A Strategy to Prioritize Emerging Drugs of Abuse for Analysis: Abuse Liability Testing Using Intracranial Self-Stimulation (ICSS) in Rats and Validation with α-Pyrrolidinohexanophenone (α-PHP)

Abstract

Novel psychoactive substances (NPS) threaten public health and safety while also straining the limited resources of forensic laboratories. To efficiently allocate the finite resources available, we propose a new strategy for prioritizing NPS with abuse liability testing using a preclinical behavioral procedure in rats known as intracranial self-stimulation (ICSS). To validate this assay, the recently-scheduled synthetic cathinone α-PHP was compared to cocaine, a mechanistically similar drug of abuse, as a positive control and saline as a negative control. Male Sprague-Dawley rats (n=6) were implanted with electrodes targeting the medial forebrain bundle and trained to respond by lever-press for electrical brain stimulation. The rats were tested with doses of 0.32, 1.0, and 3.2 mg/kg α-PHP as well as 10 mg/kg of cocaine and saline administered by intraperitoneal injection. Neither saline nor 0.32 mg/kg α-PHP altered ICSS response rates compared to baseline levels of responding; however, doses of 1.0 and 3.2 mg/kg α-PHP and 10 mg/kg cocaine facilitated ICSS responding. This ICSS profile suggests that α-PHP has high abuse potential, with a rapid onset of effects and a long duration of action, and supports the decision to schedule this compound. This study demonstrates the ability of ICSS to distinguish between compounds of low and high potential for abuse. A strategy is proposed here to screen NPS using ICSS and classify emerging drugs into four priority categories for further analysis.

1. INTRODUCTION

The proliferation of novel psychoactive substances (NPS) is a continuing threat to public health and safety^[1–5]. NPS are generally structural analogs of traditionally abused drugs. Clandestine chemists synthesize and market these drugs to avoid the legal consequences of the possession of controlled substances as well as to subvert forensic laboratories by making detection and identification a challenge^[5–7]. However, in addition to retaining the abuse potential of the parent compound, these structural modifications can also have substantial consequences on the pharmacological properties of the drug^[8–10]. These can include altering the potency, time course, off-target effects, and other drug properties. Each of these can result in unexpected impairment or unintentional overdose and threaten public health. In order to monitor these threats, it is necessary for clinical and forensic laboratories to have the capability to examine and detect NPS. In the absence of widespread and comprehensive detection methods for NPS, accurate assessment of the full scope of abuse is impossible to determine; however, their usage trends may parallel the increases of their more traditional counterparts. For example, in the United States, an analysis of urine drug tests from more than one million patients indicated drastic increases in positivity rates from 2013 to 2019 for methamphetamine (486%) and fentanyl (333%)^[11]. This surge in the appetite for known psychoactive substances combined with concurrent increases in reported NPS to the United Nations Office on Drugs and Crime over that same period highlights the urgency and importance of surveillance and intervention^[12].

In 2018 alone, 55 new NPS were reported to the European Union Early Warning System^[1]. Similarly, NPS comprised 4 of the top 25 drugs reported to the National Forensic Laboratory Information System (NFLIS) in the United States that year^[13]. The rise of NPS challenges the capability of laboratories to detect these drugs and subsequently assess their culpability in cases of intoxication or overdose death. The lack of availability of certified reference materials (CRM) for newly emerging compounds and the burden on laboratories to develop and validate new analytical methods create significant hurdles that potentially lead to under-detection and under-reporting of responsible chemicals. Furthermore, few NPS have graduated to widespread problem, rendering the need to develop CRM and analytical methods for every NPS moot ^[14]. Laboratories could benefit from a prioritization strategy to efficiently allocate their limited resources toward those NPS which are most likely to proliferate. We propose that validated preclinical behavioral procedures for abuse liability assessment in laboratory rats can serve this purpose and be used to triage NPS to target the largest threats first. Leveraging preclinical abuse liability assessment as a predictive tool would allow the forensic community to divert resources toward priority targets, leading to faster method development and earlier control of the most dangerous NPS by regulatory bodies. Pharmacological characterization of NPS is typically accomplished after preliminary identification and synthesis of reference standards. However, rapid abuse liability testing could be accomplished with seized samples that are unable to be identified by current methods or on newly identified compounds that have yet to be mechanistically or pharmacologically characterized. By performing this behavioral testing in parallel to chemical testing, valuable predictive information can allow for earlier, more efficient interventions.

