Buprenorphine is a weak dopamine releaser relative to heroin, but its pretreatment attenuates heroin‐evoked dopamine release in rats

Abstract Aims The United States of America is currently in an opioid epidemic. Heroin remains the most lethal opioid option with its death rate increasing by over 500% in the last decade. The rewarding and reinforcing effects of heroin are thought to be mediated by its ability to increase dopamine concentration in the nucleus accumbens shell. By activating Gi/o‐coupled μ‐opioid receptors, opioids are thought to indirectly excite midbrain dopamine neurons by removing an inhibitory GABAergic tone. The partial μ‐opioid receptor agonist buprenorphine is a substitution‐based therapy for heroin dependence that is thought to produce a steady‐state level of μ‐opioid receptor activation. But it remains unclear how buprenorphine alters dopamine release relative to heroin and how buprenorphine alters the dopamine‐releasing effects of heroin. Because buprenorphine is a partial agonist at the μ‐opioid receptor and heroin is a full agonist, we predicted that buprenorphine would function as a weak dopamine releaser relative to heroin, while functioning as a competitive antagonist if administered in advance of heroin. Methods We performed fast‐scan cyclic voltammetry in awake and behaving rats to measure how heroin, buprenorphine HCl, and their combination affect transient dopamine release events in the nucleus accumbens shell. We also performed a complimentary pharmacokinetic analysis comparing opioid plasma levels at time points correlated to our neurochemical findings. Results Both buprenorphine and heroin produced changes in the frequency of transient dopamine release events, although the effect of buprenorphine was weak and only observed at a low dose. In comparison with vehicle, the frequency of dopamine release events maximally increased by ~25% following buprenorphine treatment and by ~60% following heroin treatment. Distinct neuropharmacological effects were observed in the high‐dose range. The frequency of dopamine release events increased linearly with heroin dose but biphasically with buprenorphine dose. We also found that buprenorphine pretreatment occluded the dopamine‐releasing effects of heroin, but plasma levels of buprenorphine had returned to baseline at this time point. Conclusion These findings support the notion that low‐dose buprenorphine is a weak dopamine releaser relative to heroin and that buprenorphine pretreatment can block the dopamine‐releasing effects of heroin. The finding that high‐dose buprenorphine fails to increase dopamine release might explain its relatively low abuse potential among opioid‐dependent populations. Because high‐dose buprenorphine decreased dopamine release before occluding heroin‐evoked dopamine release, and buprenorphine was no longer detected in plasma, we conclude that the mechanisms through which buprenorphine blocks heroin‐evoked dopamine release involve multifaceted pharmacokinetic and pharmacodynamic interactions.


| INTRODUC TI ON
The United States of America is currently in an opioid epidemic that is producing over 100 deaths per day. Fentanyl and heroin are the most lethal opioid options, with their respective death rates increasing by 520% and 533% between 2009 and 2016. 1 The rewarding and reinforcing effects of opioids are principally mediated through their action at the μ-opioid receptor. 2 Following the development of opioid dependence, Subutex ® and Suboxone ® are commonly prescribed pharmacotherapies for its treatment. 3 Subutex ® is buprenorphine; Suboxone ® is a 4:1 ratio of buprenorphine to naloxone. The partial μ-opioid receptor agonist buprenorphine is thought to act as a substitution therapy in opioid dependence; the μ-opioid receptor antagonist naloxone is primarily added to Suboxone ® in order to prevent intravenous (IV) buprenorphine abuse. 4,5 Heroin produces its rewarding and reinforcing effects by increasing dopamine concentration in the mesolimbic dopamine pathway, [6][7][8][9] , but also see Ettenberg et al, 1982. 10 By activating μ-opioid receptors on midbrain GABA neurons, opioids are thought to disinhibit dopamine neurons in the ventral tegmental area (VTA). 6 In the awake and behaving animal, dopamine neurons in the VTA-the origin of the mesolimbic pathway-fire in two distinct patterns 11 when at rest, dopamine neurons fire in a slow (2)(3)(4)(5) pacemaker pattern that contributes to a steady-state dopamine tone. This low-concentration tone is thought to activate high-affinity dopamine receptors (eg, D2) in terminal regions of the mesolimbic pathway. 11 Dopamine neurons also fire in phasic bursts (>20 Hz) under a variety of conditions, including when animals are presented with rewarding and motivationally salient stimuli. These phasic bursts of neural activity contribute to transient dopamine release events that are sufficient in concentration to occupy low-affinity dopamine D1 receptors. 11,12 Because only D1-expressing neurons in the nucleus accumbens (NAc)-the primary terminal field of the mesolimbic pathway-undergo dendritic plasticity following repeated drug exposure, 13 studying how opioids alter high-concentration dopamine transients may be particularly important for the neurobiology of addiction.
By investigating the effects of opioids on dopamine transients, we sought to build upon the existing microdialysis, electrochemistry, and electrophysiology literature. Previous publications show that heroin increases dopamine concentration in the NAc shell [6][7][8][9] and that buprenorphine pretreatment blocks several behavioral and neurochemical effects of heroin 14,15 -including its ability to increases accumbal dopamine concentration. This latter finding is also supported by the behavioral pharmacology literature, which shows that buprenorphine can function as a competitive antagonist in the presence of other opioid ligands. 16,17 Because buprenorphine is a partial agonist at the μ-opioid receptor, when administered in advance it should compete with and therefore obstructs the full agonist 6-acetylmorphine (ie, the active metabolite of heroin) from evoking dopamine release. Therefore, we predicted that buprenorphine would function as a relatively weak dopamine releaser in the absence of heroin, but its pretreatment would block the dopamine-releasing effects of heroin.
To test these predictions, we used fast-scan cyclic voltammetry (FSCV) to measure drug-induced changes in dopamine release in the awake and freely moving rat. While monitoring transient release events in the NAc shell, we treated rats with ascending intravenous (IV) doses of either: heroin alone or buprenorphine followed by heroin ( Figure 1). To further investigate how opioid plasma dynamics relate to opioid-evoked changes in dopamine transients, we performed a complimentary pharmacokinetic (PK) study. We selected our dose ranges from a clinical investigation 18 that studied how ascending doses of IV buprenorphine and heroin alter the subjective drug experience-observations that are impossible to assess in the rat. Our FSCV findings are discussed within the context of these previously reported dose-dependent subjective effects. that buprenorphine pretreatment occluded the dopamine-releasing effects of heroin, but plasma levels of buprenorphine had returned to baseline at this time point.

