Progress in Agonist Therapy for Substance Use Disorders: Lessons Learned from Methadone and Buprenorphine

Abstract

Substance use disorders (SUD) are serious public health problems worldwide. Although significant progress has been made in understanding the neurobiology of drug reward and the transition to addiction, effective pharmacotherapies for SUD remain limited and a majority of drug users relapse even after a period of treatment. The United States Food and Drug Administration (FDA) has approved several medications for opioid, nicotine, and alcohol use disorders, whereas none are approved for the treatment of cocaine or other psychostimulant use disorders. The medications approved by the FDA for the treatment of SUD can be divided into two major classes – agonist replacement therapies, such as methadone and buprenorphine for opioid use disorders (OUD), nicotine replacement therapy (NRT) and varenicline for nicotine use disorders (NUD), and antagonist therapies, such as naloxone for opioid overdose and naltrexone for promoting abstinence. In the present review, we primarily focus on the pharmacological rationale of agonist replacement strategies in treatment of opioid dependence, and the potential translation of this rationale to new therapies for cocaine use disorders. We begin by describing the neural mechanisms underlying opioid reward, followed by preclinical and clinical findings supporting the utility of agonist therapies in the treatment of OUD. We then discuss recent progress of agonist therapies for cocaine use disorders based on lessons learned from methadone and buprenorphine. We contend that future studies should identify agonist pharmacotherapies that can facilitate abstinence in patients who are motivated to quit their illicit drug use. Focusing on those that are able to achieve abstinence from cocaine will provide a platform to broaden the effectiveness of medication and psychosocial treatment strategies for this underserved population.

1. Introduction

Substance use disorder (SUD) is a serious public health problem that affects millions of people worldwide. Substances having the highest addiction liability include opioids (most recently prescription and other new synthetic opioids, such as fentanyl), psychostimulants (such as cocaine and methamphetamine), and nicotine. Nearly 13.5 million people use opioids worldwide (WHO, 2018). In the United States, over two million people are diagnosed with opioid use disorder (OUD), resulting in economic costs that exceed 500 billion dollars each year (HHS, 2018). Notably, synthetic opioids have been driving the recent increase in cocaine overdose-induced deaths (Khatri et al., 2018). Cocaine-related overdose deaths increased nearly 60% from 2010 to 2015 (1.35 to 2.13 per 100,000 individuals; McCall Jones et al., 2017). Today, cocaine remains the leading cause of overdose deaths among African Americans (CDC, 2018). Although many patients with SUD manage to remit without pharmacotherapy, many still require or benefit from medication assistance. Despite extensive research, pharmacotherapies for SUD have advanced slowly and only a handful of drugs have been approved by the U. S. Food and Drug Administration (FDA) for the treatment of opioid, nicotine, and alcohol use disorders (FDA, 2018). Currently, there are no approved medications for the treatment of cannabis, cocaine or other psychostimulant use disorders.

Medications currently approved by the U.S. FDA for SUD can be classified into two major categories – agonist replacement therapies (such as methadone and buprenorphine for OUD, and nicotine replacement therapy [NRT] and varenicline for smoking cessation) and antagonist therapies (such as naloxone and naltrexone; FDA, 2018). Of these, methadone is a mu opioid receptor full agonist, while buprenorphine and varenicline are partial agonists that bind at mu opioid and α4β2 nicotinic acetylcholine receptors, respectively. Full agonist therapies for tobacco use disorder, in the form of nicotine replacement (e.g., nasal sprays, patches, chewing gum, and inhalers that deliver nicotine) have also shown efficacy in promoting smoking cessation (Farsalinos and Niaura, 2019; Hajek et al., 2019; Stead et al., 2012). Although agonists themselves may have inherent abuse potential, they are highly effective in the prevention of withdrawal and the reduction of drug craving and relapse. Abuse potential can be further minimized by extended release and depot medication formulations (Blanco-Gandia and Rodriguez-Arias, 2018). As such, methadone and buprenorphine have become first-line pharmacological maintenance treatments for OUD (Stotts et al., 2009).

In contrast to agonist therapies, the opioid receptor antagonists naloxone and naltrexone lack abuse potential. Naloxone in particular is a first-line treatment for opioid overdose due to its efficacy in reversing opioid-induced respiratory depression. However, naloxone and naltrexone precipitate acute opioid withdrawal (Buajordet et al., 2004; Kim and Nelson, 2015) and thereby increase the risk of relapse. Clinical observations indicate low clinical success rates and poor compliance to naltrexone as a treatment for opioid abuse, as well as mecamylamine and DHβE (dihydro-β-erythroidine hydrobromide) nicotinic receptor antagonist treatments for nicotine use disorder (Blanco-Gandia and Rodriguez-Arias, 2018; Jordan and Xi, 2018).

In this review, we focus primarily on agonist therapeutic strategies in the treatment of OUD with the assertion that this strategy may be particularly cogent for developing medications to treat cocaine use disorders. We first describe the neural mechanisms underlying opioid reward and addiction and the rationale of agonist therapy in treatment of OUD, particularly agonist therapies that produce slow-onset, long-lasting effects that achieve functional outcomes such as limiting the rewarding efficacy and withdrawal effects of these drugs of abuse while producing minimally reinforcing effects by themselves. We then describe the preclinical and clinical evidence supporting the utility of agonist-based medications for OUD, and, briefly, the successful use of agonist replacement therapy with varenicline in the treatment of nicotine use disorder. Finally, we discuss recent progress in medication development of agonist-like therapies for cocaine use disorders based on lessons learned from methadone and buprenorphine. We contend that the successful agonist approach to treatments of OUD can be applied to the development of new treatment strategies for cocaine use disorders, for which no approved pharmacotherapeutics currently exist.

2. Neural Substrates of Opioid Reward

2.1. Opioid receptor mechanisms

The ideal strategy for treatment discovery in SUD is to understand the neural mechanisms underlying drug reward and addiction and subsequently develop mechanism-based pharmacotherapies that target those same neural substrates. Opioids act by binding to opioid receptors located in the peripheral and central nervous systems. There are four primary types of opioid receptors: mu (μ), kappa (κ), delta (δ), and opioid-receptor like-1 (ORL1) or the nociception (NOP) receptor. Each receptor has seven transmembrane domains and is coupled to inhibitory G-proteins (Gαi) which, when activated, inhibit neuronal activity (Al-Hasani and Bruchas, 2011). Recent insights into biased agonism or functional selectivity indicate that some opioid receptor ligands preferentially recruit intracellular G-protein vs. β-arrestin signaling pathways. With respect to the mu opioid receptor, the β-arrestin pathway is associated with adverse effects of exogenous opioids, including tolerance, reduced analgesia, enhanced respiratory suppression, and constipation (Bohn et al., 2000; Bohn et al., 1999; Madariaga-Mazon et al., 2017; Raehal et al., 2005). The delta and kappa opioid receptors also involve β-arrestin signaling that promotes tolerance and receptor internalization (Ho et al., 2018; Vicente-Sanchez et al., 2018). All three opioid receptors, as well as opioid receptor-like 1 (ORL1) may play a role in the induction and development of addiction by differentially mediating the rewarding and euphoric vs. aversive and withdrawal effects of opioids that drive the stages of compulsive drug use and relapse. For the purposes of this review we focus primarily on the mu opioid receptor, but refer readers to additional reviews of kappa and delta opioid receptors, as well as ORL1 involvement, in addiction elsewhere (Castro and Berridge, 2014; Helal et al., 2017; Karkhanis et al., 2017; Margolis et al., 2017; Poznanski et al., 2017; Schank et al., 2012).

