Reinforcing effects of phenethylamine analogs found in dietary supplements

Abstract

Rationale:

Synthetic phenethylamine (PEA) analogs, such as β-methylphenethylamine (BMPEA) and N,α-diethylphenethylamine (DEPEA), are often found in dietary supplements, despite regulations prohibiting their sale. PEA analogs are structurally related to amphetamine, and we have shown that BMPEA and DEPEA produce cardiovascular stimulation mimicking the effects of amphetamine. However, few studies have examined behavioral effects of BMPEA, DEPEA, and other PEA analogs.

Objectives:

Here, we examined the reinforcing effects of α-ethylphenethylamine (AEPEA, 1 mg/kg/injection), DEPEA (1 mg/kg/injection), and BMPEA (3 mg/kg/injection) as compared to amphetamine (0.1 mg/kg/injection) using a fixed-ratio 1 self-administration paradigm in male rats.

Methods:

Male rats were trained in self-administration chambers containing 2 nose-poke holes. A nose-poke response in the active hole delivered drug or saline, whereas a nose-poke response in the inactive hole had no programmed consequence. Four groups of rats were initially trained for 10 days with the doses noted above. Upon acquisition of drug self-administration, a dose-effect function was determined by training rats on 3 additional doses for 3 days each. A separate group of rats was trained with saline.

Results:

Male rats self-administered each PEA analog and amphetamine, as shown by significant increases in active responses versus inactive responses. Subsequent dose-response testing showed clear differences in potency of the compounds. Amphetamine showed a typical inverted U-shaped dose effect function, peaking at 0.1 mg/kg/injection. AEPEA and DEPEA also showed inverted dose-effect functions, with each peaking at 0.3 mg/kg/injection. BMPEA did not show an inverted U-shaped dose-effect function, but active responding slowly increased up to a dose of 6 mg/kg/injection.

Conclusions:

Taken together, our findings indicate that dietary supplements containing PEA analogs may have significant abuse liability when used recreationally.

Introduction

Amphetamine-like compounds are often added to dietary supplements marketed for weight loss or athletic performance enhancement, even though these compounds are rarely listed on the product labels (e.g., Eichner et al., 2016; Zhao et al., 2018). Many of these amphetamine-like additives are structural analogs of phenethylamine (PEA) (see Figure 1 for chemical structures) that could pose significant health risks. For example, exposure to the supplement additive β-methylphenethylamine (BMPEA) has been linked with serious cardiovascular complications in human users, including hemorrhagic stroke (Cohen et al., 2015; Amatto et al., 2021). Due to the public health risks associated with BMPEA, this compound has been banned by the Food and Drug Administration (FDA) of the United States of America (USA), and dietary products containing this compound are prohibited from sale (Pawar and Grundel, 2017). Nevertheless, BMPEA and a host of structurally related compounds continue to be found in dietary supplements sold in retail markets (Cohen et al., 2021; Duiven et al., 2021).

Fig. 1.

Chemical structures of amphetamine and PEA analogs used in this study.

Similar to BMPEA, the PEA analogs α-ethylphenethylamine (AEPEA) and N,α-diethylphenethylamine (DEPEA) have been identified in dietary supplements (Santillo et al., 2014; Cohen et al., 2014; ElSohly et al., 2015; Cohen et al., 2016; Yun et al., 2017). We previously reported that BMPEA, AEPEA, and DEPEA display pharmacological properties comparable to amphetamine in laboratory rats (Schindler et. al., 2019; Schindler et. al, 2021). For example, all three PEA analogs act as substrate-type releasers at the rat norepinephrine transporter (NET), which leads to increases in blood pressure and heart rate. The PEA compounds are about 10-fold less potent than amphetamine at stimulating cardiovascular parameters. Rickli et al. (2019) recently confirmed that BMPEA interacts with human NET expressed in transfected cells. We also found that BMPEA, AEPEA, and DEPEA interact with the dopamine transporter (DAT), the site of action implicated in the rewarding properties of stimulant drugs like cocaine and amphetamine (Kuhar et al., 1991; Amara and Sonders, 1998; Howell and Kimmel, 2008). In particular, DEPEA is a potent uptake inhibitor at DAT that increases locomotor activity in rats, suggesting the potential for maintaining drug-seeking behavior as a result of increased dopamine concentrations in the brain (Di Chiara et al., 2004; Leviel, 2011).