Conclusion:
These findings support the notion that low-dose buprenorphine is a weak dopamine releaser relative to heroin and that buprenorphine pretreatment can block the dopamine-releasing effects of heroin. The finding that high-dose buprenorphine fails to increase dopamine release might explain its relatively low abuse potential among opioid-dependent populations. Because high-dose buprenorphine decreased dopamine release before occluding heroin-evoked dopamine release, and buprenorphine was no longer detected in plasma, we conclude that the mechanisms through which buprenorphine blocks heroin-evoked dopamine release involve multifaceted pharmacokinetic and pharmacodynamic interactions.

| Subjects and surgery
Catheterized male Long-Evans rats, supplied by Charles River, were singly housed under a 12:12-h light-dark cycle with a 10 AM to 10 PM active period (dark phase). All experiments were conducted in the active phase. Rats (275-325 g 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 dopamine concentration are most pronounced in this region. 19,20 In addition, a Ag/AgCl reference electrode was implanted on the contralateral side of the brain. Rats were given three days to recover before experiments were conducted.

| 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 (−0.4 to 1.3 V; 400 V/s) was applied which allowed for the detection of dopamine from cyclic voltammograms taken every 100 ms. To increase electrode sensitivity, the waveform was first applied at 60 Hz for ~30 minutes but reduced to 10 Hz before experimentation. To extract the

| Transient analysis
As previously described, 24 for every 60 seconds recording, a peakthreshold ("cutoff") polynomic line was fitted to each set of dopamine concentration data using the following equation ( Figure 2): The coefficients (p 1 , p 2 , p 3 ) with the largest R 2 value were assigned for each individual 60 seconds dopamine concentration file. The degree of the polynomial was determined by finding the lowest Akaike information criterion (AIC) score with the following equation: A second fitted line was generated to establish a relative zero ("baseline") using the following equation: To reduce type 1 error, only peaks above 1.5 standard deviation from the "baseline" that were greater than 0.5 seconds apart were analyzed. If multiple peaks occurred within the 0.5 seconds period, the highest concentration event was reported.

| Pharmacology
Heroin and buprenorphine were prepared in sterile saline and briefly sonicated. In all experiments, opioids were administered intravenously using a cumulative-dosing approach. For the heroin-only condition, we administered vehicle (ie sterile saline) and then 0.27, 0.55, 1.12, and 2.24 heroin (mg/kg IV). Cumulative doses were administered every 10 minutes as illustrated in Figure 1A.
In the buprenorphine-heroin condition, we administered vehicle (ie sterile saline) and then ascending doses of buprenorphine: 0.011, 0.044, 0.177, 0.708 ( Figure 1B). Continuing in 10-minutes intervals, we then administered the same doses of heroin used in the heroin-only condition. Buprenorphine HCl was purchased from Sigma-Aldrich, and heroin was provided by the NIDA drug supply program.
Doses of buprenorphine were calculated as the weight of the HCl salt.