2.2. Neural circuits underlying opioid reward

Mu opioid receptors are the primary mediators of the euphoric effects of opioids and are also involved in the rewarding efficacy of other drugs of abuse, including cannabinoids, alcohol and nicotine (Madariaga-Mazon et al., 2017). Mu opioid receptors are highly expressed on GABAergic neurons in regions of the brain implicated in reward and addiction, including the thalamus, amygdala, anterior cingulate cortex, striatum (including the nucleus accumbens or NAc) and midbrain (Jaferi et al., 2011; Svingos et al., 1997; Xi and Stein, 2000). In the ventral tegmental area (VTA) of the midbrain, GABAergic interneurons and afferents tonically modulate dopamine (DA) neuron firing (Figure 1; Li, 2016). By binding to mu opioid receptors expressed on GABAergic interneurons or afferents within the VTA, opioids are assumed to disinhibit VTA DA neurons and increase DA release to the NAc, an effect that underlies the subjective experience of reward and euphoria induced by all drugs of abuse (Chartoff and Connery, 2014; Fields and Margolis, 2015; Kosten and George, 2002; Wise, 2008). In addition, activation of mu opioid receptors on GABAergic medium spiny neurons in the NAc, which share reciprocal projections with the VTA, may further release GABAergic inhibition of VTA DA neurons and augment DA release (Li, 2016; Matsui et al., 2014; Ross and Peselow, 2009; Xi and Stein, 2000). Opioids also act on other areas of the brain involved in reward-related learning and memory, drug craving, tolerance and relapse that contribute to the cycle of compulsive drug abuse (Kosten and George, 2002).

Figure 1:

A summary diagram illustrating the mesolimbic DA reward system, and targets of heroin, nicotine, cocaine, and compounds used as agonist therapies for the treatment of substance use disorder.

Evidence for mu opioid receptor involvement in addiction is derived from both in vitro and in vivo studies. In vivo, opioid conditioned place preferences (CPP) and opioid self-administration are blocked by mu opioid receptor antagonists, administered either systemically or directly within the VTA (Britt and Wise, 1983; Olmstead and Franklin, 1997). Accordingly, selective genetic deletion of mu opioid receptors in the VTA abolishes opioid CPP (Zhang et al., 2009). Further, mu opioid receptor agonists sustain CPP and are voluntarily self-administered into the VTA, NAc shell, hypothalamus, amygdala, and periaqueductal grey, but not the NAc core or dorsal striatum (Bals-Kubik et al., 1993; Bozarth and Wise, 1981; David and Cazala, 1994; Olmstead and Franklin, 1997; Phillips and LePiane, 1980; Steidl et al., 2015; Zangen et al., 2002), suggesting mu expression in regions beyond the VTA also participate in opioid reward. In sum, these results strongly implicate mu opioid receptors as attractive targets for the treatment of OUD.

3. Agonist Therapies for Opioid Use Disorder

3.1. Rationale of agonist therapy

Pharmacological manipulations of mu opioid receptors can be accomplished using full agonists, partial agonists, or antagonists. While antagonists can block the effects of opioids and are effective overdose reversal agents, they also precipitate withdrawal symptoms and are used less frequently for long term abstinence, due to lack of adherence (Ndegwa et al., 2016). Nonetheless, there remain strong advocates for the use of extended-release naltrexone (XR-NTX) for the treatment of OUD, a formulation which has improved treatment retention (Sullivan et al., 2019). In contrast, the use of agonists and partial agonists (known as maintenance or substitution therapy) for OUD is based on the rationale that replacement of abused opioids with pharmacotherapeutics that exert a similar mechanism of action in the brain (Figure 1) will facilitate abstinence and relapse prevention efforts by mitigating withdrawal symptoms and craving. Notably, pharmacotherapeutics that are mildly-to-moderately reinforcing mediate increased attendance/adherence with treatment. Because full agonists effectively mimic the effects of abused opioids, they may carry significant abuse liability. Partial agonists bind to their receptor targets, but do not elicit the maximum receptor response of a full agonist. Therefore, in the presence of a full agonist, such as heroin, a partial mu agonist will behave as an antagonist and attenuate heroin’s effects. However, in the absence of other opioids, a partial agonist will function as an agonist and mitigate withdrawal symptoms. In addition, cross-tolerance may represent an additional mechanism by which agonist therapies reduce illicit opioid abuse. For example, full mu opioid receptor agonists such as methadone tend to induce rightward shifts in heroin dose-response curves, suggesting cross-tolerance (Volavka et al., 1978; Zaks et al., 1971), whereas partial mu opioid receptor agonists, such as buprenorphine, can produce both rightward and downward shifts in heroin dose-response functions, suggesting antagonist effects (Greenwald et al., 2003). Partial agonists can therefore block opioid reward, minimize craving and reduce withdrawal, thereby promoting abstinence and relapse prevention for OUD.

Converging preclinical and clinical evidence point to the utility of three opioids as efficacious medication-assisted therapies for OUD: 1) methadone, a full opioid agonist, 2) buprenorphine, a partial opioid agonist, and 3) levo-alpha-acetylmethadol (LAAM). LAAM is also a mu opioid receptor full agonist approved by the U.S. FDA in 1993 for the treatment of OUD. However, LAAM’s indication has since been revised due to hERG (human Ether-à-go-go-Related Gene) potassium channel activity and adverse side effects, such as prolonged QTc-intervals and potentially fatal cardiac arrhythmia (Clark et al., 2002; Kang et al., 2003; Wieneke et al., 2009). As such, marketing in the U.S. ceased in 2003 and LAAM is currently not available in Canada or the European Union. Therefore, in the present review we focus on evidence from animal and human studies supporting the utility of methadone and buprenorphine medications in the treatment of OUD.

