Cannabinoid Ligands and Alcohol Addiction: A Promising Therapeutic Tool or a Humbug?

Agonists

The discovery of Δ⁹-THC and the knowledge of its properties and availability opened a new path to developing highly selective and potent cannabinoid receptor agonists. The intensive synthesis of new compounds, the structures of which are based on the active principal of cannabis, resulted in a great number of active cannabimimetics. Interestingly, several of these novel ligands have a much greater potency than Δ⁹-THC, suggesting that other compounds that are present in herbal marijuana (or other herbal extracts) may influence Δ⁹-THC activity.

Dronabinol is a synthetic Δ⁹-THC formulation and one of the first cannabinoid receptor agonists that displays affinity for CB₁ and CB₂ receptors. However, it behaves only as a weak agonist for CB₂ receptors and is used as an appetite stimulant in HIV patients with anorexia. As shown by Beal et al. (1995), administration of capsules containing 2.5 mg of dronabinol twice a day caused an average weight gain of 0.1 kg versus an average loss of 0.4 kg in the placebo group (Beal et al. 1995). Additionally, several papers have demonstrated that this novel compound is an efficient analgesic and is a useful therapy for chronic non-cancer pain patients who take stable doses of opioids (Narang et al. 2008). Finally, agonist substitution has been an effective strategy for promoting abstinence from many substances of abuse (e.g., nicotine, opioids), and dronabinol has been proposed to serve as a replacement drug in cannabis addiction (Vandrey et al. 2013), especially in cases where withdrawal can be a significant barrier to cessation. Like nabilone it is used as an antiemetic drug in vomiting induced by cytotoxic agents (Tramèr et al. 2001).

Nabilone (Cesamet^®) is another synthetic cannabinoid characterized by an affinity of 2.2 nM for human CB₁ and 1.8 nM for CB₂ receptors. This compound is used to treat nausea associated with cancer chemotherapy (Maida et al. 2008). Nabilone also appears to be a promising candidate agonist for substitutive medication for cannabinoid agonist replacement treatment of cannabis use disorders (Lile et al. 2010). In addition, consumption of nabilone resulted in drug-appropriate responses in subjects who learned to discriminate from Δ⁹-THC (Lile et al. 2010). Several other studies have indicated that nabilone is a potent and effective drug for the treatment of various disorders, mainly related to pain. For example, in a randomized, double-blind placebo-controlled trial, nabilone suppressed pain and increased functional capacity in fibromyalgia patients (Skrabek et al. 2008).

Similar analgesic properties are also observed with Sativex, a cannabinoid extract oral spray containing both Δ⁹-THC and CBD, which is commercially available for neuropathic pain and spasticity in 25 countries primarily in Europe (Myers and Shetty 2008; Nurmikko et al. 2007; Palmer 2014). A selective CB₂ cannabinoid receptor agonist, O-3223, is an additional compound that can reduce pain and inflammation without apparent cannabinoid-like behavioral effects (Kinsey et al. 2011).

Considering the pharmacological profile of other cannabinoid ligands, WIN 55,212-2 appears to be very interesting. Aside from its antinociceptive activity, this nanomolar affinity cannabinoid receptor agonist (K_i values of 62.3 and 3.3 nM at the human recombinant CB₁ and CB₂ receptors, respectively) has also been reported to reduce endothelial cell (EC) inflammatory responses induced by bacterial lipopeptide and TNFα as well as to promote neural repair processes after neonatal hypoxia–ischemia (Fernandez-Lopez et al. 2010; Wilhelmsen et al. 2014). Importantly, such a neuroprotective activity was also observed in both global and focal cerebral ischemia and was blocked by a selective CB₁ receptor antagonist, SR141716A (Nagayama et al. 1999; Shen et al. 1996). Unfortunately, this antagonist also caused several side effects and enhanced locomotor responses when combined with other psychoactive drugs (e.g., heroin, amphetamine) (Lamarque et al. 2001). On the contrary, CP-55,940, another cannabimimetic, did not enhance sensitivity to the behavioral effects of cocaine. Moreover, its co-administration with cocaine reduced the locomotor hyperactivity produced by the psycho-stimulant (Arnold et al. 1998). Despite the fact that this compound is found to be almost 45 times more potent than Δ⁹-THC and is considered to be a full agonist at both CB₁ (K _i  = 0.58 nM) and CB₂ (K _i  = 0.68 nM) receptors, it has not yet been approved for human medical use, as the adverse effects and long-term damage and addiction potential of CP-55,940 are not yet known.

