Chronic cannabinoid exposure produces tolerance to the dopamine releasing effects of WIN 55,212-2 and heroin in adult male rats

Abstract

Synthetic cannabinoids were introduced into recreational drug culture in 2008 and quickly became one of the most commonly abused drugs in the United States. The neurobiological consequences resulting from synthetic cannabinoid repeated exposure remain poorly understood. It is possible that a blunted dopamine (DA) response may lead drug users to consume larger quantities to compensate for this form of neurochemical tolerance. Because the endogenous cannabinoid and opioid systems exhibit considerable cross-talk and cross-tolerance frequently develops following repeated exposure to either opioids or cannabinoids, there is interest in investigating whether a history of synthetic cannabinoid exposure influences the ability of heroin to increase DA release. To test the effects of chronic cannabinoid exposure on cannabinoid- and heroin-evoked DA release, male adult rats were treated with either vehicle or a synthetic cannabinoid (WIN55-212-2; WIN) using an intravenous (IV) dose escalation regimen (0.2–0.8 mg/kg IV over 9 treatments). As predicted, WIN-treated rats showed a rightward shift in the dose-response relationship across all behavioral/physiological measures when compared to vehicle-treated controls. Then, using fast-scan cyclic voltammetry to measure changes in the frequency of transient DA events in the nucleus accumbens shell of awake and freely-moving rats, it was observed that the DA releasing effects of both WIN and heroin were significantly reduced in male rats with a pharmacological history of cannabinoid exposure. These results demonstrate that repeated exposure to the synthetic cannabinoid WIN can produce tolerance to its DA releasing effects and cross-tolerance to the DA releasing effects of heroin.

1. Introduction

Synthetic cannabinoids (trade names: K2, Spice) were introduced into the United States recreational drug culture approximately ten years ago under the guise of herbal incense. To confuse the consumer, synthetic cannabinoids are dissolved into a soluble solution that is then sprayed onto organic plant material (Ammann et al., 2011; Berry-Cabán et al., 2012; Johnson et al., 2011). Far from a harmless potpourri, many of these synthetic cannabinoids are much more dangerous than cannabis itself. While synthetic compounds resembling classical cannabinoids exist (e.g., HU210), the most commonly abused versions (e.g., JWH-018/JWH-073) are structurally and functionally distinct aminoalkylindole cannabinoids (Howlett et al., 2002; Howlett et al., 2004). Aminoalkylindole compounds generally act as full agonists at cannabinoid type 1 receptors (CB1Rs), whereas the primary psychoactive component of cannabis, delta-9-tetrahydrocannabinol (THC), is only a partial agonist. THC is ~20% as effective at activating the CB1R G_i/o protein complex relative to aminoalkylindole cannabinoids, including WIN 55,212–2 (WIN) (Sim et al., 1996). In comparison to THC, repeated aminoalkylindole cannabinoid treatment is more effective at inducing spontaneous withdrawal (Aceto et al., 2001) and tolerance (Henderson-Redmond et al., 2020). Synthetic cannabinoids are also widely abused. Despite the fact that recreational synthetic cannabinoids were not detected in the United States until 2008 (Trecki et al., 2015), they were already the second most abused class of drugs amongst adolescents by 2012—trailing only cannabis itself (Patrick et al., 2016; Wood, 2013). In addition to being considered a safe alternative to cannabis, many adult professionals choose to abuse synthetic cannabinoids because they are not detected by standard drug screening assays (Ammann et al., 2011; Berry-Cabán et al., 2012; Johnson et al., 2011).

Drugs of abuse are thought to produce their rewarding effects by increasing DA concentration in the shell region of the nucleus accumbens (NAc) (Di Chiara et al., 2004). Through this mechanism, abused drugs mimic endogenous patterns of DA release that normally strengthen adaptive behavior (Volkow and Morales, 2015; Volkow et al., 2017). In the behaving animal, midbrain DA neurons fire in phasic bursts (>20Hz) under a variety of conditions (Redgrave et al., 2016; Sharpe and Schoenbaum, 2018), including the presentation of rewarding stimuli (Stauffer et al., 2016). These phasic bursts of neural activity contribute to transient DA release events in the NAc that are sufficient in concentration to occupy low affinity DA D1 receptors (Dreyer et al., 2010). Studying these high-concentration transient release events is particularly important for the neurobiology of addiction because only D1-expressing medium spiny neurons in the NAc undergo dendritic plasticity following repeated drug exposure (Barrientos et al., 2018).

Despite an abundance of empirical evidence, it remains controversial whether cannabinoids increase brain DA levels (Nutt et al., 2015; Oleson and Cheer, 2012).. The negation of cannabinoid-induced increases in DA is primarily based on two PET imaging studies, findingthat neither 2.5mg intravenous nor 10mg THC per os was sufficient to alter raclopride binding in the dorsal striatum (Barkus et al., 2011; Stokes et al., 2009). Similarly, a microdialysis study by Castaneda et al (Castañeda et al., 1991) reported that 10mg/kg THC delivered via gavage into freely moving rats failed to increase DA concentrations in either the dorsal striatum or the NAc. Despite the negative conclusion drawn from these studies, a substantially larger number of scientific reports indicate that cannabinoids are effective DA releasers. A wealth of electrophysiology (Diana et al., 1998; Gessa et al., 1998), microdialysis (Chen et al., 1993; Chen et al., 1990; Fadda et al., 2006; Pistis et al., 2002; Tanda et al., 1997), brain imaging (Bossong et al., 2015; Bossong et al., 2009; Voruganti et al., 2001), and fast-scan cyclic voltammetry (FSCV) studies (Cheer et al., 2004; Oleson et al., 2012; Oleson et al., 2014) suggest that a variety of cannabinoids, including endocannabinoids (Oleson et al., 2012; Solinas et al., 2006), THC (Bossong et al., 2009; Chen et al., 1993; Chen et al., 1990; Diana et al., 1998; Gessa et al., 1998; Pistis et al., 2002; Voruganti et al., 2001), and synthetic cannabinoids (Diana et al., 1998; Fadda et al., 2006; Gessa et al., 1998; Oleson et al., 2014; Tanda et al., 1997), increase brain DA levels in lab rats (Fadda et al., 2006; Gessa et al., 1998; Pistis et al., 2002) and humans (Bossong et al., 2015; Bossong et al., 2009; Voruganti et al., 2001) alike. Furthermore, cannabinoids alter brain reward thresholds (Gardner et al., 1988; Trujillo-Pisanty et al., 2011; Xi et al., 2008) and function as reinforcers during intravenous drug self-administration (Justinová et al., 2014; Smoker et al., 2019; Spencer et al., 2018). Cannabinoid abuse and dependence is also well-documented in the modern clinical literature (Copeland and Swift, 2009; Freeman and Winstock, 2015; Teesson et al., 2002).