Intracranial self-stimulation (ICSS) is a preclinical behavioral procedure that can be readily conducted in laboratory rats and that may be especially useful as a bioassay for NPS^[15,16]. In ICSS, rats equipped with an electrode targeting a brain-reward area are trained to emit an operant response (e.g. pressing a lever) to receive pulses of electrical brain stimulation. Many drugs of abuse increase (or “facilitate”) ICSS responding, and ICSS can be used to rapidly stratify the magnitude, potency, and time course with which a novel compound might produce ICSS facilitation as a signal of abuse potential. ICSS is a well-established procedure, and drug effects on ICSS has been characterized across numerous pharmacological classes^[16]. These effects can be examined in drug-naïve or drug-experienced subjects, and the test drug can be administered by any route of administration. In addition to ICSS facilitation as a signal of abuse potential, drug-induced disruption of ICSS can also be used as a quantitative measure of general behavioral toxicity, and emergence of other potential toxic effects (e.g. convulsions) can also be documented during testing. Lastly, sensitivity of drug effects to blockade by receptor-selective antagonists (e.g. opioid or cannabinoid receptor antagonists) can be used to provide initial information on drug class and receptor mechanism of action of an NPS.

The goal of the present study was to illustrate the potential of ICSS for NPS assessment by testing the synthetic cathinone α-pyrrolidinohexanophenone (α-PHP) and comparing it to a more well-known abused stimulant, cocaine. Synthetic cathinones have emerged as novel stimulants that exert their effects by altering monoamine signaling in the brain^[17,18]. Although α-PHP was identified as an NPS as early as 2014^[19,20], it was uncontrolled in the United States until its placement in Schedule I as of July 2019^[21]. Given the extensive timeframe between its initial appearance and the scheduling action, α-PHP was chosen as a candidate for abuse potential assessment using ICSS to see if early assessment would have successfully predicted its abuse liability and informed the scheduling decision. α-PHP acts as a reuptake inhibitor at the dopamine transporter (DAT)^[8]. The magnitude, potency, and time course of α-PHP effects on ICSS were determined and compared to vehicle and to cocaine as negative and positive controls, respectively. Additionally, the experimental design in this study with α-PHP was used as the foundation for a more general testing algorithm that could be used to evaluate any NPS.

2. METHODS

2.1. Subjects

Subjects were six (6) adult male Sprague-Dawley rats (Envigo, Indianapolis, IN, USA) weighing between 318 and 365 g at the time of surgery. Prior to the present study, these rats were used in studies of methcathinone and two methcathinone analogs ^[22]; however, there was a three-week drug-free washout period between completion of these prior studies and initiation of the present study. The rats were housed individually in cages with free access to food and water at all times except during ICSS sessions, and were exposed to a 24-hour light/dark cycle with lights on from 6:00 am to 6:00 pm. Animals resided in a facility accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC). Animal care and research complied with the National Institutes of Health guidelines for the care and use of animal subjects in research, and all protocols were approved by the Virginia Commonwealth University Institutional Animal Care and Use Committee.

2.2. Surgical Procedure

Surgical and experimental procedures were similar to those described previously^[9,10,23]. Anesthesia was induced by inhalation of 2.5–3.0% isoflurane (Zoetis Inc., Kalamazoo, MI, USA) in oxygen until the rat was unresponsive to toe pinch, and this level of anesthesia was maintained throughout surgery. A stainless steel, bipolar electrode (Plastics One, Roanoke, VA, USA) was implanted via stereotaxic surgery by inserting the cathode into the left medial forebrain bundle at the lateral hypothalamus (2.8 mm posterior to bregma, 1.7 mm lateral to the midsagittal suture, and 8.8 mm ventral to the exterior surface of the skull). The anode was grounded by coiling around one of three screws anchored into the dorsal surface of the skull. The electrode, screws, and grounding wire were permanently affixed to the skull using dental acrylic resin. Ketoprofen (5 mg/kg) was administered for post-operative analgesia immediately and 24 h after surgery, and rats were allowed to recover for at least 7 days prior to commencing ICSS training.

