Summations
∙ Psychostimulants induce microglia activation and the expression of different pro-inflammatory cytokines.
∙ Modulation of the endocannabinoid system is involved in the control of drug abuse and neuroinflammation.
∙ In the present review, we explore the hypothesis that a disruption in endocannabinoid signalling and CB2 activation lead to microglial activation and neuroinflammation, which might contribute to the process of addiction.
Considerations
∙ Evidence suggests the involvement of neuroinflammatory mechanisms mediated by endocannabinoids in the process of drug addiction, but this possibility has not been tested extensively.
∙ CB1 receptors play a role in the potentiation of reward and drug intake, and this receptor is highly expressed in the brain. Therefore, its participation in the present hypothesis should carefully studied.
Cannabinoids and addiction
The Diagnostic and Statistical Manual of Mental Disorders in its fourth edition defines substance dependence as a chronically relapsing disorder that is characterised by neurobiological changes that lead to the loss of control in restricting the intake of a substance, compulsion and negative emotional states that are induced by motivational withdrawal syndrome when drug taking is prevented (1,Reference Koob and Volkow2).
Marijuana (Cannabis sativa) is the most widely used illicit drug worldwide. The 2006 Annual Report on Drug Abuse estimates that 4% of the adult world’s population consumes Cannabis regularly (3). Cannabis can induce transient psychotic symptoms in healthy individuals (Reference Bloomfield, Morgan, Egerton, Kapur, Curran and Howes4,Reference Iversen5) and possibly increase the risk of psychotic disorders, such as schizophrenia, in a dose-dependent manner in individuals with special vulnerability (Reference Murray, Morrison, Henquet and Di Forti6). The primary psychoactive ingredient, Δ9-tetrahydrocannabinol (THC), is largely responsible for the subjective effects of C. sativa, but other phyto- and synthetic cannabinoids may also induce psychotomimetic effects (Reference Wiley7). Lower doses of THC (e.g. 1.0 mg/kg) enhance electrical brain stimulation reward in laboratory animals with electrodes implanted in the ventral tegmental area-medial forebrain bundle-nucleus accumbens reward axis (Reference Gardner, Paredes and Smith8–Reference Xi, Gilbert and Campos11).
The very first hypothesis regarding the mechanism underlying the subjective effects of THC was based on the ability of this substance to perturb the membrane permeability of neural cells (Reference Hillard, Bloom and Houslay12). However, a specific receptor for cannabinoids was proposed in the late 1980s (Reference Devane, Dysarz, Johnson, Melvin and Howlett13), and this receptor was later cloned and termed cannabinoid CB1 receptor (Reference Matsuda, Lolait, Brownstein, Young and Bonner14). The CB1 receptor is the major pharmacological target that is responsible for cannabinoid effects, including the subjective effects related to THC abuse (Reference Murray, Morrison, Henquet and Di Forti6,Reference Mechoulam and Hanus15). CB1 receptors are widely expressed in pre-synaptic terminals where they regulate excitatory and inhibitory transmission in the brain (e.g. GABAergic, glutamatergic, serotonergic, cholinergic and dopaminergic neurotransmission) (Reference Freund, Katona and Piomelli16,Reference Katona and Freund17). Endogenous ligands for the CB1 receptor (termed endocannabinoids) have been identified, and the most extensively studied ligands are the arachidonic acid derivatives arachidonoyl ethanolamide (anandamide – AEA) and 2-arachidonoyl glycerol (2-AG). The activity of these ligands is terminated by the enzymes fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL), respectively (Reference DI Marzo, Bifulco and De Petrocellis18). A second cannabinoid receptor (CB2) has also been characterised, and it was initially proposed to be absent in the central neurons. However, later research changed this notion (Reference Onaivi19).
The reward circuitry of the mammalian brain consists of synaptically interconnected neurons that link several brain regions through mainly by dopaminergic circuitries (Reference Gardner and Vorel20). GABAergic and glutamatergic neural inputs also contribute to the core reward system, which have come to be recognised as critically important in the regulation of reward processes and reward-driven behaviours. Cannabinoids were largely considered a separate class from other addictive drugs in terms of addictive potential and the neurobiological substrates that are involved during cannabinoid drug abuse effects (Reference Xi, Gilbert and Campos11,Reference Gardner and Vorel20). However, accumulating evidence now implicates brain endocannabinoid signalling in the aetiology of drug addiction (Reference Serrano and Parsons21), and several studies support the view that the endocannabinoid system represents a new candidate for the control of drug rewarding properties (Reference Maldonado, Valverde and Berrendero22), because there is dense CB1 receptor expression in brain regions that are involved in the motivational and addictive properties of abused drugs, including the ventral tegmental area, nucleus accumbens and prefrontal cortex (Reference Herkenham, Lynn, Johnson, Melvin, De Costa and Rice23). The endocannabinoids AEA and 2-AG (Reference Hillard, Bloom and Houslay12), and the enzymes responsible for their catabolism (Reference Ueda, Goparaju, Katayama, Kurahashi, Suzuki and Yamamoto24), are expressed in dopaminergic neurons. Moreover, endocannabinoids can be released following the depolarisation of neurons in brain areas related to reward circuitry (Reference Atwood and Mackie25).