| Comparative dosing calculations
In the present study, we took advantage of a unique clinical investigation in which human participants were given ascending doses of intravenous buprenorphine and heroin-in addition to several other opioids, and then asked to subjectively rate the resulting experience/affect (eg "I feel good"). 18 In addition, the propensity to selfadminister each dose was assessed. To convert these previously compared opioid doses in the human literature for use in the rat, we applied the following allometric scaling equation. 25 The Km used for human = 37; the Km used for rat = 6.2. 25

| Pharmacokinetic analysis
To determine circulating concentrations of buprenorphine, heroin, and its metabolites, we performed a pharmacokinetic analy-
50 µL of blank Rat plasma and 50 µL of acetonitrile with 0.1% formic acid were added to each tube and vortexed for 5 minutes. The tubes were then centrifuged for 15 minutes at 20 000 g and transferred to an autosampler with inserts.

| Experimental sample preparation
Five microlitre of 50/50 ACN/Milli-Q, 5 µL of 100 ng/mL fentanyl standard, 50 µL of unknown experimental rat plasma sample, and 50 µL of ACN with 0.1% formic acid were added to fresh 1.5-mL microcentrifuge tubes. The tubes were vortexed for 5 minutes, centrifuged for 15 minutes at 20 000 g, and transferred to an autosampler vials with inserts.

| Description of the statistical analysis
Because parametric statistical tests such as ANOVA assume that experimental samples have equal variance with data that are normally distributed, we first performed tests of normality and equal variance. If data were determined to normally distribute and show equal variance, we assessed for significant differences in the data using parametric statistics (eg, ANOVA). In contrast, if data were determined to not fit these criteria, we proceeded to assess for significant differences in the data using nonparametric statistics. Final statistical analyses were performed using SigmaPlot 11.

| Determination of animal equivalent heroin dosage for rat
To assess the effects of heroin, buprenorphine, and their interac-

| Buprenorphine is a weak dopamine releaser vs. heroin and its pretreatment blocked heroin-evoked dopamine release
As illustrated in Figures 3 and 4, buprenorphine and heroin increased the frequency of dopamine release events by ~25% and ~60%, respectively. Amplitude did not change in a significant or lawful dose-dependent manner; however, it is possible that relatively high variability in this measure might be masking dose-dependent effects. Frequency data first passed tests of normality (Shapiro-Wilk) and equal variance; thus, we proceeded with parametric analyses. Criterion for significant was predetermined to be P < .05. All doses were tested cumulatively, with 10 minutes elapsing between each IV treatment. A one-way repeated-measures ANOVA revealed that heroin F 4,29 = 14.12, P < .05 ( Figure 3) and buprenorphine  Figure 5). Figure 6 illustrates working electrode lesion sites, thereby confirming that all recordings occurred in the NAc shell.

| Pharmacokinetic analysis of plasma opioid content suggests that buprenorphine is metabolized into an active metabolite that may contribute to our dopamine observations
To measure just how much buprenorphine and 6-acetylmorphine (6-AM; active heroin metabolite) is in plasma at different time points in our cumulative-dosing scheme, we performed an additional pharmacokinetic (PK) assay. We first treated naïve rats with the same cumulative doses of heroin used in our FSCV study ( Figure 1A) and withdrew blood to analyze heroin plasma content every 10 minutes.
We withdrew blood every 10 minutes to track concentration over the duration which we measured dopamine transients between cumulative treatments. Our data show that cumulative heroin dosing lawfully increases 6-AM plasma concentration ( Figure 7A). One-way ANOVA on ranks revealed a significant main effect of heroin dose naïve rats (n = 6) with cumulative doses of buprenorphine followed by cumulative doses of heroin ( Figure 1). We withdrew blood after vehicle and after each entire round of cumulative drug treatments.
Our data show that cumulative dosing of buprenorphine followed by heroin resulted in a comparable increase in 6-AM to that observed when heroin was administered by itself ( Figure 7B). An unpaired t test