In the present study, we first sought to determine whether repeated treatment with an aminoalkylindole cannabinoid produces tolerance to its ability to evoke transient DA release events, as tolerance to several well-characterized cannabimimetic effects have been shown following chronic cannabinoid exposure. Four behavioral/physiological effects composing the ‘tetrad test’ are typically quantified to test whether a drug functions as a cannabinoid; these include antinociception, catalepsy, hyopthermia, and hypomotility (Little et al., 1988; Wiley and Martin, 2003). While it is known that tolerance rapidly develops to these cannabimimetic effects following cannabinoid use (Hama and Sagen, 2009; Nealon et al., 2019), contention surrounds whether it develops to the drug’s rewarding/reinforcing and DA releasing effects.. Because the degree of tolerance that develops to specific cannabimimetic effects varies as a result of CB1R desensitization occurring in a brain region-dependent manner (Breivogel et al., 1997; Whitlow et al., 2003), it is possible that midbrain CB1Rs show resistance to tolerance. If DA mediates the rewarding effects of cannabinoids, then their DA-releasing effects should also be resistant to tolerance in accordance to findings that regular cannabinoid use does not result in diminished rewarding nor motoric effects of either synthetic or plant-based cannabinoids (Perez-Reyez et al., 1991; Solymosi and Köfalvi, 2017)(Fanarioti et al., 2015) (Mavrikaki et al., 2010). . In support of this supposition, French et al. (2000) reported that THC induces the same degree of burst firing in putative VTA DA neurons irrespective of whether rats received THC for the first or fifteenth time. However, a synthetic cannabinoid like WIN might be more effective at producing tolerance as it is ~80% more effective at activating CB1Rs than THC (Sim et al., 1996), and emerging evidence suggests that the molecular basis of WIN- vs. THC-induced tolerance is distinct. One molecular mechanism leading to CB1R desensitization and downregulation involves the recruitment of beta-arrestin2 to GRK-phosphorylated CB1Rs (Jin et al., 1999; Nguyen et al., 2012). However, a JNK-mediated form of molecular tolerance might also mediate CB1R desensitization in an agonist specific manner (Nealon et al., 2019). It was recently reported that disrupting JNK signaling prevents several forms of behavioral tolerance induced by THC, but not by WIN (Henderson-Redmond et al., 2020). Thus, it is possible that aminoalkylindole cannabinoids like WIN produce separate effects on molecular, cellular, and/or behavioral tolerance in comparison to THC.

Cannabinoids and opioids increase DA release through comparable overlapping mechanisms, and multiple reports of cannabinoid-opioid cross-tolerance exist. Similar to cannabinoids, the historical pharmacology literature suggests that opioid abuse does not involve brain DA systems (Ettenberg et al., 1982; Pettit et al., 1984). Incongruent with these historical lesion studies, recent evidence supports that heroin increases DA neural activity in a manner that mediates its reinforcing effects by activating G_i/o coupled μ-opioid receptors on midbrain GABA neurons, a pharmacodynamic drug action capable of indirectly disinhibiting DA neurons to promote their release events (Corre et al., 2018). . Like opioids, cannabinoids are also thought to indirectly increase DA release by activating G_i/o coupled CB1Rs on GABA neurons to remove an inhibitory GABAergic tone (Melis et al., 2004; Peters et al., 2020). The cannabinoid and opioid systems also show substantial overlap at the molecular and cellular level (Vigano et al., 2005), leading to speculation that cannabinoids might influence the abuse liability of opioids. The first report to demonstrate neurochemical cross-talk used microdialysis to show that μ-opioid receptor antagonism is sufficient to block the DA releasing effects of both cannabinoids and heroin (Tanda et al., 1997).