CB1 receptors are abundant in the brain reward circuitry, and they participate in the addictive properties of different drugs of abuse (Reference Hermann, Marsicano and Lutz26). CB1 receptors in these circuits are expressed on glutamatergic and GABAergic interneurons in the reward circuitry of the mesolimbic system, which modulate the firing of dopaminergic neurons. In vivo microdialysis experiments showed that CB1 receptor activation increases dopamine (DA) release in the nucleus accumbens (Reference Tanda, Munzar and Goldberg27). The CB2 receptor, which is believed to control neuroinflammatory mechanisms, is also expressed in microglia and neurons (Reference Atwood and Mackie25) in the striatum and midbrain, which are areas related to reward and addiction (Reference Onaivi19). However, the presence of CB2 receptor in neurons is controversial because of questions regarding the specificity of CB2R antibodies (Reference Onaivi28).
It has been difficult to establish sensitive paradigms to evaluate the reward properties of cannabinoids. In contrast to other drugs, cannabinoids produce false negative results in some behavioural methods for investigations of abused drugs (Reference Wiley7). For example, some investigations failed to establish cannabinoid intravenous self-administration in non-human primates (Reference Mansbach, Nicholson, Martin and Balster29). However, a new reliable method was developed recently to induce self-administration of the phytocannabinoid Δ9-THC (Reference Tanda, Pontieri and Di Chiara30) and the endocannabinoids AEA or 2-AG by monkeys (Reference Justinova, Solinas, Tanda, Redhi and Goldberg31,Reference Justinová, Yasar, Redhi and Goldberg32). Justinova et al. (Reference Justinova, Solinas, Tanda, Redhi and Goldberg31) trained monkeys with no history of exposure to other drugs to learn to press a device connected to pump that delivered THC intravenously in a constant ratio. The self-administered THC doses in these studies are comparable to doses found in human marijuana users (Reference Tanda and Goldberg33,Reference Justinova, Tanda, Redhi and Goldberg34). The effects of cannabinoid self-administration were prevented by pretreatment with rimonabant, which shows involvement of CB1 receptors in reward. Intravenous AEA administration increases extracellular DA levels in the nucleus accumbens of rats in a CB1 receptor-dependent manner (Reference Solinas, Yasar and Goldberg35), which contributes to the reinforcing effects of exogenous AEA.
The abuse liability of drugs that inhibit endocannabinoid hydrolysis has also been investigated. Mice given daily injections of the FAAH inhibitors, URB597 or PF-3845, for 6 days and challenged with the CB1 antagonist SR141716A displayed no withdrawal symptoms (Reference Schlosburg, Carlson and Ramesh36). However, treatment for 6 days with high doses of the selective MAGL inhibitor JZL184 significantly reduced CB1 function and expression and SR141716A-induced somatic withdrawal symptoms (Reference Schlosburg, Carlson and Ramesh36). These findings suggest that inhibitors of endocannabinoid hydrolysis may present a lower abuse liability compared with direct CB1 agonists (Reference Serrano and Parsons21,Reference Maldonado, Valverde and Berrendero22,Reference Pava and Woodward37).
In addition to the obvious role of the endocannabinoid system on C. sativa abuse, this system may also been involved in the control of the intake of other drugs of abuse (e.g. cocaine, alcohol). Moreover, evidence in the literature support the modulation of the endocannabinoid system as a possible target for the treatment of drug abuse. The present review summarised the evidence that supports the hypothesis that neuroinflammatory mechanisms promote an integrative concept that links cannabinoids to the addictive properties of different drugs.
