“Weeding out” violence? Translational perspectives on the neuropsychobiological links between cannabis and aggression

Abstract

Recent shifts in societal attitudes towards cannabis have led to a dramatic increase in consumption rates in many Western countries, particularly among young people. This trend has shed light on a significant link between cannabis use disorder (CUD) and pathological reactive aggression, a condition involving disproportionate aggressive and violent reactions to minor provocations. The discourse on the connection between cannabis use and aggression is frequently enmeshed in political and legal discussions, leading to a polarized understanding of the causative relationship between cannabis use and aggression. However, integrative analyses from both human and animal research indicate a complex, bidirectional interplay between cannabis misuse and pathological aggression. On the one hand, emerging research reveals a shared genetic and environmental predisposition for both cannabis use and aggression, suggesting a common underlying biological mechanism. On the other hand, there is evidence that cannabis consumption can lead to violent behaviors while also being used as a self-medication strategy to mitigate the negative emotions associated with pathological reactive aggression. This suggests that the coexistence of pathological aggression and CUD may result from overlapping vulnerabilities, potentially creating a self-perpetuating cycle where each condition exacerbates the other, escalating into externalizing and violent behaviors. This article aims to synthesize existing research on the intricate connections between these issues and propose a theoretical model to explain the neurobiological mechanisms underpinning this complex relationship.

1. THE COMORBIDITY OF CANNABIS USE AND PATHOLOGICAL AGGRESSION

Cannabis is the most widely used illegal psychoactive substance globally, with its usage estimated to be around 200 million people (Bahji & Stephenson, 2019). Over the past few years, societal and legal attitudes all around the world have led to the decriminalization and/or legalization of cannabis. At the time of this writing (December 2023), recreational consumption of hemp products was legal in 8 countries and 24 US states, and it was decriminalized in many additional jurisdictions in Western Europe and the Americas. This phenomenon has been paralleled by a significant uptick in cannabis use, particularly among young adults, and a surge in cannabis use disorder (CUD) prevalence (Cerdá et al., 2020; Hasin et al., 2016; Melchior et al., 2019). Indeed, approximately 22% of cannabis users meet diagnostic criteria for CUD (Leung et al., 2020). These trends have raised concerns given the adverse psychiatric events associated with cannabis use, particularly among adolescents (Lowe et al., 2019).

A significant yet frequently overlooked comorbidity in cannabis users is pathological aggression, an extreme and often persistent form of aggression that is usually out of proportion to its trigger and frequently involves physical violence. According to Håkansson & Jesionowska (2018), cannabis ranks third in terms of its association with violence, following alcohol binge drinking and sedative drug use. Recent meta-analyses have identified a mild to moderate increased risk of violence in specific vulnerable groups, including adolescents, young adults, and individuals with severe mental disorders (Dellazizzo, Potvin, Athanassiou, et al., 2020). Notably, a moderate link (Odds ratio: 2.11) was found between cannabis use and the perpetration of physical violence among youth (Dellazizzo, Potvin, Dou, et al., 2020). Early research indicates a dose-response correlation between cannabis consumption and aggressive behavior. This link persists even after adjusting for potential confounders such as socio-demographic factors, psychopathic traits, and the use of other substances (Macleod et al., 2004).

Consistent and early use of cannabis is linked with heightened mental health and behavioral issues, including aggression and delinquency (Brook et al., 2014; Huas et al., 2008; Schoeler, Monk, et al., 2016; Windle & Wiesner, 2004). Additionally, a moderate correlation (Odds Ratio: 3.02) was observed between cannabis use and violence in individuals with schizophrenia, bipolar disorders, and major depression (Dellazizzo et al., 2019). However, understanding the causes of this comorbidity is challenging due to the methodological constraints of most existing studies (Sorkhou et al., 2022), including their retrospective and cross-sectional nature. Despite these limitations, several articles suggest that cannabis use in early adolescence may predispose individuals to violence in young adulthood (Harris et al., 2010; Reynolds et al., 2011; Walton et al., 2009). Recent longitudinal studies indicate that cannabis use frequency and severity increase the risk of violent behavior (Beaudoin et al., 2020, 2023; Reingle et al., 2013). However, other studies do not support a clear link between cannabis use in adolescence and violence risk (Green et al., 2010; Pedersen & Skardhamar, 2010).

Complementary to these findings, additional research posits that the link between cannabis consumption and aggressive behavior may be bidirectional. For example, several studies have documented that engaging in physical violence during adolescence and early adulthood heightens the risk of later cannabis initiation (Herrenkohl et al., 2009; Lim & Lui, 2016; Schoeler, Theobald et al., 2016; White et al., 1999).

The already complex relationship between cannabis use and aggression becomes even more intricate when taking into account the diverse effects associated with different CUD stages. Several studies have indicated that cannabis withdrawal is highly associated with aggression, irritability, and anger (Bonnet & Preuss, 2017; Budney et al., 2003; Budney & Hughes, 2006; Kouri et al., 1999; Livne et al., 2019; P. H. Smith et al., 2013), which become particularly evident during the initial week following cessation of cannabis use (Budney et al., 2003). Aggressive responses during cannabis withdrawal are characterized by sudden onset, followed by a swift intensification (Levin et al., 2010). Collectively, these findings suggest that, at least in a subset of CUD-affected individuals, cannabis consumption may mitigate, rather than induce, aggression, and its sudden discontinuation may therefore lead to aggressive outbursts. Consistent with this framework, aggression and anger have been identified as key predictors of future relapses in CUD (Allsop et al., 2012; Copersino et al., 2006; Levin et al., 2010), and contributors to the failure of treatments aimed at cannabis detoxification (Budney et al., 2008). Furthermore, some research suggests that individuals with aggressive inclinations might resort to cannabis as a means to mitigate antisocial tendencies and alleviate negative emotional states (Arendt et al., 2007).

Epidemiological data aside, a connection between cannabis use and aggressive behavior is also confirmed by findings from studies based on experimental laboratory paradigms such as the point-subtraction aggression paradigm (PSAP). Specifically, cannabis use was reported to decrease subjective aggression following exposure to aggressive stimuli and reduce aggressive responses during the PSAP compared to a placebo (De Sousa Fernandes Perna et al., 2016). The PSAP was also found to elicit higher aggression scores among individuals with cannabis withdrawal syndrome (Kurt Tunagur et al., 2022).

One of the reasons behind the mixed findings on the link between aggression and cannabis use is likely due to the lack of distinctions between subtypes of aggressive behaviors in most previous research (Ostrowsky, 2011). While statistically related, a substantial body of research supports the existence of two functions of aggression, specifically reactive (or impulsive) and proactive (or instrumental) (Dodge & Coie, 1987; Fite et al., 2006, 2016, 2023; Houston et al., 2003; Kempes et al., 2005; Poulin & Boivin, 2000; Vitaro & Brendgen, 2012). Reactive aggression occurs in response to a perceived threat, and this impulsive and retaliatory behavior appears to be best explained by the frustration-aggression hypothesis in which increasing frustration and anger lead to aggressive outbursts (Berkowitz, 1993; Crick & Dodge, 1996; Dodge, 1991). In contrast, proactive aggression is goal-oriented and appears consistent with social learning theory, in which behaviors are maintained through the anticipation of rewards (Bandura, 1986; Crick & Dodge, 1996; Dodge, 1991; Kempes et al., 2005). Of these two functions of aggression, reactive is the most common (Fite et al., 2023; Fite et al., 2016; Vitaro & Brendgen, 2012), with approximately 60% of the 4.5 million violent crimes committed in the United States each year being reactive or impulsive (Brouwers et al., 2010). These aggressive acts lead to significant societal costs, including property damage, court and imprisonment, psychological treatment, medical services, and lost wages (Foster & Jones, 2005). Available data appear to support that most aggressive acts linked to cannabis use are indeed reactive (see below).

The intricate connection between pathological reactive aggression and cannabis is influenced by numerous elements, encompassing shared genetic and environmental susceptibilities, psychological traits, and individual variations in the endocannabinoid system within the brain. In addition, the effects of cannabis consumption on aggression are shaped by the specific cannabis consumption regimen (in terms of both dosage and frequency of use) and the chemotypes of cannabis, which reflect the relative concentrations of its ingredients. In the subsequent sections, we will examine how these diverse factors can impact the association between aggression and the use of cannabis.

2. SHARED CAUSES OF CANNABIS USE AND REACTIVE AGGRESSION

2.1. Environmental predispositions.

Ample evidence indicates that the propensity to engage in reactive aggression and consume cannabis is primarily influenced by socioeconomic and other environmental factors, particularly throughout childhood and adolescence. This section will examine the main environmental factors that increase the risk for both problems.

