Receptor mechanisms underlying the CNS effects of cannabinoids: CB₁ receptor and beyond

Abstract

Cannabis legalization continues to progress in many US states and other countries. Δ⁹-tetrahydrocannabinol (Δ⁹-THC) is the major psychoactive constituent in cannabis underlying both its abuse potential and the majority of therapeutic applications. However, the neural mechanisms underlying cannabis action are not fully understood. In this chapter, we first review recent progress in cannabinoid receptor research, and then examine the acute CNS effects of Δ⁹-THC or other cannabinoids (WIN55212-2) with a focus on their receptor mechanisms. In experimental animals, Δ⁹-THC or WIN55212-2 produces classical pharmacological effects (analgesia, catalepsy, hypothermia, hypolocomotion), biphasic changes in affect (reward vs. aversion, anxiety vs. anxiety relief), and cognitive deficits (spatial learning and memory, short-term memory). Accumulating evidence indicates that activation of CB₁Rs underlies the majority of Δ⁹-THC or WIN55121-2’s pharmacological and behavioral effects. Unexpectedly, glutamatergic CB₁Rs preferentially underlie cannabis action relative to GABAergic CB₁Rs. Functional roles for CB₁Rs expressed on astrocytes and mitochondria have also been uncovered. In addition, Δ⁹-THC or WIN55212-2 is an agonist at CB₂R, GPR55 and PPARγ receptors and recent studies implicate these receptors in a number of their CNS effects. Other receptors (such as serotonin, opioid, and adenosine receptors) also modulate Δ⁹-THC’s actions and their contributions are detailed. This chapter describes the neural mechanisms underlying cannabis action, which may lead to new discoveries in cannabis-based medication development for the treatment of cannabis use disorder and other human diseases.

1. Introduction

Cannabis is the most commonly abused drug worldwide and accounts for half of all drug seizures by law enforcement (WHO, 2021). Since the 2000s, the general public has reported less perceived risk from cannabis, while diagnoses of cannabis use disorder (CUD) climb (Carliner, Brown, Sarvet, & Hasin, 2017). An estimated 147 million people in the world are cannabis users WHO, 2021. Recreational use remains the most prevalent (53.4%), but a growing community of individuals report purely medical use (10.5%) or a combination of medical and recreational use (36.1%; Schauer, King, Bunnell, Promoff, & McAfee, 2016). Therapeutic use of cannabis has a long history and an accumulating body of work has supported cannabinoids in the treatment of chronic spasticity and pain (Whiting et al., 2015). In this chapter, we explore the acute effects of cannabis from a neurobiological viewpoint. The majority of work in this vein has focused on Δ⁹-tetrahydrocannabinol (Δ⁹-THC), the primary psychoactive phytocannabinoid in cannabis that underlies its rewarding effects, but also the majority of therapeutic uses. Δ⁹-THC was first isolated from hashish by Rafael Mechoulam in 1964 (Gaoni & Mechoulam, 1964). This compound produces a number of physiological and behavioral changes in preclinical animal models including the classic tetrad effects (analgesia, catalepsy, hypothermia, hypolocomotion), a change in affective state either positive (reward, anxiety relief) or negative (aversion, anxiogenesis), and deleterious effects on cognition. Systemic reviews on the endocannabinoid system, pharmacology of cannabinoids, and their involvement and implications in various human diseases have previously been conducted and are beyond the scope of this chapter (Alexander, 2016; Leung, 2011; Mechoulam & Parker, 2013; Pertwee, 2005, 2006). Here we focus on research progress investigating the neural mechanisms underlying the behavioral effects of Δ⁹-THC and other cannabinoids in experimental animals. We first describe the receptor systems where cannabinoids bind followed by detailed region- and cell type-specific receptor mechanisms underlying Δ⁹-THC’s CNS effects.