Cannabinoids and opioids are also well known to produce cross-tolerance to several shared neurobehavioral effects (Gerak et al., 2015; Hine, 1985; Manzanares et al., 1999; Thorat and Bhargava, 1994; Vigano et al., 2005). Exposure to THC produces tolerance to the analgesic and hypothermic effects of morphine in a manner independent from changes to CNS μ-opioid receptors (Thorat and Bhargava, 1994). It is possible that cross-tolerance between the two drug classes is mediated by cannabinoid receptors and endogenous cannabinoid signaling as it has been reported that morphine-tolerant rats exhibit a reduction in cannabinoid receptor levels and receptor functionality, as well as alterations in endocannabinoid levels (Viganò et al., 2003; Vigano et al., 2005). Furthermore, multiple studies confirm the role of various opioid receptors in cannabinoid dependence and withdrawal (Castañé et al., 2003; Lichtman and Martin, 2005). While these findings support the existence of cross-talk and cross-tolerance between opioids and cannabinoids, others report cross-sensitization can develop to drug-evoked stereotypy (Cadoni et al., 2001). Thus, it is important to consider that cannabinoids and opioids may produce tolerance to some neurobehavioral effects, but sensitization to others. Most relevant to our study is the question of whether cannabinoids can produce a cross-tolerance or cross-sensitization to the rewarding properties of opioids. Repeated THC exposure and CB1R antagonism is sufficient to induce tolerance to and block morphine-induced place preference, respectively (Jardinaud et al., 2006; Navarro et al., 2001). Additionally, a microdialysis study by Cadoni et al. found that rats with a history of THC exposure exhibit a reduced DA response to morphine in the NAc shell but an increased response in the NAc core (Cadoni et al., 2008). . However, due to issues of temporal resolution and differences in pharmacodynamics, it remains unknown whether repeated aminoalkylindole cannabinoid exposure alters opioid-induced changes in transient DA release events. Again, we believe it is particularly important to study DA transients in isolation because this unique form of DA transmission is sufficient in concentration to occupy low-affinity D1Rs and promote drug-induced neural plasticity (Barrientos et al., 2018; Dreyer et al., 2010).

Due to outstanding historical controversy on the subject, we first sought to clarify whether WIN and heroin evoke DA transients in the awake and behaving rat. Then, we tested whether a history of WIN exposure produces tolerance to either its own or heroin’s DA-releasing effects. Needing more of a drug to achieve a desired effect (i.e., tolerance) is a DSM-V criterion used to diagnose substance dependence (O’Brien, 2011). If DA is important for drug reward (Di Chiara et al., 2004) or to motivate drug seeking (Volkow et al., 2017), a diminished ability to evoke DA release could promote the use of larger quantities and more potent doses—thereby facilitating the addiction process. To accomplish this goal, we applied FSCV to measure DA transients in the NAc shell of freely moving and behaving adult rats. We chose WIN because it is an aminoalkylindole cannabinoid that is structurally comparable to JWH-018/JWH-073 and, it is one of the most frequently studied synthetic cannabinoids in the scientific community.

2. Materials and methods

2.1. Subjects and Surgery

All experiments were conducted in catheterized Long-Evans adult male rats, supplied by Charles River. While many studies are being conducted in adolescent rats, it is also worthwhile to study the neurobiological consequences of cannabinoid exposure in the developed brain. Cannabis abuse amongst adults is increasing globally (Copeland and Swift, 2009) and adult professionals are known to turn to synthetic cannabinoids, in particular, to evade positive drug tests (Ammann et al., 2011; Berry-Cabán et al., 2012; Every-Palmer, 2011; Johnson et al., 2011). Subjects were housed individually in a temperature-controlled environment on a 12 h light/dark schedule with access to food and water ad libitum. All experiments were conducted in the active phase. Rats (275–325g at the time of surgery) were placed under isoflurane anesthesia (5% induction, 2% maintained) for surgery conducted in a stereotaxic apparatus. A guide cannula that mates with a micromanipulator was implanted to be aimed at the NAc shell (+1.7AP, +0.8ML relative to bregma). The shell region of the NAc was targeted because the initial effects of drugs on DA concentration are most pronounced in this region (Di Chiara et al., 2004). In addition, an Ag/AgCl reference electrode was implanted on the contralateral side of the brain. Rats were given three days to recover before experiments were conducted. The University of Maryland Baltimore or the University of Colorado Denver Institutional Animal Care and Use Committee approved all experiments and procedures in advance.

2.2. General pharmacology

WIN (Tocris) was freshly suspended in a 1:1:18 ratio (in 20 total parts) of ethanol, emulphor (Alkamuls EL‐620; Rhodia Cranbury, NJ) and saline (0.9%) the morning of each treatment day. The aforementioned 1:1:18 ratio solution was used for all cannabinoid vehicle treatments. During electrochemical assessments, cumulative doses were administered every 15min as illustrated in Figures 4A and 5A. Heroin (diamorphine HCl; provided by the NIDA drug supply program) was dissolved in 0.9% sterile saline at a concentration of 17mg/ml to create a stock solution. Individual injections were prepared at point body weight the morning of FSCV experimentation. During FSCV recordings, cumulative doses were administered every 10min as illustrated in Figure 6A. All drugs were administered intravenously (IV). Pharmacology and FSCV interaction considerations: In addition to previously showing that this vehicle fails to alter DA release [28,32], we have repeatedly shown that multiple vehicle injections do not influence DA release events [21,31] and confirmed [31] accounts from the historical literature [33,34] that electrode sensitivity remains stable over the course of our in vivo recording sessions.

2.3. WIN treatment history and dose selection rationale

Rats received intravenous (IV) WIN twice daily for five consecutive days; recording sessions occurred at the time the tenth injection would occur. Dosage increased across days as follows: day 1 (0.2 and 0.2mg/kg), day 2 (0.4 and 0.4mg/kg), day 3 (0.6 and 0.6mg/kg), day 4 (0.8 and 0.8mg/kg), day 5 (0.8mg/kg and recording). The first injection occurred at the onset of the active cycle (i.e., 10am). The second injection and/or recording sessions occurred 6hrs into the active cycle. (i.e, 4pm). The chronic WIN exposure design was adapted from the existing literature (Mavrikaki et al., 2010; Moranta et al., 2009); for example, (Moranta et al., 2009) reported that 9 treatments of 2–8mg/kg WIN IP administered twice/d was sufficient to produce unique changes in gene expression vs. acute WIN treatment. We then converted the IP dose range of 2–8mg/kg to an IV dose range of 0.2–0.8mg/kg because the IV route produces more comparable effects on bioavailability resulting from inhalation vs. the IP route. In addition, similar to the inhalation route, the IV route of administration reduces lag-time. We sought to administer WIN in a manner that minimized lag-time because the rapid onset of abused drugs increases their motivational value and is necessary to produce neural adaptations that accompany the development of dependence (Samaha and Robinson, 2005).