3.2. Methadone for OUDs

3.2.1. Historical discovery of methadone agonist therapy

For centuries, OUD was considered a disease of the mind, due to criminal or deviant behavior and a weak personality. As a result, a major approach to managing OUD was incarceration and punishment. The 1960’s heralded a revolution in addiction management as a “metabolic disease” of the brain with resultant behaviors of “drug hunger” and drug self-administration, which required pharmacological intervention rather than punishment (Dole and Nyswander, 1966a; Dole and Nyswander, 1966b). Beginning in 1959, the Canadian Department of Health approved a series of experiments by Vancouver specialist Dr. Robert Halliday, to utilize methadone, an orally effective synthetic opioid, in the management of acute opioid withdrawal (Fischer, 2000). This short-term experiment quickly transitioned to a “prolonged withdrawal” program, which emphasized psychosocial support alongside methadone treatment with the goal of harm reduction rather than abstinence. Halliday likened his opioid maintenance program to insulin treatment for diabetes (Fischer, 2000). With a resurgence of heroin abuse in New York in the 1960s, Dole and Nyswander soon afterwards began prescribing methadone, which was initially assumed to be a short-acting opioid, for opioid detoxification by administering multiple doses per day, followed by rapid tapering without further treatment. Due to nursing limitations, a handful of patients were given only one or two doses of methadone in a day. Surprisingly, these patients showed significant reductions in opioid withdrawal severity, suggesting that methadone might have a longer-acting profile than previously assumed (Kreek, 2000; Kreek et al., 2004). Nonetheless, these anecdotal findings inspired researchers to study whether methadone would be effective in the treatment of heroin abuse.

3.2.2. Pharmacology and preclinical studies with methadone

Methadone is a mu opioid receptor full agonist:

In vitro receptor binding and functional assays indicate that methadone itself is a mu opioid receptor full agonist with Ki values of 1.7, 435, and 405 nM for mu, delta, and kappa opioid receptors, respectively, similar to morphine (Ki values of 1.4, 145, 23.4 nM for mu, delta, and kappa opioid receptors, respectively (Codd et al., 1995). Of note, methadone also acts as a glutamate NMDA receptor antagonist (Ebert et al., 1998; Oxenham and Farrer, 1998), although it is unclear whether or not this off target action is relevant to its therapeutic utility for treating OUD. Heroin itself exhibits relatively low affinity for the mu opioid receptor (Ki = 483 nM) (Inturrisi et al., 1983), but when administered systemically heroin works as a prodrug, rapidly entering the brain and metabolizing from 6-acetyl-morphine to morphine itself, thereby producing euphoric, analgesic, and anxiolytic effects (Sawynok, 1986). In vivo, both morphine and methadone increase expression of c-Fos, an immediate early gene indicating neuronal activation, in the somatosensory and insular cortices (Taracha et al., 2008). Systemic administration of heroin or methadone produces similar increases in NAc DA and locomotor hyperactivity in a dose-dependent manner (Fig. 2 A–D). Chronic administration of methadone also produces physical dependence and withdrawal symptoms that mirror those associated with morphine, including wet dog shakes, weight loss, diarrhea, ptosis and teeth chattering (Ling et al., 1984). However, methadone withdrawal has more profound effects than morphine on gene expression in the pineal gland, melatonin synthesis, and regulation of circadian rhythms (Pacesova et al., 2016). Methadone overdose also causes respiratory depression similar to morphine and heroin (Lewanowitsch et al., 2006). Together, these findings support comparable mu opioid receptor agonist profiles of heroin, morphine and methadone, both in vitro and in vivo.

Figure 2.

Characterization of the neurochemical and behavioral effects of heroin and methadone in vivo in rats. A, B: Systemic administration of heroin or methadone produced significant and dose-dependent increases in extracellular NAc DA, with methadone displaying a longer-duration of action than heroin. C, D: Systemic administration of heroin or methadone dose-dependently increased open-field locomotor activity. Again, methadone displays a long-acting profile. E, F: Systemic administration of heroin produced a dose-dependent increase in intracranial brain-stimulation reward (BSR) maintained by electrical stimulation of the medial forebrain bundle of the hypothalamus, while methadone produced a modest increase in BSR only at 3 mg/kg. G, H: Heroin priming induced robust reinstatement of heroin-seeking behavior in rats extinguished from previous heroin self-administration, while methadone did not induce reinstatement at either dose tested. *p<0.05, **p<0.01, ***p<0.001, compared to baseline before heroin or methadone injection or compared to vehicle control group. (Some data are replotted from Peng et al., 2010).

Methadone displays less rewarding and addictive potential than heroin:

Preclinical studies in experimental animals indicate that methadone is less rewarding and has lower addictive potential than heroin. As shown in Figure 2 (A, B), systemic administration of heroin produces a rapid increase in extracellular NAc DA, while methadone produces a slow-onset, long-lasting increase in extracellular NAc DA (Peng et al., 2010; Preshaw et al., 1982). Behaviorally, methadone induces dose-dependent increases in open-field locomotor activity with a longer duration of action than heroin (Fig. 2 C, D), possibly due to methadone’s relatively long half-life (24–36 h), slow metabolism and high fat solubility (Eap et al., 2002).

In intracranial self-stimulation (ICSS) maintained by electrical stimulation of the medial forebrain bundle of the hypothalamus, heroin produces a robust, dose-dependent increase in brain-stimulation reward, while methadone does not, except at a moderate 3 mg/kg dose (Fig. 2 E, F). Following extinction from heroin self-administration, methadone priming fails to reinstate drug-seeking behavior in rats, in contrast to morphine, heroin and oxycodone (Fig. 2 G, H; Leri et al., 2004; Stewart et al., 1996; Werner et al., 1976; You, 2018; You et al., 2017). During substitution testing following heroin self-administration, methadone sustains a higher rate of self-administration but elicits progressive decreases in drug intake over time, suggesting that methadone may have lower reinforcing value than heroin and that the higher rate of methadone self-administration could be a compensatory response to reduced reward (Peng et al., 2010). Methadone also appears to have lower transgenerational abuse liability than morphine. Offspring of dams exposed to chronic morphine voluntarily drink more morphine solution than controls and show greater reinstatement to morphine consumption after an abstinence period. In contrast, offspring of methadone-exposed dams show no differences in methadone consumption (Hovious and Peters, 1985). Together, these findings suggest that methadone not only prevents opioid withdrawal but is less rewarding and addictive than other opioids such as heroin (Peng et al., 2010), supporting its utility in the treatment of OUD.