2.3. Apparatus

Behavioral sessions were conducted in modular operant test chambers (29.2 × 30.5 × 24.1 cm) (Med Associates, St. Albans, VT, USA) constructed of stainless steel and clear polycarbonate with a grid floor and housed inside ventilated sound-attenuating cubicles. The test chamber contained a response lever located 3 cm above the floor and 7.6 cm below three stimulation lights (red, yellow, and green), a 2 W house light, and an ICSS stimulator (see Figure 1A). A bipolar cable and swivel commutator (Model SL2C; Plastics One, Roanoke, VA, USA) connected the ICSS stimulator to the electrode. Data collection and ICSS programming were accomplished using a computer system running Med-PC IV software (Med Associates, St. Albans, VT, USA).

Figure 1.

Panel A shows a simplified schematic of the test chamber, including the ICSS stimulator attached via a bipolar cable to the implanted electrode. Panel B shows a frequency-rate curve for a trained rat (simulated data), with responding absent at low brain-stimulation frequencies and a gradual, frequency-dependent increase in responding toward high frequencies of brain stimulation. Abuse liability is indicated by leftward shifts in this frequency-rate curve, a phenomenon known as “ICSS facilitation.” Abscissa: Frequency of electrical brain stimulation in Hz. Ordinate: Percent maximal control reinforcement rate (%MCR) during each 1-min frequency trial.

2.4. Training

Following initial shaping of lever-press responding, rats were trained under a fixed-ratio 1 (FR 1) schedule of brain stimulation. During behavioral sessions, each lever press resulted in the delivery of a 0.5 s train of square wave cathodal pulses (0.1 ms pulse duration) and illumination of the stimulus lights over the lever. Stimulation intensity and frequency were set at 150 μA and 126 Hz, respectively, during initial 60-min training sessions. Stimulation intensity was then individually adjusted for each rat until ICSS rates > 30 stimulations/min were observed. This intensity was then held constant, and frequency manipulations were introduced. Sessions involving frequency manipulations consisted of three sequential 10 min components. During each component, a descending series of 10 frequencies (158 to 56 Hz in 0.05 log increments) was presented, with each frequency available for a 1 min trial. Each frequency trial consisted of a 10 s time-out, during which five non-contingent “priming” stimulations were delivered at the frequency of stimulation that would be available during that trial, followed by a 50 s “response” period, during which responding produced electrical stimulation under a FR 1 schedule as described above. The chamber remained dark until the response period, during which the house light would illuminate as a discriminative stimulus that electrical brain stimulation could be received contingent upon pressing the response lever. The stimulation intensity was again adjusted to yield baseline frequency-rate curves that had high rates of reinforcement for the upper frequencies and minimal responding at the remaining frequencies (see Figure 1B). This intensity (120–340 μA across rats) was then held constant throughout the study. Training continued until frequency-rate curves were not statistically different over three days of training as indicated by lack of a significant effect of “day” in a two-way analysis of variance (ANOVA) with frequency and day as the two variables (see data analysis).

2.5. Testing

ICSS testing was performed on α-PHP at doses of 0.32, 1.0, and 3.2 mg/kg. These doses were selected following pilot studies that gradually increased dosage in 0.5 log units in individual rats until a behaviorally active dose was identified. ICSS testing was also performed using a 10 mg/kg dose of cocaine and saline as positive and negative controls, respectively. Test sessions consisted of three consecutive baseline components followed by intraperitoneal (i.p.) drug injection and then by pairs of consecutive test components beginning after 10, 30, 100, and 300 min. For 3.2 mg/kg α-PHP, ICSS testing was also performed after 1440 min (24 hr). Treatments were counterbalanced across rats using a Latin-Square design. Testing was generally performed on Tuesdays and Fridays, and three-component training sessions were conducted on other weekdays. Data from training sessions were monitored to ensure a return to ICSS rates consistent with pre-drug levels to confirm drug wash-out. All testing was performed during the light phase of the light/dark cycle.

