The two-hit hypothesis of schizophrenia states that a combination of genetic and environmental risk factors will cause the development of schizophrenia (Bayer et al., 1999; Rapoport et al., 2005; Caspi and Moffitt, 2006). Scores of genetic risk factors have been suggested for schizophrenia (Allen et al., 2008) and mouse mutants have been developed for most of those candidates (Desbonnet et al., 2009). Neuregulin 1 (NRG1) is one of the more promising schizophrenia candidate genes as associations with schizophrenia have been found in several studies (Stefansson et al., 2002; Tosato et al., 2005; Munafo et al., 2006). However, similar to many other schizophrenia susceptibility genes, recent genome wide associations studies suggest that it is more important to consider an interplay of different genetic and environmental risk factors for schizophrenia to understand the etiology of the disorder (Sanders et al., 2008). Thus, many environmental factors have been considered as risk factors for schizophrenia (Van Os et al., 2010) and cannabis use has been the focus of an ever larger growing list of studies. Cannabis appears to be a component/cumulative cause for schizophrenia and increases the overall risk of developing the disorder by 2-fold (Henquet et al., 2005). Importantly, an increasing number of researchers believe that this risk might be elevated for cannabis users with a genetic vulnerability to schizophrenia (Caspi and Moffitt, 2006). Indeed, a functional polymorphism in the gene for catechol-O-methyl transferase (COMT) was implicated in conferring vulnerability to cannabis-induced psychosis (Caspi et al., 2005; Henquet et al., 2006). Subsequently, a genetic mouse model for COMT was treated with the main psychoactive component of cannabis (Δ⁹-tetrahydrocannabinol: THC) during adolescence and exhibited a greater behavioral sensitivity to the long-term effects of THC (O'Tuathaigh et al., 2010) and a genotype-specific response in dopaminergic and GABAergic pathways as well as in the protein expression of cannabinoid 1 receptors (CB1) (Behan et al., 2012). Based on NRG1's established role in schizophrenia and the availability of validated mouse mutants for Nrg1 (Duffy et al., 2008), we investigated over the last decade if Nrg1 represents a second candidate for gene-cannabis interactions in schizophrenia. The mini review will outline how our mouse research has been instrumental in discovering Nrg1-cannabis interactions relevant to schizophrenia and in deciphering potential mechanisms. Our studies not only considered THC but also the cannabinoid cannabidiol (CBD), which is devoid of psychoactive properties and has been reported to block or reverse effects of THC and have antipsychotic properties (Arnold et al., 2012).

The protein Nrg1 influences key neurodevelopmental processes such as myelination, synaptogenesis, neuronal migration, and is involved in the expression and activation of N-methyl-D-aspartic acid (NMDA) receptors (Harrison and Law, 2006; Mei and Xiong, 2008). Importantly, a number of genetic mouse models have been developed for the different isoforms of Nrg1 (Duffy et al., 2008; Mei and Xiong, 2008; Karl et al., 2011). Among those, the heterozygous transmembrane domain Nrg1 mutant mouse (Nrg1 HET) has shown compelling face, construct, and partial predictive validity for schizophrenia research (Stefansson et al., 2002; Karl et al., 2007, 2011; Van Den Buuse et al., 2009; Duffy et al., 2010; Chesworth et al., 2012a,b). Thus, our team has utilized this model to determine the nature of Nrg1-cannabis interactions in great detail (for review see Arnold et al., 2012). The clinical relevance of this research has recently been highlighted by a genome-wide linkage and single nucleotide polymorphism association analysis, which discovered NRG1 as a major candidate for the development of cannabis dependence in African Americans (Han et al., 2012). The findings of our earlier mouse model research will be outlined in the following.