Cannabinoids and alcohol
The endocannabinoid system may play a role in addictive properties of different drugs of abuse. Alcohol (or ethanol) is a widely abused substance that is associated with diverse social problems. A growing body of biochemical and pharmacological evidence has established a role for the endocannabinoid system in the neurobiology of alcohol (Reference Colombo, Orrù and Lai38). Rats exposed chronically to alcohol had increased AEA content in the limbic forebrain, which is a key area for the reinforcement of psychoactive drugs (Reference Gonzalez, Fernandez-Ruiz, Sparpaglione, Parolaro and Ramos39–Reference Gonzalez, Valenti and De Miguel41). FAAH knockout (KO) mice display increased preference for alcohol and consume more ethanol than wild type mice (Reference Basavarajappa, Yalamanchili, Cravatt, Cooper and Hungund42). These studies suggest that CB1 receptors are involved in ethanol addiction. Acute CB1 receptor agonist exposure increases motivation for drinking beer (Reference Gallate, Saharov, Mallet and Mcgregor43). Similarly, the genetic deletion of CB1 receptors reduces alcohol consumption in rodents. CB1 receptor KO mice exhibit reduced voluntary alcohol consumption and do not release DA in the nucleus accumbens after alcohol consumption (Reference Hungund, Szakall, Adam, Basavarajappa and Vadasz44). Further, there is a reduction in ethanol self-administration and ethanol conditioned place preference in mice lacking CB1 receptors (Reference Thanos, Dimitrakakis, Rice, Gifford and Volkow45).
Preclinical evidence indicates that the CB1 antagonist rimonabant suppresses alcohol-related behaviours, such as alcohol drinking and seeking behaviour, and alcohol self-administration in rats and mice (Reference Colombo, Orrù and Lai38,Reference Serra, Carai and Brunetti46,Reference Gessa, Serra, Vacca, Carai and Colombo47). In constrast, clinical studies reveal controversial data. Subjects in one study were treated with rimonabant (20 mg/day for 12 weeks), and no significant benefits were observed, except a delayed time to have a first drink and relapse were noted (Reference Soyka, Koller and Schmidt48). Another study showed that rimonabant in the same treatment regimen did not change alcohol self-administration or endocrine measures during a laboratory session in non-treatment-seeking heavy alcohol drinkers (Reference George, Herion and Jones49). However, all clinical trials of rimonabant were discontinued due to adverse psychiatric effects.
As an alternative, CB2 receptors have also emerged as a potential target for alcohol abuse. CB2 receptor activation enhances alcohol intake in stressed mice, and a CB2 antagonist may induce the opposite behaviour (Reference Onaivi19). However, a more recent work shows that deletion of the CB2 receptor gene increased preference for and vulnerability to ethanol consumption (Reference Ortega-Álvaro, Ternianov, Aracil-Fernández, Navarrete, García-Gutiérrez and Manzanares50).
Therefore, the picture is not entirely clear, and the endocannabinoid system may favour or counteract neural changes that mediate alcohol abuse through a mechanism that is dependent on a predominant activity of CB1 or CB2 receptors.
Cannabinoids and psychostimulants
Several studies suggest the involvement of the endocannabinoid system in behaviours related to psychostimulants (Reference Arnold51,Reference Wiskerke, Pattij, Schoffelmeer and De Vries52). CB1 and CB2 receptors are expressed in glutamatergic and GABAergic interneurons in the reward circuitry of the mesolimbic system, which modulates dopaminergic neurons that are responsible for most effects of cocaine and amphetamine (Reference Atwood and Mackie25,Reference De Vries and Schoffelmeer53–Reference Solinas, Justinova, Goldberg and Tanda55).
One important animal model that is currently used to study the addictive behaviour of psychostimulants is behavioural sensitisation (Reference Sanchis-Segura and Spanagel56). This test is characterised by a progressive increase in a particular response, such as locomotion, after repeated exposures to a drug (Reference Sanchis-Segura and Spanagel56). Motor sensitisation to cocaine is impaired in CB1-deficient mice or after pharmacological blockade of these receptors (Reference Corbille, Valjent and Marsicano57,Reference Filip, Golda and Zaniewska58). Furthermore, genetic ablation of CB1 receptors decreases cocaine self-administration (Reference Thanos, Dimitrakakis, Rice, Gifford and Volkow45). Treatment with antagonists impairs self-administration behaviour and inhibits cocaine-enhanced brain stimulation reward (Reference Soria, Mendizabal and Tourino59,Reference Xi, Spiller and Pak60). The stress-induced reinstatement of cocaine seeking is also prevented by blockade of CB1 receptors (Reference Vaughn, Mantsch and Vranjkovic61). Reductions in CB1 receptor expression and signalling in the prefrontal cortex from human cocaine addicts and animal rodents have also been reported (Reference Gonzalez, Fernandez-Ruiz, Sparpaglione, Parolaro and Ramos39,Reference Alvaro-Bartolome and Garcia-Sevilla62).
Evidence also suggests that CB2 receptors modulate processes related to cocaine addiction. Recent studies showed a decrease in cocaine motor sensitisation and self-administration in mice overexpressing cannabinoid CB2 receptors, which suggests that this receptor is involved in cocaine-evoked behaviours (Reference Alvaro-Bartolome and Garcia-Sevilla62,Reference Aracil-Fernandez, Trigo and Garcia-Gutierrez63). Moreover, Xi et al. (Reference Xi, Peng and Li64) demonstrated that CB2 receptors modulate the rewarding and locomotor activity of cocaine via a dopaminergic neurotransmission-dependent mechanism in KO mice. However, new studies are necessary to investigate the mechanisms of the role of CB2 receptors in cocaine reward, and whether this mechanism is applicable to other drugs of abuse.