2.4. Behavioral Pharmacology

Separate groups of either vehicle or WIN-treated rats were tested in a sequential series of behavioral and pharmacological tests that are known to measure cannabimimetic effects (Figure 1A). After allowing 30min to acclimate to the testing room, baseline temperature and antinociception values were determined. Temperature scores were reported as change from baseline; antinociception scores were reported as %maximal possible effect. The %maximal possible effect was calculated as the percentage difference between the post-drug response and the baseline response, divided by the difference between the maximum response and the baseline response. For temperature assessments, rats were individually restrained while core temperature was rectally measured as previously described [28]. For antinociception assessments, rats were placed on a hot plate (55°C) for a maximum of 30s; time was stopped when the rat withdrew both hind paws from the hot plate. After baseline assessments were taken, rats were placed back into their home cage for 15min before receiving a single injection (IV) of WIN. Vehicle treated rats received either 0.02, 0.08, or 0.2mg/kg IV WIN; WIN-treated rats received either 0.2, 0.6, or 0.8 mg/kg IV WIN. Ninety seconds after IV injection, antinociception was reassessed and the rat was returned to their home cage. After 120s, catalepsy was assessed by placing the rats front arms on a metal bar (1 cm in diameter, positioned 10 cm above the ground) as previously described [29,30]. The amount of time until both front paws returned to the ground was measured, with a 3min maximum time allowed. The following rating scale was then applied: 1, 0–25 s; 2, 26–60 s; 3, 61–110 s; 4, 111–180 s; 5, >180s to determine a catalepsy score [29]. After allowing for a maximum of 3min, a final core temperature measurement occurred.

2.5. FSCV

Voltammetric recordings were conducted by lowering a glass-encased carbon fiber microelectrode using a micromanipulator that fits inside the implanted guide cannula. An initial waveform (−0.4V to 1.3V; 400V/s) was applied which allowed for the detection of DA from cyclic voltammograms taken every 100ms. To increase electrode sensitivity the waveform was first applied at 60Hz for ~30min but reduced to 10Hz before experimentation. Data was continuously collected in 60s files for the duration of each recording. Recording continued for 15min following each WIN injection and 10min following each heroin injection. To extract the DA component, principle component regression was applied to the raw voltammetric data. Specifically, we used recording-specific training sets (n=7/analyte; DA and pH) to produce pH and background subtracted (10 consecutive scans) DA concentration files for transient analysis. To quantify DA concentration, calibration factors were pre-determined using linear regression as previously described [31]. The resulting calibration factors allow us to calculate molar concentrations of DA in a session-specific manner using total background current. The color plots and concentration x time files in illustrative figure were smoothed using the built in TarHeelCV smoothing option (eight-point nearest neighbor smoothing kernel); voltammetric data were not smoothed prior to data quantification. At the end of experimentation, rats were terminally anesthetized with an overdose of urethane (3 g/kg), and long high-amplitude, constant-current pulses (0.6 166 mA for 10s) were applied to each carbon fiber microelectrode to confirm that recordings occurred in the NAc shell. Finally, rats were transcardially perfused with 4% paraformaldehyde and their brains prepared for histological reconstruction.

2.6. Transient Analysis:

As previously described (Schelp et al., 2018), for every 60s recording (Figure 2A), a peak-threshold polynomic line (Figure 2B) was fitted to each set of DA concentration data using the following equation:

$$p(x)=p_1{x}^2+p_2{x}^2+p_3$$

Figure 2:

Signal processing and quantification of DA release events. Representative color plot (A) and corresponding DA concentration trace (B) show transient WIN evoked DA release events. In the color plot (A), voltammetric current (z-axis) is plotted against applied scan potential (y-axis) and time (x-axis). To quantify the frequency and amplitude of drug-induced changes in DA release, a peak-threshold polynomic line is first fit to the data (B, red line). A second line is then fit to the data in order to set a relative zero (C, red line). Only DA release events with concentrations that were 1.5 standard deviations above relative zero (inverted orange triangles) were included in the data analysis.

The coefficients (P1, P2, P3) with the largest R2 value were assigned for each individual 60s DA concentration file. The degree of the polynomial was determined by finding the lowest Akike information criterion (AIC) score with the following equation:

$$AIC=n\ast log\left(\frac{RSS}{n}\right)+2K$$

A second fitted line (Figure 2C) was generated to establish a relative zero (figure 2B) using the following equation:

$$p(x)=p_1{x}^n+p_2{x}^{n-1}+\cdots p_{n}x+p_n+1$$

To reduce type 1 error, only peaks above 1.5 standard deviation from the ‘baseline’ that were greater than 0.5s apart were analyzed (Figure 2C; orange triangle). If multiple peaks occurred within the 0.5s period, the highest concentration event was reported.

2.7. Data analysis and statistics

We first assessed whether data were normally distributed (Shapiro-Wilk). For behavioral pharmacology tests, if data were normally distributed, we proceeded with either one-way ANOVA (>2 groups; i.e., within group dose comparison) or unpaired t-test (2 groups; i.e., between group 0.2mg/kg dose comparison). If data were not normally distributed, we proceeded with either one-way ANOVA on ranks (>2 groups) or Mann-Whitney U test (2 groups). For DA assessments, we performed either paired t-test (vehicle group within group dose comparison), repeated one-way measures ANOVA (WIN group within group dose comparison), or unpaired t-test (between group comparison a the 0.2mg/kg dose). Criterion for significance was predetermined to be p<0.05.

