Anxiety and trauma-related disorders will affect up to one in four people in their lifetime. Diseases like phobias and post-traumatic stress disorder (PTSD) show abnormal persistence of fear memories and can be debilitating, causing a huge societal and economic burden. Psychological treatments like exposure therapy are used to treat these disorders but they often show only limited or temporary effectiveness. Medications are also available but these are not fully effective in a significant proportion of patients, can have adverse side effects, and may even interfere with the efficacy of psychological therapy. There is therefore an urgent need for better treatments for these disorders (Fineberg et al., 2013; Baldwin et al., 2014).

A promising area of research in this field focuses on repurposing existing and developing novel drugs to enhance the effectiveness of psychological therapies in alleviating fear-related symptoms (Myers and Davis, 2007; Steckler and Risbrough, 2012). One drug showing broad therapeutic potential in various psychiatric diseases is cannabidiol (CBD), the main non-psychotropic constituent of the Cannabis sativa plant (Izzo et al., 2009; see chemical structure of CBD in Figure 1A). Studies in humans demonstrate the promise of CBD for treating anxiety (Blessing et al., 2015) and preclinical studies in rodents are elucidating the pharmacological mechanisms underlying its acute anxiolytic effects. These mechanisms include potentiation of serotonin (5-HT) transmission via 5-HT_1A receptor (5-HT_1AR) activation and elevation of endocannabinoid levels via inhibition of their metabolism and re-uptake, which indirectly facilitates cannabinoid receptor type1 (CB1R) activation (for a review see Campos et al., 2016).

As well as reducing anxiety in behavioral tests of unconditioned fear, emerging evidence indicates that CBD regulates fear learning and memory in paradigms that are translationally relevant to diseases such as phobias and PTSD, along with their psychological treatment. In this paper we review the recent studies on CBD regulation of fear memory processing, which have focused on contextual fear. We also present novel data on CBD regulation of auditory fear memory and its extinction, which forms the theoretical basis for exposure therapy. We then outline future directions for research on this topic to gain a broader perspective on the neural circuit, psychological, pharmacological, and cellular bases of the regulation of learned fear by CBD.

Recent evidence indicates that CBD modulates fundamental neurobiological processes involved in Pavlovian fear conditioning, a form of associative learning by which certain stimuli or environments become predictive of threat and therefore enhance survival. During acquisition a neutral conditioned stimulus (CS) is associated with an aversive unconditioned stimulus (US), such as a mild footshock. The CS can be either discrete (i.e., cued), such as a light or tone, or the environment (i.e., context) where the US was presented. CS re-exposure after conditioning initially induces a fear response, which has frequently been inferred from behavioral (e.g., freezing) and/or autonomic (increased heart rate/blood pressure, decreased body temperature) changes (Fendt and Fanselow, 1999; Resstel et al., 2009). After its acquisition the CS-US association is consolidated into long-term fear memory. Later retrieval can render fear memory labile through destabilization of the memory trace, allowing for maintenance or updating of the memory through its reconsolidation (Lee, 2009). Extinction of fear memory occurs with longer durations or repeated sessions of retrieval. This form of inhibitory learning results in the encoding of a new CS-no US association which suppresses fear expression by competing with the original fear memory (Myers and Davis, 2007). Figure 1B depicts the different phases of fear memory and its possible reconsolidation or extinction after retrieval.

Accumulating evidence indicates that CBD regulates different contextual fear memory processes. An initial study by Resstel et al. (2006) showed that systemic CBD administration decreases the freezing response and autonomic changes induced by exposure to an aversively conditioned context; this effect was similar to the positive control diazepam. Subsequent studies confirmed the CBD-induced reduction in conditioned freezing expression with acute administration before retrieval (Lemos et al., 2010) or acquisition (Levin et al., 2012). In contrast, ElBatsh et al. (2012) showed that repeated daily injections (14 days) of CBD increased freezing expression during contextual fear retrieval. Chronic treatment with CBD has, however, been shown to facilitate adult hippocampal neurogenesis (Wolf et al., 2010; Campos et al., 2013), which is involved in aversive learning and memory processing as its facilitation enhances contextual discrimination and related fear expression (Efstathopoulos et al., 2015). Mice with reduced neurogenesis, on the other hand, presented less contextual fear (Pan et al., 2012; Denny et al., 2014). Both associative (through facilitation of associative learning) and non-associative (by buffering non-associative, anxiogenic effects of the aversive experience) mechanisms seem to play a role in regulating fear learning by adult hippocampal neurogenesis (Seo et al., 2015). Since the animals were conditioned during chronic CBD treatment, facilitated contextual fear learning due to enhanced hippocampal neurogenesis could explain the results reported by ElBatsh et al. (2012). This idea is supported by studies which showed that chronic CBD treatment rescues deficits in other types of memory in animal models of cognitive impairment (Fagherazzi et al., 2012; Cheng et al., 2014).