2. Cannabinoid receptors

There are at least two types of cannabinoid receptors (CB₁R and CB₂R) identified. Δ⁹-THC, synthetic cannabinoids (WIN55,212-2, CP55940, HU-210), and the endocannabinoids (anandamide, AEA; 2-arachidonoyl glycerol, 2-AG) have high binding affinities at both the GPCR-coupled receptors. Cannabinoids also bind and activate other putative cannabinoid receptors, including G protein-coupled receptor 55 (GPR55), transient receptor potential vanilloid 1 (TRPV1) channel, and peroxisome proliferator-activated nuclear receptors (PPARs) (Fig. 1). In addition, cannabinoids may also indirectly act on other receptor systems such as opioid, adenosine, and serotonin receptors by diverse ways.

Fig. 1.

Major targets of Δ⁹-THC based on their receptor binding and functional assays as shown in Table 1.

Table 1 shows the receptor binding profiles of those commonly used cannabinoids to both CB₁ and CB₂ receptors and other putative receptors as shown in Fig. 1. In brief, these compounds are classified into four categories based on their chemical structures: classical, nonclassical, eicosanoid, and aminoalkylindole (Pertwee, 2008a, 2008b; Pertwee et al., 2010). The classical category includes dibenzopyran derivatives such as Δ⁹-THC and HU-210. Δ⁹-THC is a weak partial agonist at both the CB₁R and CB2R with greater CB1R affinity and activity at GPR55 and PPARγ (Pertwee, 2008a, 2008b). HU-210 is a synthetic analog of Δ⁸-THC with 100–800-fold greater potency than Δ⁹-THC at the CB₁R and CB₂R, a prolonged duration of action, and activity at GPR55 and TRPV1 (Devane et al., 1992; Pertwee et al., 2010). CP55940 fits within the nonclassical nomenclature, containing compounds that are Δ⁹-THC derivatives and lack a pyran ring (Howlett et al., 2002). CP55940 has marginally lower affinity than HU-210 at the CB₁R and CB₂Rs and binds to GPR55, TRPV1 and PPARγ (Pertwee et al., 2010). Within the eicosanoid classification are the two main endocannabinoids: 2-AG and AEA. AEA is a partial agonist at both the cannabinoid receptors with even lower CB₂R affinity than Δ⁹-THC (Pertwee, 2005). 2-AG has high affinity at the CB₁R and somewhat less at the CB₂R with greater efficacy observed at CB₁Rs than CP55940. Both AEA and 2-AG bind to GPR55, TRVP1, and PPARγ (Pertwee et al., 2010). The fourth and final category, the aminoalkylindoles, have the most distinct chemical structures relative to the other subtypes (Ferraro et al., 2001). The aminoalkylindole WIN 55,212-2 is a full agonist at both the CB₁Rand CB₂Rwith greater affinity than Δ⁹-THCattheCB₁R andactivity at TRVP1, PPARα and PPARγ (Howlett et al., 2002).

Table 1.

Receptor binding profiles of several major cannabinoids on CB1, CB2 and other putative cannabinoid receptors.

Drug	CB1 (Ki, nM)	CB2 (Ki, nM)	GPR55 (EC50, nM)	TRPV1 (EC50, μM)	PPARγ (EC50, μM)
Anandamide (AEA)	61–543	279–1940	18	0.16–1.15	8–10
2-Arachidonoylglycerol (2-AG)	58–472	145, 1400	3	8.4–26	10–30
Δ9-Tetrahydrocannabinol (THC)	5.05–80.3	3.13–75.3	8	>100	0.3
WIN55,212-2 (WIN)	1.89–123	0.28–16.2	N.D.	>100	10
CP55940	0.5–5.0	0.69–2.8	5	>100	10
HU-210	0.06–0.73	0.17–0.52	26	1.2	N.D.

Based on Pertwee (2008a) and Pertwee et al. (2010).