3. Results

3.1.1. Behavioral tolerance to the cannabimimetic behavioral and physiological effects of WIN:

To assess whether the selected dosing regimen was sufficient to produce tolerance, we tested separate groups of either vehicle or WIN-treated rats in a sequential series of behavioral and pharmacological tests that are known to measure cannabimimetic effects (Figure 1A). In vehicle-treated rats, a relatively low-dose range of WIN (0.06–0.2mg/kg IV) changed scores of antinociception (F_(2,20)=84.6, p<0.001), catalepsy (F_(2,20)=50.1, p<0.001), and hypothermia (F_(2,20)=19.6, 231 p<0.001; Figure 1B-D). Tukey-post hoc tests showed that these main effects occurred in a dose-dependent manner. In the antinociception test the 0.08mg/kg effect was significantly greater than that observed at 0.06 or 0.02mg/kg, and the 0.06mg/kg effect was significantly greater than that observed at 0.02mg/kg. In the catalepsy test the 0.08mg/kg effect was significantly greater than that observed at 0.02, and the 0.06mg/kg effect was significantly greater than that observed at 0.02mg/kg. In the hypothermia test the 0.08mg/kg effect was significantly greater than that observed at 0.02mg/kg, and the 0.06mg/kg effect was significantly greater than that observed at 0.02mg/kg. In WIN-treated rats a higher dose-range of WIN (0.2mg/kg-0.8mg/kg IV) was necessary to produce comparable changes in antinociception (F_(2,23)=808, p<0.001), catalepsy (H₍₂₎=16.9, p<0.001), and hypothermia (F_(2,23)=64, p<0.001; Figure 1B-D). As in the vehicle treated rats, Tukey-post hoc analysis revealed the dose-dependency of our main effects. In the antinociception test the 0.8mg/kg effect was significantly greater than that observed at 0.6 or 0.2mg/kg, and the 0.6mg/kg effect was significantly greater than that observed at 0.2mg/kg. In the catalepsy test the 0.8mg/kg effect was significantly greater than that observed at 0.2, and the 0.6mg/kg effect was significantly greater than that observed at 0.2mg/kg. In the hypothermia test the 0.8mg/kg effect was significantly greater than that observed at 0.2mg/kg, and the 0.6mg/kg effect was significantly greater than that observed at 0.2mg/kg. Because the 0.2mg/kg dose was common between groups, we performed either an unpaired t-test or Mann-Whitney U to determine if this dose was less effective at inducing cannabimimetic effects in WIN-treated vs. vehicle-treated rats. Our analysis revealed that 0.2mg/kg WIN was significantly more effective at inducing cannabimimetic effects in tests of antinociception (U=84; p<0.001), catalepsy (U=84; p<0.001), and hypothermia (t₍₁₃₎=5.15; p<0.001) in vehicle vs. WIN-treated rats.

Figure 1:

Tolerance developed to several cannabimimetic effects: After receiving either chronic vehicle (blue) or WIN (orange) treatment, three behavioral/physiological tests occurred sequentially (top, timeline). WIN was less potent in tests of antinociception (A), catalepsy (B), and hypothermia (C). in WIN-treated (orange) vs. vehicle-treated (blue) rats. Comparable maximal effectiveness was observed in WIN-treated rats when we increased WIN dosage.

3.1.2. Neurochemical tolerance to the DA-releasing effects of WIN:

To investigate whether repeated exposure to an aminoalkylindole cannabinoid produces tolerance to its DA releasing effects we tested WIN exposed rats and their vehicle controls with IV doses of WIN while concurrently performing FSCV. All recordings were done in awake and behaving rats. When applicable, doses were administered in a cumulative, ascending order. Figure 3 shows the results of an illustrative recording performed in a WIN-treated rat. The frequency and amplitude of transient DA release events visually increased as WIN was administered in a cumulative, ascending manner.

Figure 3:

Illustrative recording session in which WIN is administered to a WIN-treated rat. Stitched color plots [voltammetric current (z-axis) x applied scan potential (y-axis) x time (x-axis)] are shown above corresponding DA concentration traces. Vehicle (A), 0.2mg/kg (B), and 0.8mg/kg WIN (C) are administered in cumulative, ascending IV doses while FSCV measurements of DA release events occur in near real-time. Dose dependent increases in thee frequency and amplitude of DA release events can be observed by the larger and more frequent green dots at a potential of +0.6V in the color plots.

3.1.3. Tolerance develops to the frequency at which WIN evokes DA transients

WIN significantly changed the frequency of DA release events in both vehicle (F_(1,11)=24.43, p<0.001) and WIN-treated (F_(2,29)=47, p<0.001) rats (Figure 4). In comparison to vehicle, Tukey post-hoc tests revealed that the 0.2mg/kg WIN dose significantly increased the frequency of DA release events in vehicle-treated (p<0.001) but not in WIN-treated (n.s.) rats. However, we did find that WIN still effectively increased the frequency of DA release when we increased the dose to 0.8mg/kg IV (p<0.001). We then used Kruskal-Wallis One Way ANOVA on Ranks followed by Dunn’s post-hoc method to compare the effects of the common WIN dose (0.2mg/kg IV) between treatment groups. 0.2mg/kg WIN significantly increased the frequency of DA release events in vehicle vs. WIN treated rats (H_(6,10)=4.706 with 1 df; p=0.030); Dunn’s Q=2.169; p<0.05).

Figure 4.

Chronic WIN treatment produces tolerance to the frequency at which DA release events occur. A. Cumulative dosing scheme for chronic vehicle (VEH; orange), and WIN (blue) treatment. The colored numbers denote cumulative dose whereas the black numbers denote additive dose and time between treatments (15min). B. WIN increased the frequency of DA release events but was less potent in chronically WIN-treated rats. A higher dose of WIN (0.8 vs. 0.2 mg/kg IV) was required to produce a significant increase in DA release vs. vehicle-treated rats. Vehicle also increased the frequency of DA release events in WIN-treated vs. vehicle-treated rats. Horizontal bars indicate within group dose comparison; vertical bars indicate between group comparison. *p<0.05.