2.1.1. Child maltreatment.

The global prevalence of child maltreatment (including physical, sexual, and emotional abuse, as well as physical and emotional neglect) is substantial, with 50% of children suffering some form of violence each year globally (Hillis et al., 2017). Ample evidence has documented that child maltreatment is a key risk factor for reactive aggression (Dambacher et al., 2022; McRae et al., 2021), likely due to emotion dysregulation (Shields & Cicchetti, 1998). In turn, this process may explain the high rate of juvenile delinquency and adult crime in individuals with a history of child abuse and neglect (Norman et al., 2012; Springer et al., 2007; Thornberry et al., 2010; Wilson & Widom, 2009). For example, prospective and longitudinal studies from the 1980s and 1990s have supported the ‘cycle of violence’ hypothesis, suggesting that maltreated children are more likely to exhibit antisocial behavior later in life (Ehrensaft et al., 2003; Maxfield & Widom, 1996; Ogloff et al., 2012; C. A. Smith et al., 2005; Topitzes et al., 2012; Widom, 1989). A recent meta-analysis of longitudinal studies showed a strong association between childhood and adolescent maltreatment and an increased likelihood of antisocial behavior in adulthood. This association was found to be stronger for individuals suffering from a longer duration of maltreatment (Braga et al., 2018).

Child maltreatment has also been shown to be a key risk factor for several substance use disorders (Cicchetti & Handley, 2019; Halpern et al., 2018). Cross-sectional studies have shown strong correlations between child maltreatment and cannabis use (Afifi et al., 2012, 2023). Using prospective official data in a large birth cohort study, Mills et al. (2017) also uncovered a strong link between various forms of child maltreatment and the onset of cannabis use in early adulthood. A recent meta-analysis concluded that there is robust evidence of a relationship, particularly between physical (OR: 1.39) and sexual abuse (OR: 1.29) in childhood and increased likelihood of adolescent cannabis use (De la Peña-Arteaga et al., 2021).

Several theories explain the link between maltreatment and both antisocial behavior and cannabis use. Social learning theories suggest children learn violent behaviors from their environments (Bandura, 1977; Dodge & Crick, 1990; Stith et al., 2000), while general strain theory posits that maltreatment creates emotional strain leading to antisocial acts (Agnew, 1992, 2001). The developmental psychopathology perspective focuses on the interaction between abnormal developmental processes and environmental factors (Cicchetti & Toth, 2005; Cicchetti & Valentino, 2006; Toth & Cicchetti, 2013). On the other hand, adverse childhood experiences could heighten susceptibility to addiction (Kim et al., 2017). The self-medication hypothesis posits that substance use, like cannabis, could be a coping mechanism for psychological distress stemming from such adverse childhood experiences (Khantzian, 1997). Alternatively, cannabis use may reflect broader risk-taking or externalizing behavior patterns, often linked to mental health disorders and a history of childhood maltreatment (Fergusson & Lynskey, 1997; Grogan-Kaylor et al., 2008).

2.1.2. Negative Life Events (NLEs).

NLEs are life-changing experiences requiring one to adjust to cope with the event(s) that have occurred (Fink, 2010; Nou, 2009). NLEs include a variety of experiences, such as witnessing or being the victim of a crime or assault, living with a severe illness, and having a family member with a mental/emotional, drug, or alcohol problem (Swearingen & Cohen, 1985). NLEs have been associated with a variety of externalizing behaviors (Buehler & Gerard, 2013; Katz et al., 2012), including both reactive (but not proactive) aggression (Fite et al., 2016; McRae et al., 2021; Silvern & Griese, 2012) and cannabis use (Kilpatrick et al., 2000; Van der Pol et al., 2013). Moreover, Fite and colleagues (2015) found that NLEs partially accounted for the link between reactive aggression and cannabis use initiation in a sample of Latino adolescents. The influence of NLEs on reactive aggression and cannabis use appears to be consistent with the stress-process framework, which suggests that NLEs may result in poor behavioral and emotional regulation (Roosa et al., 2010; Thoits, 1983; Turner & Finkelhor, 1996), and poor emotion regulation is associated with both reactive aggression (Fite et al., 2016) and cannabis use (Bonn-Miller et al., 2011; Simons et al., 2005; Wills et al., 2006).

2.1.3. Poor Social Relationships.

Reactively aggressive individuals experience peer difficulties (Fite et al., 2023; Fite et al., 2016; Vitaro & Brendgen, 2012). More specifically, reactive (but not proactive) aggression has been consistently associated with peer rejection and low levels of peer acceptance (Day et al., 1992; Dodge et al., 1997; Fite et al., et al., 2012; Prinstein & Cillessen, 2003; Raine et al., 2006). Further, reactively aggressive behavior is associated with high levels of peer conflict, low friendship satisfaction, and a low likelihood of being selected as a best friend (Poulin & Boivin, 1999). Similarly, social difficulties, including peer rejection and social anxiety, are associated with an increased risk of cannabis use (e.g., Buckner et al., 2007; Krygsman & Vaillancourt, 2022; Prinstein & La Greca, 2004). For example, in a sample of middle school-age youth, high levels of peer victimization were associated with elevated levels of cannabis use for both males and females (Wormington et al., 2013). Moreover, in a series of studies, Deckman and colleagues (2014) found that cannabis use can counteract the negative effects of experiencing loneliness and social exclusion among adolescents and adults. Thus, cannabis use may be a way to escape social discomfort.

2.1.4. Poor Academic Achievement.

Poor academic achievement is the precursor to a host of lifelong difficulties, including poor job satisfaction, inadequate work competence, contact with law enforcement, and poor romantic relationships (Masten et al., 1995, 2005; Roisman et al., 2004). Furthermore, it has been linked with externalizing problem behaviors (e.g., Dodge & Pettit, 2003; Fite et al., 2012; Hinshaw, 1992; Loveland et al., 2007; Masten et al., 2005). In particular, a small body of research suggests a unique association between reactive (but not proactive) aggression and poor academic performance (Day et al., 1992). For instance, one study found that high levels of reactive, but not proactive, aggression were uniquely associated with low levels of academic performance in a sample of school-age youth (Fite et al., 2013). Additionally, a link between poor academic performance and cannabis use is evident in the literature (Brook et al., 1999; Windle & Wiesner, 2004). For example, Farhat et al., (2011) found that academic achievement was negatively associated with cannabis use, particularly for males. Academic achievement may be associated with both reactive aggression and cannabis use due to poor executive functioning and cognitive deficits, including symptoms of ADHD, that have been associated with both reactive aggression (Arsenio et al., 2009; Connor et al., 2003; Dodge et al., 1997; Jambroes et al., 2018; Thomson & Centifanti, 2018) and cannabis use (Flory et al., 2003; Vitulano et al., 2014; Winters et al., 2007).

2.2. Genetic predispositions.

Recent studies have begun to examine the potential genetic contributors to CUD and pathological aggression. Investigations in these areas are still in their early stages, with limited data regarding the genetic influence on the link between cannabis use and different aggression constructs. Nonetheless, recent findings point to several genetic elements that could play a role in predisposing individuals to the concurrence of cannabis consumption and several behavioral traits and domains related to pathological aggression, such as externalizing behaviors or reactive aggression.

2.2.1. CADM2.

Recent extensive genome-wide association studies (GWASs) have identified a gene on chromosome 3, known as cell adhesion molecule 2 (CADM2), which is linked to lifetime cannabis use (Pasman et al., 2018; Stringer et al., 2016). Later studies confirmed this finding (Arends et al., 2021; Levey et al., 2023). Importantly, the same gene has been shown to influence impulsivity (Sanchez-Roige et al., 2023) and propensity to engage in risky behaviors (Clarke et al., 2017; Day et al., 2016). CADM2 proteins play a role in glutamate signaling, γ-amino-butyric acid (GABA) transport, and neuron cell-cell adhesion, particularly in the prefrontal cortex (PFC) (Ibrahim-Verbaas et al., 2016). This area is crucial for cognitive control and motivational importance, impacting impulse control and self-regulation (Allman et al., 2001; Hyman et al., 2006). While the specific link between CADM2 and aggression has not been documented in humans, knockout (KO) mice for this gene have been shown to engage in aggressive behavior (Tanabe et al., 2013), opening the possibility that this gene may also contribute to the ontogeny of reactive aggression.

