Interplay between endocannabinoid and endovanilloid mechanisms in fear conditioning

Abstract

Objective:

The transient receptor potential cation channel, subfamily V (vanilloid), member 1 (TRPV1) mediates pain perception to thermal and chemical stimuli in peripheral neurons. The cannabinoid receptor type 1 (CB₁), on the other hand, promotes analgesia in both the periphery and the brain. TRPV1 and CB₁ have also been implicated in learned fear, which involves the association of a previously neutral stimulus with an aversive event. In this review, we elaborate on the interplay between CB₁ receptors and TRPV1 channels in learned fear processing.

Methods:

We conducted a PubMed search for a narrative review on endocannabinoid and endovanilloid mechanisms on fear conditioning.

Results:

TRPV1 and CB₁ receptors are activated by a common endogenous agonist, arachidonoyl ethanolamide (anandamide), Moreover, they are expressed in common neuroanatomical structures and recruit converging cellular pathways, acting in concert to modulate fear learning. However, evidence suggests that TRPV1 exerts a facilitatory role, whereas CB₁ restrains fear responses.

Conclusion:

TRPV1 and CB₁ seem to mediate protective and aversive roles of anandamide, respectively. However, more research is needed to achieve a better understanding of how these receptors interact to modulate fear learning.

Summations

• The tripartite system, anandamide, TRPV1, and CB1, may be an important player in regulating fear responses.

• Anandamide either facilitates or restrains fear, by acting upon TRPV1 or CB1, respectively.

• The intensity of aversive stimulus and doses of cannabinoids or vanilloids ligands are major determinants in detecting anti-aversive effects.

Considerations

• This is a narrative review of the interactions among anandamide, TRPV1, and CB1 in the modulation of fear learning. We carefully selected studies contributing to the field.

• We recognize that some relevant work might not have been cited in this review.

• Here we suggest potential mechanisms and pathways involved in the regulation of fear by anandamide, TRPV1 and CB₁.

Introduction

Fear can be defined as a coordinated reaction, involving autonomic, behavioural and cognitive changes in response to innate or learned threatening stimuli (Fanselow and Pennington, 2018). Fear learning enables an individual to assign motivational content to cues previously paired with threatening stimuli, allowing the prediction of danger and the elaboration of an adaptive response (Krause & Domjan, 2017). Thus, conditioned fear is crucial for survival and well-being (Krause & Domjan, 2017). However, beyond certain limits, a mechanism sustaining health may become maladaptive and predispose to psychiatric disorders (Milton, 2019). Indeed, alterations in conditioned fear processing have been related to anxiety, depression and post-traumatic stress disorders (Foa et al., 1989; Luyten et al., 2011; Conoscenti and Fanselow, 2019; Bienvenu et al., 2021).

In experimental animals, learned fear can be studied using protocols in which a neutral stimulus, such as a context or an auditory cue, is paired with an aversive one (unconditioned, such as a footshock). Thereafter, the formerly neutral stimulus, when subsequently presented, functions as a conditioned stimulus and triggers a conditioned response. Contextual fear conditioning involves the use of a contextual element as a neutral stimulus, while cue fear conditioning often relies on auditore stimuli (a tone). The sound can terminate simultaneously or be delayed in relation to the duration of the aversive stimuli. Alternatively, the aversive stimuli can be delivered after a certain time, characterising the trace-auditory fear conditioning. One advantage of fear conditioning models is the possibility of dissecting the mechanisms underlying specific learning phases. The whole learning and memory process can comprise the following phases: The acquisition phase, when the two stimuli are presented simultaneously (e.g. shock + context, shock + tone) and the processing and initial encoding of memory takes place. This is immediately followed by memory consolidation, a series of complex molecular processes crucial for the duration of the memory (Asok et al., 2019; de Oliveira and Do-Monte, 2021). Synaptic consolidation implies the stabilisation of the memory trace (Asok et al., 2019; de Oliveira and Do-Monte, 2021), which is encoded by an assembly of neurones called engram (Josselyn et al., 2015). Later, when the animals are re-exposed to the previously neutral stimulus, they will display the conditioned response, for which the underlying phenomenon is called memory retrieval, which involves the activation of the engram encoding the memory previously acquired (Josselyn et al., 2015; Josselyn and Tonegawa, 2020). Likewise, the exposition to the conditioned stimulus can trigger different processes, depending on the duration of this exposition, among other factors (Auber et al., 2013). Long or repeated expositions to the neutral stimuli can induce extinction, in which a new memory trace is formed, decreasing the intensity of the conditioned response. After extinguished, conditioned responses can be restored, by reinstatement, renewal or spontaneous recovery (Asok et al., 2019; de Oliveira and Do-Monte, 2021).

