It is well established that NL3^R451C mice display autism-relevant behaviors (Chadman et al., 2008; Etherton et al., 2011; Rothwell et al., 2014; Tabuchi et al., 2007). In the current study, mixed-genotype juvenile WT/NL3 pairs showed increased jumping compared with WT/WT and NL3/NL3 pairs, which could be an indication of hyperactivity as has been demonstrated in another mouse model of autism (Shank2 KO mice; Schmeisser et al., 2012), although this only occurred in mixed genotype pairs. Interestingly, recent advances in automated analyses of behavioral data have enabled accurate tracking of mouse group dynamics (Chaumont et al., 2018). In assessing two genetic mouse models of ASD, Shank2 and Shank3 KO mice, this approach also identified atypical social behavior that disrupted group interactions when mixed genotypes were present but not for groups of mice with the same genotype. With greater uptake of such technology, underlying differences in mixed genotype behavioral changes compared to genotype-matched mouse groups may be better understood in the context of the NL3^R451C model.

We have previously shown that NL3^R451C mice demonstrate a robust aggressive phenotype when bred on a mixed-background strain (Burrows et al., 2015). Here, we show that the aggressive phenotype persists on a congenic C57Bl/6 background strain as seen by increased incidence and duration of attacks. Attack latency was also reduced. In addition to these attack related measures, we also observed increased mounting behavior and frequency of tail rattles in mice bred on a pure C57Bl/6 background strain expressing the R451C mutation further confirming a role for NL3 in aggressive behavior. This study therefore provides clear evidence that the aggressive phenotype is not due to susceptibility genes inherent in the background strain or solely environmental effects as the phenotype is also present on the pure C57Bl/6 strain. Further, the aggressive phenotype shown in the C57Bl/6 mice is more easily compared across laboratories than a mixed strain background which would show quite different genetic drift with breeding over time.

The NL3 R451C mutation alters cannabinoid signaling in specific neuronal subtypes and thus impacts the cannabinoid neuronal regulation pathway (Földy et al., 2013; Speed et al., 2015). To determine whether cannabinoid modulators affect aggressive behavioral parameters in NL3^R451C mice, we investigated changes in aggression interactions in the presence of the CB1 receptor agonist WIN 55, 212–2. Overall, administration of WIN 55, 212–2 at 0.3 and 1 mg/kg rescued the aggressive phenotype in NL3^R451C mice. CB1 receptors are highly expressed in the adult mouse amygdala (Mackie, 2005; Marsicano and Lutz, 1999), further supporting the potential involvement of this brain region in regulating heightened aggression in NL3^R451C mice. Although the CB1 receptor is broadly expressed in multiple brain regions, its role is modulating the neural circuitry relevant to aggression in NL3^R451C mice is unknown. The integrated activity of brain regions generates complex behaviors such as aggression, therefore in addition to changes in the basolateral amygdala it is likely that other brain regions will also be altered by the R451C mutation. The involvement of cannabinoids in modulation of aggressive behavior in NL3^R451C mice is in line with previous reports showing that CB1 agonists strongly decrease aggressive behavior as well as findings that CB1 KO mice are more aggressive than WT control mice (Rodriguez-Arias et al., 2013).

In order to obtain further understanding of the mechanisms underlying the behavioral phenotype in these mice, we examined neuronal activity in the amygdala, a brain region associated with aggression. Our data demonstrates altered synaptic activity in a brain region that has not previously been investigated in these mice. Specifically, we have demonstrated decreased inhibitory postsynaptic activity alongside an increase in excitatory post-synaptic activity onto pyramidal cells in the basolateral amygdala. Previous reports of abnormal synaptic activity in these mice strongly suggest that alterations in activity are cell-type and synapse-type dependent (Földy et al., 2013). For example, Földy et al. (2013) showed specific changes in synaptic activity in CCK-immunoreactive neurons synapsing with pyramidal neurons in the hippocampus and that these alterations are mediated by tonic inhibition due to changes in endocannabinoid neurotransmission. In the current study, we show increased excitation and a reduction in GABAergic synaptic events in the basolateral amygdala, a brain region implicated in aggressive behaviors (Varkevisser et al., 2017; Buades-Rotger et al., 2018). We propose that this alteration in neural activity could contribute to the aggressive phenotype we have observed in these mice. Cannabinoid receptors are expressed broadly on excitatory and inhibitory neurons in the rodent brain (Kawamura et al., 2006; Hill et al., 2007). The finding that aggressive behavior is reduced to control levels in NL3^R451C mice following administration of WIN-55, 212–2 suggests that cannabinoid agonists may alleviate abnormal synaptic activity and assist in identifying new approaches to improve challenging behaviors in the clinic when synaptic mutations are involved. These novel findings further confirm the face validity of the R451C mutation and demonstrate that the aggressive phenotype persists in mice bred on the C57Bl/6 background strain.

A previous report showed increased levels of the vesicular GABA transporter (vGAT) using immunolabelling and quantitative immunoblotting in NL3^R451C mice bred on a mixed background strain (Tabuchi et al., 2007) suggesting a potential mechanism for increased inhibitory neurotransmission via GABA_A receptors in the somatosensory cortex. In contrast, when we assessed for potential changes in GABA_A receptor density using autoradiography we did not identify any genotype specific alterations. This may indicate that potential changes in receptor density are below the detection threshold of the autoradiography technique, which gives an overall indication of expression in specific brain regions.

In summary, this study shows that the NL3 R451C mutation confers a heightened aggression phenotype in C57Bl/6 mice that is reversed by administration of the CB1 agonist, WIN-55, 212–2 and that mice in mixed genotype pairs show increased jumping; potentially indicating hyperactivity. We further demonstrate synaptic dysfunction in the basolateral amygdala, an area that may contribute to the aggressive behavioral phenotype in this model. Further studies to determine the precise effects of these synaptic changes as well as potential rescue effects of WIN-55, 212–2 on synaptic function in multiple brain regions in these mice are required in order to inform the design of novel targeted clinical therapies.