2.2.2. CHRNx genes.

Nicotinic acetylcholine receptors are ligand-gated ion channels consisting of five subunits. Five muscular subunits (α1, β, γ, ε, and δ) and eleven neuronal subunits (α2-α7, α9, α10, and β2-β4) have been identified in the mammalian genome. Nicotinic receptors are constructed from either five identical subunits (such as α7 or α9) or a combination of different subunits, which must include at least two α-type subunits. This assembly allows for the formation of numerous heteromeric receptors, each exhibiting various pharmacological and kinetic characteristics. Recent data point to several of these genes in the predisposition to substance use. For example, the gene cluster CHRNA5/CHRNA3/CHRNB4 on chromosome 15 has gained significant interest for its repeated links to tobacco smoking, nicotine dependence, and early initiation stage (Berrettini et al., 2008; Caporaso et al., 2009; Furberg et al., 2010; Greenbaum et al., 2006; Greenbaum & Lerer, 2009; Jean Bierut et al., 2008; Keskitalo et al., 2009; Liu et al., 2010; Rose, 2007; Schlaepfer et al., 2008a; Sherva et al., 2008; Stevens et al., 2008; Thorgeirsson et al., 2008, 2010). Additionally, associations have been found between the CHRNA5/CHRNA3/CHRNB4 locus and alcohol dependence, alcohol response levels, and age of alcohol initiation (Joslyn et al., 2010; Schlaepfer et al., 2008b; Sherva et al., 2010) and two SNPs in CHRNA5, rs16969968 and rs684513, have been associated with cocaine dependence (Jean Bierut et al., 2008; Sherva et al., 2010). Recent studies have also shown that CHRNA2 and CHRNA3 are some of the top genetic vulnerability factors for cannabis use disorder (Demontis et al., 2019; Levey et al., 2023; Lubke et al., 2012). Unsurprisingly, genetic studies have shown that CHRNA5/CHRNA3/CHRNB4 gene cluster variations also predict externalizing behaviors (Stephens et al., 2012). Notably, the α7 subunit, encoded by the CHRNA7 gene, has been identified as a regulator of aggressive behavior in mice (Lewis et al., 2017; Picciotto et al., 2015). Additionally, treatment with agonists of α7-containing nicotinic receptors has been shown to reduce physical aggression in mice, further underscoring its role in modulating aggressive behavior (Lewis et al., 2017). CHRNA7 is located on chromosome 15q13.3. Individuals carrying heterozygous deletions of this region display aggression (Miller et al., 2009; Shinawi et al., 2009; Van Bon et al., 2009; Vu et al., 2011). Overall, these data may point to the idea that different stoichiometric arrangements of α subunits of nicotinic receptors may condition both the risk for externalizing behavior, aggressiveness, and cannabis use.

2.2.3. DRD2.

The DRD2 gene, which codes for the dopamine D₂ receptor, is an important gene associated with dopaminergic signaling. Reduced dopaminergic activity is consistently associated with impulsive violence (Seo et al., 2008). Pharmacological alterations in striatal dopamine functioning have induced aggression in mice (Rodriguiz et al., 2004; Couppis & Kennedy, 2008) and humans (Rocca et al., 2002). In children, DRD2 variants that lead to reduced dopamine function, such as the G allele for the DRD2 A-241G and the T allele for the TaqIA polymorphism, are overrepresented in children displaying aggressive behaviors, including bullying, anger expression, and cruelty (Zai et al., 2012). Most importantly, (Vaske et al., 2011) indicated that offenders with a history of violent victimization have a greater likelihood of carrying specific DRD2 alleles than offenders who were not victimized.

Emerging evidence also indicates that this gene is associated with the predisposition to develop CUD (Conner et al., 2005; Levey et al., 2023). These findings are in line with converging evidence indicating that the DRD2 gene is a key determinant of reward deficiency syndrome (Blum et al., 2023), a condition characterized by a dysfunctional reward function due to hypodopaminergic features, which predisposes to substance use as well as impulsive behaviors (Blum et al., 2000). Indeed, previous studies have indicated an association of DRD2 polymorphisms with impulsivity and substance use (Amadéo et al., 1993; Blum et al., 1990; Comings et al., 1994; Conner et al., 2005; Esposito-Smythers et al., 2009; Jung et al., 2019). However, it is worth highlighting that other existing evidence has not demonstrated any link between DRD2 polymorphisms and cannabis use in adolescents (Creemers et al., 2011). Similarly, previous studies failed to show a clear association between DRD2 variants and alcohol consumption (Gelernter et al., 1991), even though the lack of positive findings in these studies may be partially attributed to several significant limitations in the experimental design (Blum et al., 2023). Taken together, these findings suggest that specific genetic predispositions associated with specific DRD2 alleles may hinder reward processes, potentially leading to impulsive and compulsive behaviors, including aggressive reactions and the misuse of various substances, such as cannabis. In line with this idea, DRD2 polymorphisms have been proposed to influence aggression by increasing sensation seeking (Chester et al., 2016).

2.2.4. GABRx genes.

The major receptor enabling neuronal inhibition, GABA-A, is a chloride-gating channel comprising three subunits selected from a pool of 19 gene products. These gene products are encoded by the genes GABRA1-GABRA6 (representing α1-6 subunits), GABRB1-GABRB3 (β1-3 subunits), GABRG1-GABRG3 (γ1-3 subunits), GABRR1-GABRR3 (ρ1-3 subunits), GABRD (δ subunit), GABRE (ε subunit), GABRP (π subunit), and GABRQ (θ subunit). Most synaptic receptors are formed by combining two α subunits, two β subunits, and one γ subunit (Olsen & Sieghart, 2009). Several lines of evidence implicate several SNPs of these genes in the predisposition to pathological aggression and cannabis use. For example, the gene GABRA2 has been shown to moderate peer influence on externalizing behavior (Dick et al., 2009; Villafuerte et al., 2014). Kiive and colleagues (2017) showed that subjects with specific variants of this gene and a history of stressful life events had higher physical aggression scores, pointing to GABRA2 as a critical factor in aggression ontogeny. In the same research, it was also found that the same interaction predisposed to alcohol use in adolescence. Interestingly, GABRA2 has also been involved as a key factor for adolescent substance abuse (Trucco et al., 2014), as well as cannabis dependence (Philibert et al., 2009). Another gene related to the predisposition to cannabis use and anger/externalizing behavior is GABRB1. In a recent survey, Levey and colleagues (2023) reported a significant association between GABRB1 and cannabis use disorder. The same gene was identified as a predisposition factor for hostility in a GWAS conducted in young Finns (Merjonen et al., 2011).

2.2.5. MAOA.

Ample evidence has shown that the MAOA gene, encoding the enzyme monoamine oxidase A, is a critical genetic factor interacting with child abuse and neglect to predict long-term behavioral outcomes in the externalizing spectrum, including aggression (Kolla & Bortolato, 2020) and substance use (Fite et al., 2018, 2019; Nilsson et al., 2011). Monoamine oxidase A catalyzes the degradation of serotonin, norepinephrine, and dopamine (Bortolato et al., 2008). The best-characterized MAOA functional polymorphism is a 30-bp variable number tandem repeat located in its promoter region (uVNTR) (Sabol et al., 1998). This genotype features six alleles with different numbers of repeats (2, 3, 3.5, 4, 5, and 6) (Huang et al., 2004), which are associated with different transcriptional efficiency and enzyme activity. Specifically, variants featuring 2 or 3 repeats have been associated with low activity (MAOA-L). The interplay between these alleles and early-life adversity is one of the best-documented gene × environment (G × E) interactions in the pathophysiology of antisocial behavior and pathological aggression (Caspi et al., 2002; Kolla & Bortolato, 2020). Complementary to these lines of evidence, mouse models of Maoa deficiency exhibit striking similarities with clinical phenotypes (Bortolato et al., 2018; Bortolato & Shih, 2011; Cases et al., 1995; Godar et al., 2011). In particular, a mouse model of MAOA hypomorphism, Maoa^Neo, was found to exhibit overt aggression following exposure to early-life stress (Godar et al., 2019). A wealth of evidence also shows the impact of MAOA uVNTR variants on the predisposition to substance use. MAOA-L alleles, for instance, are correlated with an earlier onset of alcohol dependence (Vanyukov et al., 1995, 2004) and antisocial alcoholism (Samochowiec et al., 1999) in males. A history of maltreatment elevates the risk of alcohol use in female carriers of high-activity alleles (MAOA-H) or male carriers of MAOA-L (Nilsson et al., 2011). Consistently with these findings, we showed that the interaction between childhood maltreatment and MAOA-L variants in males or MAOA-H alleles in females predicts greater polysubstance use risk (Fite et al., 2019). Furthermore, we showed that the interaction of physical and emotional abuse with MAOA-H alleles predicted lifetime cannabis use in female but not male college students (Fite et al., 2020). Building on this evidence, in a recent study, we found that the same G×E interaction could predict the amount of cannabis consumed by emerging adults with a history of child maltreatment. Our data (Figure 1) show that, although more robust in males, experiencing child maltreatment was associated with high quantities of cannabis use for MAOA-L but not MAOA-H variants in males and females.