2.1. CB₁ receptor

As stated above, Δ⁹-THC acts as a partial agonist at G-protein coupled CB₁R (Iwamura, Suzuki, Ueda, Kaya, & Inaba, 2001). This receptor recruits G_i/o proteins and inhibits adenylate cyclase while increasing mitogen-activated protein kinase (Howlett, 2005; Pertwee, 2008a, 2008b). CB₁R can also inhibit N-type and P/Q-type calcium currents, stimulate A-type outward potassium channels, and use G_s proteins to signal (Howlett et al., 2002; Jarrahian, Watts, & Barker, 2004).

Regional distribution of CB₁R:

The first CB₁R distribution studies used autoradiography with [³H]-CP55,940, a tritiated CB₁R agonist (Herkenham et al., 1991, 1990) and found extraordinarily high levels of CB1Rs in the substantia nigra, globus pallidus, hippocampus, cerebellum, and cortex (Fig. 2). Autoradiographic studies using [³H]-WIN55,212-2 further confirmed this pattern in rat (Jansen, Haycock, Ward, & Seybold, 1992) and human brains (Glass, Faull, & Dragunow, 1997; Mato, Del Olmo, & Pazos, 2003). In situ hybridization (ISH) and immunohistochemistry (IHC) assays corroborated the autoradiographic reports and revealed that CB₁Rs are highly expressed in a restricted set of forebrain neurons, particularly in the cortex, amygdala, and hippocampus (see reviews by Galaj & Xi, 2019; Hu & Mackie, 2015). These neurons project widely throughout the CNS, resulting in a dense network of CB₁-positive axons (Bodor et al., 2005; Mackie, 2008). Double-label immunostaining and ISH experiments revealed that the cells expressing CB1Rs in the forebrain are primarily GABAergic and CCK-positive interneurons (Katona et al., 1999; Tsou, Mackie, Sañudo-Peña, & Walker, 1999).

Fig. 2.

[³H]CP55,940 autoradiography demonstrating CB₁R distribution in the rat brain. A high density of CB₁Rs is expressed in the SNR, GP, Hi, and cerebellum. Fr, Frontal cortex; FrPaM, frontal primary motor cortex; PO, pre-olfactory bulb; Tu, olfactory tubercle; Hi, hippocampus; VP, ventral pallidum; Me, Median eminence; fi, fimbria of the hippocampus; ic, internal capsule; LP, lateral post thalamus nuclei; SC, superior colliculus; IC, inferior colliculus; Cb, cerebellum; CbN, cerebellar nuclei; CC, corpus cal losum; GP, globus pallidus; EP, entopeduncular nucleus (homolog of GPi); SNR, substantia nigra pars reticulata; PCRt, parvicellular reticular nuclei. (This image was provided by Dr. Miles Herkenham at NIMH, USA)

Neuronal CB₁R:

RNAscope ISH is a highly sensitive and selective assay that we and others have used to characterize the cellular distributions of CB₁Rs in the brain. High densities of CB₁ mRNA have been detected in the cell bodies of both GABA and glutamate neurons in multiple brain regions including the cortex, thalamus, midbrain and cerebellum (Fig. 3; Han et al., 2017; Humburg et al., 2021; Vickstrom et al., 2021). This CB₁ mRNA signal is highly specific as selective deletion of CB₁Rs from either GABAergic neurons or glutamatergic neurons abolished CB₁ mRNA staining in the corresponding cell types. Functional studies provide further information regarding cellular localization of CB₁R. For example, electrophysiological assays demonstrate that CB₁R activation inhibits GABA release in the midbrain, which may lead to postsynaptic (dopamine) neuron disinhibition (or activation; Lupica & Riegel, 2005; Szabo, Siemes, & Wallmichrath, 2002). This suggests that activation of GABAergic CB1Rs has functional consequences. Electrophysiological assays also demonstrate functional CB1R expression in glutamatergic neurons or their terminals in the midbrain and many other brain regions (Melis, Gessa, & Diana, 2000; Melis et al., 2004). We have examined CB₁R expression in midbrain dopamine (DA) neurons. Under high magnification, CB₁-immunostaining was found mainly in cell membranes and nerve fibers, but not in neuronal cell bodies (Han et al., 2017). Since nerve fibers from different neuronal types are always intertwined, IHC assays alone are not sufficient to identify whether midbrain DA neurons express CB₁Rs. However, RNAscope ISH assays indicate that a subpopulation of midbrain DA neurons expresses CB₁ mRNA (Fig. 3).