In addition to reducing contextual fear acutely when given before retrieval, CBD has also been shown to have enduring effects on fear expression when given in conjunction with memory reactivation (leading to reconsolidation) or extinction. Systemic CBD administration immediately after briefly retrieving a contextual fear memory disrupted its reconsolidation, resulting in a lasting reduction in fear expression during later retrieval (Stern et al., 2012, 2015; Gazarini et al., 2014). This effect depended on the (indirect) activation of CB1Rs rather than 5-HT_1ARs as prior antagonism of CB1Rs, but not 5-HT_1ARs, blocked the disruptive effect of CBD on reconsolidation (Stern et al., 2012). In contrast to its effect on the reconsolidation of contextual fear memory, CBD potentiated contextual fear extinction but this also results in reduced fear expression at later retrieval. Intracerebroventricular infusion of CBD before three longer retrieval sessions facilitated the extinction of contextual fear, an effect mediated indirectly via CB1R activation as it was blocked by CB1R antagonist pretreatment (Bitencourt et al., 2008). Systemic CBD administration before a single long retrieval session also affects contextual fear extinction but this depends on the strength of prior fear conditioning. Whereas, CBD enhanced the extinction of contextual fear resulting from strong conditioning, it impaired the contextual fear extinction induced by weaker conditioning (Song et al., this issue). Table 1 and Figure 1C summarize the reported effects of systemic CBD administration on contextual fear memory processing.

Recent studies have begun to determine the neural substrates mediating the effects of CBD on contextual fear memory processing. Lemos et al. (2010) initially investigated the brain sites in which CBD acts when attenuating the expression of learned fear. They reported that systemic CBD pretreatment prevented the increase in c-Fos expression induced by re-exposure to the conditioned context in the prelimbic (PL) and infralimbic (IL) subregions of the medial prefrontal cortex (mPFC) and the bed nucleus of the stria terminalis (BNST), brain areas which play key roles in fear regulation (Dejean et al., 2015; Lebow and Chen, 2016). Subsequently, the effect of direct CBD infusion into these brain regions on contextual fear memory retrieval was investigated. There was reduced conditioned freezing expression after its infusion into the PL or BNST (Lemos et al., 2010; Gomes et al., 2012; Fogaça et al., 2014). When infused into the IL, however, CBD produced the opposite effect (Lemos et al., 2010; Marinho et al., 2015). Convergent evidence suggests that the PL and IL present distinct neuroanatomical connections and functional roles (Hurley et al., 1991; Condé et al., 1995; Vertes, 2004), including the regulation of learned fear expression (Vidal-Gonzalez et al., 2006; Fenton et al., 2014). Contrasting effects of CBD after PL or IL infusion have also been reported in unlearned tests of anxiety (e.g., elevated plus maze), in which it induced anxiogenic- or anxiolytic-like effects, respectively (Fogaça et al., 2014; Marinho et al., 2015). Restraint stress pre-exposure reversed these effects, indicating that the stress experience could modulate the response of these mPFC subregions to aversive stimuli (Fogaça et al., 2014; Marinho et al., 2015). CBD regulation of contextual fear expression in these regions depends on 5-HT_1AR activation as it is prevented by local 5-HT_1AR antagonist pretreatment (Gomes et al., 2012; Fogaça et al., 2014; Marinho et al., 2015). CBD disruption of fear memory reconsolidation depended on CB1R but not 5-HT_1AR activation in the PL (Stern et al., 2014). CBD infused into the IL cortex facilitated contextual fear extinction in a CB1R-dependent manner (Do Monte et al., 2013). Table 2 summarizes the reported effects of central CBD infusion on contextual fear memory processing.