Previous research demonstrates that PEA itself is self-administered in animal models. Risner and Jones (1977) first showed that dogs will self-administer intravenous (i.v.) PEA in doses from 1.5 to 6.0 mg/kg/injection (see also Shannon and DeGregorio, 1982; Shannon and Thompson, 1984). Non-human primates will also self-administer i.v. PEA (Bergman et al., 2001) in doses from 0.32 to 10 mg/kg/injection, with the effects potentiated by pretreatment with MAO-B inhibitors. Recently, Ryu et al. (2021) showed that rats self-administer a dose of 1 mg/kg/injection i.v. PEA, and pretreatment with the dopamine D1 antagonist SCH 23390 attenuates this effect. In contrast to PEA, very little information is available about the rewarding or reinforcing effects of PEA analogs found in dietary supplements.

Oberlender and Nichols (1991) showed that AEPEA partially substitutes for amphetamine in a rat discriminative stimulus procedure, which indicates this analog may engender subjective effects akin to amphetamine. It is well established that amphetamine is readily self-administered by rats (Yokel and Wise, 1976; Carrol and Lac, 1997), non-human primates (Griffiths et al., 1976; Wee and Woolverton, 2006) and humans (de Wit et al., 1986; Comer et al., 1996). Due to the structural similarities between PEA analogs and amphetamine (see Fig. 1), research investigating the reinforcing properties of PEA analogs is essential for determining the possible health risks of these substances. To this end, we examined the ability of AEPEA, DEPEA, and BMPEA to support intravenous drug self-administration in male rats. A separate group of rats was trained to respond for amphetamine to provide a basis for comparative potency assessments of PEAs versus a known standard stimulant compound.

Materials and Methods

Subjects

Male Sprague-Dawley rats (Envigo, Frederick, MD, USA) weighing 300-400 grams at the start of training were individually housed in a temperature (22.2 ± 1.1°C) and humidity (45% ± 10%) controlled room, under a 12-hour reverse light/dark cycle (lights off 0700). Water was freely available, but food was restricted to keep weights constant or slowly increasing over the course of the study. All animal procedures were carried out in accordance with Animal Study Proposals approved by the Animal Care and Use Committee of the National Institute on Drug Abuse (NIDA), Intramural Research Program (IRP). Animal facilities were fully accredited by the American Association for Accreditation of Laboratory Animal Care.

Surgery procedure

Rats were anesthetized with a combination of 75 mg/kg ketamine and 5 mg/kg xylazine (i.p.). Jugular catheters were surgically implanted, and the distal end of the catheter was attached to a back mount that could be connected to a fluid swivel (Instech Laboratories, Plymouth Meeting, PA, USA), as previously described (Schindler et. al., 2016). Rats received ketoprofen (2.5 mg/kg, s.c., Butler Schein, Owings Mills, MD) for pain management immediately following surgery. Rats were given at least 7 days to recover before starting experimental sessions. Catheters were flushed 5 days a week with gentamicin saline solution (0.2 mL, 5 mg/mL, Butler Schein, Owings Mills, MD) prepared from stock gentamicin.

Drugs

Racemic BMPEA, AEPEA, and DEPEA were synthesized using standard organic chemical reactions and techniques. BMPEA was synthesized as a fumarate salt, whereas AEPEA and DEPEA were prepared as HCl salts. Each compound was chemically characterized with appropriate high-resolution mass spectral and 400-MHz NMR analyses, and with appropriate combustion analyses for carbon, hydrogen, and nitrogen. Each compound was chromatographically homogenous by thin-layer chromatography. Chemical purity for each compound was estimated to be greater than 98%. The corresponding α-methyl comparator compound (S)-amphetamine sulfate (i.e., amphetamine) was obtained from the pharmacy at the NIDA, IRP, in Baltimore, MD.