Acute and chronic cocaine administration also promotes alterations in levels of AEA and 2-AG (Reference Centoze, Battista and Rossi65–Reference Justinova, Panlilio and Goldberg67). These data suggest that the cannabinoid system attempts to modulate cocaine-induced changes. However, pharmacological interventions that culminate with increases in endocannabinoids levels have yielded controversial results. For example, inhibition of MAGL or FAAH does not alter sensitisation behaviour to cocaine. Furthermore, blockade of FAAH does not affect self-administration (Reference Justinova, Panlilio and Goldberg67,Reference Adamczyk, Mccreary, Przegalinski, Mierzejewski, Bienkowski and Filip68). However, inhibition of AEA hydrolysis prevented reinstatement of cocaine-seeking behaviours (Reference Adamczyk, Mccreary, Przegalinski, Mierzejewski, Bienkowski and Filip68).
The involvement of endocannabinoid systems in the reinforcing effects of amphetamines has also been studied. Analogous to the results with cocaine, pharmacological blockade of CB1 receptors reduces amphetamine self-administration and decreases the reinstatement of drug seeking (Reference Vinklerova, Novakova and Sulcova69,Reference Anggadiredja, Nakamichi and Hiranita70). Depletion of CB1 receptors also attenuates drug-induced acute hyperlocomotion. Furthermore, CB1 KO mice did not sensitise to the locomotor stimulant effects of amphetamine (Reference Murray, Morrison, Henquet and Di Forti6). Results with CB1 antagonists in motor sensitisation were controversial because both inhibition and potentiation of this behaviour occurred (Reference Thiemann, van der Stelt, Petrosino, Molleman, Di Marzo and Hasenohrl71,Reference Runkorg, Orav, Koks, Matsui, Volke and Vasar72).
Acute or chronic amphetamine treatment increase anandamide concentrations in the dorsal striatum and decrease AEA and 2-AG levels in the ventral striatum (Reference Thiemann, van der Stelt, Petrosino, Molleman, Di Marzo and Hasenohrl71). In contrast, methamphetamine administration reduced MAGL activity and increased 2-AG levels in the limbic forebrain 7 days after neurotoxic doses (Reference Gutierrez-Lopez, Llopis, Feng, Barrett, O'Shea and Colado73). A recent study also showed that facilitation of anandamide neurotransmission attenuated amphetamine-induced behavioural sensitisation (Reference Eisenstein, Holmes and Hohmann74).
Cannabinoids and opiates
Several studies demonstrated that the endocannabinoid system modulates distinct opioid-induced responses, such as pain, anxiety and reward (Reference Scavone, Sterling and Van Bockstaele75,Reference Tucci76). A functional and bidirectional interaction between the endocannabinod and the opioid system is observed for reward (Reference Scavone, Sterling and Van Bockstaele75). Blockade of opioid receptors reverses the effects of THC, and conversely, blockade of cannabinoid receptors prevents the development of morphine self-administration and conditioned place preference in rodents (Reference Solinas, Zangen, Thiriet and Goldberg77,Reference Chaperon, Soubrie, Puech and Thiebot78). Consistent with these findings, the CB1 receptor antagonist, rimonabant, reduces the reinforcing effects of self-administered heroin and inhibits reinstatement to this opiate (Reference Navarro, Carrera and Fratta79,Reference Fattore, Spano, Cossu, Deiana and Fratta80). A very similar panorama is found in CB1 KO mice, which show reduced behavioural sensitisation, conditioned place preference and self-administration induced by opiates (Reference Cossu, Ledent and Fattore81,Reference Ledent, Valverde and Cossu82). The pharmacological and genetic blockade of CB1 receptors also attenuates opioid withdrawal syndrome (Reference Ledent, Valverde and Cossu82).
Interestingly, one recent study suggested that facilitation of endocannabinoid signalling also reduces withdrawal in morphine-dependent mice (Reference Ramesh, Haney and Cooper83). Although this effect might seem controversial, a clinical study noted that moderate cannabis use is associated to better naltrexone treatment adherence in opiate-dependent patients (Reference Raby, Carpenter and Rothenberg84). These data demonstrated that enhancement of endocannabinoid levels and blockade of CB1 receptors ameliorate reward, reinstatement and withdrawal promoted by opiates.