2.6. Data Analysis

The primary dependent variable was reinforcement rate in stimulations per minute during each frequency trial. To normalize these data, raw reinforcement rates from each trial in each rat were converted to percent maximum control rate (%MCR), with MCR defined as the mean of the maximal rates observed during the second and third baseline components for that rat on that day. Thus, %MCR values for each trial were calculated as (reinforcement rate during a frequency trial) / (MCR) x 100. For each test session, data from the second and third baseline components were averaged to yield a baseline frequency-rate curve, and data from each pair of test components were averaged to generate test frequency-rate curves. Baseline and test curves were then averaged across rats to yield mean baseline and test curves for each manipulation. For statistical analyses, results were compared by repeated measures two-way ANOVA with ICSS frequency as one factor and either dose or time as the second factor. A significant ANOVA was followed by the Holm-Sidak post hoc test, and the criterion for significance was set at P < 0.05. Frequency-rate curves for vehicle (saline), the negative control, were compared to baseline frequency-rate curves, with the expectation of no significant difference. All drug treatments were compared to vehicle. Cocaine was expected to significantly facilitate ICSS responding compared to vehicle^[23,24].

To provide a secondary summary measure of drug effects, the total number of stimulations delivered across all 10 frequency-trials of each component was determined for each drug dose at each time point. Mean number of stimulations per component for baseline and vehicle test components at all time points were compared to evaluate the negative control using a repeated measures ANOVA, with significance criterion set at P < 0.05. Summary test data from each day were then normalized to individual baseline data from that day using the equation % baseline total stimulations per component = (mean total stimulations per test component) / (mean total stimulations per baseline component) x 100. Data were then averaged across rats and plotted as a function of time for each drug dose. Results were compared using a repeated measures two-way ANOVA using time as one factor and drug dose as the second factor. A significant ANOVA was followed by the Holm-Sidak post hoc test, with the criterion for significance set at P < 0.05. Statistical analyses were performed using Prism 8.1.2 (GraphPad Software, San Diego, CA, USA).

2.7. Drugs

α-PHP oxalate was synthesized as previously reported^[8] and (−)-cocaine HCl was supplied by the National Institute on Drug Abuse Drug Supply Program (Bethesda, MD, USA). Drugs were dissolved in sterile saline (Hospira Inc., Lake Forest, IL, USA) and doses are expressed as the salt form of the drug.

3. RESULTS

Across all rats in the study, the mean ± standard error of the mean (SEM) MCR was 53.1 ± 3.6 stimulations per trial, and the mean ± SEM total stimulations earned per baseline component was 200 ± 14.0 stimulations per component. Under baseline conditions, electrical brain stimulation maintained a frequency-dependent increase in ICSS rates. Two-way ANOVA of baselines for the five experimental days indicated stable baseline responding with no significant main effect of day (F(1.373, 6.864) = 1.637, P=0.2534) and no significant day x frequency interaction (F(3.470, 17.35) = 1.274, P=0.3166), but a significant main effect of frequency (F(1.890, 9.448) = 130.3, P<0.0001). Baseline ICSS rates were minimal for low frequencies of electrical stimulation (56–79 Hz), increased at intermediate frequencies (89–112 Hz), and reached a maximum rate at high frequencies (126–158 Hz).

Figure 2 shows the effects of our negative control, saline, on ICSS performance. Figure 2A compares frequency-rate curves for baseline components and test components after injection of saline with a 10-minute pretreatment time. Statistical analysis of these curves shows no significant effect of saline treatment (F(1, 5) = 0.6856, P=0.4454) and no significant frequency x treatment interaction (F(2.928, 14.64) = 0.2439, P=0.8601). Figure 2B shows the summary measure of ICSS from 10 to 300 min after saline injection. A one-way ANOVA comparing mean total number of stimulations per component indicates no significant difference between baseline ICSS responding and ICSS responding after saline injection for all four experimental time-points (F(2.307, 11.53) = 2.068, P=0.1675).

