport was initiated by adding [³H]serotonin, [³H]dopamine, or [³H]norepinephrine at a final concentration of 5 nM to the mixture. After 10 min, uptake transport was halted by the transfer of 100 μl of the cell mixture to 500 μl microcentrifuge tubes containing 50 μl of 3 M KOH and 200 μl silicon oil (1:1 mixture of silicon oil types AR 20 and 200; Sigma-Aldrich). The tubes were centrifuged for 3 min at 13,200 rpm, to allow the transport of the cells through the silicon oil layer into the KOH layer. The tubes were frozen in liquid nitrogen and the cell pellet was cut into 6 ml scintillation vials (Perkin-Elmer) containing 0.5 ml lysis buffer (1% NP-40, 50 mM NaCl, 0.05 M TRIS-HCl, 5 mM EDTA and deionized water). The samples were shaken for 1 h before 3 ml of scintillation fluid (Ultima Gold, Perkin Elmer, Schwerzenbach, Switzerland) was added. Monoamine uptake was then quantified by liquid scintillation counting on a Packard Tri-Carb Liquid Scintillation Counter 1900 TR. Nonspecific uptake in the presence of selective inhibitors was subtracted from the total counts.

All calculations and analyses were performed using Prism 7.0a (GraphPad, San Diego, CA, United States). IC₅₀ values of the radioligand binding were determined by calculating nonlinear regression curves for a one-site model using at least three independent 10-point concentration-response curves for each substance. The K_i values correspond to the dissociative constant for the inhibitor and were calculated using the Cheng-Prusoff equation. Nonlinear regression concentration-response curves were used to determine EC₅₀ values for 5-HT_2A and 5-HT_2B receptor activation. Maximal activation activity (efficacy) is expressed relative to the activity of 5-HT, which was set to 100%. Monoamine uptake of four independent experiments was compared to control using 1-way ANOVA analysis of variance followed by a Dunett’s multiple-comparison test. Monoamine uptake of MDMA was included as comparison. Receptor affinity binding (K_i) < 50 nM was defined as high affinity binding, K_i < 500 nM as moderate affinity binding, while K_i > 1,000 nM was defined as low affinity binding. Activation efficacy (max %) < 85% was defined as partial agonism while max % > 85% was defined as full agonism.

Monoamine receptor and transporter binding affinities are listed in Table 2. None of the examined compounds activated the human TAAR1 (EC₅₀ > 10,000 nM). At the rat TAAR1, most phenethylamine derivatives bound within a micromolar range (K_i = 1,000–3,000 nM) with the exception of DFM (12), TFM (13), TFP (26), and BZ (33), which bound at submicromolar concentrations (K_i = 110–910 nM). FE (16), IP (22), and DFIP (23) did not bind to the rat TAAR1 at the concentrations tested (K_i > 4,000 nM). The amphetamine derivative 3C-DFM (14) bound with a K_i of 380 nM to the rat TAAR1. TMA (6), 3C-DFE (19), 3C-P (28), and 3C-AL (31) bound in the micromolar range (K_i = 3,200–3,900 nM); for 3C-FE (20), 3C-E (21), and 3C-FP (27) no binding was observed at the rat TAAR1 at examined concentrations (K_i > 4,700 nM).

At the mouse TAAR1, the phenethylamine derivatives MDFM (11), TFM (13), MAL (32), and BZ (33) bound in the micromolar range (K_i = 1,900–3,900 nM) while none of the remaining derivatives bound to the receptor (K_i > 4,200 nM). The amphetamine derivatives TMA (6), 3C-DFM (14), 3C-DFE (19), 3C-P (28), and 3C-AL (31) bound in the micromolar concentration range to the mouse TAAR1 (K_i = 1,000–3,300 nM); 3C-DFE (19), 3C-FE (20), and 3C-E (21) did not bind at examined concentrations (K_i > 4,200 nM).

At the adrenergic α_1A receptor, the phenethylamine derivatives MDFM (11), DFM (12), and TFM (13) were the only derivatives showing any affinities at tested concentrations (K_i = 3,200–8,000 nM). At the α_2A receptor, all phenethylamine derivatives bound in the micromolar range (K_i = 1,200–3,700 nM) except for TFM (13), which bound with moderate affinity (K_i = 450 nM). The only amphetamine derivatives that bound to the α_2A receptors were TMA (6), 3C-DFM (14), and 3C-P (28) (K_i = 2,600–4,600 nM).

None of the compounds examined bound to the dopaminergic D₂ receptor (K_i > 6,300 nM) or any of the monoamine transporters (K_i > 7,500 nM). Furthermore, none of the investigated compounds significantly inhibited any of the monoamine uptake transporters (IC₅₀ > 10,000 nM).

Taken from the extensive SAR of 2,4,5-trisubstituted derivatives, small lipophilic substituents at the 4-position of 2,5-dimethoxy substituted phenethylamines and amphetamines lead to derivatives that have agonistic properties, while derivatives with large lipophilic substituents at the 4-position lead to antagonistic effects at the 5-HT_2A/2C receptors (Dowd et al., 2000). Furthermore, hydrophilic substituents at the 4-position attenuate 5-HT_2A receptor affinity and in vivo potency (Nelson et al., 1999; Braden 2007). In line with functional properties of ligands with lipophilic 4-substituents, a similar trend could be observed when reviewing the active doses of these compounds as psychedelics in man; when surpassing a certain steric bulkiness, compounds tend to lose their psychedelic properties (Shulgin and Shulgin 1991; Trachsel et al., 2013). Thus, smaller lipophilic 4-substituents not only yield agonists/partial agonists but also lead to the most potent psychedelics.