3.1.4. Tolerance develops to the amplitude of WIN-evoked DA transients

Similar trends were observed when the amplitude of each DA release event was quantified (Figure 5). In WIN-treated rats, One-Way Repeated Measures ANOVA revealed that WIN significantly changed the amplitude of DA release events (F_(2,29)=27.5, p<0.001). Tukey post-hoc tests showed only the 0.8mg/kg dose produced a significant increase in amplitude vs. vehicle (p<0.001); this dose also significantly increased the amplitude of transient events vs. the 0.2mg/kg dose. By contrast, in vehicle-treated rats the 0.2mg/kg WIN dose significantly increased the amplitude of DA release events vs. vehicle (F_(1,11)=28.997, p<0.001). Thus, we also performed a Kruskal-Wallis One Way ANOVA on Ranks followed by Dunn’s post-hoc method to compare the effect of this dose on amplitude between groups. 0.2mg/kg WIN significantly increased the amplitude of DA release events in the vehicle vs. the WIN-treated rats (H_(6,10)=6.233 with 1 df; p=0.013); Dunn’s Q=2.495; p<0.05).

Figure 5.

Chronic WIN treatment attenuates the concentration of DA observed in each WIN-evoked transient release event. A. Cumulative dosing scheme for chronic vehicle (VEH; orange), and WIN (blue) treatment. The colored numbers denote cumulative dose whereas the black numbers denote additive dose and time between treatments (15min). B. WIN increased the amplitude of DA release events but was less potent in chronically WIN-treated rats. A higher dose of WIN (0.8 vs. 0.2 mg/kg IV) was required to produce a significant increase in the amplitude of release events vs. vehicle-treated rats. Horizontal bars indicate within group dose comparison; vertical bars indicate between group comparison. *p<0.05.

3.1.5. Cross-tolerance develops to the frequency at which heroin evokes DA transients

According to One-Way Repeated Measures ANOVA, heroin significantly changed the frequency of DA release events in both vehicle (F_(4,24)= 8.483, p=0.009) and WIN-treated (F_(4,29)=3.644, p=0.02) groups (Figure 6A,B). In vehicle treated rats, Tukey post-hoc tests revealed that the 1.0 mg/kg heroin dose significantly increased the frequency of DA release events vs. vehicle (p=0.013) and 0.1 mg/kg heroin (p=0.017). However, in WIN treated rats, there was not a tested dose of heroin that significantly increased the frequency of DA release events vs. vehicle (n.s.). Two-way repeated measures ANOVA revealed main effects of dose (F_(4,36)=11.65; p<0.001) and group (F_(4,36)=7.0; p=0.027) without significant interaction. Tukey post-hoc analysis did not reveal a significant different between groups at any individual heroin dose.

Figure 6.

Chronic WIN treatment attenuates the frequency of heroin-evoked DA transients. A. Cumulative dosing scheme for heroin used in the present FSCV study. Black numbers denote cumulative dose; white numbers denote additive dose; 10min were allowed between injections. B. Heroin dose-dependently increased the frequency of DA release events but was less effective in chronically WIN-treated rats. In WIN-treated rats, heroin did not significantly increase the frequency of DA transients vs. vehicle at any dose tested. C. Heroin did not significantly increase the amplitude of DA transients in a dose-dependent manner. *p<0.05.

3.1.6. Heroin does not dose-dependently increase the amplitude of DA release events

Unlike WIN, One-Way Repeated Measures ANOVA revealed that heroin did not increase the amplitude of DA release events in a dose-dependent manner (Figure 6A,C). Two-Way Repeated Measures ANOVA did not reveal any main effect or interaction.

4. Discussion

In the present study, we first demonstrated that WIN and heroin increase dopamine transient release events in the NAc shell of awake and freely moving rats. Whereas both drugs increase the frequency of DA release events in a dose-dependent manner, only WIN significantly increased their amplitude. We have previously found that unique drug classes differentially alter the frequency and amplitude of DA release events. For example, the benzodiazepine valium dose-dependently increases the frequency while decreasing the amplitude of DA transients (Schelp et al., 2018). And, heroin has been previously shown to not increase the amplitude of DA release events as detected by FSCV (Pattison et al., 2012).

We then determined that a chronic WIN dosing regimen was sufficient to produce tolerance to several well-accepted cannabimimetic effects (Little et al., 1988; Wiley and Martin, 2003). In confirmation of previous studies (Hama and Sagen, 2009; Nealon et al., 2019), WIN-treated rats displayed a rightward shift in the dose-response relationship (0.002–0.8 mg/kg IV) across all behavioral/physiological measures when compared to vehicle-treated controls. Thus, we used FSCV to investigate whether the same pharmacological history produced tolerance to the DA releasing effects of WIN and cross-tolerance to the DA releasing effects of heroin. After chronic WIN exposure, both drugs were less effective at increasing the frequency of DA release events in the NAc shell of adult male rats. WIN was also less effective at increasing the amplitude of DA transients following chronic cannabinoid exposure.