Fig. 3.

RNAscope ISH results, illustrating the cellular distributions of CB₁Rs in the midbrain ventral tegmental area. CB₁ mRNA was detected in GAD1-labeled GABAergic neurons (A), VgluT2-labeled glutamate neurons (B) and a small population of TH-labeled DA neurons (C) in the midbrain of WT, but not CB₁-KO mice (D).

Endocannabinoids regulate physiological functions in the brain mainly through activation of CB₁Rs that inhibit presynaptic GABA or glutamate release via a retrograde endocannabinoid-CB₁R mechanism (Castillo, Younts, Chávez, & Hashimotodani, 2012; Piomelli, 2003). Specifically, presynaptic neuronal excitation increases glutamate release at excitatory synapses by activation of voltage-dependent Ca⁺⁺ channels, which subsequently activates postsynaptic AMPA and NMDA receptors and depolarizes post-synaptic neurons. Meanwhile, glutamate may also activate postsynaptic mGluR1 or mGluR5, causing an increase in 2-AG synthesis in postsynaptic neurons. Postsynaptic neuronal depolarization may also elevate intracellular Ca⁺⁺ and elicit 2-AG production. After being released from postsynaptic neurons, 2-AG retrogradely travels across the synapse to activate presynaptic CB₁Rs. Presynaptic CB1Rs are G_i/o protein-coupled receptors (Howlett, 2005; Pertwee, 2008a, 2008b). Their activation leads to inhibition of presynaptic glutamate or GABA release (Hoffman, Laaris, Kawamura, Masino, & Lupica, 2010; Howlett et al., 2002; Howlett, Blume, & Dalton, 2010; Jarrahian et al., 2004; Laaris, Good, & Lupica, 2010). This neuronal CB₁R-mediated inhibition leads to several types of short-term or long-term synaptic plasticity, such as depolarization-induced suppression of excitation at excitatory synapses, depolarization-induced suppression of inhibition at inhibitory synapses, or long-term depression, which are associated with endocannabinoid involvement in various brain functions (Galaj & Xi, 2019).

Glial CB₁R:

CB₁R has also been detected on non-neuronal cells such as astrocytes (Djeungoue-Petga & Hebert-Chatelain, 2017; Han et al., 2012; Oliveira da Cruz, Robin, Drago, Marsicano, & Metna-Laurent, 2016; Stella, 2010). Astrocytes were traditionally thought to provide nutrients to neurons and to maintain a functional homeostasis for neuronal functions. However, recent studies have indicated that astrocytes can regulate synaptic transmission and brain functions. For example, electrical stimulation of adjacent neurons can increase intracellular Ca⁺⁺ levels in hippocampal astrocytes that express CB₁R (Navarrete & Araque, 2008). This effect is mediated by a Gαq protein-phospholipase C signal pathway, rather than the Gα_i/o protein-cAMP signal pathway observed in neurons. The increase in astrocyte Ca⁺⁺ induces gliotransmitter release (Metna-Laurent & Marsicano, 2015; Mothet et al., 2000) and results in hetero-synaptic potentiation at excitatory or inhibitory synapses. Thus, glial CB₁R-mediated neuronal excitation differs significantly from neuronal CB₁R-mediated homosynaptic inhibition.