In terms of neural circuitry, fear learning requires an extensive network of central structures involved in cognition, emotion, and their corresponding autonomic and behavioural responses. The hippocampus (HPC) is responsible for encoding the contextual information associated with an aversive stimuli (Hennings et al., 2022; Marks et al., 2022; Lee and Kaang, 2023). Bidirectional projections between the HPC and the amygdala (AMG) are involved in sustaining the emotional valence of the conditioned response (Marks et al., 2022). The various AMG nuclei process information related to conditioned cues (Li et al., 2023) through neuronal connections with other structures, such as the periaqueductal grey (PAG) and the parabrachial nucleus and the thalamus (Marks et al., 2022). In addition, cortical structures, especially the prefrontal cortex (PFC), receive substantial connections from subcortical brain regions such as the AMG and the HPC, integrating the fear-learning circuit (Thomas et al., 2002; Alexandra Kredlow et al., 2022). Other structures outside the classic circuit have also been investigated in the last years for their contribution to fear conditioning. For instance, the nucleus accumbens (NAC) core is involved in the evaluation of threat degree (Ray et al., 2020). However, its role in fear associated with discrete cues remains uncertain (Thomas et al., 2002).

Thus, different structures and pathways work in concert along the complex process of fear learning, in order to keep the balance between allostasis and allostatic overload (Milton, 2019). Various neurochemical mechanisms have been implicated in this process, including the endocannabinoid system (ECS) (Lutz et al., 2015), which is briefly described in the following section.

The endocannabinoid system

The ECS comprises several molecular components, including the cannabinoid type-1 (CB₁) (Matsuda et al., 1990) and type-2 (CB₂) (Munro et al., 1993) receptors; the endocannabinoids (eCB), N-arachidonoylethanolamide (anandamide) (Devane et al., 1992) and 2-arachidonoylglycerol (2-AG) (Mechoulam et al., 1995); the enzymes responsible for their synthesis, N-acyl phosphatidylethanolamine phospholipase D (NAPE-PLD) and diacylglycerol lipase (DAGL), respectively; and those responsible for their hydrolysis, fatty acid amide hydrolase (FAAH) (Deutsch & Chin, 1993) and monoacylglycerol lipase (Dinh et al., 2002), respectively. One particularity of this system is that the eCBs are produced and released on demand; anandamide can be released by either the pre- or the postsynaptic neurones, while 2-AG seems to be released mostly from the postsynaptic neurones (Howlett et al., 2002; Piomelli and Mabou Tagne, 2022). Once released, eCBs can act as retrograde messengers or as autocrine modulators (Uchigashima et al., 2007; Busquets-Garcia et al., 2018). CB₁ seems to be expressed mostly in presynaptic terminals, although their postsynaptic presence has also been suggested and described as regulating neuronal self-inhibition (Bacci et al., 2004). Its functions seem to be related to the regulation of dendritic excitability mediating long-term potentiation (LTP), which is necessary for cognition and spatial memory (Maroso et al., 2016; Busquets-Garcia et al., 2018). In addition to these classic members, a more comprehensive description of the ECS may include other targets under the umbrella term of the expanded ECS (Cristino et al., 2020). The list includes enzymes and receptors modulated directly by phytocannabinoids, eCB or other products of their biosynthetic pathways, such as the transient receptor potential cation channel subfamily V (vanilloid), member 1 (TRPV1) (Cristino et al., 2020).

TRPV1 and CB₁ share some important features. Both were originally discovered as targets for phytochemicals. In the case of CB₁, its prototypical agonist is delta-9-tetrahydrocannabinol, the main psychoactive compound of Cannabis sativa (Devane et al., 1988); as for TRPV1, its main agonist is capsaicin, a substance present in certain species of chilli peppers and responsible for the burning pain associated with their intake (Caterina et al., 1997). After their discoveries as orphan receptors, both CB₁ and TRPV1 were found to share anandamide as a common endogenous agonist (Devane et al., 1992; Zygmunt et al., 1999). TRPV1 is preferentially activated at high temperatures (>42’C) (Caterina et al., 1997), whereas anandamide binds to this channel with low affinity, as compared to CB₁ (Devane et al., 1992; Ross, 2003). These particularities, however, do not refute the possibility that anandamide acts as an endogenous TRPV1 agonist. Indeed, TRPV1 activity is controlled by various other mechanisms in addition to temperature, including calmodulin (Numazaki et al., 2003), ATP (Lishko et al., 2007), calcineurin (Docherty et al., 1996) and several kinases, such as PKA or PKC (Premkumar and Ahern, 2000; De Petrocellis et al., 2001; Numazaki et al., 2003). The action of these enzymes on TRPV1 may modify its response to its ligands (e.g. anandamide) and enable its activation at physiologic temperatures (Premkumar and Ahern, 2000; De Petrocellis et al., 2001; Numazaki et al., 2003). Therefore, it is not surprising that anandamide was initially described as a full or partial agonist (Zygmunt et al., 1999; Ross, 2003) depending on the conditions determining TRPV1 conformation.