Figure 1:

Association between child maltreatment and cannabis quantity in a population of 470 college students (239 females and 231 males) (Fite et al., 2020). Descriptively, approximately 80.8% of participants reported experiencing at least one traumatic event. Further, roughly 45% of participants reported some level of cannabis use (range of responses = 0 to 12 times per day), with 32.8% of individuals reporting using cannabis on average one time on the days they use cannabis. A significant 3-way interaction effect was found for cannabis quantity (B = 1.11, P =0.03). For males, childhood trauma was positively associated with cannabis quantity at MAOA-L genotype (B = 0.91, P =0.00) but unrelated at MAOA-H genotype (B = −0.41, P =0.18). In a similar but less robust pattern, for females, childhood trauma was positively associated with cannabis quantity in female participants with at least one MAOA-L allele (B =0 .48, P = 0.00) but not for MAOA-H homozygous counterparts (B = 0.26, P = 0.40).

2.3. Sex and gender differences.

Gender differences in the specific link between reactive aggression and cannabis use have yet to be evaluated. However, there is some empirical support for gender differences. Some evidence suggests that reactive aggression may be more prevalent in males (Baker et al., 2008; Kempes et al., 2005); although other studies have found no differences in rates of reactive aggression between males and females (e.g., Connor et al., 2003). Furthermore, males have been shown to engage in cannabis use more frequently than females (Hasin et al., 2019). Some risk factors seem to affect males and females differently; for instance, chronic cannabis use and abstinence have been associated with more pronounced impairments in psychomotor skills and cognitive processing speed among males (Lisdahl & Price, 2012).

However, there are mixed results regarding gender differences in the influence of exposure to negative life events, such as abuse and maltreatment. While some literature suggests boys are more susceptible to behavioral difficulties after a traumatic event (Jaffe et al., 1986), other literature indicates that girls are more susceptible to poor behavioral adjustment outcomes following the events (Cummings et al., 1999; Holden & Ritchie, 1991; Steinberg, 1998). Still, other research suggests no gender differences in the influences of these adverse events on adjustment outcomes (Grych et al., 2000). For example, Cullerton-Sen et al., (2008) found that maltreatment was specifically linked to physical aggression for boys and relational aggression for girls. Specific to substance use, Lansford and colleagues (2010) found that physical abuse was a risk factor for adolescent substance use for girls but not boys.

Although not always the case (O’Donnell et al., 2012), most studies suggest that peers may be more influential on boys’ than girls’ acting out behavior (Crosnoe et al., 2002; Mears et al., 1998; Piquero, 2005), including substance use (Duncan et al., 2005; Erickson et al., 2000). Moreover, there is evidence suggesting that engaging in high levels of delinquency is more strongly associated with depressive symptoms for girls than boys (Diamantopoulou et al., 2011; Wiesner & Kim, 2006). However, gender differences in the risk factors associated with comorbid reactive aggression and cannabis use have not been readily evaluated. Thus, further research examining gender differences in these shared correlates of reactive aggression and cannabis use is warranted.

3. PSYCHOLOGICAL CORRELATES OF AGGRESSION AND CANNABIS USE

Preliminary evidence suggests that associations between aggression and cannabis use may depend on the function of aggression. Specifically, reactive aggression is more strongly associated with cannabis use than proactive aggression in samples of youth (Fite et al., 2008, 2014, 2022; Fite, Schwartz, et al., 2012; Fite, Stoppelbein, et al., 2009). Some evidence also supports this relationship in adulthood (Arendt et al., 2007; Fite et al., 2011; Hayatbakhsh et al., 2009; Muntaner et al., 1990).

3.1. Impulsivity.

Reactive, but not proactive, aggression and cannabis use have both been repeatedly associated with impulsivity. In fact, impulsivity is the hallmark feature of reactively aggressive behavior (Fite et al., 2016, 2023; Vitaro & Brendgen, 2012). Further, impulsivity is a well-established risk factor for use of cannabis and other substances (e.g., Guy et al., 1994; Hayaki et al., 2011; Rinehart & Spencer, 2021; Vangsness et al., 2005). Thus, it appears that both reactive aggression and cannabis use may be the result of behavioral disinhibition.

3.2. Negative Emotions.

Evidence links reactive, but not proactive, aggression to observable and physiological markers of negative emotion (De Castro et al., 2005; Hubbard et al., 2004, 2010; Thomson et al., 2021). On the flip side, teaching youth emotion regulation strategies has been found to reduce the association between anger and reactive aggression (Calvete & Orue, 2012). It is believed that a temperamental predisposition for overly emotional responding contributes to the development of reactive aggression (Arsenio & Lemerise, 2001; Dodge, 1991; Fite et al., 2023). Cannabis use is also associated with poor emotion regulation (e.g., Dvorak & Day, 2014; Simons & Carey, 2002; Zvolensky et al., 2009). For example, Shrier and colleagues (2014) found that levels of negative affect were higher just before cannabis use compared to other periods. Moreover, a common reason individuals give for using cannabis is to cope with the negative emotions they are experiencing (Buckner, 2013; Buckner et al., 2014; Patrick et al., 2011), as cannabis can alter mood and induce more positive affect (Hart et al., 2010).

3.3. Anxiety.

Reactive, but not proactive, aggression has been linked with symptoms of anxiety throughout childhood and adolescence and into adulthood (Card & Little, 2006; Fite et al., 2023; Raine et al., 2006; Seah & Ang, 2008). For example, in a longitudinal study, Fite et al. (2009) analyzed 335 males at two different time points 10 years apart, and found that reactive aggression predicted anxiety in adulthood, while proactive aggression did not. Reactively aggressive behavior is linked with feelings of shame and remorse that produce symptoms of anxiety (Dodge et al., 1997). In contrast, proactive aggression has been linked to callous and unemotional traits (Barry et al., 2007; Fite et al., 2023; Fite, Raine, et al., 2009; Fite, Stoppelbein, et al., 2009). A link between anxiety and cannabis use has also been established (e.g., Buckner et al., 2011), with individuals reporting that they use cannabis to help cope with their symptoms of anxiety (Kosiba et al., 2019; Zvolensky et al., 2009). In turn, among frequent daily cannabis users, increased levels of irritability and anxiety are commonly reported when abstaining from cannabis use (Kouri & Pope, 2000). Thus, a vicious cycle of experiencing anxiety and engaging in cannabis use can evolve.

3.4. Sleep Difficulties.

A small but growing body of literature indicates a specific association between sleep difficulties and reactive (not proactive) aggression (Becker, 2014; Dodge et al., 1997; Fite et al., 2015). For instance, Becker (2014) found that teacher reports of reactive aggression were positively associated with child-reported sleep problems, even when considering other behavioral issues (i.e., attention-deficit/hyperactivity disorder and oppositional defiant disorder symptoms) known to be associated with sleep difficulties. The link between reactive aggression and sleep problems seems to be emotional arousal, which elicits reactive aggression (Hubbard et al., 2004; 2010) and is strongly tied to sleep difficulties (Dahl, 2006; El-Sheikh & Buckhalt, 2005; Gregory & Sadeh, 2012). Along the same lines, given that cannabis use can reduce negative affect (Hart et al., 2010), some individuals use cannabis to cope with sleep difficulties and improve sleep quality (Bolla et al., 2010).

4. THE ENDOCANNABINOID SYSTEM AND ITS ROLE IN REACTIVE AGGRESSION

4.1. Circuitry of reactive aggression.

Reactive aggression, characterized by a defensive response to perceived threats, is part of the defense cascade, which involves high arousal and negative emotions like anger and fear. Lesional studies and electric activation studies in rodents and nonhuman primates (NHPs), as well as neuroimaging analyses in humans, have led to the definition of a complex corticolimbic circuit governing reactive aggression (for comprehensive reviews, see Bartholow, 2018 and Nelson & Trainor, 2007). The convergence of these studies has ascertained that threatening or negatively valenced (such as social provocation or frustration) stimuli trigger the activation of the medial amygdala. This, in turn, initiates a pattern of activation that travels along the stria terminalis to reach the medial hypothalamus and the periaqueductal gray (PAG) (Blair, 2004, 2016). These neural pathways are influenced by other brain areas like the prefrontal cortex, lateral septum, and monoaminergic nuclei (Gregg & Siegel, 2001). In particular, two major compartments of the prefrontal cortex, namely the orbitofrontal cortex (OFC) and the ventromedial prefrontal cortex, are critical in modulating the function of this limbic circuit. The OFC assesses the contrast between actual and anticipated rewards, making it relevant to experiences of frustration (Blair, 2004). Furthermore, this region plays a role in identifying breaches of social norms (Blair & Cipolotti, 2000), which can serve as potential triggers for reactive aggression. Lesions to the OFC resulted in increased aggression in rodents, NHPs (De Bruin et al., 1983; Machado & Bachevalier, 2006), and humans (Anderson et al., 1999; Blair & Cipolotti, 2000). Furthermore, focal OFC injuries are specifically linked to increased aggression when compared to other frontal areas (Brower & Price, 2001). The ventromedial prefrontal cortex (vmPFC) is also believed to regulate the acute threat response system. This region encodes potential rewards and costs of actions, influencing decision-making, and its dysfunction is conjectured to lead to improper representations of the consequences of impulsive aggression (Blair, 2016). Thus, a functional impairment of the OFC and vmPFC may lead to inadequate weighing of costs and benefits (Blair, 2008, 2013), promoting a greater propensity for aggression in reaction to perceived threats. Indeed, neuroimaging studies showed that the activity in these regions is typically reduced in individuals with high levels of reactive aggression (Blair, 2004).