In an initial study we exposed Nrg1 HET mice to acute doses of THC before testing them in an array of schizophrenia-relevant behavioral paradigms (Powell and Miyakawa, 2006). Nrg1 mutant mice exhibited an increased sensitivity to the locomotor-suppressant and anxiogenic effects of THC compared to wild type-like littermates (WT). Surprisingly, the mutants also showed improved sensorimotor gating following THC challenge as measured by prepulse inhibition of the startle response (PPI) (Boucher et al., 2007a). Increased PPI is often detected after treatment with antipsychotic drugs which normalize PPI deficits of schizophrenia patients (Geyer et al., 2001). Recent human data suggest that NRG1 may also confer increased behavioral sensitivity to THC [although NRG1 polymorphisms worsened THC-induced information processing dysfunction rather than improving it (Stadelmann et al., 2010)]. It is possible that the effects of heterozygous deletion of Nrg1 in mice might be opposite to that conferred by NRG1 polymorphisms in patients. Future studies may examine whether mice overexpressing Nrg1 protein display exaggerated THC-induced PPI deficits. Follow-up experiments revealed that the enhanced behavioral response of Nrg1 HETs to acute THC was sex-specific as female mutants showed no enhanced susceptibility to acute THC and actually developed resistance to aspects of THC-induced social withdrawal (Long et al., 2010a). It is unclear as to why Nrg1-cannabinoid interactions are sex-specific. Gender influences the actions of cannabinoids (McGregor and Arnold, 2007) and interactions between gonadal hormones and neuregulin have been demonstrated [(Lacroix-Fralish et al., 2006); but also see Taylor et al., 2011]. Future studies could examine whether Nrg1 expression regulates the modulatory effects of gonadal hormones on cannabinoid receptor sensitivity.

The increased behavioral susceptibility of male Nrg1 HET mice to THC was accompanied by elevated neuronal activation as measured using c-Fos immunohistochemistry (Boucher et al., 2007b). THC selectively increased c-Fos expression in the ventral part of the lateral septum (LSV) of Nrg1 mutants. No corresponding effect was observed in control littermates. Interestingly, drugs, which modulate PPI, whether they are pro-psychotic drugs that impair PPI, or anti-psychotic drugs that facilitate PPI, all increase c-Fos expression in the lateral septum (Sumner et al., 2004). Furthermore, Nrg1 HET mice exhibited a more pronounced enhancement of c-Fos levels in stress-related brain regions (i.e., paraventricular nucleus of hypothalamus and central nucleus of amygdala). In summary, these animal studies provided the very first evidence for an interaction between the schizophrenia risk gene Nrg1 and cannabis and implied that stress and gender may also influence these interactions.

Chronic cannabis use is more relevant in cannabis-induced psychosis than acute exposure. Thus, our team continued this line of research and determined the neuro-behavioral response of Nrg1 mutants to long-term cannabinoid exposure. Nrg1 mutant and control mice were treated chronically with the synthetic CB1 receptor agonist CP 55,940 (Boucher et al., 2011). Nrg1 hypomorphic mice developed tolerance to the hypothermic and locomotor-suppressant effects of CP 55,940 more rapidly than WT mice. Interestingly, tolerance development toward the anxiogenic effects of the cannabinoid was only observed in control mice whereas Nrg1 HETs maintained persistent THC-induced anxiety with repeated CP 55,940 dosing. All mice developed tolerance to the genotype-specific effects of acute CP 55,940 on PPI (i.e., impairment in WT and facilitation in Nrg1 HET mice). Mutant mice showed a selective increase in CP 55,940-induced FosB/ΔFosB expression in the LSV, which is a marker for long-term neuroadaptive changes. These findings suggest that Nrg1 is not only involved in the acute neuro-behavioral response to cannabinoids but also modulates neuroadaptive responses to long-term cannabinoid challenge. Furthermore, it confirms the LSV as an important brain region for Nrg1–cannabinoid interactions. This could be related to the fact that the lateral septum shares reciprocal projections with the hypothalamus and the amygdala and receives cognitive input from the hippocampus and the prefrontal cortex (Sheehan et al., 2004). These brain areas are important in schizophrenia and are characterized by high expression levels of Nrg1, its main receptor ErbB4 and CB1 (Law et al., 2004; Kofalvi, 2008; Neddens and Buonanno, 2011). Future studies should examine in more detail the role of the LSV in mediating the neuro-behavioral effects of cannabinoids in Nrg1 HET mice and define in particular the involvement of CB1 and ErbB4 receptors.