Notwithstanding these similarities, CB₁ and TRPV1 differ in several aspects. Firstly, although both CB₁ and TRPV1 are expressed in brain structures related to emotion and cognition, CB₁ is usually expressed in presynaptic neurones (Katona et al., 1999), whereas TRPV1 is thought to predominate in postsynaptic neurones (Tóth et al., 2005; Zhao and Tsang, 2017). Remarkably, despite different cellular locations, TRPV1 and CB₁ are often co-expressed in the same synapsis. Regarding the mechanisms, CB₁ is one of the most highly expressed G-protein-coupled receptors in the brain (Tsou et al., 1998; Busquets-Garcia, et al., 2018), usually coupled to a Gαi/o protein (Busquets-Garcia et al., 2018). Thus, when activated, CB₁ inhibits adenylate cyclase, activates inwardly rectifying K⁺ channels and decreases neurotransmitters release (Howlett et al., 2002), regulating depolarisation-induced suppression of both inhibition and excitation (Ohno-Shosaku et al., 2001; Wilson and Nicoll, 2001; Uchigashima et al., 2007). Conversely, TRPV1 is a non-selective cation channel, highly permeable to Ca²⁺ (Caterina et al., 1997). Once activated, it promotes an increase in intracellular Na⁺ and Ca²⁺, with subsequent increase in neuronal activity (Marinelli et al., 2005; Starowicz et al., 2007). In addition, since eCB synthesis and release can be triggered by Ca²⁺ (Alger, 2002), TRPV1 activation may in turn increase eCB tonus (Maccarrone et al., 2008). Finally, although both CB₁ and TRPV1 are activated by anandamide, this compound has at least twenty times more affinity for the former (Ross, 2003; van der Stelt et al., 2005).

This body of evidence endorses the hypothesis that anandamide, CB₁ and TRPV1 configure a tripartite system regulating neuronal activity in the brain (Fig. 1). In synapsis expressing both receptors, low levels of anandamide activate presynaptic CB₁ receptors, decreasing neurotransmitter release, whereas higher levels of anandamide may also recruit TRPV1 receptors, increasing neuronal activity and counterbalancing CB₁-mediated effects. Here, we make the case for the possibility that the dual action of anandamide upon CB₁ and TRPV1 may also participate in the modulation of learned fear. Our hypothesis will be built upon pharmacological and genetic studies to investigate the role of each receptor in fear-conditioned paradigms, considering the various phases of fear-learning processing.

Role of CB₁ in fear conditioning

Studies using either genetic or pharmacological approaches support the involvement of CB₁ in the regulation of fear conditioning. Indeed, initial studies by Marsicano and colleagues revealed that CB₁-deficient mice presented impaired fear extinction when subjected to a tone previously paired with footshock (Marsicano et al., 2002). Regarding the effects of pharmacological interventions, the modulation of fear by drugs targeting CB₁ and other molecular components of the ECS may differ depending on the specificity of the compound, dosage and intensity of conditioning. As elaborated below, it also varies according to the memory phase in which the intervention occurred (acquisition, consolidation, expression and extinction).

Acquisition of fear memory

Several studies, using pharmacological approaches, have implicated CB₁ in the acquisition of fear memory. The administration of CB₁ antagonists and inverse agonists, such as AM4113 (6mg/kg) and AM251 (4 and 8 mg/kg), respectively, impaired fear acquisition in an auditory fear conditioning task (Sink et al., 2010). However, in another study, AM251 (5mg/kg) was found to enhance the subsequent fear responses in both trace (HPC-dependent) and delayed (amygdala-dependent) fear conditioning; these effects were further increased when the drugs were administered before both conditioning and expression (Reich et al., 2008). Similarly, Sink and colleagues observed that AM251 enhanced the acquisition of fear conditioned to a context, while no effect was noticed after AM4113 treatment (Sink et al., 2010). Regarding cannabinoid agonists, the non-selective CB₁, agonist WIN55,212-2 (2.5 and 5mg/kg), impaired contextual but not auditory fear conditioning (Pamplona and Takahashi, 2006). These effects were prevented by pre-treatment with selective CB₁ antagonists, such as SR141716A or SR147778 (Pamplona and Takahashi, 2006).