These premises indicate that reactive aggression is a likely by-product of dysfunctional corticolimbic connectivity between the OFC/vmPFC and the central amygdala. The projections from the OFC and vmPFC onto these regions are largely based on glutamatergic projections from pyramidal cells and the control of local GABAergic interneurons (primarily parvalbumin- or cholecystokinin-positive). Functional neuroimaging studies indicate that the activation of the vmPFC is negatively correlated to that of the amygdala, and this negative functional coupling is inversely correlated with GABA levels in the PFC (Delli Pizzi et al., 2017). Similarly, animal studies have shown that GABA receptors in the amygdala enable inhibitory gating on the information flow from the PFC (Chang, 2017; Chang & Grace, 2018).

4.2. The role of endocannabinoids in the circuitry of reactive aggression.

The endocannabinoid system is a complex neuroregulatory network that plays a significant role in the regulation of emotional behavior and mood (Marsicano et al., 2002; Valverde & Torrens, 2012), particularly through its control of corticolimbic connectivity. In this section, we will overview the main constituents of this system, namely its receptors, endogenous ligands, and attending biosynthetic and metabolic enzymes, as well as the current knowledge on their implication in aggression pathophysiology.

4.2.1. Cannabinoid receptors.

The two best-characterized cannabinoid receptors, CB₁ (Matsuda et al., 1990) and CB₂ (Munro et al., 1993) are coupled to Gi/o proteins (Howlett et al., 1989) and mediate intracellular responses through various changes that impact signaling cascades, including the inhibition of adenylyl cyclase, activation of G-protein-activated inwardly rectifying potassium channels (GIRKs), and phosphorylation of extracellular signal-related kinases (ERKs) (Demuth & Molleman, 2006; Ho et al., 1999).

CB₁ receptors are highly abundant in GABAergic and glutamatergic neurons across major brain structures involved in behavioral regulation, such as the PFC, amygdala, septo-hippocampal system, striatum, thalamus, and brainstem nuclei, among others (Deutch & Charney, 1996; Glass et al., 1997; Herkenham et al., 1990, 1991; Katona et al., 2001). Although CB₁ receptors have been identified in postsynaptic locations (Salio et al., 2002; Tsou et al., 1998), they are typically located on presynaptic terminals (Freund et al., 2003; Mackie, 2005), where they regulate neurotransmitter release and synaptic plasticity. CB1 receptors mediate two well-characterized forms of short-term synaptic plasticity: the depolarization-induced suppression of inhibition (DSI) and depolarization-induced suppression of excitation (DSE). These processes are based on the reduction of GABA and glutamate release, respectively, from presynaptic boutons following stimulation of postsynaptic terminals (Morishita & Alger, 1999; Ohno-Shosaku et al., 2001; Varma et al., 2001; Wilson & Nicoll, 2001). CB₁ activation has also been shown to inhibit the neurotransmission of other mediators, including glycine, acetylcholine, norepinephrine, and serotonin (Szabo & Schlicker, 2005). However, the underlying mechanisms of these effects remain incompletely understood. Additionally, CB₁ receptors have been implicated in short- and long-term synaptic depression in relation to phasic or tonic endocannabinoid release (for a comprehensive review of these topics, see Lovinger, 2008).

All the brain regions included in aggression circuitry exhibit an abundant density of CB₁ receptors. Given the key role of these receptors in modulating the release of both GABA and glutamate, alterations in the distribution or function of these receptors may disrupt the connectivity between PFC and amygdala.

Importantly, CB₁ KO mice have been directly employed to investigate the role of these receptors in aggressive behaviors. While CB₁ KO mice behave normally under standard conditions, exposure to adverse environmental conditions may lead to abnormal behaviors and anxiogenic-like phenotypes (Valverde & Torrens, 2012), potentially altering the sensitivity of mice and promoting aggressive behaviors. It has been observed that resident CB₁ KO mice display increased aggression toward intruders compared to their wild-type (WT) counterparts, accompanied by elevated anxiogenic-like responses in the light/dark box test. This increased aggressive behavior is primarily observed during the first session of testing and dissipates in subsequent sessions, suggesting that heightened basal anxiety levels may contribute to the initial aggression (Martin et al., 2002). CB₁ KO mice exhibited increased aggressiveness compared to WT littermates; however, isolation increased aggression only in WT mice (Rodriguez-Arias et al., 2013). Indeed, constitutive deletion and pharmacological antagonism of CB₁ receptors increase defensive coping strategies, including avoidance, freezing, and risk assessment, in response to unfamiliar conspecifics, indicating an increase in anxiety-like behaviors (Litvin et al., 2013). Additionally, reduced levels of social interaction have been observed in CB₁ KO mice when tested in an unfamiliar and stressful environment (Haller et al., 2004).

Taken together, these data suggest that aggression may be accompanied by a reduction in CB₁ receptors, which would lead to the reduction of endocannabinoid signaling in this area, causing a likely modification of the excitatory/inhibitory balance in the cortex, which is under control of CB₁ receptors (Boon et al., 2014). In turn, this mechanism may reduce the activity of the pyramidal cells, controlling local inhibition of the amygdala or other subcortical areas.

CB₂ receptors are abundantly expressed in most peripheral organs, particularly in immune cells that regulate cytokine secretion and modulate cell trafficking (Walter & Stella, 2004). In the brain, their presence appears to be limited to microglial cells. However, they have also been identified in various regions of the rat brain, including the cerebral cortex, striatum, amygdala, thalamus, cerebellum, spinal nucleus, olfactory nucleus, and hippocampus (Gong et al., 2006; Onaivi, 2006; Onaivi et al., 2008; Van Sickle et al., 2005). It is worth noting that CB₂ receptors are typically detectable in the brain during states of neuroinflammation (Benito et al., 2008).

Some studies suggest that neuronal CB₂ receptors may predominantly reside in postsynaptic terminals (Brusco et al., 2008; Onaivi, 2006). However, their functional role in neurons remains largely elusive and awaits further characterization.

Recently, the role of CB₂ receptors in aggression has also been investigated using the social interaction test and the resident-intruder paradigm. CB₂ KO mice displayed higher levels of offensive aggression when housed in groups, similar to CB₁ KO mice, and social isolation had no significant effect on CB₂ KO mice (Rodríguez-Arias et al., 2015). The study outcomes (Rodríguez-Arias et al., 2015) indicated that group-housed CB₂ KO mice engaged in threat and attack behaviors for a longer time and initiated more attacks compared to their WT counterparts. Increased aggression was evident in both social interaction tasks.

Taken together, these studies involving CB₁ and CB₂ receptor knockout animals suggest that a decreased activation of this system might be linked to the emergence of aggressive behavior in animal models. Conversely, the activation of CB₁ and CB₂ receptors seems to have a calming effect. For instance, researchers tested the impact of synthetic cannabinoid agonists on isolated Oncins France 1 (OF1) mice, a strain bred for aggressive behavior. Both the CB₁ agonist Arachidonyl-2’-chloroethylamide (ACEA) and the CB₂ agonist JWH133 significantly reduced aggression in OF1 mice, emphasizing the crucial role of these two major cannabinoid receptors in regulating aggressiveness (Rodriguez-Arias et al., 2013; Rodríguez-Arias et al., 2015).

Beyond CB₁ and CB₂ receptors, the endocannabinoid system encompasses additional receptors like the transient receptor potential vanilloid type 1 (TRPV1) and two G-protein-coupled receptors: GPR55 (Di Marzo, 2006; Hillard, 2015; Marco et al., 2014; Piomelli, 2003), which has also been identified as the target of lysophosphatydylinositol (Oka et al., 2007); and GPR18, which binds to anandamide, its metabolite N-arachydonyl glycine (McHugh et al., 2010), and resolvin D2, a metabolic product of the omega-3 fatty acid docosahexaenoic acid (Chiang et al., 2015). Nothing is currently known about the involvement of these receptors in the ontogeny of reactive aggression.