Human research suggests that adolescence is a time of increased vulnerability to the detrimental effects of cannabis on the development of psychosis (Caspi et al., 2005). Thus, our team exposed adolescent WT and Nrg1 HET mice to chronic THC (Long et al., 2013). Surprisingly, Nrg1 mutants appeared less susceptible to THC-induced suppression of investigative social behaviors than control mice. However, adolescent THC exacerbated the hyperlocomotive phenotype characteristic for adult Nrg1 mutant mice (Karl et al., 2007; Long et al., 2013). Nrg1 deficiency also modulated the effects of adolescent THC on neurotransmitter systems involved in the pathophysiology of schizophrenia. Radioligand binding analyses found genotype-specific THC effects on CB1 expression in the substantia nigra: Nrg1 HET mice exhibited reduced CB1 levels drug-free whereas CB1 binding was decreased in WT and increased in Nrg1 mice post THC challenge. Lower CB1 expression levels in the substantia nigra might be responsible for the observed decreased susceptibility of adolescent Nrg1 mutant mice (whereas binding studies in adult Nrg1 HETs found increased levels of CB1 in the same brain region; manuscript currently being submitted). Interestingly, ErbB4 is localized in dopaminergic neurons in the substantia nigra (Abe et al., 2009). Thus, ErbB4 and CB1 might interact in the substantia nigra and thereby regulate the hyper-locomotor phenotype of Nrg1 mutant mice. Nrg1 also conferred opposing effects of THC on 5-HT2A receptor expression in the insular cortex and NMDA receptor binding was selectively increased in the hippocampus and cingulate cortex of Nrg1 HET mice (Long et al., 2013) (for a better mechanistic understanding of Nrg1-THC interaction on NMDA receptor expression in the hippocampus see Spencer et al., 2013).

The cannabis sativa plant is a mix of over 60 different cannabinoids, one being THC, another being CBD, which blocks, or reverses the effects of THC and other psychotropic drugs such as methamphetamine (Long et al., 2010b). Varying levels of THC and CBD in different cannabis strains could modify the consequences of long-term cannabis consumption and also shift the nature of gene-cannabis interactions such as the one reported here for Nrg1. Thus, our team characterized the neuro-behavioral response of Nrg1 HET mice to acute and chronic CBD (Long et al., 2012) to investigate its potentially therapeutic-like effects in animal models for schizophrenia. CBD did not alter schizophrenia-relevant behaviors such as hyperlocomotion or PPI deficits in our Nrg1 HET mouse model (Long et al., 2012). Nevertheless, high dose CBD selectively increased social interaction of Nrg1 mutant mice, which are normally characterized by diminished investigative social behaviors (i.e., social withdrawal) at baseline. Furthermore, chronic CBD also increased GABAA receptor binding in the granular retrosplenial cortex of mutant mice suggesting that Nrg1 may not only modulate neuro-behavioral actions of THC but also of CBD in a task- and brain region-specific manner. Further research using a variety of CBD doses thereby considering dose and age-effects will address the issue of varying or even opposing effects of different cannabinoids more comprehensively.

In summary, the transmembrane domain Nrg1 mouse model has enabled the detailed analysis of acute vs. chronic effects of cannabinoids at different stages of brain development. Nrg1 modulated the behavioral sensitivity of mice to cannabinoids differentially during adolescence and adulthood providing evidence for a role of Nrg1-cannabis interactions in schizophrenia. Furthermore, insights into the molecular and neurobiological mechanisms of Nrg1-cannabinoid interactions (involvement of CB1, 5-HT2A and NMDA receptors in particular) would not have been possible without utilizing these mouse mutants. Future research will extend on our initial findings and address sex specificity and the opposite effects of CB1 stimulation in adolescence (Nrg1 mutant less susceptible) and adulthood (Nrg1 mutants more susceptible) in greater detail. Finally, models of cannabinoid addiction should be considered given the significant comorbidity of schizophrenia and drug dependence.