Consolidation

Studies focusing on the role of the ECS on fear memory consolidation observed that anandamide release restrained this process, an effect prevented by a subeffective concentration of AM251 (Scienza-Martin et al., 2022). Also, the CB₁ agonists, HU-210 (Maćkowiak et al., 2009) and ACPA (Nasehi et al., 2016), impaired fear consolidation in both contextual and auditory fear conditioning. As expected, HU-210 effect was blocked when co-administered with AM251 (Maćkowiak et al., 2009). Similarly, the phytocannabinoid cannabidiol impaired memory consolidation via CB₁, since its effect was prevented by AM251, although also by the CB₂ antagonist, AM630 (Stern et al., 2017). This effect was observed either with systemic or local (dorsal HPC) administration (Stern et al., 2017).

Retrieval

In studies investigating the role of CB₁ in fear memory retrieval, the CB₁ antagonist, SR141716 (1, 5 mg/kg), was ineffective (Mizuno et al., 2022). However, the non-selective agonist WIN55,212-2 (0.25 mg/kg) decreased contextual fear responses when animals were subjected to a more intense, but shorter protocol (1 × of 1.5mA) (Pamplona et al., 2008). When the intensity of the aversive stimulus was lower, but its duration was longer (3 × of 0.75 mA), WIN55,212-2 enhanced fear memory retrieval at both doses tested (0.075mg/kg and 0.75 mg/kg) in males, but just at the highest one in females (Mizuno et al., 2022).

Regarding compounds that indirectly facilitated the ECS, JZL184, an inhibitor of 2-AG hydrolysis, increased freezing in females (8 mg/kg), but not in male mice, an effect mediated by CB₁, but not CB₂ receptors. The FAAH inhibitor URB597 (0.3, 1, 3 mg/kg) was ineffective (Mizuno et al., 2022).

Concerning the brain regions involved, no effect was verified by administering AM251 into the ventromedial PFC (Lisboa et al., 2010; Simone et al., 2015), although this CB₁ antagonist did increase freezing in animals exposed to a less aversive conditioning protocol (Lisboa et al., 2010). Local injection of anandamide (5 pmol/200 nl) or the anandamide transport inhibitor, AM404 (50 pmol/200 nl), into this region, attenuated the fear-conditioned responses, a result prevented by local pre-treatment with AM251 (100 pmol/200 nl). No effect was also reported in auditory fear conditioning after ACEA, another CB₁ agonist (Simone et al., 2015). Finally, when administered locally into the PAG, 2-AG decreased freezing response, an effect prevented by AM251 (Brianis et al., 2022).

Extinction

The importance of CB₁ in the extinction of aversive memories was observed through several strategies. For instance, CB₁-deficient mice showed impaired short- and long-term extinction, and this was mimicked by the administration of the CB₁ antagonist, SR141716A, to wild-type animals (Marsicano et al., 2002). The extinction of the auditory fear conditioning was impaired by the administration of another CB₁ antagonist/inverse agonist, AM251 (3 mg/kg), and facilitated by the selective agonist, ACEA (0,1 and 0,5 mg/kg) (Simone et al., 2015).

Accordingly, a selective CB₁ antagonist disrupted extinction, while administration of WIN55,212–2 resulted in the opposite effect (Pamplona et al., 2006). The facilitation of fear extinction was also observed in the extinction of remote memories, an effect prevented by a CB₁ antagonist (Pamplona et al., 2006). In another study, however, the administration of WIN55,212-2 (0.075, 0.75 mg/kg) inhibited fear extinction in both sexes (Mizuno et al., 2022). Regarding the direct modulation of endocannabinoid hydrolysis, extinction impairments were reported in animals treated with a DAGL inhibitor as well as in animals lacking DAGL (DAGLα^−/− mice) (Cavener et al., 2018). In addition, 2-AG or anandamide facilitation, with JZL184 (4 or 8 mg/kg) and URB597 (0.3, 1, 3 mg/kg), respectively, inhibited fear extinction on day 5; however, such differences disappeared on the last day of extinction (day 28) for both sexes (Mizuno et al., 2022).

Therefore, the predominant effect of facilitating endocannabinoid signalling is the inhibition of fear responses, whereas its blockade tends to induce the opposite response. These effects tend to occur particularly in protocols in which the animals are exposed to aversive stimuli of moderate intensity.