Drug self-administration

Ten training chambers were used for self-administration sessions (ENV-008CT, Med Associates, St. Albans, VT, USA). Each chamber was enclosed in a sound-attenuation cubicle equipped with a fan to provide ventilation and stable background noise. Each chamber had a grid floor and two nose-poke response holes (ENV-114BM, Med Associates), one on each side of a food trough (not used in this study). The nose-poke holes were illuminated from inside the hole by a dim yellow light. Experimental events were controlled by a MED-PC computer system (Med Associates).

Active nose poke responses resulted in delivery of a constant volume of drug (0.2 mL) over a period of 2 s on a fixed-ratio (FR) 1 schedule, followed by a 20 s timeout. Responses during the timeout were recorded but had no programmed consequences. Sessions lasted for two hours every weekday. The concentration of drug in the infusion solution was adjusted according to the weight of the rat, which was taken daily. This ensured accurate mg/kg dosing upon delivery of the fixed 0.2 mL infusion volume. When weight changed by 20 grams or a rat progressed onto a new dose, a new drug solution was prepared. Acquisition doses for the compounds were 0.1 mg/kg/injection for amphetamine, 1 mg/kg/injection for AEPEA, 1 mg/kg/injection for DEPEA, and 3 mg/kg/injection BMPEA. Training doses were chosen based our previous findings examining the cardiovascular effects of the drugs in rats and DAT uptake inhibition potencies (Schindler et. al., 2019; Schindler et. al., 2021). A separate group of rats was given saline as the self-administration drug for 10 days. Rats were trained for 10 days, however, whether a rat continued to dose-effect testing was determined by the criteria of more than 10 active responses and at least 75% active responses during the last two days of training. Following 10 days of acquisition training, dose-effect testing began for those rats trained with amphetamine, BMPEA, AEPEA, or DEPEA that had acquired the self-administration criteria. Three additional doses were tested in a non-systematic order. Each dose was tested for only 3 days, and responses for days 2 and 3 were averaged for statistical analysis. The rate of responding for each subject adjusted quickly to the change in dose and was stable over these 2 days. We initially trained 6 rats per drug, however with catheter loss (assessed with a methohexital test) early in dose-effect training more rats were added to some groups. As a result, not all rats were tested at all doses and group sizes varied.

Statistical Analysis

Data were analyzed using GraphPad Prism 9 software (GraphPad Scientific, San Diego, CA, USA). Acquisition analysis was performed for each drug and saline and employed day and responses (active/inactive) as factors, and post-hoc analysis was conducted using Sidak’s multiple comparisons test. Rats were included in the acquisition analysis if they completed at least 7 days of acquisition. Dose-response analysis was performed for each drug using dose as the factor, and post-hoc analysis was conducted using Dunnett’s multiple comparisons test to the averaged active responses of the last 2 days of saline self-administration. For a dose to be included in the dose-effect analysis, a rat had to complete all 3 days of testing on that dose. The average of active responses over the last two days of acquisition was used for the training dose.

Results

Figure 2 shows the data for acquisition of self-administration behavior. Amphetamine and all three PEA analogs showed acquisition of self-administration behavior as indicated by significant differences in active versus inactive responses (Amphetamine F_1,276 = 73.39, p<0.0001; AEPEA F1_,170 = 105.0, p<0.0001; DEPEA F_1,100 = 38.64, p<0.0001; BMPEA F_1,276 = 73.39, p<0.0001). Post-hoc analysis confirmed that active responses were higher than inactive over several days towards the end of training (represented by filled symbols). In contrast, although animals trained with saline showed some initial separation of active and inactive responses (F_1,99 = 7.0, p = 0.009), post-hoc analysis failed to reveal any individual day on which active and inactive responses were significantly different. For amphetamine, 13 of 14 rats reached the criteria of at least 10 active responses with over 75% total responses in the active hole. For AEPEA (6/8), DEPEA (4/6), and BMPEA (11/14) about two-thirds of rats reached the acquisition criteria. Only 1 of 6 rats trained on saline reached criteria.