4.2.2. Endocannabinoids and biosynthetic/metabolic enzymes:

The two best-characterized endogenous activators of cannabinoid receptors are N-arachidonoylethanolamine (AEA; also commonly known as anandamide, derived from the Sanskrit word “ānanda,” meaning bliss) (Devane et al., 1992) and 2-arachidonoylglycerol (2-AG) (Mechoulam et al., 1995; Sugiura et al., 1995). Both AEA and 2-AG are derived from arachidonic acid, an unsaturated C20 fatty acid with four double bonds, which also serves as the precursor of other eicosanoids, including prostaglandins and leukotrienes. A comprehensive schema of the biosynthesis, metabolism, and function of AEA and 2-AG is provided in Figure 2.

Figure 2:

Synoptic presentation of biosynthesis, metabolism, and function of the endocannabinoids anandamide (AEA, N-arachydonoylethanolamine) and 2-AG (2-arachydonoylglycerol). For details, see text.

AEA, found in picomolar concentrations, acts as a high-affinity partial agonist for both CB₁ and CB₂ receptors. It is synthesized on demand through enzymatic hydrolysis of the membrane phospholipid N-arachidonoyl phosphatidylethanolamine (NAPE), a process catalyzed by several phospholipases (C. Liu & Walker, 2006; Okamoto et al., 2004; Sun et al., 2004). Following its release and activation of CB receptors, anandamide is rapidly removed from the synaptic cleft by a carrier-mediated system and subsequently hydrolyzed by the membrane enzyme fatty acid amide hydrolase (FAAH) (Cravatt et al., 1996; Hillard et al., 1995; Ueda et al., 1995). FAAH also contributes to the catabolism of other substrates, including oleoylethanolamine (OEA) and palmitoylethanolamine (PEA). While these compounds do not activate CB₁ receptors, they may slow down anandamide degradation by competing for FAAH (Lo Verme et al., 2005).

Evidence indicates that AEA’s impact on aggressive behavior is dose-dependent. Sulcova et al. (1998) showed that lower systemic doses of AEA did not influence agonistic behavior in aggressive mice. In contrast, the highest dose administered significantly diminished aggression in aggressive mice while inducing timidity. Conversely, AEA produced diametrically opposed effects in timid mice, with low doses causing aggressive behavior and high doses having no impact. These data suggest that AEA’s role in aggression is modulated by individual temperament differences, possibly through differential expression of CB receptors or FAAH within distinct regions of the aggression circuitry. In line with this idea, neuroimaging studies have documented an increase of FAAH (as measured by binding to the radioligand [¹¹C]CURB) in the PFC of borderline personality disorder patients, which was positively correlated with measures of hostility (Kolla et al., 2020). In contrast, FAAH density was found to be decreased in the amygdala of antisocial individuals (Kolla et al., 2021). Interestingly, FAAH inhibitors increase prosocial behavior in rats, suggesting that different experimental conditions and pharmacological approaches may influence the effect of AEA on aggression (Cassano et al., 2011; Manduca et al., 2014; Moise et al., 2008).

2-AG is much more abundant than anandamide, reaching nanomolar concentrations in most tissues, and acts as a full agonist of both CB receptors. It is produced from 1,2-diacylglycerol (DAG) by diacylglycerol lipase (DAGL) (Bisogno et al., 2003) and is mainly degraded by monoacylglycerol lipase (MAGL) (Dinh et al., 2002), which is found near CB₁ receptors in presynaptic terminals (Gulyas et al., 2004). However, other enzymes may contribute to this process (Kozak et al., 2005).

The distinct neurochemical profiles of anandamide and 2-AG underscore their different physiological roles. While our current understanding of the specific functions of each endocannabinoid is still limited, the development of FAAH and MAGL inhibitors has been instrumental in elucidating their roles in synaptic and neurochemical regulation. 2-AG is the retrograde mediator of DSI and DSE, while anandamide may serve as an activity-dependent regulator of monoaminergic transmission. There is evidence suggesting a potential biological antagonism between anandamide and 2-AG. However, emerging evidence also points to similar roles for both endocannabinoids in the regulation of anxiety and pain (although acting on different receptors). The development of JZL195, a potent FAAH/MAGL inhibitor, has further revealed that the behavioral effects of CB₁ receptor agonists can be recapitulated by the combination of both endocannabinoid-mediated functions (Long et al., 2009).

2-AG signaling has recently emerged as a potent negative modulator of aggressiveness. The impact of 2-AG neurotransmission on aggressive behavior was investigated in a recent study employing both a MAGL inhibitor (JZL184) and a CB₁ receptor antagonist (AM251). MAGL inhibition (JZL184; administered at 8 and 16 mg/kg via intraperitoneal injection) was shown to reduce the number of bites initiated by resident mice and increase the number of bites received during the resident-intruder paradigm. At higher MAGL inhibitor doses (16 mg/kg), mice received more bites than they initiated, while their defensive behavior remained unchanged. The addition of a CB₁ receptor antagonist (AM251; administered at 0.5 mg/kg via intraperitoneal injection) did not attenuate the effects of MAGL inhibition, indicating that the outcomes of MAGL inhibitor administration were not mediated through CB₁ receptors (Aliczki et al., 2015). Among intruders treated with the MAGL inhibitor, mice received more bites than they delivered, and they engaged in more defensive behavior as opposed to offensive strategies. These observations were consistent across residents and intruders, regardless of their hierarchical positions in the employed paradigm. It is noteworthy that these findings differed from previous studies linking cannabinoids to aggression, as they were influenced by factors such as the stressfulness of experimental manipulations, timing of testing, and treatment duration. Conversely, MAGL inhibition markedly differed from other cannabinoid interventions in that it abolished both biting and offensive behaviors in treated mice. It is important to consider future studies involving the replication of these results in the same and other species, employing a variety of experimental manipulations, to establish MAGL inhibitors as potent negative modulators of aggression.

In addition to AEA and 2-AG, other lipids have been proposed as potential endocannabinoids, including 2-arachidonoylglycerylether (noladin ether) and O-arachidonoylethanolamine (virodhamine). Recent evidence has also identified peptidic ligands, such as hemopressin and its derivatives, as potential modulators of CB receptors. To the best of our knowledge, nothing is known about the role of these modulators in aggression.

4.2.3. Sex Differences in the Corticolimbic Endocannabinoid System:

While several studies have begun to delve into the disparities of the endocannabinoid system between males and females, available results reveal a complex scenario. Human studies have only been limited to the investigations of CB₁ receptor distribution using diverse PET radiotracers. In particular, most investigations employing [¹¹C]OMAR point to higher CB1 receptor densities in women compared to men across many regions (Neumeister et al., 2013; Normandin et al., 2015). Conversely, male rats exhibit heightened CB1 receptor expression in several brain regions, including the PFC, amygdala (Castelli et al., 2014), anterior pituitary (González et al., 2000), and brainstem (de Fonseca et al., 1994), while showing no such elevation in the hypothalamus, striatum, and limbic forebrain (de Fonseca et al., 1994). Conversely, female rats manifest increased levels of 2-AG and AEA in specific brain regions compared to males, with fluctuations observed across the estrous cycle (Bradshaw et al., 2006; González et al., 2000). These sex disparities likely stem from the distinct roles of ovarian hormones, as evidenced by alterations in receptor expression across various brain regions following ovariectomy in an estrogen-sensitive manner (Castelli et al., 2014; de Fonseca et al., 1994; S. González et al., 2000).

5. EFFECTS OF CANNABIS INGREDIENTS ON AGGRESSION.

Cannabis sativa contains over 120 distinct terpenophenolic compounds collectively referred to as phytocannabinoids (Morales et al., 2017). The two most prevalent and best-characterized phytocannabinoids are Δ⁹-tetrahydrocannabinol (Δ⁹-THC) and cannabidiol (CBD). In this section, we will review the mechanism of action of phytocannabinoids and their influence on aggressive responses.

5.1. Pharmacology of phytocannabinoids.

The main psychoactive ingredient in cannabis, Δ⁹-THC, is a highly lipophilic compound primarily synthesized in the leaves, flowers, and glandular trichomes of the plant. Most pharmacological effects induced by hemp products, encompassing alterations in emotional and cognitive states, analgesia, hypothermia, and appetite stimulation, are attributed to the actions of Δ⁹-THC as a partial agonist of cannabinoid CB₁ and CB₂ receptors (see below). In addition to these targets, Δ⁹-THC interacts with other receptors, including GPR55 and GPR18 (Console-Bram et al., 2014; Lauckner et al., 2008), activates PPARγ (O’Sullivan et al., 2005) and glycine receptors (Xiong et al., 2011), and interacts with TRPV2, TRPV3, TRPV4, TRPA1, and TRPM8 channels (De Petrocellis et al., 2011; de Petrocellis et al., 2012; Muller et al., 2019; Neeper et al., 2007). Finally, it has been proposed to modulate opioid receptors (Cichewicz, 2004) and block serotonin 5HT_3A receptors (Yang et al., 2010) as well as acetylcholinesterase (Brown, 1972; Mishima et al., 2002; Yoshimura et al., 1974).