The list of potent agonists at either CB₁ or CB₂ receptors is relatively long. Despite their reported efficacy in many biological conditions, most cannabinoid agonists are not currently used in clinic and are instead used as tools for cannabinoid research.

Antagonists

SR141716A, also known as rimonabant, is the most potent and orally active antagonist of the CB₁ receptor. This compound, discovered by the laboratory of Rinaldi-Carmona, displays nanomolar affinity for CB₁ (K _i  = 1.98 nM) and micromolar for human CB₂ receptors expressed in CHO cells (Rinaldi-Carmona et al. 1994). Intensive studies have suggested its anorectic activity (Freedland et al. 2000). Indeed, several papers demonstrated that intake of different diets seemed to be sensitive to dose-dependent, antagonist-induced cannabinoid receptor blockade, which does not significantly alter other normal behaviors (Freedland et al. 2000; Tucci et al. 2006). Accordingly, other well-known antagonists are able to exert similar effects as rimonabant, including AM251 [N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide], which is a selective and potent CB₁ receptor antagonist with well-characterized anorectic effects in freely fed animals (Hildebrandt et al. 2003); and novel antagonists LH21 (Pavon et al. 2006), and URB477 (LoVerme et al. 2009).

Due to the anti-obesity activity of CB₁ receptor antagonists, they may be considered to be promising tools to treat obesity disorders. However, application of these compounds leads to several serious adverse effects. For instance, the study by Moreira and Crippa (2009) showed that CB₁ receptor blockade impaired fear elimination (Moreira and Wotjak 2010). Additionally, an increased incidence of depression and aggression has been reported, including an increase in suicidal thoughts (Christensen et al. 2007).

On the contrary, the first CB₂ receptor antagonist SR144528, having K _i values of 0.6 nM at CB₂ and 400 nM at CB₁ (Rinaldi-Carmona et al. 1998), as well as other recently discovered CB₂ receptor antagonists (e.g., JTE-907) can markedly reduce ear swelling in chronic contact dermatitis induced either by repeated challenge with oxazolone (Oka et al. 2006) or by repeated exposure to dinitrofluorobenzene in mice (Ueda et al. 2005). However, in comparison with some of the CB₁ receptor agonists and antagonists, none of the existing CB₂ receptor antagonists are listed as controlled substances worldwide.

Of importance, because many CB₁ antagonists are, in fact, partial or inverse agonists (e.g., AM281, AM251, SR141716A), they may produce significantly different effects via different mechanisms of action. These effects may also depend on the state of endocannabinoid signaling pathway. Therefore, further studies are warranted to establish the complete pharmacological profile of such compounds.

Preclinical Studies

Chronic alcohol intoxication evoked a reduction in eCB levels in the midbrain and an increase in AEA levels in the limbic forebrain in rats (Gonzalez et al. 2002a). The increased AEA levels in the limbic area, which is a key region for drug reinforcement processes, may be involved in the eCB-induced elevation of synaptic transmission, which leads to the addictive properties of alcohol. At the same time, reduced eCB levels in the midbrain may be a compensatory response induced by the activation of the negative feedback regulatory loop from the limbic forebrain to the ventral tegmental area (Gonzalez et al. 2002a). In another study, AEA levels increased in the cortex in mice simultaneously with decreasing cortical FAAH activity (Vinod et al. 2006). The rise in AEA levels may be associated with the activation of phospholipase A₂ (PLA₂), which is a crucial enzyme for eCB formation as shown in an animal model of chronic alcohol consumption (Basavarajappa et al. 1998b). It was also documented that chronic alcohol exposure might increase the levels of eCBs by inhibiting the AEA transporter in neurons exposed to chronic alcohol, and chronic alcohol additionally prevented AEA inactivation (Basavarajappa et al. 2003).

It was speculated that a rise in the levels of eCBs led to down-regulation of the CB₁ receptors. In fact, chronic exposure to alcohol evoked a reduction in CB₁ receptor density in mice (Basavarajappa et al. 1998a; Basavarajappa and Hungund 1999b; Vinod et al. 2006) and in rats (Ortiz et al. 2004). However, this phenomenon was not observed in earlier research in the rat brain (Gonzalez et al. 2002b; Rimondini et al. 2002) likely due to differences between the two experimental designs (alcohol intake, experimental time, etc.).