Cannabinoids and nicotine
Nicotine is the main psychoactive constituent of tobacco, and it is responsible for the development of dependence (Reference Xi, Spiller and Gardner85). Notably, frequent concomitant consumption of marijuana and tobacco is reported (Reference Peters, Schwartz, Wang, O'Grady and Blanco86), which may reflect a possible interaction between these systems. Preclinical studies revealed that co-administration of non-effective doses of nicotine and THC produced significant conditioned place preference in mice (Reference Valjent, Mitchell, Besson, Caboche and Maldonado87). In addition, alterations in endocannabinoid levels in distinct brain regions was observed in animals chronically treated with nicotine (3).
Cross-talk between nicotine addiction and the endocannabinoid system was confirmed in experiments that showed that treatment with a CB1 antagonist reduced nicotine self-administration and place preference in rodents that was associated with a decrease in DA release in the nucleus accumbens (Reference Cohen, Perrault, Voltz, Steinberg and Soubrie88,Reference Foll, Goldberg and Rimonabant89). In agreement with these results, CB1 KO mice do not express behaviours related to nicotine-induced CPP or nicotine self-administration (Reference Cohen, Kodas and Griebel90,Reference Castane91). Similar responses were observed with pharmacological and genetic blockade of CB2 receptors (Reference Ignatowska-Jankowska92).
One clinical study also suggested rimonabant as a potential therapeutic tool for relieving the symptoms of smoking cessation. Rates of smoking cessation for subjects who received a major dose of rimonabant were double the rates of patients who received placebo (Reference Gelfand and Cannon93). Another study suggested that rimonabant and nicotine patch treatment also decreases smoking compulsion (Reference Rigotti, Gonzales, Dale, Lawrence, Chang and Group94).
Facilitation of endocannabinoid signalling may also impact nicotine reward. Pharmacological and genetic FAAH disruption in mice enhances nicotine reward and withdrawal (Reference Muldoon, Lichtman, Parsons and Damaj95). However, pharmacological blockade of FAAH significantly inhibits nicotine reward but has no effect on nicotine withdrawal in rats (Reference Muldoon, Lichtman, Parsons and Damaj95). These latter symptoms were not modified after blockade of cannabinoid receptors (Reference Tanda, Munzar and Goldberg27).
Neuroinflammation and addiction: role of (endo)cannabinoids
The exact mechanisms of the genesis of addiction are not well understood, but it is possible to speculate that neuroinflammation plays a role in the pathophysiology of this condition. In fact, evidence in the literature suggests that psychostimulants induce microglia activation and the expression of different cytokines, such as tumour necrosis factor (TNF)α and interleukins (IL) and nitric oxide in animal models, and these results may also be present in humans (Reference Luque-Rojas, Galeano and Suarez96–Reference Sadasivan, Pond, Pani, Qu, Jiao and Smeyne98). Moreover, these cytokines could, per se, facilitate addiction development. For example, IL-1β increases mRNA expression and activity of serotonin transporters in human JAR choriocarcinoma cells. This increased activity of serotonin transporters could enhance the effect of psychostimulants that target these transporters.
Five-day intraperitoneal (i.p.) injections of IL-2 in male BALB/c mice enhance the sensitivity of animals to a selective DA uptake inhibitor 5 weeks after cytokine treatment (Reference Ramamoorthy, Ramamoorthy and Prasad99). These long-lasting changes induced by IL-2 might be important for central nervous system (CNS) abnormalities that are induced by addictive drugs. Similar to the long-lasting effect observed with IL-2 treatment, 5-day treatment with IL-6 increased the sensitivity to the locomotor-stimulating effects of an amphetamine administered 14 days after the last i.p. administration of IL-6 (Reference Zalcman, Savina and Wise100). Moreover, IL-2 and interferon (IFN)-α potentiated the response induced by a psychostimulant drug in a protocol of drug discrimination behavioural effect using d-amphetamine (Reference Zalcman101,Reference Ho, Lu and Huo102). In contrast, TNF-α might play a different role in addiction. Methamphetamine increases the expression of this cytokine, which could attenuate the rewarding effects and discriminative stimulus effects of this psychostimulant in rats (Reference Nakajima, Yamada and Nagai103,Reference Yamada and Nabeshima104). Serum levels of IL-10 are decreased in human cocaine abusers but TNF-α expression is increased (Reference Solinas, Zangen, Thiriet and Goldberg77). Cocaine infusions rapidly increase the production of IFN-γ and decrease IL-10 secretion from polymorphonuclear cells (Reference Gan, Zhang and Newton105). Cocaine withdrawal in rats is associated with increases in plasma levels of corticoids (Reference Avila, Morgan and Bayer106). Amphetamine increases the number of circulating neutrophils but decreases circulating lymphocytes (Reference Llorente-Garcia, Abreu-Gonzalez and Gonzalez-Hernandez107).
