Our final step was to determine whether direct amygdala administration of AM3506 itself produced effects on extinction. We surgically prepared another set of mice and bilaterally infused 0.1 μg/μl AM3506 (or equivalent volume of 0.5 μl per hemisphere vehicle) into the amygdala 30 min before extinction training. Drug treatment did not affect baseline or CS-induced freezing during extinction training (ANOVA effect of treatment: P>0.05, n = 8–11) (Supplementary Figure S6c), but fear was reduced during retrieval test in amygdala-AM3506-treated mice, compared with vehicle-infused controls (t-test: t(17) = 2.77, P<0.05, n = 8–11) (Figures 3d and e). Thus, taken together this series of experiments demonstrate that augmenting endocannabinoid signaling at CB1R in the amygdala is both necessary and sufficient to facilitate extinction. It should be noted that this conclusion, and indeed all of the preclinical data in this study, is based on one compound and one model of impaired extinction, and it will be important to extend thesefindings to other rodent models, as well as other specific FAAH inhibitors.