Role of TRPV1 in fear conditioning

The literature on TRPV1 and fear is scant, in comparison to the more robust evidence discussed for CB₁. One of the earliest studies implicating TRPV1 in fear memory reported phenotypic differences between TRPV1 KO and wild-type animals in the auditory fear conditioning (Marsch et al., 2007); knock out animals displayed lower levels of freezing when evaluated shortly or remotely after conditioning, while unconditioned fear responses and pain threshold remained unchanged as compared to controls (Marsch et al., 2007). The responses seemed to be dependent on the intensity of the conditioning (Marsch et al., 2007). This early study suggested that TRPV1 may be involved in fear memory, but whether its role was relevant for acquisition, consolidation or retrieval was still unknown.

Acquisition

Pharmacological studies revealed no effect after the administration of a TRPV1 blocker, capsazepine, before conditioning, although administration of capsaicin, a TRPV1 agonist, had biphasic effects – low doses enhanced acquisition and high doses impaired it (Almeida et al., 2019). A potential explanation for this biphasic profile is the fast desensitisation of TRPV1 induced by even single doses of capsaicin (Szallasi and Di Marzo, 2000; Almeida et al., 2019).

Consolidation

In agreement with the results obtained from experiments in knockout mice, intra-hippocampal administration of capsazepine impaired memory consolidation when the animals were exposed to high intensities of the aversive stimulus, although no effect was observed after capsaicin administration (Genro et al., 2012). Interestingly, this effect was replicated in a more recent study (Scienza-Martin et al., 2022).

Retrieval

Local administration of TRPV1 blockers (capsazepine or 6-iodonordihydrocapsaicin) into the ventromedial PFC decreased behavioural and autonomic responses to contextual fear learning (Terzian et al., 2014). However, the effect of 6-iodonordihydrocapsaicin was not observed when the animals were exposed to aversive stimuli of low intensity (Terzian et al., 2014). In spite of this, at this intensity, this compound was able to prevent the enhancement of conditioned responses induced by central administration of capsaicin (Terzian et al., 2014). The augmentation of behavioural and autonomic responses induced by capsaicin administration into the ventromedial PFC was replicated in another study (Uliana et al., 2020). Finally, TRPV1 blockers directly administered into the dorsal HPC impaired memory retrieval when the mice were exposed to footshocks of moderate-to-high, but not low, intensities (Iglesias et al., 2023).

Extinction

The potent TRPV1 inhibitor iodoresiniferatoxin administered before conditioning, retrieval and extinction enhanced fear extinction without affecting other phases (Laricchiuta et al., 2013). However, in the auditory fear conditioning, an acute administration of SB366791 had no effect (Llorente-Berzal et al., 2015).

In summary, the scarce literature suggests that activation of TRPV1 channels exacerbates fear memory with a biphasic profile, possibly associated with fast desensitisation of TRPV1 channels at certain doses of agonists (Szallasi and Di Marzo, 2000; Almeida et al., 2019). On the contrary, genetic deletion as well as pharmacological blockade of TRPV1 promotes anti-aversive responses. This effect depends on the intensity of the aversive experience, probably due to the tonus of the ECS, as will be discussed below.

TRPV1 and (endo)-cannabinoid interactions

Although the aforementioned studies focused on CB₁ and TRPV1 separately, we argue that these receptors function in concert to mediate opposite functions of anandamide. In this section, we built on the hypothesis that anandamide, CB₁ and TRPV1 form a tripartite system modulating fear memory based on three sets of evidence: First, CB₁ and TRPV1 are co-localized in fear-related brain regions; second, they interfere with common downstream pathways involved in fear memory. Finally, they depend on each other to mediate the effects of selective pharmacological interventions.

CB₁ and TRPV1 are co-localized in fear-related structures

The presence of CB₁ in brain circuits modulating fear learning is supported by an enormous and consistent body of evidence (Tsou et al., 1998; Wilson-Poe et al., 2012; Lazenka et al., 2013; Gomes-de-Souza et al., 2021). Moreover, histological studies with specific antibodies found CB₁ to be co-expressed with TRPV1 in several synapses of fear-related regions (Cristino et al., 2006). Additional studies in specific brain regions observed TRPV1 and CB₁ colocalization in the PFC (Fogaça et al., 2012; Diniz et al., 2019), the PAG (Casarotto et al., 2012) and the dorsomedial hypothalamus (Dos Anjos-Garcia & Coimbra, 2019). In the dorsal HPC, the use of high-resolution confocal microscopy and z-stack three-dimensional analysis also revealed co-expression of CB₁ and TRPV1 (Iglesias et al., 2023). Therefore, studies applying histological immuno-histochemical techniques, along with microscopy analysis, support the possibility that CB₁ and TRPV1 can be simultaneously activated by anandamide in fear-related brain regions, provided the local synaptic concentration of these endocannabinoids reaches levels high enough to bind both receptors.