The mechanisms underlying the differential addictive liability of heroin and morphine vs. methadone are unclear. One reason may be that chronic morphine induces minimal, while methadone induces robust mu opioid receptor internalization in a dose-dependent manner (Liao et al., 2007). A second reason may be related to the dynamic changes in extracellular DA after heroin versus methadone administration, since a drug’s rewarding efficacy is positively correlated with the dynamic change induced by the drug in extracellular DA. The faster the rise and subsequent fall in extracellular DA, the higher the presumed drug-induced reward and locomotor activation (Busto and Sellers, 1986; Kimmel et al., 2008; Kimmel et al., 2007; Volkow et al., 1995; but see Li et al., 2011; Peng et al., 2010). As shown in Figure 2, systemic administration of methadone leads to a slow-onset, long-lasting increase in extracellular NAc DA compare to heroin (Peng et al., 2010), which may be related to its unique pharmacokinetic profiles such as high lipophilicity with rapid GI absorption, large initial volume of distribution and slow tissue release, and long half-life (Ayonrinde et al., 2000; Eap et al., 2002). The unique pharmacokinetic profile of methadone may not only explain in part why oral administration of methadone has reduced addictive liability in humans compared to heroin, but also why systemic (i.p.) administration of methadone is less rewarding than heroin in rats.

Methadone treatment attenuates illicit opioid action:

In vitro and in vivo evidence demonstrates that co-administration or pretreatment with methadone attenuates the pharmacological and behavioral effects of illicit opioids such as morphine and heroin. At the cellular level, co-administration of methadone blocks morphine-induced inhibition of adenyl cyclase, desensitizes the mu opioid receptor response to morphine, and inhibits morphine-enhanced cAMP formation and accumulation caused by forskolin (Blake et al., 1997). In vivo, pretreatment with methadone dose-dependently blocks heroin-enhanced extracellular DA in the NAc (Fig. 3 A, B) and heroin-enhanced electrical brain-stimulation reward in rats (Fig. 3 C, D) (Peng et al., 2010; Preshaw et al., 1982). Methadone also reduces intravenous heroin self-administration under both fixed-ratio (FR1; Fig. 3 E) and progressive-ratio reinforcement schedules (Peng et al., 2010), and blocks heroin-induced reinstatement of drug-seeking behavior in rats (Fig. 3 F; Leri et al., 2004). In rhesus monkeys, methadone blocks the shift to heroin choice over food during heroin withdrawal, while other medications, including dopamine agonists and corticotrophin releasing factor (CRF) or kappa receptor antagonists, fail to reliably reduce heroin choice (Negus and Banks, 2018). Such attenuation produced by methadone may be partially due to cross-tolerance to illicit opioids, competitive receptor binding and the relatively long half-life of methadone (Eap et al., 2002).

Figure 3:

Methadone pretreatment attenuates heroin action in animal models of addiction. A: Effects of methadone pretreatment on heroin-induced increase extracellular NAc DA; B: Heroin-induced increases in extracellular DA are blocked by methadone pretreatment (data highlighted in gray were normalized over the baseline before heroin injection); C: Methadone dose-dependently attenuated heroin-enhanced BSR; D: Methadone, at 5 mg/kg, attenuated heroin-enhanced BSR produced by multiple heroin doses; E: Methadone inhibited intravenous heroin self-administration in rats in a dose-dependent manner; F: Methadone, administered 30 min prior to heroin, dose-dependently attenuated 0.25 mg/kg heroin-induced reinstatement of drug-seeking behavior. *p<0.05, ***p<0.001, compared to vehicle control group. (Some data are replotted from Peng et al., 2010).

3.2.3. Clinical studies with methadone

The systematic study of methadone treatment for OUD began in the 1960’s. Early reports from Canada and the U.S. identified significant reductions in opioid craving and withdrawal symptoms in subjects given only one to two doses (5–10 mg) of methadone (Ferguson et al., 1965; Kreek, 2000; Kreek et al., 2004), providing a foundation for the first large-scale studies on methadone for the treatment of heroin abuse. This seminal 1963 study, involving 214 heroin-addicted patients, found that methadone achieved three key hallmarks: 1) prevention of opioid withdrawal, 2) reduction of opioid craving, and 3) normalization of physiological functions (e.g. gastrointestinal) that were perturbed by chronic opioid abuse. These observations led to a decade of clinical studies demonstrating the safety and efficacy of methadone in reducing opioid use and relapse (for comprehensive reviews, see Kreek et al., 2000, 2004). Around the same time, the Ontario Addiction Research Foundation began the first methadone treatment program (Ferguson et al., 1965; Fischer, 2000). By the late 1960’s, methadone was a widely accepted treatment for OUD in Canada (Fischer, 2000). Methadone was approved later by the U.S. FDA as an agonist therapy for OUD in 1972 (Blanco-Gandia and Rodriguez-Arias, 2018; Joseph and Woods, 2018; Kreek, 2000; Kreek et al., 2004; Kreek and Vocci, 2002). Since then, a great number of studies have further confirmed the safety and effectiveness of methadone pharmacotherapyfor OUD (Ali et al., 2017; van Dorp et al., 2007).

Today, tapered methadone treatment (the process of slowly decreasing a replacement opioid agonist) is used for opioid detoxification to mitigate withdrawal and craving during the onset of long-term abstinence and treatment programs (Li, 2016). Typically, methadone is administered in progressively decreasing doses over a long period of time. The detoxification and/or tapering process is associated with fewer withdrawal effects, reduced heroin use, and improved treatment retention compared to non-pharmacological detoxification programs (Amato et al., 2013; Mattick et al., 2009), although psychosocial support is almost always necessary to sustain long-term abstinence (Lobmaier et al., 2010; Mattick et al., 2009; Veilleux et al., 2010).

In addition to detoxification, methadone maintenance therapy is frequently used as a first-line treatment for heroin abuse, and substantially reduces healthcare costs compared to non-pharmacologic therapies (Blanco-Gandia and Rodriguez-Arias, 2018). However, methadone maintenance programs are associated with high degrees of social stigma (Woods and Joseph, 2018), variable attrition rates that are contingent upon dosage (Maxwell and Shinderman, 2002), and withdrawal symptoms upon cessation of use. Although beyond the scope of the current review, behavioral and psychosocial support is a key aspect in maintaining remission during maintenance programs (Dugosh et al., 2016). In addition, as in experimental animals, methadone alone produces euphoric effects in human subjects and is susceptible to abuse and overdose (Jasinski and Preston, 1986; Li, 2016). Thus, although methadone may be less rewarding compared to heroin or morphine, it nonetheless remains a Schedule II drug with relatively high abuse liability. For these reasons, partial agonist therapy, such as that provided by buprenorphine, may be superior in long-term treatment plans.

3.3. Buprenorphine for OUDs

Buprenorphine was discovered in 1966 by chemists at Reckitt & Colman located in Hull, England, and was introduced in phase 1 clinical safety trials in the late 1960’s (Campbell and Lovell, 2012). Reckitt subsequently supplied buprenorphine to the U.S. Addiction Research Center located in Lexington, Kentucky for efficacy testing in opioid-dependent humans, leading to the first publication in 1978 on buprenorphine’s improved safety and reduced addiction liability compared to methadone (Campbell and Lovell, 2012; Jasinski et al., 1978). Thereafter, buprenorphine slowly attained approval in other countries for pain and later the treatment of OUD (see Campbell and Lovell, 2012 for detailed history; Kumar et al., 2009; Lintzeris et al., 2004). The U.S. FDA approved buprenorphine in 2002 as an office-based treatment for OUD (Campbell and Lovell, 2012; FDA, 2002).

