How the length of a chemical chain in 2,5-dimethoxyamphetamines changes their effects on serotonin receptors in the lab versus psychedelic-like effects in living animals

Many derivatives of phenethylamine are potent 5-hydroxytryptamine 2 (5-HT₂) receptor ligands with therapeutic potential and usefulness in serotonergic systems research [1–3]. In addition, a variety of substituted phenethylamines have appeared on the recreational drug market due to their reported psychedelic and euphoric properties [4, 5]. Even though the risk of physical harm from use of phenethylamine psychedelics is low [4–7], overdose can result in a variety of adverse effects [4, 5, 8].

Psychedelic phenethylamines, and other classes of psychedelics such as lysergamides and tryptamines, act as partial to full agonists at the 5-HT_2A receptor subtype [9–16] and mediate 5-HT_2A-dependent behaviors in vivo, such as the head-twitch response (HTR) in mice [17–25]. A recent study suggests that the HTR in mice is related to 5-HT_2A-G_q efficacy of a given drug, whereas β-arrestin 2-biased agonists block psychedelic-like effects and can induce receptor downregulation [26]. Based on the aforementioned study, it has been postulated that a certain threshold level of 5-HT_2A-G_q activity is required to achieve psychedelic-like effects, and low efficacy agonists do not elicit HTRs [26]. In addition to 5-HT_2A agonism, most phenethylamine psychedelics are agonists at 5-HT_2B and 5-HT_2C receptors [2]. Agonist actions at 5-HT_2B receptors may pose risk for cardiac mitogenesis and valvulopathy when psychedelics are taken chronically, such as repeated microdosing [27], but more clinical studies are needed to address this issue. Agonist actions at 5-HT_2C receptors can functionally oppose expression of 5-HT_2A-mediated behavior after high drug doses, which contributes to the biphasic dose–response curves exhibited for HTR and locomotion in mice [2, 28–30]. On the other hand, some evidence indicates that 5-HT_2C activity is required for psychedelic-like effects, including HTR [30, 31].

Many 5-HT_2A agonists also interact with 5-HT_1A receptors, and this activity modulates pharmacological effects of the drugs. For example, 5-HT_1A receptors are colocalized with 5-HT_2A receptors on cortical pyramidal neurons, where the receptor types exert opposing effects [2, 32, 33]. Various tryptamine psychedelics have agonist activity at 5-HT_1A receptors that can dampen their psychedelic-like effects in rodents and humans [24, 34–41]. In general, phenethylamine psychedelics are more selective for 5-HT_2A vs. 5-HT_1A receptors [4, 36, 42, 43], though their psychedelic-like effects can still be modulated by 5-HT_1A agonism in some cases [43, 44].

Interestingly, 2,5-DMA is devoid of robust psychoactive effects [48], while its more potent 4-methyl derivative DOM was a popular recreational drug in the late 1960s under the street name “STP” [48, 51]. Among the 4-alkylated derivatives of 2,5-DMA, the reported psychoactive effects are optimal for two- and three-carbon chain lengths (DOET and DOPR, respectively), with decreased effects for four-carbon and five-carbon chain lengths (DOBU and DOAM, respectively) [52]. Enantioselective activity has been reported for 2,5-DMA derivatives, with the R isomers exhibiting higher activity in animals and humans [43, 48]. Here, we aimed to further study the structure–activity relationships of 4-alkylated 2,5-DMA derivatives to better understand the relationship between receptor interactions in vitro and psychedelic-like behavioral effects in vivo. Specifically, we studied three signaling cascades linked to 5-HT_2A receptor activation as well as interactions with other serotonergic and non-serotonergic targets. Furthermore, we assessed the HTR in mice as a behavioral proxy for psychedelic-like effects and quantified drug concentrations in post-mortem plasma and brain tissue. To assess extracellular drug concentrations in the brain, we conducted in vivo microdialysis experiments.