CB₁ and TRPV1 functions modulate common fear memory-related pathways

Neurotrophic signalling exerts several functions in the brain, including modulation of fear memory (Notaras & van den Buuse, 2020). Activation of tyrosine receptor kinase B (TrkB) by brain-derived neurotrophic factor (BDNF) leads to the regulation of several downstream pathways involved in plasticity, such mTOR, Akt, CamKII or ERK, all of them crucial for the consolidation of fear memory (Minichiello, 2009). The link between neurotrophic and anandamide-CB₁ signalling has been extensively explored in recent years. On the one hand, BDNF and TrkB activation enhances endocannabinoids release (Yeh et al., 2017; Wu et al., 2020); on the other hand, certain effects derived from CB₁ activation are mediated by BDNF (Blázquez et al., 2015; Navabpour et al., 2021). As for TRPV1, its relation with BDNF remains unknown. However, one study showed that TRPV1-mediated synaptogenesis in the HPC seems to require BDNF (Hurtado-Zavala et al., 2017). Despite the scarcity of data, one could hypothesise that, if neurotrophic signalling increases endocannabinoid tonus, this may facilitate the recruitment of TRPV1 and its involvement in some of the BDNF-related effects. Indeed, an in vitro study showed that anandamide enhances TrkB phosphorylation (Diniz et al., 2019). Interestingly, the mechanism underlying this effect is dose-dependent, with low doses of anandamide acting through CB₁, whereas at higher doses, anandamide action occurs through TRPV1 activation (Diniz et al., 2019).

Another important player in fear and memory is the nitric oxide (NO) pathway (Susswein et al., 2004; Medeiros et al., 2022). In postsynaptic neurones, calcium influx and calmodulin activation promote neuronal nitric oxide synthase (nNOS) activity, with the subsequent retrograde effects of NO and enhancement in neurotransmitter release (Huang, 1997). Several pieces of evidence point to the involvement of nNOS/NO in plasticity and specifically in the modulation of certain fear memory phases (Sadeghi et al., 2022). More important for the scope of this review, the ECS and the nitric oxide pathway seem to interact to modulate learned fear. For instance, the enhancement of the endocannabinoid tonus, by means of FAAH inhibition, prevented the extinction deficits in mice with genetic deletion of inducible NOS, iNOS (Lisboa et al., 2015), suggesting that the ECS may be a downstream effector of NO or at least able to compensate for some of its effects. Moreover, facilitation of fear retrieval induced by TRPV1 agonism or CB₁ antagonism was prevented by a subeffective dose of a NO scavenger, NO inhibitors and a soluble guanylate cyclase inhibitor (Uliana et al., 2016). Similar results were observed in the ventromedial PFC (Uliana et al., 2020). In addition, NOS inhibition blocked the LTP induced by a TRPV1 agonist in the AMG, the same effect being elicited by a CB₁ antagonist (Zschenderlein et al., 2011). These data suggest that TRPV1 and CB₁ act in opposite directions upon the NO pathway, which may partially explain their contrasting role in modulating fear memory.

Furthermore, CB₁ and TRPV1 interact in the modulation of electrophysiological processes underlying memory and neuronal plasticity. For instance, in the neocortex and striatum, spike time-dependent LTP was blocked by both TRPV1 and CB₁ antagonism (Cui et al., 2015, 2018). Similarly, anandamide induced long-term depression (LTD) through both CB₁ and TRPV1 in the Nac (Grueter et al., 2010). In the dentate gyrus, CB₁ and TRPV1 agonists seem to modulate excitatory postsynaptic field potentials and LTP in opposite directions (Tahmasebi et al., 2015). However, this intersection between CB₁ and TRPV1 remains controversial. For example, anandamide facilitated LTD through TRPV1 but not CB₁ (Yang et al., 2014), while other reports indicated that anandamide rescued impaired hippocampal LTP through CB₁ activation (Basavarajappa et al., 2014).

Some of these processes may be explained by the capacity of TRPV1 and CB₁ to modulate glutamatergic neurotransmission. For example, CB₁ activation in the hypothalamus decreases glutamate release, while the opposite goes for TRPV1 (Jamieson et al., 2022). Similarly, CB₁ and TRPV1 presented opposite effects on NMDA-induced autonomic responses (Lagatta et al., 2018) and plasticity (Back & Carobrez, 2018). These examples suggest that TRPV1 and CB₁ may act upon common mechanism in the regulation of fear memory, usually leading to opposite outcomes.

Anandamide, CB1 and TRPV1 interact to modulate fear responses

Complementing the histological and neurochemical evidence, the last section will focus on studies using pharmacological interventions in animals exposed to fear conditioning. Administration of anandamide itself, compounds that inhibit its hydrolysis (by inhibiting FAAH blockers, such as URB597) or compounds that exert dual TRPV1 and FAAH blockade (e.g., AA-5-HT) provide evidence of opposite functions for CB₁ and TRPV1 in mediating the actions of anandamide in different phases of fear responses.