3.3.1. Pharmacology and preclinical studies with buprenorphine

Buprenorphine is a long-acting partial agonist at mu opioid receptors due to a unique slow receptor association/dissociation profile, and is therefore anticipated to have lower abuse liability. As a partial agonist, buprenorphine functions as an antagonist in the presence of other opioids by competitively binding to the mu opioid receptor. However, in the absence of opioids, buprenorphine elicits partial activation of the mu opioid receptor and therefore mitigates withdrawal symptoms in chronic users. Notably, buprenorphine also acts as an antagonist at kappa opioid receptors, which may contribute to its antidepressant effects (Falcon et al., 2016). Like methadone, in HEK cells expressing cloned mouse mu opioid receptors, buprenorphine blocks morphine inhibition of adenyl cyclase, desensitizes the mu opioid receptor, and blocks forskolin-induced cAMP increases by morphine, providing a cellular basis by which buprenorphine may help to treat OUDs (Blake et al., 1997). Buprenorphine alone activates VTA DA neurons in a manner similar to morphine and enhances basal extracellular DA release in the NAc, but attenuates NAc DA responses to heroin, consistent with a partial agonist profile (Grant and Sonti, 1994; Sorge et al., 2005; Sorge and Stewart, 2006). Buprenorphine, like methadone, also reverses altered brain glucose metabolism during morphine withdrawal in the thalamus, insular cortex and periaqueductal gray, albeit in a sex-dependent manner (Santoro et al., 2017).

In behaving animals, buprenorphine induces locomotor excitation similar to morphine, and morphine-treated rats show cross-tolerance and cross-sensitization to buprenorphine (Bartoletti et al., 1999; Bartoletti et al., 1993; Galici et al., 2005). At low doses buprenorphine produces modest conditioned place preference, but at high doses buprenorphine produces modest conditioned place aversion, suggesting some biphasic effects (Stinus et al., 2005). Like methadone and other commonly abused opioids, buprenorphine increases sensitivity to electrical ICSS, particularly in opioid-dependent subjects (Bruijnzeel et al., 2007; Hubner and Kornetsky, 1988). These findings suggest buprenorphine has minimal abuse liability but may also alleviate dysphoria associated with opioid withdrawal. Accordingly, buprenorphine prevents spontaneous withdrawal from fentanyl and reduces naloxone-precipitated physical withdrawal symptoms in fentanyl-dependent rats (Bruijnzeel et al., 2007). In morphine-dependent adults as well as rat pups, buprenorphine blocks withdrawal syndrome and withdrawal-induced conditioned place aversion (Stinus et al., 2005; Stoller and Smith, 2004). Chronic buprenorphine delivered via osmotic minipump (1.5 or 3.0 mg/kg/day) is ineffective in reducing ongoing heroin self-administration under fixed-ratio or progressive-ratio schedules over the course of daily, 3-hour tests, but buprenorphine does increase the latency to respond to heroin-associated cues at the onset of self-administration sessions (Sorge and Stewart, 2006) and reduces heroin seeking during extinction and heroin-primed reinstatement (Sorge et al., 2005). In contrast, buprenorphine dose-dependently reduces heroin intake in rats with unlimited heroin self-administration access (Chen et al., 2006), indicating that the animals’ self-administration history impacts the efficacy of buprenorphine in reducing heroin intake. Notably, the effective dose of buprenorphine capable of reducing opioid withdrawal in rats (80 μg/kg) may be different than that required to suppress opioid self-administration (40 μg/kg) (Chen et al., 2006; Sorge and Stewart, 2006), although comparisons across these preclinical studies may be limited due to methodological differences in dosing regimens, timing, rat handling, etc.

Rhesus monkeys voluntarily self-administer buprenorphine over saline, but do not show withdrawal symptoms or physical dependence upon cessation of buprenorphine use (Mello et al., 1981; Mello and Mendelson, 1985; Yanagita et al., 1982). Buprenorphine also blocks monkeys’ shift to heroin choice over food during withdrawal, although methadone is more effective than buprenorphine in blocking heroin choice (Negus, 2006). In macaques, buprenorphine produces dose-dependent reductions in intravenous heroin self-administration, whereas methadone is ineffective in suppressing heroin intake in 4 out of 5 subjects. Buprenorphine also has fewer toxic side effects (e.g., seizures, respiratory suppression, sedation) than methadone (Mello et al., 1981). While other studies in macaques have reported that buprenorphine is a more potent reinforcer than methadone at low doses under a progressive-ratio schedule of reinforcement, both medications are less efficacious reinforcers than heroin (Mello et al., 1984; Mello et al., 1988).

3.3.2. Clinical studies with buprenorphine

In human subjects, buprenorphine produces mild analgesia and subjectively reinforcing effects. However, unlike methadone, heroin, and other opioids, buprenorphine produces fewer physical dependence or withdrawal symptoms upon cessation of use (Comer et al., 2002; Houde, 1979; Jasinski et al., 1978; Mello et al., 1984; Mello et al., 1982), although some studies have reported difficulties with buprenorphine tapering (Fiellin et al., 2014), and in practice adjunctive medications (such as clonidine or sleep-promoting agents) are often co-prescribed to attenuate opioid withdrawal during buprenorphine dose tapering.

In early studies involving human heroin users, buprenorphine was efficacious as a maintenance therapy for OUDs, reducing heroin intake by 69–98% compared to controls (Mello and Mendelson, 1980; Mello et al., 1982), and these findings have been replicated since (Burchenal, 1977; DiPaula et al., 2002; Kakko et al., 2003; Sung and Conry, 2006). Moreover, unlike methadone, prolonged tapering of buprenorphine treatment (7 days vs. 28 days) does not convey additional benefits in promoting abstinence, potentially reducing the need for long-term maintenance programs (Ling et al., 2009). In contrast, a later review of 28 studies suggested that buprenorphine taper duration (ranging from 0 to 365 days) is positively associated with varying degrees of abstinence (Dunn et al., 2011). However, when medium to high doses are used, buprenorphine and methadone appear to be equally effective in treatment retention and reducing illicit opioid use (Mattick et al., 2014). Buprenorphine is less efficacious than methadone in treatment retention when flexible or low fixed doses are used, and in subjects with high opioid tolerance (Mattick et al., 2014).