In contrast to contextual fear memory processing, few studies have examined the effects of CBD on learned fear related to discrete cues. One study in humans showed that CBD enhanced the extinction of visual fear memory when given immediately after, but not before, extinction (Das et al., 2013). More recently it was found in rats that CBD infused into the nucleus accumbens shell impaired the encoding of olfactory fear memory, an effect blocked by 5-HT_1AR antagonism (Norris et al., 2016). To investigate this issue further we determined the effect of systemic CBD administration on the expression and extinction of auditory fear memory. Male Lister hooded rats were habituated to two distinct contexts (A and B), subjected to auditory fear conditioning (tone habituation and tone-shock pairings) in context A, treated with CBD (0, 5, 10, or 20 mg/kg, i.p.; n = 10–11/group) 30 min before undergoing extinction training (tone presentations) in context B, and underwent extinction recall testing (tone presentations) in context B (Figure 2A).

We found no differences in freezing during fear conditioning between the groups to receive different doses of CBD before extinction training (Figure 2B). Despite the relatively low levels of freezing during tone-shock pairings, conditioning resulted in robust fear memory encoding as indicated by the high freezing levels in response to tones during early extinction in vehicle-treated rats (Figure 2C). During extinction the high dose of CBD significantly decreased tone-induced freezing early in the session, compared to vehicle and the low dose. However, there were no differences in freezing during tone presentations later in the session or during extinction recall testing the next day (Figure 2D). This suggests that the high CBD dose decreased auditory fear memory expression without affecting its extinction. Although extinction occurred in a separate context to fear conditioning, vehicle-treated rats did show freezing in the period before tone presentations, indicating contextual fear expression at the start of the session (Figure 2E). We found that all three doses of CBD significantly decreased this contextual fear in comparison to vehicle. However, there were no differences in contextual fear expression before extinction recall testing (Figure 2F), suggesting that CBD acted acutely to reduce contextual fear during extinction.

Taken together, our results broadly confirm previous findings demonstrating that acute CBD treatment reduces contextual fear memory expression and extend them by showing that CBD also has a similar effect on the expression of auditory fear memory. Given that previous studies have shown a bell-shaped dose-response curve in relation to CBD regulation of anxiety and fear memory processing (Guimarães et al., 1990; Lemos et al., 2010; Stern et al., 2012), it is worth noting that we found that (1) the effect of CBD on contextual fear expression showed no dose-dependency, and (2) the dose of CBD needed to reduce auditory fear expression was much higher than for contextual fear. Moreover, CBD had no effect on auditory fear extinction. Although, this result agrees with the study of Das et al. (2013), which showed that CBD had no effect when given before cued fear extinction, it contrasts with the reported facilitatory effects of CBD on contextual fear extinction. The reasons for these discrepancies remain unclear but they could involve differences in the neural circuit and/or pharmacological mechanisms underlying CBD regulation of contextual vs. auditory fear memory expression (see below).

A growing body of literature provides compelling evidence that CBD has anxiolytic effects and recent studies have established a role for CBD in regulating learned fear by dampening its expression, disrupting its reconsolidation, and facilitating its extinction. The opposing effects of CBD on fear memory reconsolidation and extinction make it particularly attractive as a potential adjunct to psychological therapy as both may lead to lasting reductions in learned fear expression. Our novel data also suggests that CBD reduces the expression of fear memory related to both discrete and contextual cues. Although we found no effect of CBD on auditory fear extinction, decreasing fear expression during extinction without interfering in its encoding is still a useful property that has clinical implications. In this respect CBD might be an improvement over other available drugs used for treating the fear-related symptoms of phobias and PTSD, which can impair extinction (e.g., benzodiazepines) or have a less favorable side effect profile (e.g., antidepressants). As such, further research investigating the mechanisms underpinning CBD regulation of learned fear is warranted.

The study was conducted in accordance with ethical review by the Animal Welfare Ethical Review Board at the University of Nottingham and the Animals (Scientific Procedures) Act 1986, UK.

RJ: Conducted the experiment investigating the effects of CBD on auditory fear memory expression and extinction; RJ, HD, and CS: Analyzed the data from this experiment. FG, JL, LB, and CS: Drafted the paper. All authors contributed to interpreting the original data, revising the draft and approving the final version of the paper.