Psychostimulants may also increase the activation of transcription factors. For example, methamphetamine increases the activation of nuclear factor kappa B (NF-κB) and activator protein 1 (AP-1) in endothelial cells (Reference Lee, Hennig, Yao and Toborek108). Activation of these transcription factors induces the expression of different inflammatory mediators, which might be important for the development of drug dependence.
The immune system is also altered in chronic alcohol abusers. Changes in circulating immunoglobulin (Ig) levels were the first described immune system link with alcohol use disorder (Reference Bogdal, Cichecka, Kirchmayer, Mika and Tarnawski109). Alcohol alters immune function in part by its effects on neurotransmitter, neuroendocrine, behavioural and autonomic pathways (Reference Redwine, Dang, Hall and Irwin110–Reference Pettinati, O'brien and Dundon114). However, many studies reported that cytokine levels were altered in alcohol user disorder patients with or without liver disease (Reference Szabo, Mandrekar, Petrasek and Catalano115–Reference Nicolaou, Chatzipanagiotou, Tzivos, Tzavellas, Boufidou and Liappas117). Alcohol consumption in moderate or higher levels is also associated with increased IL-10, type 1 helper T cell (Th1) activation, decreased macrophage-derived chemokine concentrations and suppression of the NF-κB (Reference Chiva-Blanch, Urpi-Sarda and Llorach118–Reference Mandrekar, Bellerose and Szabo120).
The connection between the immune system and alcohol abuse is even more pronounced in animal models. Chronic alcohol administration in rodents increases TNF-α, IL-1β and IL-6 levels in the brain (Reference Zhao, Wang, Fan, Ping, Yang and Wu121). Chronic alcohol consumption is also associated with increases in pro-inflammatory cytokines in glial cells for the activation of toll-like receptor 4 (TLR-4). One study in mice lacking TLR-4 showed that this receptor prevents neuroinflammation-induced damage after chronic alcohol consumption (Reference Pascual, Fernandez-Lizarbe and Guerri122). Apparently, these receptors are responsible for the recognition of several molecules derived from microorganisms and by stimulation of innate immune responses (Reference Szabo, Mandrekar, Oak and Mayerle123). Receptor recognition triggers the execution of a sequence of signals, and genes that encode the pro-inflammatory cytokines TNF-α, IL-1β and IL-2 are expressed (Reference Fang, Wang, Zhou, Wang and Yang124).
Another possible link are microglial cells. Microglia are an important source of inflammatory mediators in the CNS, and these cells might be associated to addiction-related processes. Importantly, it has been hypothesised that primed microglia could release inflammatory mediators that are induced by different stressors during the early recovery period from addiction, which induces sickness behaviour syndromes that could function as a first step for relapse behaviour (Reference Luque-Rojas, Galeano and Suarez96). Microglia activation and the production of inflammatory mediators might also play important roles in the plasticity that accompanies the development and maintenance of drug abuse (Reference Agrawal, Hewetson, George, Syapin and Bergeson125–Reference Kovacs127). Moreover, increases in dopaminergic neurotransmission are also involved in microglial activation. DA depletion caused by α-MPT prevented the activation of mesencephalic microglia and the subsequent TH neuron loss induced by an intranigral injection of LPS. Moreover, psychostimulants, such as methamphetamine, induce neurotoxic effects by increasing microglial activation and inflammation that is dependent on DA release (Reference Thomas, Francescutti-Verbeem and Kuhn128,Reference Espinosa-Oliva, De Pablos and Sarmiento129).
Recently, Zhao et al. (Reference Zhao, Wang, Fan, Ping, Yang and Wu121) suggested that exposure to short cycles of alcohol administration (4 days) and periods of withdrawal (6 days) increases the activation of microglia, and the neurodegeneration that is associated with declines in learning and memory processes. In addition, methamphetamine increases the production of IL-6 and IL-18 in cultures of human foetal glial cells (Reference Shah, Silverstein, Singh and Kumar130). However, this result is not clear and deserves further investigation.
Therefore, it could be assumed in this context that (endo)cannabinoids could modulate addiction through its effects on microglia activation and the production of inflammatory mediators. Importantly, different papers have demonstrated the presence of CB1 and CB2 receptors on immune cells (Reference Galiegue, Mary, Marchand, Dussossoy and Carriere131) and that cannabinoids reduced the binding of the respective transcription factors to CRE and NF-κB in these cells (Reference Kerr, Harhen and Okine132). Cannabinoids may act as immunomodulators by inhibiting cytokine and chemokine production, the cell proliferation and expansion of regulatory T cells, and the induction of apoptosis in these cells (Reference Klein, Friedman and Specter133,Reference Klein, Newton and Larsen134).