Although buprenorphine has modest abuse potential in recently-detoxified opioid users and in non-opioid-dependent subjects (Comer et al., 2002; Comer et al., 2005), both animal and human studies indicate reduced abuse liability of buprenorphine compared to other opioids. Combination therapies of buprenorphine and naloxone have emerged to reduce illicit abuse and also have significant efficacy in reducing opioid withdrawal symptoms, craving and relapse (Wang et al., 2018), although combination therapy appears to have similar abuse liability as buprenorphine alone in recently detoxified heroin users (Comer and Collins, 2002). Alternative formulations of buprenorphine, including depot injections, sustained-release subdermal implants and sublingual tablets, are now promoting the convenience, utility and availability of buprenorphine in the treatment of OUD (Haight et al., 2019; Harricharan and Farah, 2017; Saxon et al., 2013; Strain et al., 2011; Walsh et al., 2017).

Co-administration of buprenorphine in emergency room visits for opioid overdose may also improve long-term treatment outcomes (Johns et al., 2018). Systematic comparison of buprenorphine to methadone and other pharmacotherapies illustrates the superiority of buprenorphine in detoxification, treatment retention, reduction of illicit drug use and drug cravings, and minimizing adverse reactions and side effects (Ling et al., 2005; Ling and Wesson, 2003). In addition, total healthcare costs of patients maintained on buprenorphine are up to 49% lower than those of patients maintained on methadone (Baser et al., 2011). Therefore, individual differences and needs must be considered when choosing the best pharmacotherapeutic strategy for OUD.

4. Agonist Therapies for Other Substance Use Disorders

4.1. Varenicline for tobacco use disorder

Like buprenorphine treatment for OUD, partial agonist therapies have strong potential for the treatment of other drug use disorders, including tobacco use (Jordan and Xi, 2018). The euphoric and addictive properties of tobacco are attributed to nicotine’s effects in the mesolimbic reward system, where nicotine is thought to increase VTA DA neuron activity by binding to and activating α4β2 nicotinic acetylcholine receptors (nAChRs). In addition to various formulations of nicotine itself (e.g. NRT; nicotine patch, gum), Varenicline (Chantix) was developed as a smoking cessation agent that selectively targets α4β2 nAChRs. As a potent partial agonist, varenicline attenuates nicotine-induced DA release in the NAc and thereby reduces the reinforcing value of nicotine. Because varenicline elicits partial activation of the α4β2 nAchR in the absence of nicotine, it also mitigates withdrawal symptoms during abstinence. Varenicline therefore conveys significant advantages over full agonists such as nicotine replacement therapies, which do not eliminate nicotine use disorder, and antagonists, which precipitate withdrawal symptoms in the presence of nicotine. Both preclinical and clinical evidence support the efficacy of varenicline in reducing tobacco use and promoting smoking cessation (for review see Jordan and Xi, 2018). It is worth noting that bupropion, a medication that is not a nAChR agonist or antagonist, is also prescribed for tobacco use disorder, although it is typically less efficacious than agonist therapies such as varenicline (Jordan and Xi, 2018).

4.2. Agonist therapy for cocaine use disorder

4.2.1. Classical DAT inhibitors

Cocaine is the third most commonly abused illicit drug (behind opioids and cannabinoids), and currently, there are no FDA-approved medications available for the treatment of cocaine or other psychostimulant use disorders. Cocaine’s abuse potential derives from blockade of monoamine transporters, most notably dopamine (DA) transporters (DAT), which leads to rapid and dramatic increases in extracellular NAc DA. Based on the rationale that methadone and buprenorphine are long-acting mu opioid receptor agonists successful in the treatment of opioid dependence, low dose, slow-release monoamine transporter or DAT inhibitors have been proposed as agonist therapies for cocaine use disorder (Fig. 1). To this end, oral cocaine has shown efficacy in reducing the subjective and physiological effects of low doses of intravenously administered cocaine (Walsh et al., 2000). Accordingly, several classical stimulants that have similar DAT binding profiles as cocaine have since been tested in humans and experimental animals.

d-Amphetamine:

Dextroamphetamine (d-Amphetamine) is the active enantiomer of amphetamine, a potent psychostimulant that is prescribed for the treatment of attention deficit hyperactivity disorder (ADHD) and narcolepsy. Pharmacologically, d-Amphetamine is a full agonist at the trace amine-associated receptor 1 (TAAR1) and a vesicular monoamine transporter 2 (VMAT2) inhibitor that causes release of DA, serotonin, and norepinephrine (Sitte and Freissmuth, 2015). In early clinical studies, sustained-release d-Amphetamine reduced illicit cocaine use and improved treatment retention (Grabowski et al., 2004). However, in a randomized, placebo-controlled, double-blind clinical trial in treatment-seeking individuals with methamphetamine-use disorder, oral d-Amphetamine reduced withdrawal symptoms and craving but failed to reduce methamphetamine use (Galloway et al., 2011). In turn, sustained-release methamphetamine itself was found to reduce cocaine use and craving (Mooney et al., 2009). Given that both d-Amphetamine and methamphetamine are highly potent psychostimulants that carry significant abuse liability, their viability and potential for attaining FDA approval for the treatment of cocaine use disorder is limited.

Methylphenidate:

Methylphenidate (Ritalin) is an FDA-approved psychostimulant used for treatment of ADHD and narcolepsy. Pharmacologically, methylphenidate acts as a DA-norepinephrine reuptake inhibitor (Childress and Sallee, 2013). Early studies suggested that methylphenidate was not effective in reducing cocaine use (Grabowski et al., 1997), and more recent studies have confirmed the inability of methylphenidate to reduce psychostimulant abuse or improve addiction treatment retention (Miles et al., 2013).

CTDP-31,345:

The success of methadone in treating OUD suggests that long-acting monoamine transporter or DAT inhibitors may be similarly useful for treating cocaine use disorder. Our laboratory has examined this hypothesis using the slow-onset long-acting monoamine transporter inhibitor, CTDP-31,345 (trans-[4-(3,4-dichlorophenyl)-1,2,3,4-tetrahydronaphthalen-1-yl]dimethylammonium chloride) in a variety of cocaine abuse-related animal models. This indatraline analog was effective in antagonizing cocaine reward and relapse (Peng et al., 2010; Xi and Gardner, 2008). However, by itself this slow-onset long-acting monoamine transporter inhibitor displayed similar abuse liability to cocaine, as evidenced by increased electrical brain-stimulation reward and extracellular NAc DA after systemic administration and substitution for cocaine in intravenous self-administration protocols. CTDP-31,345 also failed to alter cocaine-induced increases in extracellular NAc DA, but dose-dependently inhibited cocaine self-administration (Peng et al., 2010), suggesting CTDP-31,345 is more cocaine-like with limited translational potential, similar to other slow-onset, long-acting monoamine transporter inhibitors (e.g. CTDP 30,640; Gardner et al., 2006). Ideally, agonist therapies for cocaine abuse should functionally antagonize cocaine’s action while having low addictive potential on their own.