Fig. 2.

Acquisition of self-administration for all 4 drugs tested and saline. Active (circles) and inactive (triangles) responses are shown, with filled circles indicating a significant difference from inactive responses on that day. Each point represents n = 14-15 for Amphetamine; n = 8-10 for AEPEA; n = 6 for DEPEA; n = 14-15 for BMPEA; n = 6 Saline.

Rats that reached acquisition criteria continued onto dose-effect testing. Those results are shown in Figure 3, where the average number of responses for doses tested are plotted along with averaged responses for the saline group (black point above S) over the last 2 days of training. As expected, amphetamine (green curve) produced a typical inverted U-shape dose-effect function and statistical analysis revealed a significant effect of dose (F_4,32 = 6.068, p=0.0009), with responding significantly above saline at 0.1 mg/kg/injection (p<0.01) and 0.03 mg/kg/injection (p<0.01).

Fig. 3.

Dose-effect curves for all 4 drugs tested compared to the last 2 days of saline training (point above S). Filled symbols indicate significant differences from saline. Amphetamine, AEPEA and DEPEA all supported infusion rates above saline levels. While infusion rates increased with dose, no dose of BMPEA was significantly different from saline. Each point represents n = 6-13 for Amphetamine; n = 4-6 for AEPEA; n = 4 for DEPEA; n = 7-11 for BMPEA; n = 6 Saline.

Both AEPEA (blue) and DEPEA (red) also showed dose-effect functions following an inverted-U shape, as well as a significant effect of dose (AEPEA F_4,21 = 2.962, p=0.0437; DEPEA F_4,17 = 10.03, p=0.0002). Post-hoc analysis revealed that the only dose significantly different from saline for AEPEA was 0.3 mg/kg/injection (p<0.01), whereas DEPEA showed significant difference from saline at both 0.3 mg/kg/injection (p<0.01) and 0.1 mg/kg/injection (p<0.01) with max responding at 0.3 mg/kg/injection. Although BMPEA resulted in acquisition of drug-taking behavior, BMPEA (orange) did not produce a significant effect of dose (F_4,33 = 1.634, p=0.1653) during dose-effect testing, although responding appeared to increase with dose.

The overall amount of responding per session varied across drugs, with DEPEA appearing to support the highest amount of responding and thus the highest rate of intake. Because the sensitivity to the drugs may vary among animals, we investigated this difference further by looking at the highest rate of responding for each animal regardless of the drug dose. This analysis again showed that DEPEA had the highest amount of responding per session (73.5 ± 10.0 responses/session), followed by AEPEA (55.3 ± 13.8). Amphetamine (40.9 ± 4.5) and BMPEA (40.1 ±7.7) supported lower levels of responding. Despite these differences, statistical analysis failed to show a difference between drugs in the amount of responding per session (F_3,30 = 2.6, p = 0.07).

Discussion

Dietary supplements are often marketed for weight loss or enhancement of athletic performance. In recent years, supplement products have been adulterated with synthetic amphetamine-like compounds, including a variety of PEA analogs (Cohen et al., 2021; Duiven et al., 2021). These PEA analogs can mimic many of the effects of amphetamine, including stimulation of the cardiovascular system and increased locomotor activity (Schindler et al., 2019, 2021). However, whether these PEA analogs might also support drug self-administration behavior has not been investigated. The purpose of the present study was to examine whether PEA analogs found in dietary supplements induce reinforcing effects comparable to those produced by amphetamine, an established reinforcing agent in rats (Yokel and Wise, 1976; Carroll and Lac, 1997). Our findings show that AEPEA, DEPEA, and BMPEA support intravenous drug self-administration behavior in male rats, providing evidence that these compounds have reinforcing properties and may display significant abuse liability in humans.