These results bear several timely implications. In addition to confirming that a member of the aminoalkylindole class of synthetic cannabinoids dose-dependently increases the frequency of transient DA release events, we show that tolerance and cross-tolerance can develop to this effect in a manner that might advance the addiction process. The DSM-V lists a need for more drug to achieve a desired effect as one of the requisite criteria for a diagnosis of substance dependence (O’Brien, 2011). If DA mediates the rewarding (Di Chiara et al., 2004) or motivational (Volkow et al., 2017) properties of abused drugs, then a diminished ability to increase brain DA levels may lead animals to compensate for a loss in drug-induced hedonia by increasing drug consumption. While sensitized DA responses can occur under a variety of conditions, mounting evidence suggests that a hypo-responsive DA release state develops coincident with chronic drug abuse. The electrophysiology literature shows that chronic drug exposure blunts the ability of abused drugs to increase DA neural activity (Bailey et al., 2001; Diana et al., 1993; Melis et al., 2005). In agreement with these studies, several microdialysis and FSCV studies have reported that abused drugs are less effective at increasing accumbal DA concentration following chronic exposure to a variety of abused drugs (Budygin et al., 2007; Calipari et al., 2014; Koob, 1992; Oleson et al., 2009). However, we acknowledge that many variables, including: neurodevelopmental stage, sex, pharmacological history, terminal field (e.g., core vs. shell vs. tubercle), and genetics, could influence our results. The exclusion of female rats, in particular, is a limitation of the current study as they were not studied here.

Previous investigations into the neurobehavioral consequences of chronic WIN exposure reveal that repeated administration of aminoalkylindole cannabinoids produce an array of behavioral and neurobiological changes. Far from straight forward, these changes are impacted by multiple variables. Considerable attention is currently being devoted to neurodevelopment because age-dependent cannabinoid effects are increasingly recognized as a major distinguishing feature of these drugs’ influence (Hurd et al., 2019). Adolescence is a particularly sensitive neurodevelopmental timepoint during which circuit connectivity can be adapted through experience-dependent synaptic plasticity. This period may be especially vulnerable to cannabinoid exposure as it is marked by the maturation of the endocannabinoid system within mesocorticolimbic structures that are central to learning and decision making (Bossong and Niesink, 2010; Heng et al., 2011; Meyer et al., 2018). This perspective is substantiated by studies suggesting adolescent cannabis use may combine with genetic risk factors to predispose individuals to drug abuse and other psychiatric disorders (Hurd et al., 2019). Thus, the consequences of aminoalkylindole cannabinoid abuse might manifest in a unique, or more pronounced fashion if administered during this critical period. In agreement with this perspective, adolescent use of synthetic cannabinoids is correlated with increases in emergency department visits relating to CNS depression and seizures, as well as increased substance abuse vs. age matched controls (Anderson et al., 2019; Martz et al., 2018). Preclinical studies also corroborate adolescent aminoalkylindole cannabinoid use as a risk factor for potentiated use of other drugs of abuse as well as comorbid conditions. Outbred adolescent CD1 mice given 5-days WIN treatment showed an increase in their preference for ethanol and an increase in anxiogenic behavior vs. vehicle controls (Frontera et al., 2018). Adolescent exposure to cannabinoids has also been shown to dynamically affect neurobehavioral and epigenetic responses to psychostimulants that target mesocorticolimbic DA. For example, Scherma & colleagues recently reported that a 7-day treatment of WIN induces cross-sensitization to cocaine in adolescent but not adult rats, and these behavioral outcomes align with multiple epigenetic alterations in the PFC along with increased levels of mitogen-activated protein kinase and blunted decreases in AMPA receptor availability in the NAc (Scherma et al., 2020). Finally, increased impulsivity has been observed in adult mice following chronic adolescent WIN exposure (Johnson et al., 2019).

While possibly distinct from the effects of chronic exposure in adolescence, substantial neurobehavioral changes occur following chronic WIN exposure in adulthood as well. The relevance of such data may be underscored by increases in professional use of aminoalkylindole cannabinoids in order to evade standard drug screens (Ammann et al., 2011; Berry-Cabán et al., 2012; Johnson et al., 2011). In preclinical tests, the most common findings point to deficits in learning and working memory. In adult rats, working memory deficits aligning with neuronal perturbations in the hippocampus have been shown under acute WIN treatment (Goonawardena et al., 2010). Additional hippocampal alterations in morphology have been observed in rats following chronic WIN treatment such as decreased spine density, and region-specific alterations in MAP-2 staining which is used to measure changes in neuronal differentiation (Candelaria-Cook and Hamilton, 2014; Lawston et al., 2000). Likewise, adult rats chronically treated with the synthetic cannabinoid HU210 showed an increase in GABA_A receptor density in the hippocampus, an effect not observed in adolescent rats given identical treatment (Verdurand et al., 2010). Changes to prefrontal substrates and related behaviors have also been noted that differentiate adult and adolescent use. Alterations in dendritic morphology within PFC pyramidal neurons have been shown to exclusively be expressed in adult but not adolescent rats following repeated WIN treatment, revealing increased dendritic arborization and spine density (Carvalho et al., 2016). Similarly, a 7-day WIN treatment regimen in adult rats decreased extinction of fear memory while concomitantly occluding WIN-induced decreases in CB1R density and GABAergic inhibition in the PFC (Lin et al., 2008). Relating to a more systems-wide perspective, both acute and chronic WIN treatment reduce recognition memory in adult mice, with the latter correlated with decreased functional connectivity between limbic areas, including the PFC and hippocampus (Mouro et al., 2018). Other disparate yet noteworthy effects of WIN treatment in adult rodent studies bear mentioning. Chronic WIN exposure in mice during adulthood but not adolescence has been shown to disrupt reversal learning, further implicating aminoalkylindole cannabinoids in cognitive disfunction (Johnson et al., 2019). Chronic WIN self-administration has also been shown to reduce the psychotomimetic effects of PCP in a drug-induced schizophrenic rat model. It is germane to note that schizotypic positive symptoms have been correlated with striatal DA in both human and rodent studies, and together, these data suggest a cannabinoid-DA interaction that may influence such symptomology (Agid et al., 2007; Brisch et al., 2014; Spano et al., 2013). On a related note, acute WIN treatment in rats result in decreased accumulation of the DA precursor DOPA in the NAc, indicative of inhibited biosynthesis of DA within projections terminating there. Tolerance to this decrease was observed following a history of chronic WIN treatment (Moranta et al., 2009).