4.2.2. Atypical DAT inhibitors

Significant progress in DAT-based medication development indicates that not all monoamine transporter or DAT inhibitors elicit behavioral effects identical to those of cocaine. While substitution or cocaine-like effects may improve adherence to and/or retention in treatment, ideally new therapeutics for substance use disorders should have limited abuse liability alone while retaining efficacy in reducing illicit drug abuse, ameliorating withdrawal, and promoting abstinence. To this end, atypical DAT inhibitors are defined as exhibiting reduced or in some cases a complete lack of cocaine-like rewarding effects (Tanda et al., 2009). Moreover, pretreatment with these compounds can reduce cocaine-elicited behaviors in rodent models (Reith et al., 2015), suggesting translational potential for the treatment of cocaine use disorder.

JHW 007:

JHW 007 (N-Butyl-3α-[bis(4’-fluorophenyl)methoxy]tropane) emerged as a lead compound out of a series of benztropine analogues (Agoston et al., 1997; Desai et al., 2005). JHW 007 is an atypical DAT inhibitor (K_i = 25 nM, compared to 1330 and 1730 nM for NET and SERT, respectively) with a slow-onset, long-acting profile. In vivo, JHW 007 alone has minimal cocaine-like behavioral effects, while pretreatment with JHW 007 inhibits cocaine self-administration, cocaine-induced hyperactivity and cocaine locomotor sensitization in rats (Desai et al., 2014). JHW 007 also attenuates the rewarding and locomotor-stimulating effects of amphetamine and other psychostimulants (Hiranita et al., 2014; Reith et al., 2015; Velazquez-Sanchez et al., 2013). Although JHW 007 may have had drug development potential (Raje et al., 2003), it was never studied in humans and thus will likely remain as a preclinical research tool, currently available commercially.

CTDP-32476:

CTDP-32476 (2-(1-(4-chlorophenyl)-3-methylbutyl)piperidine) is a methylphenidate analog with a slow-onset, long-acting profile. In vitro binding assays indicate that CTDP-32476 is a potent and selective DAT inhibitor (Ki = 12 nM) and competitive with cocaine at DAT (cocaine Ki = 279 nM; Froimowitz et al., 2007; Xi et al., 2017). Systemic administration of CTDP-32476 alone produces a slow-onset, long-lasting (6–12 h) increase in extracellular NAc DA, locomotor activity, and brain-stimulation reward. Drug-naive rats do not self-administer CTDP-32476. In a substitution test, cocaine self-administering rats displayed a progressive reduction (i.e., extinction) in CTDP-32476 self-administration, suggesting significantly lower addictive liability than cocaine. Pretreatment with CTDP-32476 inhibited cocaine-enhanced extracellular NAc DA, cocaine self-administration, and cue-induced relapse to cocaine seeking (Xi et al., 2017). Therefore, CTDP-32476 appears to be a unique DAT inhibitor that not only satisfies drug craving through slow-onset, long-lasting DAT inhibition, but also renders subsequent administration of cocaine ineffective.

RTI-336:

RTI-336 was a lead compound emerging from a class of 3-phenyltropane cocaine analog, developed for the treatment of cocaine use, that acts as a selective dopamine reuptake inhibitor (Carroll et al., 2006b). Compared to other 3-phenyltropane analogs such as RTI-177, RTI-336 exhibited a higher LD50/therapeutic ratio, favorable oral bioavailability, induced low levels of locomotor sensitization compared to cocaine, and reduced cocaine self-administration in both rats and non-human primates (Carroll et al., 2006a; Carroll et al., 2006b). However, PET imaging revealed that high levels of DAT occupancy (>90%) by RTI-336 were required to reduce non-human primate cocaine self-administration (Carroll et al., 2006b). In another non-human primate study, RTI-336 produced fast-onset stimulant effects similar to cocaine, and a trend towards reliable self-administration, albeit at lower levels than cocaine (Czoty et al., 2010; Howell et al., 2007; Kimmel et al., 2007). Still other studies in rats revealed strain-dependent effects: while Lewis rats showed reduced cocaine intake and cocaine-induced locomotor activity following RTI-336 administration, RTI-336 increased cocaine intake in F344 rats and had no effect on cocaine-induced locomotion (Haile et al., 2005). Double-blind, placebo-controlled studies suggest RTI-336 is well-tolerated and safe in humans, potentially warranting further investigation (Carroll et al., 2018).

Modafinil:

Modafinil is a clinically available atypical DAT inhibitor with low potency (IC₅₀ = 2–4 μM in the DA reuptake assay) and downstream activities that may interfere with its efficacy for the treatment of psychostimulant use disorders (Mereu et al., 2013; Sangroula et al., 2017; Zolkowska et al., 2009). In an early human laboratory study, modafinil reduced cocaine euphoria (Dackis et al., 2003). However, subsequent clinical trials in patients with psychostimulant use disorder failed to demonstrate efficacy of modafinil over placebo (Anderson et al., 2012; Dackis et al., 2012; Shearer et al., 2009). Indeed, a recent meta-analysis indicated no evidence supporting the superiority of modafinil in promoting cocaine abstinence and treatment retention (Sangroula et al., 2017). However, post-hoc findings from these studies suggest that modafinil may be more efficacious in less severe cases of addiction. The more active R-modafinil (Armodafinil, Nuvigil) has also been examined preclinically as a potential treatment for cocaine use disorders (Loland et al., 2012), although clinical evaluation of R-modafinil in cocaine abusers has not been investigated, to our knowledge.

JJC8–016:

More recently, our laboratory developed a series of more soluble, selective and potent modafinil analogs, with the rationale that these modifications might improve effectiveness in treating psychostimulant use disorders. JJC8–016 (N-(2-((Bis(4-fluorophenyl)methyl)thio)ethyl)-3-phenylpropan-1-amine) was an early lead compound from this series with moderately high affinity for DAT (K_i=116 nM) (Okunola-Bakare et al., 2014; Zhang et al., 2017). In rats, JJC8–016 alone failed to alter extracellular NAc DA, locomotor activity, electrical brain-stimulation reward and reinstatement of drug-seeking behavior. Moreover, substitution of JJC8–016 for cocaine did not maintain self-administration in rats, and pretreatment with JJC8–016 significantly inhibited cocaine-taking and cocaine-seeking behaviors (Zhang et al., 2017). Together, these observations suggest that JJC8–016 has low abuse liability, but translational utility for the treatment of cocaine use disorder was limited by poor metabolic and pharmacokinetic profiles.