Figure 2.

Saline injection does not affect ICSS responding. Panel A shows a lack of significant effect of i.p. saline injection compared to baseline ICSS responding at 10-min post-injection. Abscissa: Frequency of electrical brain stimulation in Hz. Ordinate: Percent maximal control reinforcement rate (%MCR) during each 1-min frequency trial. Panel B shows a summary measure of the lack of effect of saline on ICSS responding over time. Abscissa: Time post-administration of saline injection. Ordinate: Percent baseline number of stimulations (% Baseline Stimulations) obtained across all frequencies during each 10-min component. All data are expressed as mean ± SEM for six rats (n=6).

Figure 3 compares the effects of saline with the effects of our positive control, 10 mg/kg cocaine, on ICSS performance. Figure 3A compares frequency-rate curves determined 10 min after saline or cocaine injection. Statistical analysis of these curves shows a significant effect of cocaine treatment (F(1, 5) = 185.4, P<0.0001) as well as a significant cocaine x frequency interaction (F(9, 45) = 9.333, P<0.0001). These data indicate significant facilitation of ICSS responding maintained by low and intermediate brain-stimulation frequencies after cocaine administration compared to saline vehicle. Figure 3B compares the time course of effects produced by saline and cocaine on the summary measure of ICSS. A two-way ANOVA comparing cocaine and saline shows a significant effect of time point (F(1, 5) = 85.69, P=0.0002) and significant time point x treatment interaction (F(3, 15) = 39.79, P<0.0001). Post hoc tests indicate total stimulation rates for cocaine treated rats were significantly elevated compared to vehicle at the 10-, 30-, and 100-min time points (P < 0.05). At the 300-min time point, total stimulation rates for cocaine treated rats were not significantly different from saline vehicle. These results show that cocaine, an established drug of abuse and positive control, produced robust ICSS facilitation with rapid onset and a duration of at least 100 min.

Figure 3.

Cocaine (10 mg/kg) facilitates ICSS. Panel A shows significant facilitation of ICSS responding at low and intermediate frequencies of electrical brain stimulation after i.p. cocaine injection (10 mg/kg) compared to vehicle at 10 minutes post-injection. Abscissa: Frequency of electrical brain stimulation in Hz. Ordinate: Percent maximal control reinforcement rate (%MCR). Panel B shows a summary measure of the effect of 10 mg/kg cocaine on ICSS responding over time. Abscissa: Time post-administration of injection. Ordinate: Percent baseline number of stimulations delivered per component. For both panels, data are expressed as mean ± SEM for six rats (n=6). Filled points denote a significant difference compared to vehicle by a two-way ANOVA followed by a Holm-Sidak post hoc test (P < 0.05).