Acquisition

The systemic administration of the FAAH inhibitor URB597 had no effect on the acquisition of fear memory (Laricchiuta et al., 2013; Balogh et al., 2019). Similarly, local administration of this drug into the ventral HPC, prelimbic PFC or AMG had no effects on memory retrieval (Balogh et al., 2019). However, direct administration of anandamide into the NAC core impaired the acquisition of contextual, but not auditory, fear (Pedroza-Llinás et al., 2013). Likewise, after systemic FAAH inhibition by the compound OL-135, an impairment in the acquisition of the contextual fear conditioning was observed (Burman et al., 2016), but no effect was detected in the auditory fear conditioning (Burman et al., 2016). Contextual and auditory fear conditioning rely on different brain structures, specifically, the dorsal portion of the HPC seems involved in contextual but not in auditory fear conditioning (Phillips and LeDoux, 1992). Even though FAAH is expressed in the amygdala, which is involved in both tasks (Gulyas et al., 2004), its inhibition might not be relevant to the acquisition of the auditory fear conditioning (Burman et al., 2016). Instead, the disruption observed in the contextual task after increasing endocannabinoid tonus (Burman et al., 2016) may be related to the action of this drug in structures such as the dorsal HPC, which is not involved in auditory fear conditioning. However, this effect seems to depend on the type of modulation, since the impairment in fear acquisition was observed with OL-135 (Burman et al., 2016), but not with URB597 (Laricchiuta et al., 2013; Balogh et al., 2019). This may be related to differences in the way how these compounds inhibit FAAH (Naidu et al., 2007) or off target enzymes (Zhang et al., 2007).

Resembling URB597 effects, the dual FAAH/TRPV1 blocker, AA-5-HT, had no effect on the acquisition of contextual fear conditioning (Gobira et al., 2017). However, the non-selective anandamide reuptake blocker and TRPV1 agonist, AM404, impaired fear acquisition, an effect dependent on both TRPV1 and CB₁, since it was prevented by capsazepine and by rimonabant (Almeida et al., 2019). Similarly, intra-HPC administration of AM404 prevented memory acquisition via CB₁ activation (Lin et al., 2011).

The discrepancies between the administration of FAAH dual blockers and AM404 may rely on two different but complementary hypotheses. First, the levels of anandamide depend on intensity/aversiveness of the experience (Morena et al., 2014; Iglesias et al., 2023), thus low intensities may not promote enough release of anandamide and its hydrolysis inhibition will not reach a substantial effect. Indeed, Gobira (2017) showed no effects of AA-5-HT, while Almeida and colleagues (2019) did observe fear inhibition after AM404 administration. However, the intensity of the conditioning was much higher in the latter study. Alternatively, AM404 may act as a partial agonist at TRPV1 (Ross, 2003).

Consolidation

During the consolidation of fear memory, increased levels of eCB were observed in the basolateral AMG and the HPC, but not in the PFC (Marsicano et al., 2002; Morena et al., 2014). In the HPC, anandamide levels seem to depend on intensity of aversive stimuli (Morena et al., 2014). However, post-training administration of OL-135 (Burman et al., 2016) or anandamide (intra-NAc) (Pedroza-Llinás et al., 2013) did not impair contextual or auditory fear conditioning. On the other hand, similarly to the acquisition, the administration of AM404, an inhibitor of anandamide reuptake, disrupted contextual fear consolidation when administered into the CA1 area of the dorsal HPC (Scienza-Martin et al., 2022).