In this study, 4-alkylated 2,5-DMA derivatives bound with highest affinity to 5-HT_2A receptors and induced partial to full agonist activities at Gα_q and Gα_i/o,12/13-mediated signaling as well as β-arrestin 2 recruitment at this receptor. Additionally, the substances displayed high affinity and full agonist properties at 5-HT_2B and 5-HT_2C receptors. Compared to the potent activity at 5-HT₂ receptors, interactions at other investigated targets were overall mostly minor. Consistent with high activity at 5-HT_2A receptors, 2,5-DMA derivatives induced dose-dependent psychedelic-like effects in mice. More specifically, the substances induced the HTR, and compounds with short to intermediate 4-alkyl chain lengths (i.e., DOM and DOPR) were the most potent and efficacious. 2,5-DMA competed for in vitro binding to 5-HT_1A and 5-HT₂ receptors with low micromolar affinity and activated 5-HT₂ receptors as a full agonist. By comparison, 4-methylation increased the affinity and activity at 5-HT₂ but not 5-HT_1A receptors. Gradually increasing the length of the 4-alkyl chain enhanced binding affinity at 5-HT_2A receptors across the series. Similarly, it has previously been shown that agonist activity of serotonergic compounds in sheep umbilical artery strips was enhanced until a three-carbon chain length and then decreased with further extensions [63]. Here, we found that the same rank order applies to all functional assays. The calculated bias factors indicated that some of the compounds were biased towards Gα_q or Gα_i/o,12/13-mediated signaling relative to 5-HT, whereas no β-arrestin 2 bias was observed. Notably, the rank order of potency determined in vitro did not match potencies to induce the HTR, where the compounds with the longest 4-alkyl chains (i.e., DOAM, DOBU) showed reduced potencies and efficacies relative to short or intermediate ones (i.e., DOM, DOPR).

Despite higher affinity of DOBU and DOAM at the 5-HT_2A receptor relative to the other compounds tested, these compounds had reduced HTR activity vs. compounds with similar (i.e., DOPR) or reduced affinity (i.e., DOM) at the 5-HT_2A receptor. High 5-HT_2A receptor affinity with diminished HTR has been reported for 2,5-dimethoxyphenylalkylamines with a bulky 4-thio substituent [22]. Similarly, elongation of the 4-alkyl chain seems to augment antagonistic properties of 2,5-dimethoxyamphetamines by increasing 5-HT_2A receptor affinity while reducing receptor activation potency. Clinical potency of psychedelics correlates relatively well with affinity at 5-HT_2A and 5-HT_2C receptors [42] as well as 5-HT_2A-mediated β-arrestin 2 recruitment [64]; however, the HTR has also been reported to correlate with G_q efficacy but not β-arrestin 2 recruitment at 5-HT_2A receptors [26]. Assay specific conditions (i.e., signaling event measured, time of measurement, temperature conditions) may therefore be important for predicting some relationships between the in vitro receptor pharmacology and in vivo behavioral readouts of psychedelics, given the known influence of these factors in cellular assays measuring 5-HT_2A functional activities [25, 65]. Moreover, weak 5-HT_2A receptor efficacy could lead to a lack of significant psychedelic effects, as observed for various recently described psychedelic analogs [25, 26, 54, 66]. All 2,5-dimethoxyamphetamines assessed in the present study displayed G_q efficacies of ≥ 83% and none of the substances was β-arrestin 2-biased. Our study demonstrates that neither 5-HT_2A receptor binding nor activation of different G protein-coupled receptor signaling pathways alone is sufficient to completely predict the in vivo potency within a small set of chemically similar psychedelics.

Due to the high structural similarity of the tested substances, ligand docking studies did not reveal any clear SAR trends. Besides the amine group, there are no functional groups that could engage in differential interactions with the binding pocket. It is well established that Asp155 forms a salt bridge with the protonated amine of various psychedelics, including phenethylamines [67–70]. Previous molecular docking studies demonstrated that a hydrophobic region between transmembrane helices IV and V can accommodate bulky 4-substituents, such as a propylthio group, on psychedelic phenethylamines [71]. This suggests that the same region may serve as a binding pocket for the 4-alkyl substituents of the 2,5-DMA derivatives examined in our study. In addition to this hydrophobic interaction, it is plausible that the aromatic ring system of these compounds engages in π–π stacking with aromatic residues in the binding pocket. Several such residues are implicated in ligand binding, including Trp151, Trp336, and Phe339. Trp151 is proposed to participate in π-interactions with ergoline-based psychedelics, although evidence for its involvement with phenethylamines is lacking [70]. Trp336, by contrast, is essential for ligand binding and receptor activation [68]. The psychedelic phenethylamine 25CN-NBOH engages with Trp336, thereby modulating the transition from the inactive to active receptor state [69, 70]. Phe339 is known to form π-cation interactions with the phenethylamine mescaline [70]; another phenylalanine residue, Phe340, has also been implicated in ligand binding for this class of compounds [72]. Interactions involving the 2,5-dimethoxy groups of phenethylamines remain poorly characterized. A hydrophobic contact between the 3’-methoxy group of mescaline and Leu229 on extracellular loop 2 has been reported, as well as a potential electrostatic interaction with Ser242 on helix V [70]. However, mutation of Ser242 to alanine does not affect DOM binding affinity at the 5-HT_2A receptor, suggesting that this interaction is not critical for 4-alkylated phenethylamines [73].