Moreover, chronic alcohol consumption may inhibit some processes related to gene expression. In Marchigian Sardinian alcohol-preferring (msP) rats, CB₁ receptor mRNA expression was increased and chronic alcohol consumption reduced CB₁ receptor mRNA levels in the caudate-putamen and pituitary, the brain areas relevant to the processing of reward and reward-associated behaviors (Cippitelli et al. 2005). In C57Bl/6J mice, chronic alcohol consumption induced a decrease in CB₁ receptor density in the hypothalamus and an increase in CB₁ density in the ventral tegmental area (Pava et al. 2012). Ten-day treatment of mice with alcohol also resulted in a reduction in the hypolocomotive, antinociceptive, and hypothermic effects of an acute dose of the CB₁ agonist WIN 55,212-2, while a 10-day withdrawal period reversed this effect (Pava et al. 2012). Together, these data suggest that CB₁ receptors play an important role in the neurochemical processes related to alcohol consumption.

Additionally, Agudelo et al. (2013) was the first to provide evidence of up-regulation of CB₂ and GPR55 (the third putative cannabinoid receptor) in monocyte-derived dendritic cells from alcohol abusers. Ethanol-treated cells demonstrated higher levels of CB₂ and GPR55 mRNA. Furthermore, it was observed that alcohol significantly modulated dendritic cells to produce higher levels of the pro-inflammatory cytokine IL-1β (Agudelo et al. 2013). Thus, these receptors may play an immunoprotective role during alcohol-induced immune dysfunction. It has also been reported that CB₂ receptor agonists have the ability to positively regulate Kupffer cells, which result in liver injury when activated by endotoxins in alcohol drinking individuals, and CB₂ receptors play a role in alcohol-induced inflammation (Louvet et al. 2011).

In vitro studies showed increased AEA (Basavarajappa et al. 2003) and 2-AG (Basavarajappa et al. Basavarajappa et al. 2000a, b) levels in cerebellar granule neurons with reduced FAAH activity (Basavarajappa et al. 2003) and increased AEA levels in neuroblastoma SK-N-SH cells after chronic alcohol exposure (Basavarajappa and Hungund 1999a) (Table 1). A rise in these lipid mediators may serve as a neuronal compensatory adaptation in the chronic presence of alcohol, and it should be noted that this increase was not dependent on alcohol-induced AEA transporter inhibition.

Alcohol abstinence also provoked changes within the eCB system. The increase in AEA levels was reversed in the mouse cortex (Vinod et al. 2006) and in the rat amygdala and caudate-putamen (Rubio et al. 2008). The reduced level of AEA was accompanied by a low glutamate concentration in the amygdala, and thus, it appears to be related to reduced NAPE-phospholipase D activity that depends on glutamate-mediated calcium influx (Hansen et al. 2000; Rubio et al. 2008). Another study indicated that eCB levels increased in the hippocampus after 40 days of abstinence. Specifically, increased 2-AG levels were rapidly observed only after 2 days of abstinence. Interestingly, CB₁ receptor levels were reduced after 2 days and increased after 40 days of alcohol abstinence (Mitrirattanakul et al. 2007). The increased eCB production was the result of chronic increases in hippocampal excitability on the terminals of GABAergic interneurons and in consequence may have reduced long-term increases in CB₁ activation on GABA release (Mitrirattanakul et al. 2007). An increase in CB₁ receptors was also observed in the cortex (Rimondini et al. 2002) and hippocampus (Vinod et al. 2006), while alcohol abstinence reversed the reduced levels of CB₁ in the cortex, striatum, and cerebellum in rats (Vinod et al. 2006). On the other hand, a reduction of CB₁ receptors was found in the globus pallidus and substantia nigra (Rubio et al. 2008). In C57Bl/6J mice, 10-day abstinence evoked normalization in CB₁ protein expression in the hypothalamus and ventral tegmental area (Pava et al. 2012). Apart from the present review, there is another excellent paper showing in details the preclinical interactions between alcohol and the brain eCB system (Pava and Woodward, 2012). Moreover, the effects of drugs of abuse (particularly cocaine) on the eCB system are described in the recent review paper of Vlachou and Panagis (2014). Based on the above preclinical knowledge of the role of the eCB system in processes related especially to alcohol consumption further studies on alcohol use disorder in which the eCB system plays a crucial role are urgently required.