JJC8–091:

JJC8–091 (1-(4-(2-((Bis(4-fluorophenyl)methyl)sulfinyl)ethyl)piperazin-1-yl)-propan-2-ol) is a recently described modafinil analog and atypical DAT inhibitor (Ki=289 nM; Cao et al., 2016; Tunstall et al., 2018). Systemic administration of JJC8–091 produced a mild slow-onset, long-duration increase in extracellular NAc DA (Keighron et al., 2018). Drug-naïve rats do not self-administer JJC8–091, and JJC8–091 fails to substitute for cocaine in intravenous self-administration studies, suggesting extremely low addictive potential (Newman et al., 2019). Strikingly, pretreatment with JJC8–091 attenuates cocaine self-administration under progressive-ratio schedule of reinforcement and blocks cocaine-induced reinstatement to drug-seeking behaviors. In addition, JJC8–091 attenuates compulsive methamphetamine self-administration and decreases escalation of methamphetamine intake in rats (Tunstall et al., 2018).

Alongside its favorable effects on stimulant abuse in preclinical models, JJC8–091 displays a metabolic and pharmacokinetic profile consistent with a viable drug development candidate. The development of the Multiparameter Optimization (MPO) algorithm to predict CNS penetration has improved prioritization of clinical candidates in drug development (Wager et al., 2010a; Wager et al., 2010b; Wager et al., 2016). Applying the CNS MPO tool to six physicochemical properties of JJC8–091: lipophilicity calculated partition coefficient (ClogP), calculated distribution coefficient at pH7.4 (ClogD), molecular weight (MW), topological polar surface area (TPSA), number of hydrogen bond donors (HBD), and pKa of most basic center, this DAT inhibitor is predicted to have optimal parameters for CNS penetration and drug safety (MPO=5.22; Table 1). Collectively, these data suggest that JJC8–091 may represent an innovative new agonist therapy for the treatment of cocaine or other psychostimulant use disorders, a fundamental and yet unmet public heath need.

Table 1.

In silico ADME calculations for JJC8–091^a

Parameter	Value	T0b
ClogP	1.49	1
ClogD	2.32	0.84
TPSA	43.78	1
MW	422.53	0.55
HBD	1	0.83
PKa	6.89	1

CNS MPO = 5.22

^aClogP, TPSA, and MW values were calculated using ChemDraw. ClogD and PKa values were calculated using Chemicalize.

^bT0 was calculated using the method published in (Wager et al., 2010b).

5. Summary

Methadone and buprenorphine are successful examples of agonist replacement therapies for the treatment of OUD. These medications are effective in reducing illicit opioid use, mitigating opioid craving and withdrawal syndromes, and promoting abstinence. The efficacy of methadone can be attributed to three specific characteristics: 1) a long-acting mu opioid receptor agonist, similar to morphine; 2) lower addictive potential than morphine and heroin; and 3) attenuation of the rewarding and withdrawal-related effects of morphine and heroin. Based on findings with methadone, two partial agonist therapies (buprenorphine, varenicline) have been subsequently developed and successfully used for the treatment of opioid and tobacco use disorders, respectively. Partial agonists convey additional benefits over full agonist therapies, acting as antagonists in the presence of abused substances while mitigating withdrawal symptoms during abstinence through partial activation at the target receptor. Nevertheless, partial agonist therapies may fail patients with severe drug use, in whom full agonist treatment may be required. A “stepped-care” approach may also be indicated in severe OUD, in which patients can be transitioned from buprenorphine to methadone if clinical efficacy is insufficient (Kakko et al., 2007). Nonetheless, the rationale of agonist/partial agonist therapies is currently being applied to the development of new atypical DAT inhibitors for the treatment of cocaine and other psychstimulant use disorders. This approach has yielded significant progress, with several new lead compounds showing promising translational potential.

6. Future Investigation

Although agonist therapies are effective in the treatment of SUD, a major remaining challenge is the prevention of relapse to drug use after a period of abstinence. Longitudinal studies on heroin users in the Amsterdam Cohort Study have shown that 86% of patients in “low-threshold” harm reduction programs relapse within 5 years of methadone-sustained abstinence (Termorshuizen et al., 2005). A similar relapse rate was reported in tobacco users within a year of combined pharmacological and behavioral treatments (Stead et al., 2012). Alternative medication strategies for future study may involve the combination of multiple pharmacotherapies. For example, dopamine D₃ receptor antagonists are highly effective in reducing drug (cocaine or opioid) reward and in preventing relapse to drug-seeking behavior (Heidbreder and Newman, 2010; Sokoloff and Le Foll, 2017; Xi and Gardner, 2007; You et al., 2018; You et al., 2017). Thus, the combination of a full/partial mu opioid receptor agonist with a D₃ receptor antagonist may yield promising results in relapse prevention and the promotion of abstinence. In addition to D3 receptor antagonists, other compounds that target brain metabotropic glutamate, GABA, or endocannabinoid systems also show promising results in preventing relapse to drug seeking (Xi and Gardner, 2008). Another emerging strategy for the treatment of OUD is the development of biased mu opioid receptor agonists, which preferentially recruit activation of intracellular G-protein signaling over β-arrestin and therefore convey reduced risk of respiratory suppression and overdose (Bohn et al., 2000; Bohn et al., 1999; Madariaga-Mazon et al., 2017; Raehal et al., 2005; Schmid et al., 2017). However, additional studies are needed to determine whether G-protein signaling preferentially mediates the rewarding or addictive liability of opioids, and whether G-protein-biased agonist therapy is superior to classical (non-biased) agonist therapies for the treatment of OUD.

Human clinical trials will be required to confirm that the “agonist” therapy approach to cocaine use disorders using the atypical DAT inhibitors or other stimulant-like medications can be applied to this patient population. While avoidance of physical withdrawal is the hallmark of opioid relapse, individuals dependent on cocaine do not experience marked physical withdrawal. As a result, it is currently unclear how best to prevent relapse in this population. Nevertheless, identifying new lead molecules that are safe and show promise in animal models is the only way to ultimately determine if this approach is viable. Undoubtedly, there will be no magic bullet for all psychostimulant abusers. Going forward, it is essential to identify pharmacotherapies that can help at least a subpopulation of those patients who are motivated to quit their illicit drug use. We must ultimately broaden the effectiveness of available treatment strategies, especially for those that suffer from cocaine use disorders, for whom there are no therapeutic options.

Highlights.

• Agonist replacement therapies have been successfully used for the treatment of opioid and nicotine use disorders, but not yet for addiction to cocaine

• Methadone is a long-acting mu opioid receptor full agonist, and buprenorphine is a mu opioid receptor partial agonist

• Methadone and buprenorphine have lower addictive liability than morphine or heroin, and pretreatment with either attenuates opioid use

• Significant progress has been made in the development of agonist-like atypical DAT inhibitors for the treatment of cocaine use disorder