When compared to drug affinity at 5-HT_2A receptors, the affinity at the 5-HT_2B receptor was similar (< 2-fold difference) for all 4-alkylated derivatives. Each compound activated the 5-HT_2B receptor as a full agonist, with slightly (~ 3-fold) increased activation potency observed for DOET and DOPR compared to DOM, DOBU, and DOAM. All compounds activated the 5-HT_2C receptor as full agonists. Activation potency was highest for derivatives with a two-carbon chain or longer. However, no clear pattern between chain length and activation was evident. Compared to 5-HT_2C receptor affinity, activation potencies were around two orders of magnitude higher, which may, at least in part, be explained by differences in the used cell lines. Overall, the 5-HT_2A vs. 5-HT_2C and 5-HT_2B receptor selectivity increased with increasing lipophilicity of the 4-substituent. Of all tested 4-alkylated 2,5-dimethoxyamphetamines, DOM least potently interacted with 5-HT_2C receptors and induced the highest amount of HTR counts. This is in line with the notion that 5-HT_2C receptor activation opposes HTR [28]. Still, the fact that 5-HT_2C receptor binding and activation of DOET and DOAM were similar in the present study, suggests that the decreased potencies and efficacies of derivatives with longer 4-alkyl chain is not solely explained by differential 5-HT_2C receptor interactions. In contrast to the 5-HT₂ receptor subtypes, 4-alkylation did not substantially affect 5-HT_1A binding. This finding is consistent with previous studies reporting a higher 5-HT_2A vs. 5-HT_1A selectivity for psychedelic phenethylamines compared to tryptamines or LSD [2, 9, 15, 74–78]. Differential affinities at 5-HT_1A receptors may account for some of the qualitative differences between phenethylamine and tryptamine psychedelics.

Given that little is known about the pharmacological interactions of 4-alkylated 2,5-dimethoxyamphetamines with non-5-HT receptors, we examined interactions at other monoaminergic targets and TAAR1. Compared to the potent interactions with serotonin 5-HT₂ receptors, the affinities at most other monoaminergic targets were rather weak (i.e., > 1 µM). Similarly, DOPR, DOBU, and DOAM activated TAAR1 at high concentrations only (EC₅₀ = 10–22 µM), whereas the other substances did not activate the receptor at the investigated concentration range (EC₅₀ > 30 µM). TAAR1 activation has been demonstrated to mediate anti-psychotic and cognition-enhancing activities of other compounds in vivo [79]. Furthermore, TAAR1 antagonism blocks the inhibitory effect of LSD on dopaminergic neurons in the ventral tegmental area [80] and reduces 5-hydroxytryptophan but not psilocybin-induced HTR in mice [81]. Still, there is currently no evidence that weak TAAR1 activity, as in case of DOPR, DOBU, and DOAM, contributes to functional differences in vivo.

DOM and DOPR induced the highest number of twitches (119 and 98, respectively), which gradually decreased with increasing length of the alkyl chain. The HTR induced by DOET was not examined in this study; however, Halberstadt and colleagues previously reported an ED₅₀ = 0.20 mg/kg and a maximum of ~122 twitches for DOET [20]. Results from the aforementioned study found the potency and maximal HTR count for DOM to be similar to what we observed here, despite differences in the salt form of the drug used (freebase presently vs. hydrochloride salt) and routes of administration (s.c. presently vs. i.p.). Further, the potency for DOBU (ED₅₀ = 1.17 mg/kg) from the Halberstadt et al. study [20] agrees with the present results for the drug (ED₅₀ = 1.29 mg/kg) despite differences in route of administration.