In contrast to Δ⁹-THC, CBD lacks psychotropic properties but has been demonstrated to play a role in modulating the behavioral effects of cannabis (Zuardi, 2008). In fact, the Δ⁹-THC: CBD ratio is a primary criterion for categorizing distinct cannabis chemotypes (de Meijer et al., 1992; Pacifico et al., 2006), and it has been proposed to contribute to the variability in the neurobehavioral outcomes of cannabis or hashish consumption (Fadda et al., 2004; Ryan et al., 2006). Interestingly, the majority of cannabis strains commonly found in illicit markets tend to contain elevated levels of Δ⁹-THC (Starks, 1990).

The contrasting characteristics of Δ⁹-THC and CBD are rooted in their distinct mechanisms of action. While Δ⁹-THC exhibits nanomolar affinity for both CB₁ (Ki = 25.1 nmol/L) and CB₂ (Ki = 35.2 nmol/L) receptors, CBD displays significantly lower affinity for these receptors (Bisogno et al., 2001; Pertwee, 1999; Showalter et al., 1996; A. Thomas et al., 2004; B. F. Thomas et al., 1998). However, research has revealed that CBD is a weak antagonist/inverse agonist of both CB receptors (A. Thomas et al., 2007), possibly through a non-competitive receptor blockade mechanism (Pertwee, 2008). Additionally, CBD is an antagonist of GPR55 and GPR18 receptors and an activator of 5HT_1A, 5HT_2A, and 5HT_3A serotonin receptors, adenosine A1A receptors, and PPARγ receptors. Finally, it modulates glycine, GABA-A receptors, purinergic P2X4 receptors, as well as transient receptor potential (TRP) channels.

In addition to Δ⁹-THC and CBD, many other phytocannabinoids have been recognized, including cannabigerol (CBG), cannabichromene (CBC), and cannabinol (CBN). CBG has a low affinity for CB₁ and CB₂ receptors but interacts with TRP channels, including TRPA1, TRPV1, TRPV2, TRPV3, TRPV4, and TRPM8. It also acts as a potent agonist for the α2 adrenoceptor and a blocker of the serotonin 5-HT_1A receptor (Cascio et al., 2010). CBC, one of the most abundant phytocannabinoids in the plant, does not have a significant affinity for CB₁ and CB₂ receptors. However, it activates TRP channels, particularly TRPA1, TRPV3, TRPV4, and TRPM8. CBN is an oxidized metabolite of Δ⁹-THC, displaying weak psychoactive effects and binding to cannabinoid receptors, with a higher affinity for CB₂. CBN also acts as a potent agonist of TRPA1 and an antagonist of TRPM8 channels. The effects of these compounds on aggression remain unknown.

5.2. Effects of phytocannabinoids on aggression.

To our knowledge, the only experimental study involving human subjects to investigate the impact of Δ⁹-THC on aggression utilized the Taylor aggression paradigm (TAP) (Myerscough & Taylor, 1985). Findings from this research revealed that at low doses (0.1 mg/kg), Δ9-THC marginally amplified participants’ propensity to escalate the magnitude of administered shocks to adversaries, whereas moderate to high doses of the compound diminished this tendency.

Apart from this particular study, most existing evidence stems from rodent studies. These investigations have demonstrated that acute Δ⁹-THC administration (0.125-4.0 mg/kg) elicits a dose-dependent reduction in the frequency of attack by resident mice, rats, and squirrel monkeys (Miczek, 1978). Furthermore, acute Δ⁹-THC administration suppresses attack behavior in dominant rats and enhances submissive reactions in subordinate opponents, implying the involvement of the endocannabinoid system in both offensive and defensive aggression (Miczek & Barry, 1977). Notably, the anti-aggressive effects of acute Δ⁹-THC or Cannabis sativa extract administration have been observed in non-mammalian species such as birds (pigeons) tested in the fixed-ratio schedule-induced aggression paradigm (Cherek & Thompson, 1973) and Siamese fighting fish (González et al., 1971). However, the anti-aggressive effect of Δ⁹-THC in the latter disappears after repeated exposures to the drug.

Subsequent research (van Ree et al., 1984) explored social contact behavior in isolated rats that received either low or high doses of intraperitoneally injected Δ⁹-THC or cannabidiol, a non-addictive phytocannabinoid derived from cannabis with potential anxiolytic effects (Crippa et al., 2011). While higher Δ⁹-THC doses (10 mg/kg) demonstrated a suppressive impact on social interactions, lower doses (1 mg/kg) reduced aggressive behaviors, including fighting, kicking, or biting. Cannabidiol had no discernible effect on social contact behaviors. Similar findings were reported in pigeons, where an inverse correlation between Δ⁹-THC dose (0.5 or 1.0 mg/kg) and aggressive responses were observed (Cherek et al., 1980).

Conversely, under stressful conditions, Δ⁹-THC or Cannabis sativa extract administration may exacerbate or even induce aggressive behavior. Studies have shown that in REM sleep-deprived or morphine-dependent rats, a single application of cannabis induces dose-dependent aggressive behavior. Chronic administration of Δ⁹-THC or cannabis extract also increases aggressive behavior in food-deprived rats (Carlini et al., 1972). Additionally, mouse-killing (muricide) behavior has been observed in group-housed rats following chronic, but not acute, Δ⁹-THC administration, while in stressed (socially isolated) rats, a single dose of Δ⁹-THC is sufficient to induce muricide behavior (Miczek, 1978; Ueki et al., 1972). For instance, in the case of aggressive, electrically shocked rats, the provision of propylene glycol and cannabis (1 mg Δ⁹-THC/kg ingested orally) heightened aggressive response (Carder & Olson, 1972). A subsequent investigation by (Ueki, 1979) revealed that group-housed rats became significantly more aggressive, exhibiting fighting among cage mates and muricide, because of chronic daily Δ⁹-THC doses (6 mg/kg administered via intraperitoneal injection). This aggression emerged approximately two weeks into the treatment regimen and persisted as long as the rats were isolated. Upon transfer to group housing, muricide rates decreased by 50%, and attacks diminished.

Notably, the dose-dependent effects of Δ⁹-THC and cannabinoids on aggression are likely linked to the ability of CB₁ agonists to increase serotonin (5-HT) at low doses while abruptly decreasing 5-HT at higher doses (Bambico et al., 2007).

However, it is worth noting that chronic cannabis administration might facilitate aggressive behaviors, especially under stressful conditions. In parkin-null, human tau overexpressing mice, chronic treatment with Sativex^® (a mixture of Δ⁹-THC and CBD) reduced abnormal behaviors related to stress and aggression (Casarejos et al., 2013).

The effects of CBD on aggression have been studied only in a rudimentary fashion. In a study on resident mice, CBD was found to reduce the number of attacks (Hartmann et al., 2019), and both antagonism of CB₁ and 5-HT_1A receptors prevented this effect. Furthermore, CBD treatment reduced the upregulation of c-Fos expression in the PAG, indicating that this region may contribute to the anti-aggressive effects of CBD. Nevertheless, it is unclear whether CBD may reduce reactive or proactive aggression, given that resident aggression is also reflective of instrumental territoriality.

6. THE BIOLOGICAL LINKS OF REACTIVE AGGRESSION AND CANNABIS USE

The background presented in the previous sections has documented that the intricate association between cannabis use and reactive aggression is shaped by several factors:

i) Both reactive aggression and cannabis use share common environmental and genetic predispositions, particularly within the context of vulnerabilities in the externalizing behavior spectrum. Extensive evidence from animal models demonstrates that early-life stress, including manipulations that heighten aggressive behaviors in later developmental stages, results in a marked downregulation of CB₁ receptor expression, particularly in the prefrontal cortex (PFC), as documented by Hill et al. (2019). Moreover, early-life stressors affect AEA and 2-AG levels in several brain regions, further complicating the long-term consequences of these manipulations.

The observed reduction in CB₁ receptor expression in the PFC, in tandem with decreased limbic responsiveness, may serve as a central framework for understanding the interplay between early stress, endocannabinoid signaling, and aggressive behavior. Notably, in Maoa hypomorphic mice, we found a markedly reduced expression of CB₁ receptors in the PFC (Figure 3), and this down-regulation is further exacerbated by exposure to early postnatal stress.

Figure 3:

The expression of CB1 receptor proteins (tested by western blotting) in the prefrontal cortex of adult mice is reduced by both early-life stress [ES; P<0.01 in comparison with non-stressed (NS) counterparts; main effect for ES in a 2-way ANOVA] and MAOA-hypomorphic genotype [NEO; P<0.01 in comparison with wild-type (WT) littermates; main effect for genotype in a 2-way ANOVA]. As a result, MAOA-A deficient mice exposed to ES (which exhibit high levels of aggression and poor impulse control; see Godar et al., 2019) display a marked reduction of this receptor in this key brain region for aggression regulation. N=8/group.