Our previous studies showed that the PEA analogs tested here induce dose-related inhibition of uptake and stimulation of release at DAT, but their potencies are much lower than that of amphetamine (Schindler et. al., 2019; Schindler et. al., 2021). Thus, we hypothesized that the PEA analogs might exhibit less robust self-administration responding when compared to amphetamine. The low DAT potency was especially relevant for BMPEA, which showed reduced cardiovascular effects and failed to induce locomotor stimulation (Schindler et al., 2019), suggesting it lacked amphetamine-like psychostimulant effects in vivo. Despite our previous findings, all three PEA analogs produced comparable rates of active nose-poke responding during acquisition training, and these rates were comparable to those produced by amphetamine.

In terms of dose-response testing, both AEPEA and DEPEA produced inverted-U shaped dose-effect functions that were right shifted from amphetamine. Amphetamine reached maximum responding at 0.1 mg/kg/injection consistent with the findings of others (Carroll and Lac, 1997), whereas the two α-ethyl analogs reached maximum responding at a dose of 0.3 mg/kg/injection. While AEPEA and DEPEA showed comparable rates of active responding to amphetamine during acquisition training, they produced higher rates of responding during the dose-effect testing, although this difference did not reach statistical significance. The increased response rates relative to amphetamine may reflect a shorter duration of action for AEPEA and DEPEA, where rats will take more drug to sustain perceived effects as the amount of circulating drug decreases for drugs possessing shorter half-lives (Panlilio et al., 2008; Tsibulsky and Norman, 1999). Unfortunately, no information is available about the half-life of AEPEA or DEPEA. Amphetamine itself has a relatively long half-life in rats after i.v. administration (approximately 63 min; Riviere et al., 1997), while PEA has a relatively short half-life in dogs after i.v. administration (Shannon et al., 1982). This short-half-life led Bergman et al. (2001) to pretreat non-human primates with an MAO inhibitor to facilitate self-administration since PEA is metabolized by MAO. If the PEA analogs have relatively short half-lives similar to PEA, it may explain the higher self-administration rates observed.

While BMPEA was able to reach significant self-administration levels during acquisition training, the ꞵ-methyl analog did not reach significance from saline at any of the doses tested following acquisition training. We suspect that the failure of BMPEA to elicit robust responding may be due to insufficient drug doses, which were too low for proper maintenance of drug self-administration. This notion is supported by the fact that maximum responding for BMPEA occurred at the highest dose tested of 6 mg/kg/injection. Additional research with higher doses is needed to determine whether an inverted U-shaped dose-effect function would also be seen with BMPEA.

The order of potency regarding PEA analog self-administration is different than what we have observed for locomotor activity measured by telemetry. Amphetamine was clearly the most potent in increasing locomotor activity (peak effect at 1 mg/kg), followed by DEPEA whose peak effect was at a minimum of 10 mg/kg (Schindler et al., 2021). However, neither AEPEA (up to 10 mg/kg, Schindler et al., 2021) nor BMPEA (up to 30 mg/kg, Schindler et al., 2019) showed significant increases in locomotor activity measured by the telemetry system. There is the possibility that higher doses of these drugs may have produced significant increases in locomotion. Nevertheless, the finding that both AEPEA and DEPEA showed similar dose effect functions for self-administration yet had statistically different dose-effect functions for locomotor activity indicates that the rewarding and locomotor effects of these drugs are not equivalent (see also Eisner-Dorman et al., 2011; Runegaard et al, 2018; Thomsen and Caine, 2011).

Accompanying their similar effects on self-administration, AEPEA and DEPEA also produced analogous effects on blood pressure (BP), with both compounds elevating BP to a similar degree at similar doses (ED₅₀ 2.54 and 3.74 mg/kg respectively; Schindler et al. 2021). BMPEA also increased BP (ED₅₀ 5.49) at slightly higher doses. Amphetamine, AEPEA, and BMPEA are fully efficacious releasers at both DAT and NET (Rothman and Baumann, 2003; Schindler et al., 2019; 2021). In release assays for NET, AEPEA and BMPEA are approximately 10 to 15 times less potent than amphetamine, suggesting that their effects on BP could be mediated at the NET. At the DAT, however, AEPEA and BMPEA are 50 and 100 times less potent than amphetamine, respectively. These potencies at DAT are in line with the relatively weak effects of AEPEA and BMPEA on locomotion but cannot as easily account for the potency of AEPEA in self-administration. DEPEA is also a full releaser at NET, but in contrast to the other analogs is only a partial releaser at DAT. Its predominant effect at DAT may be as an uptake inhibitor, where it is only approximately 5 times less potent than amphetamine. This difference in potency is in line with its effects on both locomotion and self-administration. Of course, the effects of the PEA analogs for any of the in vivo effects could be influenced by factors other than uptake or release at DAT and NET, such as direct receptor actions or pharmacokinetic effects.