Retrieval

Anandamide levels increase in the basolateral AMG after retrieval of fear memory (Gaspar et al., 2022). The same was observed in the HPC, where anandamide levels increase as a function of fear intensity (Iglesias et al., 2023). Moreover, the direct administration of anandamide into the medial PFC (Lisboa et al., 2010) or the PAG (Resstel et al., 2008) reduced freezing. Furthermore, administration of AM404 systemically (Pamplona et al., 2008), into the PFC (Lisboa et al., 2010), PAG (Resstel et al., 2008) or into the HPC (Scienza-Martin et al., 2022), mimicked anandamide effects. AM404 effects on retrieval were prevented by a CB₁ antagonists (Lisboa et al., 2010; Llorente-Berzal et al., 2015) and by a TRPV1 blocker (Llorente-Berzal et al., 2015). In addition, AA-5-HT (a dual FAAH/TRPV1 blocker) administered systemically or into the HPC impaired fear memory retrieval, an effect prevented by pre-treatment with AM251 (Gobira et al., 2017). Remarkably, AA-5-HT effects were mimicked by co-administration of subeffective doses of a FAAH inhibitor with a TRPV1 blocker (Gobira et al., 2017), supporting the hypothesis of opposite roles for CB₁ and TRPV1. In agreement with this possibility 1) the administration of a CB₁ antagonist or a TRPV1 agonist into the dorsolateral PAG induced the same effect on fear expression (Uliana et al., 2016); 2) a CB₁ antagonist prevented the retrieval deficits induced by TRPV1 blockers in the HPC (Iglesias et al., 2023) and 3) a TRPV1 blocker prevented the enhancement of memory retrieval induced by CB₁ antagonists in the PAG (Uliana et al., 2016). Altogether, these data support our proposal that CB₁ and TRPV1 act in concert to mediate opposite functions of anandamide in the control of fear responses.

Extinction

Anandamide infusion into the dorsal HPC facilitates extinction, an effect prevented by pre-treatment with AM251 (de Oliveira et al., 2008). In addition, administration of AM404 systemically (Pamplona et al., 2008), via intracerebroventricular (Bitencourt et al., 2008) or directly into the dorsal HPC (Abush and Akirav, 2010), facilitates fear extinction. These effects seem to depend on the activation of CB₁, but not TRPV1 (Bitencourt et al., 2008). However, TRPV1 involvement in AM404 effects on extinction was observed in auditory fear conditioning (Llorente-Berzal et al., 2015).

Conclusion and future directions

The evidence reviewed here supports our hypothesis that anandamide, CB₁ and TRPV1 act in concert as a neurochemical system regulating fear memory. Low-intensity aversive stimuli could promote moderate anandamide release, which activates CB₁ receptors and decreases the release of glutamatergic neurotransmission, with subsequent inhibition of fear. However, in response to highly aversive stimuli, anandamide levels would further increase; as a result, TRPV1 channels would be activated to promote Ca⁺² influx, increase neuronal firing and, finally, activate the neuronal mechanisms promoting fear.

However, some limitations should be considered. For instance, most studies have focused on male rodents as experimental subjects. Since only recently has sex been included as an experimental variable, little is known regarding differences in anandamide/CB₁/TRPV1 interactions between males and females. A recent study showed that eCB signalling facilitation had no effect on fear extinction in males, while extinction was impaired in females, probably via TRPV1 activation (Morena et al., 2021). Another biological variable to be taken into consideration is development, since TRPV1 (Huang et al., 2014) and CB₁ (Liu et al., 2003) expression change along the lifespan. Future research should address the impact of these variables in anandamide/CB₁/TRPV1 interactions in specific brain regions. Finally, a remaining question is how this tripartite system could be targeted for developing new drugs for the treatment of certain psychiatric disorders, particularly those resulting from exacerbated responses to aversive stimuli.

Figure 1. a ) CB₁ and TRPV1 receptors are co-localized in brain regions that modulate fear responses, such as the prefrontal cortex (PFC), hippocampus (HPC), amygdala (AMG) and periaqueductal grey (PAG). b ) Molecular pathways involved in fear memory modulated by CB_1. receptors and TRPV1 channels in response to low or high aversive stimuli. Under low-intensitiy, aversive stimuli (left panel) CB₁ activation by anandamide inhibits adenylate cyclase. (AC), reduces glutamate (GLU) release and activates rectifying potassium (K⁺) channels. However, as the intensity of the aversive stimulus increases (right panel), anandamide binds to TRPV1 receptors and causes calcium (Ca⁺²) influx. This leads to calmodulin (CaM) activation and neuronal nitric oxide synthase (nNOS) activity, resulting in nitric oxide (NO) production and its retrograde activity, increasing GLU release (2-AG, 2-arachidonoylglicerol; AC, adenylate cyclase; anandamide, N-arachidonoylehtanolamide or anandamide; Akt, protein kinase B; AMPA, α-amino-3-hydroxy-5-methyl- 4-isoxazolepropionic acid channel; BDNF, brain-derived neurotrophic factor; Ca⁺², calcium; CaM, calmodulin; CamKII, calcium–calmodulin (CaM)-dependent protein kinase II; CB₁, cannabinoid type 1 receptor; ERK, extracellular signal-regulated kinase; GLU, glutamate; K⁺, potassium; mTOR, mammalian target of rapamycin; NMDA, N-mehtyl-d- aspartate channel; nNOS, neuronal nitric oxide synthase; NO, nitric oxide; TrkB, receptor tyrosine kinase B; TRPV1, transient receptor vanilloid type-1 channel).