Less is known regarding CB₁R present on microglia and mitochondria. Microglia are the immune cells of the CNS. They act as macrophages and can change phenotype based on their microenvironment. Microglia in an activated state dispense a substantial amount of nitric oxide (NO). Interestingly, administration of a CB₁R agonist blocks this effect (Waksman, Olson, Carlisle, & Cabral, 1999) and microglia contain an anandamide binding site coupled to NO release (Stefano, Liu, & Goligorsky, 1996). As such, changes in NO production may mediate the effects of cannabinoids via microglial CB₁Rs. On the other hand, mitochondria are organelles responsible for a cell’s energy production. They support basic brain functioning primarily via the process of mitochondrial respiration i.e., the conversion of oxygen and nutrients into ATP. CB₁R on mitochondria (mtCB₁Rs) modulates mitochondrial respiration (Bénard et al., 2012) and are neuroprotective (Ma et al., 2015). Further, mtCB₁R mediates depolarization-induced suppression of inhibition, a form of short term synaptic plasticity in which glutamatergic neurons in the hippocampus are depolarized leading to the release of endocannabinoids and subsequent CB₁R activation and decreased GABAergic activity (Bénard et al., 2012).

2.2. CB₂ receptor

The cannabinoid CB₂R was cloned in 1993 from human leukemia cells (Munro, Thomas, & Abu-Shaar, 1993). CB2R has 44% sequence homology with CB₁Rs (Pertwee, 1997). They are G-protein coupled (G_i/o) and inhibit adenylate cyclase, leading to a decrease in cAMP signaling and neuronal inhibition (Patel, Davison, Pittman, & Sharkey, 2010). CB₂R activation can also stimulate p42/p44 MAP kinase and elevate intracellular calcium (Cabral & Griffin-Thomas, 2008). Δ⁹-THC is a partial agonist at CB₂R with relatively high affinity (Table 1; Iwamura et al., 2001).

Regional distribution of CB₂R:

CB₂R was initially referred to as “peripheral cannabinoid receptors” due to their predominant expression in peripheral tissues including immune cells, spleen, tonsils, lymph nodes, liver, and the gastrointestinal tract (Galiègue et al., 1995; Onaivi et al., 1999) and the failure to detect CB₂R in the CNS. However, more advanced techniques such as RNAscope ISH and fluorescence-activated cell sorting (FACS) followed by RT-PCR assays have unearthed CB₂R expression in the CNS including the spinal cord (Nent, Nozaki, Schmöle, Otte, & Zimmer, 2019), brain stem (Van Sickle et al., 2005), hippocampus (Li & Kim, 2015), ventral tegmental area (Zhang et al., 2017), and cerebellum (Gong et al., 2006).

Cellular distribution of CB₂R:

The specific cell types within the CNS that express CB₂R are somewhat controversial. The majority of work assumes that CB₂Rs are expressed on microglia although more direct anatomical evidence is still needed. Recent studies have revealed neuronal CB₂R expression on DA neurons in the midbrain (Zhang et al., 2015; Zhang et al., 2017), glutamate neurons in the red nucleus and hippocampus (Li & Kim, 2015; Stempel et al., 2016; Zhang et al., 2021) and GABA neurons in the striatum and cerebellum (Fig. 4; Li & Kim, 2015; Zhang, De Biase, et al., 2021) (Fig. 4). CB₂Rs in the brain are located on the postsynaptic cells (Brusco, Tagliaferro, Saez, & Onaivi, 2008a, 2008b) and are inducible, showing upregulation under neuroinflammatory conditions (Atwood & Mackie, 2010; Maresz, Carrier, Ponomarev, Hillard, & Dittel, 2005).

Fig. 4.

Conventional (A, B, C) and RNAscope (D) ISH results, illustrating the cellular distributions of CB₂Rs in mouse brain. CB₂ mRNA was detected in cortical VgluT2-labeled glutamatergic neurons (A), cerebellar GAD1-labeled GABAergic neurons (B), red nucleus VgluT2-labeled glutamatergic neurons (C), and VTA TH-labeled dopaminergic neurons (D).