Figure 4 compares the effects of three doses of α-PHP to the effects of saline on ICSS performance. Figure 4A compares the frequency-rate curves of 0.32, 1.0, and 3.2 mg/kg α-PHP against vehicle determined 10 min after α-PHP or saline injection. A two-way ANOVA shows a significant effect of dose (F(1.772, 8.859) = 26.64, P=0.0002) as well as a significant dose x frequency interaction (F(3.866, 19.33) = 5.391, P=0.0046). Post hoc tests indicate a lack of facilitation compared to saline at any individual frequency for a dose of 0.32 mg/kg α-PHP, but robust facilitation of ICSS maintained by low- and intermediate-frequencies of brain stimulation at a dose of 3.2 mg/kg. Additionally, no dose attenuated responding for high-frequency brain stimulation. Figure 4B compares summary measures of time courses of effects for 0.32, 1.0, and 3.2 mg/kg doses of α-PHP and saline. A two-way ANOVA comparing drug doses and the negative control over time shows significant main effects of drug dose (F(2.816, 14.08) = 27.05, P<0.0001) and time point (F(1.455, 7.275) = 23.77, P=0.0010), as well as a significant dose x time interaction (F(2.602, 13.01) = 11.44, P=0.0008). Post hoc tests indicated that the dose of 1.0 mg/kg α-PHP was only significantly different from saline at the 10-min time point, whereas for 3.2 mg/kg α-PHP, robust ICSS facilitation compared to baseline was significant to the 300-min time point (P<0.05). Peak facilitation occurred during the 10-min test components. An additional pair of ICSS test components were collected for the 3.2 mg/kg dose of α-PHP at 1440 min (24 hr), and the rates of ICSS seem to return to baseline by that time; however, statistical analysis was not performed as rats treated with saline were not tested at that time point. Overall, α-PHP exhibited a profile of strong ICSS facilitation manifested as a leftward shift of the frequency-rate curve with a similarly rapid onset to that of cocaine. The ICSS profile of rats treated with 3.2 mg/kg α-PHP is similar to that of rats treated with 10 mg/kg cocaine. However, cocaine and α-PHP at these doses differ in their duration of action, with α-PHP causing significant facilitation of ICSS at least to 300 min.

Figure 4.

α-PHP (1.0 and 3.2 mg/kg) facilitates ICSS. Panel A shows significant facilitation of ICSS responding at low and intermediate frequencies of electrical brain stimulation after i.p. α-PHP injection (3.2 mg/kg) compared to vehicle at ten 10 minutes post-injection, and modest facilitation of ICSS at a dose 1.0 mg/kg. Abscissa: Frequency of electrical brain stimulation in Hz. Ordinate: Percent maximal control reinforcement rate (%MCR). Panel B displays shows a summary measure of the effect of 0.32, 1.0, and 3.2 mg/kg α-PHP on ICSS responding over time. Abscissa: Time post-administration of injection. Ordinate: Percent baseline number of stimulations delivered per component. For both panels, data are expressed as mean ± SEM for six rats (n=6). Filled points denote a significant difference compared to vehicle by a two-way ANOVA followed by a Holm-Sidak post hoc test (P < 0.05).

4. DISCUSSION

This study used an ICSS procedure in rats to evaluate the abuse liability of the synthetic cathinone α-PHP. ICSS is a preclinical behavioral procedure in which laboratory animals (e.g. rats) are trained to press a response lever to receive pulses of electrical brain stimulation, and drugs of abuse produce characteristic increases in ICSS responding, a phenomenon called “ICSS facilitation”^[16]. A number of procedural variants for ICSS have emerged since the introduction of the technique, including discrete trial current intensity threshold procedures^[25]. Here we use a hybrid frequency-rate method which is quantifiable and sensitive to both abuse-related facilitation of responding maintained by low-frequency electrical stimulation as well as abuse-limiting depression of responding maintained by high-frequency electrical stimulation. In the present study, α-PHP produced a potent, robust, and sustained ICSS facilitation similar to that produced by cocaine and consistent with the recent DEA decision to emergency-schedule α-PHP. This study also illustrated the potential of ICSS for use in rapid abuse liability testing of NPSs. Below, we propose a preclinical testing algorithm that capitalizes on ICSS testing for rapid screening of NPSs to guide prioritization of resources for drug control and development of methods for drug analysis.

Our results showed baseline responding was stable over the experimental period, and the baseline frequency-rate data ranged from an absence of responding at low frequencies to maximal responding at high frequencies of electrical brain stimulation. This spectrum of baseline response rates allows the experimental procedure to detect both rate-increasing and rate-decreasing effects of test drugs. Although these rats were not drug naïve, they underwent a lengthy wash-out period of three weeks between test compounds. Furthermore, ICSS procedures are resistant to sensitization by dopaminergic drugs ^[26], and we have unpublished data also indicating that ICSS facilitation by dopamine transporter blockers and substrates is consistent with repeated dosing and in drug-naïve versus drug-experienced animals. Saline was assessed as a negative control in this experiment, and it did not alter rates of responding relative to baseline. Cocaine (10 mg/kg) was assessed as the positive control in this experiment, and as expected, it robustly facilitated ICSS response rates. This is consistent with the effects of cocaine on ICSS shown in previous experiments^[23,24]. The controls clearly demonstrated the ability of this procedure to differentiate between a substance with no abuse potential (saline) from a substance known to have high abuse potential.