Clinical Studies

In postmortem studies, elevated levels of eCBs were observed in the dorsolateral prefrontal cortex of alcoholic suicide victims and were suggested to inhibit GABAergic signaling and provoke impulsivity (Vinod et al. 2005). The increased eCBs levels may also play a role in reinforcing the effects of alcohol or may act more directly as a neuroprotective adaptation to chronic alcohol abuse. Although CB₁ receptors were increased in the dorsolateral prefrontal cortex (Vinod et al. 2005) and in the ventral striatum (Vinod et al. 2010) of alcoholic suicide victims, the latter changes were not associated with alcohol dependence, as a down-regulation of CB₁ receptors was noted in alcohol-dependent non-suicide victims (Smaga et al. 2014b; Vinod et al. 2010). An increase in FAAH activity was identified in the ventral striatum of alcohol-dependent suicide victims compared with alcohol-dependent non-suicide victims (Vinod et al. 2010). Another study showed a decrease in AEA levels in the nucleus accumbens and frontal cortex of Cloninger type 1 alcoholics with a reduced dopaminergic transmission in the accumbal reward system (Lehtonen et al. 2010).

Preclinical Studies

A large body of data suggested that a genetic predisposition to alcohol abuse depends on a disturbance of the eCB system. In fact, CB₁ receptor knockout mice generated on a CD₁ genetic background (outbred CD₁ mouse strain) demonstrated a reduction in alcohol-induced conditioned place preference (CPP) (Houchi et al. 2005; Thanos et al. 2005) and a decrease in consumption of alcoholic solution in the two-bottle choice test (Naassila et al. 2004). The same decreased voluntary intake of alcohol was demonstrated in C57BL/6J mice with genetic deletion of the CB₁ receptor gene (Lallemand and de Witte 2005; Poncelet et al. 2003; Vinod et al. 2008b; Wang et al. 2003). Additionally, genetic deletion of FAAH in C57BL/6J mice provoked an increase in voluntary alcohol intake in the two-bottle choice protocol (Basavarajappa et al. 2006; Blednov et al. 2007; Vinod et al. 2008a). These studies confirmed the involvement of CB₁ receptors in alcohol abuse, and pharmacological blockade of these receptors seems to be a rational approach to the treatment of alcohol addiction.