Evidence has been presented that 2-AG protects neurons exposed to harmful insults, such as inflammation, by acting as an endogenous inhibitor of cyclooxygenase-2 (Reference Zhang and Chen135). AEA inhibits TNF-α-induced NF-κB activation by direct inhibition of the IκB kinase (Reference Sancho, Calzado, Di Marzo, Appendino and Munoz136). However, AEA and 2-AG degradation increased the production of prostaglandins in activated glial cells (Reference Navarrete, Fiebich and De Vinuesa137,Reference Nomura, Morrison and Blankman138). This effect might be due to the hydrolysis of AEA and 2-AG to arachidonic acid, which leads to enhanced levels of substrate for the formation of prostaglandins (Reference Nomura, Morrison and Blankman138). The apparent contradictory pro-inflammatory effect of endocannabinoids seems to be mediated by a CB1/CB2 receptor-independent mechanism, and it is prevented by endocannabinoid hydrolysis inhibitors (FAAH and MAGL inhibitors) (Reference Navarrete, Fiebich and De Vinuesa137,Reference Nomura, Morrison and Blankman138). In addition, it is important to stress that the enhanced levels of prostaglandins in the brain should not be interpreted as a mere pro-inflammatory signal, because these mediators also possess anti-inflammatory effects (Reference Shi, Johansson, Woodling, Wang, Montine and Andreasson139).
Treatment with WIN55,212-2, a CB1/CB2 receptor agonist, reduced mRNA expression of pro-inflammatory cytokines, TNF-α, IL-1β, IL-6 and IFN-γ, in the CNS in a viral model of multiple sclerosis (Reference Croxford and Miller140). The pharmacological inhibition of AEA hydrolysis reduces microglial activation, nitric oxide levels and the production of several inflammatory mediators, such as TNF-α, IL-6, IL-1β and IL-12, most likely due to the activation of CB2 receptors (Reference Eljaschewitsch, Witting and Mawrin141–Reference Ortega-Gutierrez143).
Another possible mechanism of inhibition of cannabinoid receptor-dependent inflammatory response might involve the activation of peroxisome proliferator-activated receptors (PPARs). Several studies in the last decade reported non-CB1 and non-CB2-mediated cannabinoid effects, and several cannabinoids interact with PPARs but in a complex manner (Reference Sun and Bennet144). The PPAR family (PPARα, PPARβ and PPARγ) plays important roles in the maintenance of lipid metabolism, peroxisomal enzyme expression, insulin sensitivity, glucose homoeostasis, cell proliferation, apoptosis and inflammation (Reference Stahel, Smith, Bruchis and Rabb145). Indeed, a large body of evidence suggests that PPAR-γ mediates some of the modulatory effects of cannabinoids on neuroinflammation (Reference O’Sullivan, Bennet, Kendal and Randall146,Reference Enayatfard, Rostami, Nasoohi, Oryan, Ahmadiani and Dargahi147). Activation of PPARs inhibits the transcription of pro-inflammatory genes that prevent the signalling pathway of NF-κB (Reference Enayatfard, Rostami, Nasoohi, Oryan, Ahmadiani and Dargahi147–Reference Rockwell, Snider, Thompson, Vanden Heuvel and Kaminski150). In particular, several protective effects of PPARγ have been demonstrated in the brain (Reference Combs, Johonson, Karlo, Cannady and Landreth151,Reference Colino, Aragno and Mastrocola152). For example, it was demonstrated recently that neuroinflammatory mechanisms and PPAR-γ are involved in the behavioural sensitisation that is induced by the synthetic cannabinoid WIN55,212-2 (Reference Enayatfard, Rostami, Nasoohi, Oryan, Ahmadiani and Dargahi147). In addition, the endocannabinoid 2-AG may decrease IL-2 production via PPAR-γ activation (Reference Rockwell153). The PPAR receptors may also play a role in addiction. Coincidentally, major findings of the putative role of PPARs in addiction come from studies conducted with cannabinoids (Reference Le Foll, Di Ciano, Panlilio, Goldberg and Ciccocioppo154). Currently, several pieces of evidence suggest that PPAR-α and PPAR-γ agonists play a role in relapse, sensitisation, conditioned place preference, withdrawal and drug intake induced by psychostimulants (Reference Le Foll, Di Ciano, Panlilio, Goldberg and Ciccocioppo154).
It has been proposed that chronic activation of microglia plays a major role in disorders that are characterised by nervous tissue inflammation (Reference Matute, Alberdi, Ibarretxe and Sanchez-Gomez155). CB1 is expressed constitutively in microglia, and CB2 is expressed in microglia during activation states (Reference Ortega-Gutierrez143,Reference Matute, Alberdi, Ibarretxe and Sanchez-Gomez155,Reference Ehrhart, Obregon and Mori156). However, CB1 receptors have been suggested to modulate inflammation (Reference Albayram, Alferink and Pitsch157,Reference Mnich, Hiebsch, Huff and Muthian158), but it could also play a role in potential addictive properties of C. sativa, which is a potential contradiction. However, CB1 activation by Cannabis consumption would increase dopaminergic neurotransmission, which could lead to the activation of microglia via a CB1-independent mechanism (Reference Thomas, Francescutti-Verbeem and Kuhn128).