Thus far, the few in vitro investigated 3,4,5-trisubstituted phenethylamines and amphetamines have been shown to have the lowest 5-HT_2A receptor affinities among psychedelic phenethylamines (Monte et al., 1997; Parker et al., 2008; Trachsel et al., 2013) when compared to the 2,4,5-trisubsituted and 2,4,6-trisubsituted phenethylamines. However, initial SAR investigations of a series of 4-alkoxysubstituted 3,5-dimethoxyphenethylamines and their α-methyl congeners revealed a similar trend in that more lipophilic 4-substituents lead to higher affinities (Parker et al., 2008; Trachsel et al., 2013). Similarly, 3,5-dimethoxy derivatives with more lipophilic 4-substituents also lead to more potent compounds in man, when not surpassing a certain steric bulkiness (Shulgin and Shulgin 1991; Trachsel et al., 2013).

The derivatives with high 5-HT_2A receptor affinities (K_i < 1,000 nM), such as TFM (13), MAL (32), and BZ (33), also displayed high activation potency (EC₅₀ in the range of 27–280 nM). Structures 13, and 33 were found to be partial agonists (efficacy < 85%) and 32 had an activation efficacy of 85%, suggesting full agonist properties. In accordance to these in vitro findings, potent psychedelic effects have been described for TFM (13) and MAL (32) (Shulgin and Shulgin 1991; Trachsel et al., 2013), suggesting 33 to be potentially psychedelic in humans. The remaining substances were less potent partial to full 5-HT_2A agonists. However, for various substances a discrepancy between binding and activation was observed (i.e., activation potency was distinctively higher than affinity). It has previously been described that unlike receptor binding values, activation potency assessed with a Ca²⁺ mobilization assay does not necessarily correlate with the potency of the drug (Luethi and Liechti 2018). Functional assays based on other signaling events, for instance IP formation or β-arrestin recruitment, might better predict the clinical potency of scalines and 3C-scalines. In addition to 5-HT_2A/2C receptor activity, the head twitch response is an established method to predict the activity and potency of psychedelics (Halberstadt et al., 2011; Halberstadt and Geyer 2014; Halberstadt et al., 2019; Halberstadt et al., 2020). Halberstadt et al. (2019) recently showed that 3C-E (21) and 3C-P (28) induced a head twitch response with almost identical potency. Thus, 28 may induce psychedelic effects in humans at similar doses as 21 (Halberstadt et al., 2019).

Among all tested substances, the only potent partial 5-HT_2B agonists were the phenethylamine derivatives MDFM (11), TFM (13), and TFE (18) (EC₅₀ of 88–210 nM), and the amphetamine derivatives 3C-DFM (14), 3C-DFE (19), 3C-FE (20), and 3C-E (21) (95–800 nM). However, these substances were low efficacy partial agonists (EC₅₀ = 18–45%). Endocardial fibrosis has been associated with 5-HT_2B activation and is therefore a potential adverse effect to consider for chronic use of substances interacting with this receptor (Rothman et al., 2000; Droogmans et al., 2007; Roth 2007; Doly et al., 2008; Elangbam et al., 2008; Bhattacharyya et al., 2009; Huang et al., 2009; Elangbam 2010; Dawson and Moffatt 2012). As psychedelics are typically not used chronically, endocardial fibrosis is an unlikely adverse effect for users of such substances despite a potential interaction with the 5-HT_2B receptor subtype (Luethi et al., 2021).

None of the investigated phenethylamine and amphetamine derivatives interacted with the human TAAR1, the D₂ receptor, or monoamine uptake transporters. It is unclear, however, whether co-expression of different receptors would alter a substance’s response at these targets. Still, some derivatives bound to the rat TAAR1 with moderate to high affinity and some substances additionally showed low affinity at the mouse TAAR1. These results confirm the previously observed TAAR1 affinity rank order (rat > mouse > human TAAR1) (Wainscott et al., 2007; Lewin et al., 2008; Rickli et al., 2015; Simmler et al., 2016; Luethi et al., 2018b; Kolaczynska et al., 2019; Luethi et al., 2019). TAAR1 has been shown to negatively modulate monoaminergic neurotransmission (Lindemann et al., 2008; Revel et al., 2011) but the lack of human TAAR1 activation calls into question the relevance of TAAR1 in the mechanism of action of scalines and 3C-scalines. All phenethylamines moderately to weakly interacted with the α_2A receptor (K_i = 450–3,700 nM) but only MDFM (11) and TFM (13) bound to the α_1A receptor (K_i = 3,200–4,300 nM). Among the amphetamines, only TMA (6), 3C-DFM (14), and 3C-P (28) bound to the α_2A receptor (K_i = 2,600–4,600 nM) whereas no binding to the α_1A receptor was observed. This is in line with a previously reported higher α_2A vs α_1A receptor selectivity observed for psychedelic 2,4,5-substituted phenethylamines (2C derivatives) (Rickli et al., 2015; Luethi et al., 2018b; Kolaczynska et al., 2019; Luethi et al., 2019). As observed for the 5-HT_2A and 5-HT_2C receptors, binding affinity at non-serotonergic receptors increased proportionally to the number of fluorine atoms (e.g., affinities at rTAAR1: mescaline K_i = 3,000 nM > DFM K_i = 880 nM > TFM K_i = 170 nM; DFE K_i = 2,100 nM > TFE K_i = 1,200 nM; FP K_i = 1,700 nM > TFP K_i = 910 nM).