It is difficult to directly extrapolate the self-administration findings shown here in rats to the potential abuse liability of PEA analogs in humans taking supplements. Nonetheless, if supplement products containing DEPEA are taken orally according to manufacturer recommendations, the amount of DEPEA ingested in these nutritional supplements would be 35-45 mg per serving (Cohen et al., 2014; ElSohly et al., 2015), or about 0.5 mg/kg for a 75 kg person. In humans, the threshold oral dose of amphetamine that can be discriminated is 5 mg (Chait et al., 1985), or about 0.07 mg/kg. Here, we found that DEPEA is about 3-fold less potent than amphetamine as a reinforcing agent. If the relationship between drug potencies in rat self-administration translates to drug potencies in humans, then the amount of DEPEA in a serving seems sufficient to produce subjective effects in humans akin to amphetamine. Thus, these PEA analogs may have habit-forming properties in humans at doses within the range that they appear in dietary supplements. While there is limited information about the effects of these compounds in humans, to the extent that they resemble amphetamine, potentially toxic effects may be expected with increased use. For example, Cohen et al. (2015) reported the case of a woman who ingested a dietary supplement 30 minutes prior to exercise and suffered a hemorrhagic stroke during physical activity. The patient reported taking no other drugs prior to exercise, and forensic analysis of the ingested supplement revealed the recommended serving contained 290 mg of BMPEA, suggesting the ingestion of this PEA compound led to injury.

Some limitations of the current study need to be acknowledged. All subjects were male, which may limit extrapolation to female subjects. As all previous studies of PEA and PEA analogs in rodents were performed in male subjects (Oberlender and Nicholas, 1991; Ryu et al., 2021, Schindler et al., 2019; 2020), this allowed a direct comparison to those studies. Further, while research has shown differences in rates of self-administration in male and female subjects with psychoactive drugs, these differences tend to be quantitative rather than qualitative (Becker and Koob, 2016; Flores et al., 2019; Nicolas et al, 2022). Therefore, while there may be differences in the potency of the PEA analogs in supporting self-administration in female rats, it is likely they would similarly support self-administration in females at some doses. Nevertheless, self-administration of the PEA analogs should also be studied in female rats to ensure applicability to the human population (Towers and Lynch, 2021).

Another limitation of the current study is the use of only a fixed-ratio reinforcement schedule. While the use of the fixed-ratio schedule can confirm that a drug might be reinforcing in humans, it does not provide information on the strength of those drugs as reinforcers. For example, these drugs should be further studied with schedules such as a progressive-ratio reinforcement schedule (cf., Gannon et al., 2018) where the required number of responses necessary to produce a reinforcement (or drug infusion) increases with each successive reinforcement. Thus, the more reinforcers a subject receives is an indication of how much effort they are willing to provide for the drug, allowing the reinforcement strength of drugs to be better compared. The assessment of reinforcement strength could also be strengthened by using choice procedures (Brown et al., 2022; Seaman et al., 2022) where one response choice would be a PEA analog and the other a comparison drug or reference non-drug choice. The degree to which a subject chooses the PEA analog to the known reinforcer can indicate its strength as a reinforcer.

Overall, this study shows that PEA analogs found in dietary products engender significant reinforcing effects in a preclinical rat self-administration model. Considering that these compounds are often not listed on product labels and still require further pharmacological characterization, their inclusion in recreational products warrants significant caution. This is particularly true since these compounds have also been shown to have significant cardiovascular effects (Schindler et al., 2019; 2021).