In contrast with cannabinoid agonists, cannabinoid antagonists produce opposite effects, which mediate reduction in alcohol consumption. This conclusion is based on several observations using cannabinoid antagonists. SR141716A (rimonabant), a potent and selective CB₁ antagonist, is one of the most well studied compounds and presents a link between cannabinoid receptors and alcohol consumption. The strong ability of this compound to inhibit or reduce the consumption of ethanol has been reported in alcohol-preferring rat strains, such as Sardinian alcohol-preferring (sP) rats (Colombo et al. 1998; Serra et al. 2002), Marchigian Sardinian alcohol-preferring (msP) rats (Cippitelli et al. 2005), alcohol-preferring AA rats (Hansson et al. 2007), and Warsaw High-Preferring rats (Dyr et al. 2008). Furthermore, in alcohol-preferring AA rats, a microinjection of SR141716A into the medial prefrontal cortex reduced alcohol self-administration, while microinjections into the dorsal striatum did not change the number of active lever responses for alcohol (Hansson et al. 2007). In another study, a reduction of alcohol self-administration was observed upon microinjection of SR141716A into the nucleus accumbens (Caille et al. 2007). Alcohol-preferring mice (C57BL/6J) with a genetic deletion of the functional copies of the CB₁ gene demonstrated a decrease in voluntary alcohol intake comparable to pharmacological blockade using SR141716A administration in these mice (Vinod et al. 2008b; Wang et al. 2003). The reduced alcohol self-administration was also observed in rats after acute SR141716A administration (Cippitelli et al. 2005; Economidou et al. 2006; Freedland et al. 2001; Rodriguez de Fonseca et al. 1999) with reduced conditioned reinstatement of alcohol-seeking behavior (Cippitelli et al. 2005; Economidou et al. 2006). The recent study by Marinho et al. (2014) revealed that rimonabant abolished total and peripheral locomotion only at the dose of 10 mg/kg, while lower doses did not affect alcohol-induced hyperlocomotion (Marinho et al. 2014). Nevertheless, similar inhibitory activity and effectiveness was also observed after administration of other CB₁ receptor antagonists: i) SVL330, which attenuated alcohol self-administration and reinstatement to alcohol-seeking behaviors in rats (de Bruin et al. 2011); and ii) SR147778, which reduced alcohol preference during cessation of chronic ethanol intoxication but not when co-administered with alcohol (Lallemand and De Witte 2006). This opposing behavior of SR147778 seemed to be similar to that observed for rimonabant. However, significant differences (e.g., induction of a shorter transient increase of alcohol intake) were noted. Additionally, Vasiljevik et al. (2013) reported that two of analogs of a monohydroxylated metabolite of the synthetic aminoalkylindole cannabinoid JHW-073 were shown to decrease alcohol self-administration and alcohol CPP, and, unlike rimonabant, they did not alter body weight during the treatment period (Vasiljevik et al. 2013). However, the ability to inhibit alcohol consumption was not reported for a novel CB₁ receptor antagonist, PF514273, which did not reduce the acquisition or expression of alcohol-induced CPP (Pina and Cunningham 2014). On the other hand, LH-21 reduced alcohol self-administration (Pavon et al. 2006). Although the main interest is focused on the inhibitory activity of cannabinoids on alcohol abuse, their other potential therapeutic properties have been presented. The study conducted by Jeong’s group (2008) has indicated that a CB₁ antagonist may slow down the development of alcohol fatty liver disease and thus prevent or delay its progression to more severe and irreversible forms (Jeong et al. 2008). Yang et al. (2014) proposed administration of CBD to prevent the liver damage caused by alcohol abuse. Additionally, eCB stimulation produced a neuroprotective effect on excitotoxicity induced by the cessation of chronic alcohol consumption (Rubio et al. 2011). Indeed, it has been reported that administration of HU-210, a CB₁ receptor agonist, induced a significant protection against NMDA-induced cell death during either alcohol withdrawal or alcohol exposure, while chronic administration of a CB₁ antagonist aggravated neuronal death induced by NMDA (Rubio et al. 2011) (Table 2). On the whole, these data point to the potential therapeutic application of CB antagonists in alcohol abuse.

Clinical Studies

As described above, preclinical studies provide a vast body of evidence suggesting the potential use of the CB₁ antagonists in the treatment of alcohol abuse. Unfortunately, the response of alcoholic patients to rimonabant treatment was not so beneficial. Soyka et al. (2008) examined the effect of rimonabant (20 mg per day for 12 weeks) on the rate of relapse among recently detoxified alcoholic individuals and found that rimonabant was ineffective in preventing relapse (Soyka et al. 2008). In the line with this trial, another study using rimonabant (20 mg per day for 2 weeks) in non-treatment-seeking heavy alcohol drinkers demonstrated no effect of this drug on alcohol consumption (George et al. 2010). These two groups failed to find significant differences between rimonabant and placebo in the experiments determining drug impact either on the time to first drink and relapse to heavy drinking or on overall consumption. Several aspects of this lack of effectiveness were considered, such as incorrect administration and small numbers of clinical subjects. However, in consequence, considering an analogical study with naltrexone used at higher doses of 50–150 mg/day, which significantly decreased the consumption of alcohol (Weerts et al. 2008), and the fact that both cannabinoids and naltrexone are inhibitors of the mesolimbic dopamine reward system, it has been suggested that rimonabant used at a dose 20 mg/day only partially blocked the CB₁ receptor. Therefore, a suggestion was put forth that a significant decrease in the motivation to drink alcohol can occur only as a result of a complete blockade of the signaling associated with the reward system (Cippitelli et al. 2005).

In human clinical studies, marijuana was found to exert some beneficial effects. A prospective study of heavy drinkers revealed that administration of cannabis (marijuana) reduced alcohol consumption faster than in non-cannabis users (Metrik et al. 2011). Marijuana and some cannabinoid ligands are known to have influence not only on the consumption of alcohol but also on several effects related with alcohol. Marijuana has been shown to decrease the risk of head neck cancer among moderate tobacco smokers and light alcohol drinkers (Liang et al. 2009).