The main goal of this study was to evaluate the degree to which α-PHP might produce a cocaine-like facilitation of ICSS. This study examined three characteristics of α-PHP relevant to abuse potential: the potency, time course, and magnitude of effects on ICSS. A comparison of our results to a previous study examining multiple cocaine doses^[24] suggests that α-PHP was approximately three times more potent than cocaine at facilitating ICSS. It exhibited a rapid onset, similar to cocaine, and a longer duration of action. Lastly, both α-PHP (3.2 mg/kg) and cocaine (10 mg/kg) produced similar magnitudes of effect, approximately doubling the overall rates of ICSS facilitation. The high potency, rapid onset of effects, and, most importantly, the substantial magnitude of ICSS facilitation of α-PHP are all indicative of high abuse potential. The abuse-related effects and long duration of action of α-PHP are supported by findings from Gatch et al., which demonstrated that α-PHP fully substituted for the discriminative stimulus effects of both methamphetamine and cocaine in rats, and had a long duration of action in locomotor studies (5–8 hr) in mice^[27]. Furthermore, recent reports of alpha-PHP involvement in hospitalizations illustrate the abuse potential indicated by these results^[28]. These data are also consistent with the ICSS profile exhibited by other controlled DAT inhibitors including methylphenidate^[29],3,4-methylenedioxypyrovalerone (MDPV)^[9], and the closely related analog α‐pyrrolidinopentiophenone (α-PVP)^[30]. Another DAT-selective substituted cathinone, bupropion, is a medication currently prescribed as an antidepressant and smoking cessation aid. ICSS testing of bupropion demonstrates a similar facilitation of ICSS, but at a 10-fold higher dose (32 mg/kg) and a longer pre-treatment time (30 min)^[31]. Although bupropion is not scheduled under the Controlled Substances Act, it is regulated as prescription-only, and cases of abuse have been documented ^[32]. Overall, the profile of ICSS and other behavioral effects exhibited by α-PHP is consistent with the mechanism of action of a DAT-selective inhibitor as indicated by Kolanos et al^[8]. The evidence for high abuse potential of α-PHP determined by this study support the U.S. Drug Enforcement Administration’s classification of α-PHP as a Schedule I substance^[21].

In addition to its sensitivity to abused DAT inhibitors, ICSS is also sensitive to most other major classes of abused drugs including amphetamine-like DAT substrates^{[9,10,33–37]}, fentanyl-like mu opioid receptor (MOR) agonists^[38–45], diazepam-like gamma-aminobutyric acid (GABA) A receptor positive allosteric modulators^[46], and nicotine-like nicotinic acetylcholine receptor agonists^[47]. Conversely, many drugs that are not abused (e.g. haloperidol-like DA receptor antagonists or fenfluramine-like serotonin transporter substrates) fail to facilitate ICSS, and overall, the predictive validity of ICSS is equivalent to that of drug self-administration procedures for preclinical abuse-potential assessment^[16]. Consequently, ICSS can be used as an agnostic behavioral screen to identify abuse liability of NPS from a broad array of pharmacological classes. Preliminary evidence for mechanism of action can then be collected by evaluating sensitivity of NPS effects to blockade by receptor-selective antagonists. For example, effects of MOR agonists on ICSS can be blocked by the opioid antagonist naltrexone^[43]. ICSS is usually not facilitated by two other major classes of abused drugs: marijuana-like cannabinoid receptor agonists^[48,49] and psilocybin-like serotonin 2A receptor agonist hallucinogens^[50]. However, both classes of drugs produce robust decreases in ICSS as a pharmacological signal, with these decreases manifested as downward shifts of the frequency-rate curve and decreases in reinforcement rates maintained by high brain-stimulation frequencies. The potency, time course, and magnitude of these effects can be readily quantified, and receptor mechanisms can be determined from studies with receptor-selective antagonists such as rimonabant^[48,49] and volinanserin^[51] for cannabinoid and serotonin 2A receptors, respectively. Lastly, some drug classes (e.g. salvia-like products containing kappa opioid receptor agonists^[52]) may have low overall abuse liability but may nonetheless display transient or localized patterns of abuse and toxicity^[53,54]. Again, these classes of drugs produce robust ICSS depression that permits pharmacological characterization of drug potency, time course, and receptor mechanism of action. MOR agonists are a pharmacological class of particular concern during the current drug-overdose crisis in the United States. Acute dosing with MOR agonists can induce motor impairment that obscures abuse-related ICSS facilitation; however, repeated dosing with MOR agonists for as little as one week can produce tolerance to ICSS rate-decreasing effects and reveal robust ICSS facilitation consistent with the abuse liability of these compounds. This issue is discussed at length in a recent review article^[15].