Cannabinoids and opioids similarly increase DA release by removing an inhibitory GABAergic tone from DA neurons. For a comprehensive description of the mechanisms by which eCBs are thought to control midbrain DA release, we refer the reader to Covey et al (2017) (Covey et al., 2017). A common model used to describe endocannabinoid modulation of neural activity is known as depolarization-induced suppression of inhibition (DSI). Under conditions of DSI, stimulation-dependent mobilization of endocannabinoids may decrease presynaptic transmission onto midbrain DA cells (Covey et al., 2017). In brief, as G_i/o-coupled CB1Rs are exclusively localized to GABAergic and glutamatergic terminals in the midbrain (Julian et al., 2003; Mátyás et al., 2008), their activation by retrogradely released endocannabinoids inhibit neurotransmitter release onto DA neurons. In terms of GABAergic synapses onto DA neurons, the consequence is a disinhibited cell and increased DA transmission (Melis et al., 2013; Riegel and Lupica, 2004; Szabo et al., 2002; Wang et al., 2015). Our observations of WIN-induced increases to DA transients recorded with FSCV likely result from similar mechanisms of DAergic disinhibition. Opioids also promote DA release through disinhibition by activating G_i/o coupled μ-opioid receptors located on GABA neurons projecting onto midbrain DA neurons—highlighting the important overlap between opioid and cannabinoid functional neuroanatomy (Chieng et al., 2011; Matsui et al., 2014). In light of this, it was recently reported that self-stimulation via optogenetic inhibition of VTA GABA neurons in mice is blocked by heroin injection, validating this DA disinhibition model in the context of opioid reinforcement (Corre et al., 2018).

The mechanistic convergence of the endocannabinoid and opioid reward systems is complimented by other neurofunctional interactions. The endocannabinoid system is strongly implicated in the effects of morphine as prolonged treatment with CB1R agonists or antagonists attenuates naloxone withdrawal in morphine dependent animals and CB1R agonists have been shown to potentiate acquisition of morphine induced CPP (Rashidy-Pour et al., 2013; Rubino et al., 2000; Valverde et al., 2001). Intriguingly, converging evidence suggests the CB1R is colocalized with μ-opioid receptors on inhibitory presynaptic terminals within the VTA stemming from the rostral medial tegmental nucleus (RMTg) (Jalabert et al., 2011; Lecca et al., 2012; Matsui and Williams, 2011). Patch-clamp experiments showed that both morphine and WIN depress DA cell inhibitory postsynaptic currents induced by RMTg stimulation (Lecca et al., 2012). Furthermore, naloxone prevents the antagonistic action of rimonabant at CB1Rs and rimonabant abolished that of naloxone at μ-opioid receptors in the NAc core, suggesting allosteric interaction between receptors (Schoffelmeer et al., 2006). This receptor-mediated interaction is also suggested by one study showing co-expression of cannabinoid and opioid receptors leads to significant increases in BRET signal in HEK3 cells and antagonistic interactions in receptor signaling between these receptors have been evidenced in striatal tissue via GTPyS binding assays (Rios et al., 2006).

Chronic cannabinoid exposure is associated with CB1R desensitization and down-regulation. CB1R density has been found to be significantly decreased in human cannabis dependent males compared to control subjects (D’Souza et al., 2015; Hirvonen et al., 2012). Down-regulation and desensitization of CB1Rs on GABA projections to DA neurons may very well dampen DA release in response to chronic cannabinoid exposure. A loss in CB1R function on midbrain GABA neurons would theoretically blunt the ability of exogenous CB1Rs to disinhibit DA neurons. In addition, because endocannabinoid-mediated DSI is necessary for all drugs of abuse to increase DA transients (Cheer et al., 2007), a loss in CB1Rs on midbrain GABA neurons may also explain why heroin is less effective at increasing DA release following chronic cannabinoid exposure. However, there are likely multiple mechanisms involved in cannabinoid-opioid cross-tolerance. The aforementioned mechanistic commonalities between the cannabinoid and opioid systems may potentially link such observed changes in receptor number across both receptor classes. For instance, rats that self-administer WIN exhibit decreases in both CB1R and μ-opioid receptor binding sites across multiple corticolimbic regions, including the NAc (Fattore et al., 2007). Furthermore, cross-tolerance to the antinociceptive effects of opioids and WIN is attenuated by either μ-opioid or CB1R antagonists. Altered GABA transmission paired with enhanced D2R sensitivity within the basal ganglia has also been observed following chronic WIN exposure in rats (Moreno et al., 2018). This may in turn be explained by receptor interactions on GABA cells, as CB1Rs and D2Rs are known to form heteromeric receptor complexes (Kearn et al., 2005). This same model may also hold explanatory power with our cross-tolerance effects to heroin; abnormal GABAergic firing brought on by CB1-D2R interactions may also modulate disinhibition via opioids. Future studies will be important to assess the generalization of alterations to GABAergic inhibition of DA neurons between cannabinoids and opioids. In general, multiple complex mechanistic interactions between neuronal responses within both cannabinoid and opioid systems exist (Maguma and Taylor, 2011). As much is unknown about these interactions, further experiments will be important to parse the mechanisms underlying cross-tolerance.

Highlights.

• The synthetic cannabinoid WIN dose-dependently increased dopamine release

• Repeated IV WIN treatment produced behavioral tolerance

• The same WIN treatment produced tolerance to its dopamine releasing effects

• Cross-tolerance to the dopamine releasing effects of heroin also occurred