| D ISCUSS I ON
In the present study, we used FSCV to measure how heroin, bu- The frequency of buprenorphine-evoked dopamine release events increased across the 0.01-0.04 mg/kg range before declining to a level of insignificance across the 0.18-0.7 mg/kg range. This latter observation confirms a previous microdialysis report demonstrating that buprenorphine alters accumbal dopamine concentration in a biphasic, dose-dependent manner. 29 One implication of these findings is that buprenorphine might only function as a dopamine-based agonist therapy for opioid dependence in the low-dose range. A bellshaped relationship between buprenorphine dose and dopamine release may also contribute to its relatively low IV abuse potential.
The opioid doses used in the present study were adapted from a double-blind, placebo controlled inpatient trial. We selected these doses so that we could then compare our neurochemical findings to subjective responses that are impossible to measure in ani- concentration. 30 Heroin is a morphine derivative that is rapidly deacetylated into 6-acetylmorphine in vivo. The antinociceptive and reinforcing effects of heroin are primarily attributed to 6-acetylmorphine activating μ-opioid receptor subtype1. 2,31 In addition to μ-opioid receptor subtype 1, drug self-administration studies suggest that δ-, but not κ-opioid receptors moderately influence heroin reinforcement. 2 The pharmacodynamics of buprenorphine are more complex. 30 Buprenorphine is known to bind to all major opioid receptors, although it displays a 10-fold lower affinity for δ vs μ and κ. While buprenorphine is generally described as a partial μ-opioid receptor agonist and a κ-opioid receptor antagonist, it can act as a mixed agonist and/or antagonist at all major opioid receptor classes. 16,17,30 In addition, buprenorphine acts at a recently identified opioid receptor known as the opioid-like 1 receptor (ORL-1R), which is thought to produce counter-opioid effects. 30 While our data do not provide definitive mechanistic insight into how buprenorphine biphasically alters dopamine release, we speculate that additional drug action at either the ORL-1 and/or the κ-opioid receptor may be involved because of their ability to produce counter-opioid behavioral effects.
The complimentary PK analysis offers additional insight into the complex pharmacokinetic and pharmacodynamic mechanisms through which buprenorphine alters opioid-evoked dopamine release. While the PK data confirm that our cumulative-dosing scheme lawfully increased plasma levels of 6-AM, we were surprised to find that buprenorphine was no longer detectable in plasma at a time Another possibility is that buprenorphine is metabolized into an additional active metabolite that functions as a competitive antagonist at the μ-receptor. Future studies are needed to identify if an active buprenorphine metabolite is blocking heroin-evoked dopamine release, whether this effect results from competitive action at the μ-, κ-, and/or other receptor targets, or whether the effect is due to buprenorphine's longevity in brain tissue itself. Opioid bioavailability would also be significantly lower following SL vs IV buprenorphine. Future studies should be conducted to compare the interaction of buprenorphine and heroin on opioid-evoked dopamine release in rats with a history of contingent drug self-administration because the neurochemistry of acute noncontingent administration of drugs to animals is quite different than what occurs in models of reinforcement such as self-administration. Thus, while it may be worth considering that Subutex ® /Suboxone ® prescription guidelines suggest that physicians begin treatment with 8 mg/70 kg sublingual (SL) buprenorphine-a dose that is comparable to the highest buprenorphine dose used in the present study (0.7 mg/kg AED in rat), it is likely that IV buprenorphine is more potent and effective at increasing dopamine release vs. the same dose of SL buprenorphine.
Notwithstanding these limitations, our data provide new insight into how clinically relevant doses of heroin, buprenorphine, and their combination influence dopamine release. While heroin dose-dependently increased the frequency of dopamine release events, buprenorphine only increased dopamine release in the low-dose range.
In addition, we found that buprenorphine pretreatment attenuates the dopamine-releasing effects of heroin. Because high-dose buprenorphine stopped increasing dopamine release and buprenorphine was no longer present in plasma, it is likely that its ability to blunt heroin-evoked dopamine release is not exclusively a result of buprenorphine competing with 6-AM at the μ-opioid receptor.

ACK N OWLED G M ENTS
Funding for this work was provided by NSF grant IOS-1557755, NIH grant R03DA038734, Boettcher Young Investigator Award, and NARSAD Young Investigator Award to EBO. We thank Esteban Loetz for technical assistance, the Pharmacology Shared Resource for the CU Cancer Center (P30CA046934) for performing the analysis of opioid plasma content, and the NIDA drug supply program for providing heroin.

CO N FLI C T O F I NTE R E S T
We have no conflicts to disclose.

AUTH O R CO NTR I B UTI O N S
EBO designed the experiment and obtained funding to conduct it.
DPI, RPL, TJE, HL-B, and EBO performed the experiments and statistical analysis; DPI, LRH, and EBO wrote the manuscript. LRH and EBO finalized the manuscript. All authors read and approved the final manuscript.

A PPROVA L O F TH E R E S E A RCH PROTO CO L BY A N I N S TITUTI O N A L R E V I E WER B OA R D
We need not require IRB review.

I N FO R M E D CO N S E NT
We did not require informed consent.

R EG I S TRY A N D TH E R EG I S TR ATI O N N O. O F TH E S TU DY/ TR I A L
We did not register a clinical trial for this study.

A N I M A L S TU D I E S
The University of Colorado Denver Institutional Animal Care and Use Committee approved all animal experiments and procedures in advance.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data which support these findings are available in the Oleson