Given the sensitivity of ICSS to abused drugs from a broad range of pharmacological classes, we propose the algorithm shown in Figure 5 as an initial behavioral screen to prioritize NPS for further forensic analysis. NPS from seized samples, purified samples, or newly synthesized samples would be administered to rats in an ICSS procedure. If the drug produces ICSS facilitation as a signal of high abuse liability, then the drug should be considered a significant threat and be categorized as Priority 1 for further pharmacological study as well as the development of CRM and analytical methods. If a drug fails to produce ICSS facilitation but does produce ICSS depression, then the sensitivity of that drug effect to cannabinoid and serotonin 2A receptor antagonists should be assessed to test for cannabinoid- or hallucinogen-like mechanisms of action. A positive result here would warrant assignment to Priority 2 status. If a drug produces ICSS depression resistant to cannabinoid or serotonin 2A receptor antagonists, then it is unlikely to have long-term abuse potential, but it may still pose a threat of sporadic abuse and toxicity (e.g. as with salvia-containing products), and it should be assigned to Priority 3 status. Lastly, if a drug fails to alter ICSS, then it is unlikely to be a threat for either significant abuse or toxicity, and it should be assigned to Priority 4 status. In summary, the proliferation of NPS is taxing capabilities for detection and analysis, and forensic testing would benefit from data-driven strategies to prioritize compounds and guide efficient allocation of resources. Here, we propose ICSS as a preclinical behavioral procedure in rats that can rapidly characterize abuse liability and other manifestations of behavioral toxicity across a broad range of drug classes, and the sensitivity of this procedure was illustrated in studies with the recently scheduled cathinone analog α-PHP. We further propose an ICSS testing algorithm that can be used to test unknown or known drug samples and prioritize drugs for further testing.

Figure 5.

ICSS can be used as a screening tool to rapidly prioritize NPS for further analysis. Either seized or newly synthesized NPS can be tested in an ICSS procedure and classified according to four priority categories.

Although this ICSS procedure can provide a wealth of valuable pharmacological information about NPS, there are a number of practical aspects to consider. In addition to the facilities and resources necessary for the housing and care of laboratory animals, ICSS requires specialized equipment such as electrical stimulators, operant chambers, and the computer interface, all of which can be obtained from commercial suppliers. Additionally, the procedure requires infrastructure and expertise in stereotaxic surgery for implantation of electrodes, and there is some attrition (~25%) due to missed electrode placements or poor operant learning. Initial time investments in behavioral training can be extensive, but once reliable responding for electrical brain stimulation has been established, individual rats can perform at stable levels for months. We propose this procedure as a behavioral screening tool, but its speed and throughput are dependent upon maintaining a population of trained rats. Finally, it must be recognized that pharmacological reward measured by ICSS is only one of several factors that may contribute to the incidence of abuse for a specific drug, and the predictive validity of this assay will depend at least in part upon those non-pharmacological factors. Despite these challenges, we believe that the benefits of this testing algorithm will outweigh the costs.