This reduction is accompanied by marked aggression as well as a broad reduction in the responsiveness of the stress effector system, including the hypothalamic-pituitary-adrenal (HPA) axis (coordinated by the hypothalamus) and sympathetic system (governed by the PAG), akin to that described in pathologically aggressive individuals.

Given the role of CB₁ receptors in the regulation of these effector systems, it is possible that the observed downregulation of these receptors in the PFC may contribute to the behavioral phenotype of these mice. It should be noted that Kolla et al. (2020) reported an increase in fatty acid amide hydrolase (FAAH) activity in the PFC of borderline personality individuals, providing additional evidence for dysregulation in prefrontal endocannabinoid signaling among individuals prone to reactive aggression. In general, these data suggest that environmental and genetic vulnerability factors can change the activity and expression of the endocannabinoid system, thereby altering the biological substrates upon which cannabinoids exert their effects. The interplay of genetic susceptibility factors and early-life adversities may particularly increase the vulnerability for cannabis-associated aggression and other psychopathological outcomes. To comprehensively understand the role of genetic factors in driving developmental alterations within the endocannabinoid system and their subsequent interaction with early adversities, it is imperative to undertake further investigations both in preclinical and clinical settings.

ii) Assuming that alterations in endocannabinoid signaling possess broader relevance, reduced endocannabinoid signaling in the PFC may underlie various psychological correlates observed in individuals with cannabis use issues and reactive aggression, including impulsivity, anxiety, and sleep disturbances. Notably, CB₁ knockout (KO) mice display not only a predisposition to aggression but also heightened anxiety and sleep disturbances (Silvani et al., 2014). Conversely, acute Δ⁹-THC administration, especially at low to moderate doses, has been found to facilitate sleep (Cousens & DiMascio, 1973; Feinberg et al., 1975), likely through CB₁ receptor activation. These effects have also been observed following the administration of endocannabinoid deactivation inhibitors and CB₁ receptor agonists in rodents. Recent research emphasizes the potential benefits of elevating anandamide levels through FAAH inhibition in addressing fear- and stress-related behaviors, both in humans and individuals with FAAH loss-of-function mutations (Mayo et al., 2019).

Many individuals cite improved sleep and better management of anxiety as primary motivations for using medicinal cannabis (Bonn-Miller et al., 2014). Notably, discontinuation of cannabis following chronic use has been associated with sleep disturbances and anxiety (Gates et al., 2016). In light of this evidence, it is increasingly plausible that a reduction in endocannabinoid tone, particularly in the PFC, may contribute to the mechanistic foundations of aggression and its psychological correlates, including anxiety, negative affect, impulsivity, and sleep disturbances.

Given the potential deficits in endocannabinoid tone among aggressive individuals, some may resort to cannabis use as a means of coping with negative affect, sleep disturbances, and anxiety. Low doses of Δ⁹-THC or cannabis strains with high CBD: Δ⁹-THC ratios have demonstrated anti-aggressive and anxiolytic effects mediated through CB₁ receptors (and potentially CB₂ receptors as well or even unexplored targets, such as GPR55). While these effects may be beneficial in the context of acute administration and low doses, they may promote habit formation, exacerbating the propensity for cannabis misuse, abuse, and dependence.

iii) The relationship between cannabis and aggression is shaped by numerous variables intrinsic to the use of this drug, including the Δ⁹-THC/CBD ratio, frequency, and duration of consumption, as well as method of use (smoking, vaporization, or oral ingestion), which affects its absorption. The evidence obtained from rodent models suggests that low, acute doses of Δ⁹-THC reduce aggression, while higher doses may elicit no effect or even worsen aggression. Indeed, we found that extremely low doses of Δ⁹-THC (0.03 mg/kg, administered via intraperitoneal injection) led to the ablation of aggression in our model of GxE aggression. In partial agreement with these findings, anecdotal literature on first-time users points to the calming effects of cannabis. At the same time, higher doses may lead to more violence in association with intoxication. Furthermore, it is worth noting that the activation of CB₁ receptors can lead to compensatory processes to maintain neurochemical homeostasis in the brain. In particular, CB₁ receptors can rapidly internalize through clathrin-coated pits upon binding with full agonists, leading to an enduring downregulation of their surface expression. As a result, repeated Δ⁹-THC administration likely culminates in a profound downregulation of the endocannabinoid system, further exacerbating the imbalance between excitatory and inhibitory signaling and reducing CB₁ receptor expression. The result of this process would likely be a vicious cycle in which chronic use of cannabis would worsen the original imbalances of CB₁ receptors (and/or other endocannabinoid factors). This situation may increase the probability of withdrawal, in which the insufficient activation of the endocannabinoid system may facilitate aggressive responses. This cyclic relationship is depicted in Figure 4.

Figure 4:

Proposed mechanism to explain the comorbidity of cannabis use and pathological aggression. Shared genetic and environmental predispositions may lead to a profound reduction in CB1 receptors or other alterations of the endocannabinoid system in the prefrontal cortex (PFC). In turn, these abnormalities are predicted to increase the predisposition to psychological vulnerability traits, such as anxiety and impulsivity, which facilitate externalizing behaviors, including aggression and cannabis use. Cannabis use may serve the purpose of reducing the negative affective outcomes of insufficient action of endocannabinoids in the PFC. Nevertheless, chronic cannabis consumption and perhaps execution of aggressive behavior may further compound biological defects of the endocannabinoid system in the PFC, leading to a vicious cycle.

Overall, this background suggests that the comorbidity of pathological aggression and cannabis use is an escalating spiral based on the progressive downregulation of CB₁ receptors and/or endocannabinoid signaling. Nevertheless, critical questions persist, underscoring the need for future research endeavors and emphasizing the necessity for a nuanced approach when investigating the impact of cannabinoids on aggressive behavior. For instance, despite the preponderance of evidence pointing toward the involvement of the PFC, the precise contributions of other brain regions remain less clear.

Given the distinct endocannabinoid system balances within each region, a more detailed analysis of endocannabinoid system alterations across specific brain regions, particularly in the central amygdala, anterior hypothalamus, and periaqueductal gray (PAG), is essential, given the heterogeneous nature of the brain and the intricate neural circuitry that governs aggression regulation.

Moreover, further investigations are necessary to elucidate how disparities in sex and gender might influence the complex interplay between cannabis use and reactive aggression. Rat studies have demonstrated that repeated exposure to Δ⁹-THC results in comparable down-regulation, but more pronounced desensitization, of CB₁ receptors in adolescent females, particularly in regions such as PFC, hippocampus, and PAG. This suggests a heightened susceptibility of females to disruptions in CB₁ receptor signaling induced by cannabis misuse (Burston et al., 2010). Irrespective of the biological meaning (and translatability) of these findings, any discussion on the role of gender differences in the comorbidity of pathological aggression and CUD should incorporate analyses of sex hormones in dopaminergic neurotransmission and signaling (Godar & Bortolato, 2014) when considering potential discrepancies in the effects of cannabis between males and females, as well as the observed variations in outcomes.

A comprehensive understanding of how drug-related variables influence aggression also necessitates further investigation through prospective and experimental studies. It is also crucial to probe the motives underlying cannabis use among individuals prone to aggression, discerning whether they perceive it as a means of self-therapeutic relief or as an instigator of aggression. In this context, evaluating aggressive reactions in the context of potential Δ⁹-THC withdrawal, as opposed to administration, is critical in comprehending this comorbidity.

Understanding the intricate interplay between cannabis and aggression is particularly critical considering the emergence of synthetic cannabinoids, many of which possess higher potency than Δ⁹-THC. These potent synthetic cannabinoids, such as CP47, 497, HU-210, and various JWH compounds, have regrettably gained popularity in the recreational substances market over the past two decades, often marketed under the generic brand names “Spice” or “K2.” Unlike Δ⁹-THC, which acts as a partial agonist of CB₁ receptors, these agents function as full, high-potency CB₁ receptor activators, resulting in more pronounced pro-aggressive effects, given that CB₁ receptors are principal mediators of the psychotropic actions of cannabis.

As we continue to delve into the complexities of the links between cannabis use and aggression, it is imperative that we approach this subject with the utmost caution and precision, but at the same time without political and ideological bias, in light of the socioeconomic burden imposed by this association as well as the potential therapeutic benefits offered by some cannabinoid agents, such as CBD and endocannabinoid metabolism inhibitors, in the context of mitigating violent behavior.

Highlights.

• Cannabis use disorder (CUD) is highly associated with pathological reactive aggression (PRA).

• Research reveals a bidirectional relationship between CUD and PRA

• CUD and PRA also share common genetic and environmental risk factors.

• We discuss available clinical and preclinical evidence on the comorbidity of CUD and PRA

• We propose a model to explain the neurobiological mechanisms underlying such comorbidity