Pharmacokinetic factors (i.e., absorption, distribution, metabolism, elimination) may influence the in vivo potency of a psychedelic substance [82, 83]. The summed HTR findings suggest that there may be pharmacokinetic differences across compounds with longer 4-alkyl chain lengths and therefore higher lipophilicity. Importantly, we found that brain and plasma drug concentrations varied after administration of HTR-active doses of the compounds. As an example, we found lower brain concentrations of DOBU vs. DOM when administered at the same 3 mg/kg dose, suggesting a higher volume of distribution of the more lipophilic DOBU or potentially a faster clearance. Notably, 10 mg/kg of DOAM was required to reach brain and plasma concentrations comparable to DOM given at 3 mg/kg. At equal doses, the brain vs. plasma ratio was increased for more lipophilic substances (DOBU vs. DOM, DOAM vs. 2,5-DMA); this aligns with the understanding that lipophilic compounds more readily accumulate in the brain. Nevertheless, compounds with LogP values in the range of 1.5–2.7 are found to display optimal blood–brain barrier penetration [84], whereas excessively high lipophilicity may reduce the amount of drug reaching the receptor site. 2,5-DMA has a LogP value of 1.7 when computed by XLogP3 [85]. 4-Alkylation increases the LogP value as follows: 2.2 (DOM), 2.8 (DOET), 3.4 (DOPR), 4.0 (DOBU), and 4.4 (DOAM). Elongation of the 4-alkyl chain furthermore increases the number of rotatable bonds from four in DOM to eight in DOAM and a high number of rotatable bonds is associated with decreased oral bioavailability [86, 87]. Furthermore, centrally acting substances have generally fewer rotatable bonds, with most CNS drugs having five or fewer rotatable bonds [88]. Compared to their α-desmethyl analogs, 4-substituted 2,5-DMA analogs are metabolically more stable, resulting in long durations of action in vivo [46, 48]. HTR data support this notion, showing stable psychedelic-like effects across the testing session at maximally effective doses. Regardless of stable HTR rate across the test sessions, the drug levels in brain tissue homogenate were not related to HTR activity. This suggests that other factors, such as free drug concentrations at the receptor site, play a more prominent role in pharmacodynamic differences observed for psychedelic-like effects.

To further investigate pharmacokinetic parameters relevant to receptor site exposure, we conducted a microdialysis experiment following the combined systemic administration of low doses of DOM and DOAM. The substances were co-administered at equal doses (0.5 mg/kg each) to minimize inter-animal variability and allow for a direct comparison of extracellular brain concentrations under identical physiological conditions. Consistent with the plasma and brain bioanalysis at peak HTR doses, matched 0.5 mg/kg doses produced higher levels of DOM vs. DOAM. Strikingly, the DOAM concentration in dialysate was below the detection limit of 0.04 ng/mL, indicating that dialysate differences are even more substantial than the differences observed in plasma and brain tissue. It is important to note that the microdialysis experiments were conducted using doses below those required to elicit peak HTR counts; nevertheless, the findings indicate that, as with plasma and brain tissue, higher DOAM doses are required to achieve dialysate concentrations comparable to those of DOM. The observed disconnect between in vitro and in vivo findings demonstrates that early pharmacological screening of new psychedelic drug candidates for therapeutic use should not rely solely on in vitro target and off-target interactions, but should include thorough in vivo pharmacokinetic and pharmacodynamic assessments to determine whether the drugs reach their molecular targets.

Combining data from studies in humans and animals reveals that 2,5-dimethoxyamphetamines substituted at the 4-position with a short or intermediate unbranched alkyl (i.e., methyl, ethyl, or propyl) chain are potent psychedelics. The increase in potency for 4-alkylated derivatives compared to 2,5-DMA, which lacks a 4-substituent, is a result of substantially increased affinity and activity at the 5-HT_2A receptor. However, elongation of the 4-alkyl chain to a butyl or amyl moiety reduces the potency of the respective derivatives in vivo. The observed decrease in psychedelic-like activity of the compounds with the longest 4-alkyl chain lengths does not seem related to reduced activity or efficacy at 5-HT_2A receptors or altered interactions with other molecular targets. Here, we show that compounds with the longest 4-alkyl chains required higher doses to reach comparable plasma, brain tissue, and brain dialysate levels. Overall, we propose that elongation of the 4-alkyl chain beyond ethyl or propyl reduces psychedelic effects due to pharmacokinetic factors, including poor blood–brain barrier penetration and reduced central activity, compared to short-chain analogs.

Supplementary information