CB2 is predominantly expressed in the immune system and, in our opinion, it might be a key mediator of cannabinoid regulation of immune functions during addiction via microglia activation (Reference Correa, Mestre, Docagne and Guaza159–Reference Merighi, Gessi and Varani161). Primed microglia could release inflammatory mediators and contribute to the maintenance of drug abuse (Reference Agrawal, Hewetson, George, Syapin and Bergeson125,Reference Hutchinson, Northcutt and Chao126,Reference Enayatfard, Rostami, Nasoohi, Oryan, Ahmadiani and Dargahi147), and cannabinoids would be involved in the attenuation of this effect. However, CB1 activation would control DA release due to its primarily pre-synaptic location, and CB2 activation would affect DA neurotransmission by decreasing inflammatory responses (pro-inflammatory cytokines, nitric oxide, etc.) (Reference Morales and Bonci162).
CB2 receptor stimulation reduced morphine-induced production of inflammatory mediators from activated microglia (Reference Merighi, Gessi, Varani, Fazzi, Mirandola and Borea163). Deletion of the CB2R gene increases the preference for ethanol consumption. This effect could be mediated by the increase in mRNA expression of tyrosine hydroxylase and µ opioid receptors in the ventral tegmental area and nucleus accumbens, respectively (Reference Ortega-Álvaro, Ternianov, Aracil-Fernández, Navarrete, García-Gutiérrez and Manzanares50). However, other mechanisms might also contribute to this action. Moreover, activation of CB2 receptors also reduced the rewarding and locomotor-stimulating effects of cocaine. One possibility to explain this effect is that activation of CB2 receptors on astrocytes or microglia could alter the production of pro-inflammatory mediators, which would inhibit DA release from the nucleus accumbens (Reference Xi, Peng and Li164). Therefore, we speculate that impaired CB2 receptor signalling would contribute to the reinforcing effects of different drugs because CB2 receptors regulate the expression of inflammatory mediators, which possess important roles in addiction. Therefore, it is reasonable to suggest that the effects of cannabinoids on drug abuse might be mediated by neuroinflammatory mechanisms via CB2 receptors on microglia.
Perspectives and conclusions
The present paper reviewed the possible role of neuroinflammation in the mechanism of cannabinoids on drug addiction. We explored the hypothesis that a disruption in cannabinoid signalling during drug addiction processes would involve microglial activation and a consequent neuroinflammation (Fig. 1).
Role of microglial CB2 receptor in the possible mechanism linking neuroinflammation, (endo)cannabinoids and addiction.
Several pieces of evidence suggested that abused drugs, such as psychostimulants drugs or alcohol, induce microglia activation and the expression of inflammatory mediators, such as cytokines and transcription factors. (Endo)cannabinoids may act as immunomodulators by inhibiting cytokine production and microglia activation. This latter mechanism seems particularly important because CB2 receptors on activated microglia might play a major role in neuroinflammatory processes related to addiction. Notably, several studies support a role of PPAR receptors in the anti-inflammatory effects of cannabinoids, mainly in the CNS. Activation of PPARs exerts anti-inflammatory effects by inhibiting the expression of pro-inflammatory genes and reducing the production of cytokines, metalloproteases and acute-phase proteins. An increasing body of evidence shows that (endo)cannabinoids activate PPARs, which have anti-inflammatory activities, and the activation of these nuclear receptors may represent a novel mechanism by which cannabinoids modulate inflammatory conditions. However, additional studies designed to test this hypothesis are needed to elucidate the contribution of neuroinflammation on the behavioural and neuroprotective effects of cannabinoids on drug addiction.
Acknowledgements
The authors thank Radael Júnior for his technical support in relation to design and graphic art. This work was supported by grants from CNPq, CAPES, FAPES (Programa Primeiros Projetos (PPP) number 53630408/2011) and FAPEMIG (Programa Primeiros Projetos (PPP), number CBB-APQ-04389-10; Programa de Apoio a Núcleos Emergentes de Pesquisa (PRONEM), number CBB-APQ-04625-10) and Pró-Reitoria de Pesquisa da UFMG (PRPq), Brazil. FAM and ALT are recipients of productivity research fellowships (level 2 and 1C, respectively) from CNPq. ACC is a recipient of CAPES fellowship.
Conflicts of Interest
The authors declare no conflicts of interest.