CB₂R transcripts:

An important finding in recent research is the unique distribution patterns of CB₂ transcript (mRNA) isoforms (CB_2A, CB_2B, CB_2C, CB_2D) across species and tissue types (Liu et al., 2009; Zhang et al., 2015), which may, in part, explain why early assessments failed to detect CB₂ mRNA in the brain (Galiègue et al., 1995; Munro et al., 1993; Schatz, Lee, Condie, Pulaski, & Kaminski, 1997). In humans and mice the CB_2A isoform was found primarily in the testis and brain, whereas CB_2B was expressed in the spleen and leukocytes (Liu et al., 2009). CB_2C and CB_2D isoforms were only detected in rats (Zhang et al., 2015). On the whole, CB_2A is the predominant subtype (20–30-fold higher than CB_2B; Zhang et al., 2014). However, in the mouse spleen CB_2A is only about 3-fold higher than CB_2B (Zhang et al., 2014). A direct comparison of brain and spleen CB_2A mRNA levels revealed considerably greater expression in the spleen (50–100-fold). These findings suggest that brain CB₂ mRNA is more likely to be detected with a probe that targets CB_2A rather than CB_2B transcript. However, brain CB2 expression is still detectable using riboprobes that recognize the encoding sequences on both CB_2A and CB_2B isoforms, by which CB₂ mRNA was discovered in the cortex, hippocampus, and globus pallidus of non-human primates (Lanciego et al., 2011; Sierra et al., 2015). These findings indicate that expression of the CB₂ gene is dependent on the isoform subtype and varies by species and region.

2.3. GPR55

Δ⁹-THC is an agonist at the orphan receptor GPR55 (Table 1). This receptor has been put forward as a putative “CB3 cannabinoid receptor,” given that both endocannabinoids (AEA, 2-AG) and synthetic cannabinoids (HU-210, CP55,940) are also able to bind (Table 1). However, GPR55 does not contain a quintessential cannabinoid binding pocket (Baker, Pryce, Davies, & Hiley, 2006) and has minimal receptor homology with CB1R (13.5%) or CB2R (14.4%; Elbegdorj, Westkaemper, & Zhang, 2013). GPR55 couples to G₁₂ and G₁₃ proteins and activates RhoA and Ca⁺⁺ (Henstidge et al., 2009; Ryberg et al., 2007). GPR55 is distributed throughout the nervous system. In the periphery, it was been uncovered in the GI tract (Li et al., 2013), liver (Romero-Zerbo et al., 2011), pancreas (McKillop, Moran, Abdel-Wahab, & Flatt, 2013), and adipose tissue (Imbernon et al., 2014). QT-PCR assays indicate GPR55 mRNA expression in the striatum, substantia nigra, frontal cortex, hippocampus and cerebellum (Celorrio et al., 2017; Ryberg et al., 2007; Wu et al., 2013). In cell cultures, GPR55 and microglia colocalize (Pietr et al., 2009). In striatal or substantia nigra (SN) brain tissues, GPR55 mRNA was detected in neurons (colocalized with a neuronal marker), but not in microglia or astrocytes (Celorrio et al., 2017).

2.4. Peroxisome proliferator-activated receptors (PPARs)

PPARs are nuclear receptors with 3 isoforms (α, β, γ) that regulate gene expression (O’Sullivan, 2016). Activated PPARs dimerize retinoid X receptors and bind to DNA sequences termed PPAR response elements (Bishop-Bailey, 2000). Δ⁹-THC is an agonist at PPARγ (EC50~0.3μM; O’Sullivan, Tarling, Bennett, Kendall, & Randall, 2005) (Table 1), but does not bind to PPARα (Sun et al., 2007). However, one report demonstrated that Δ⁹-THC administration increased PPARα transcriptional activity (Takeda et al., 2014). PPARs are also activated by endocannabinoids (AEA, 2-AG) and fatty acids (oleic acid, arachidonic acid) and may function as lipid sensors, monitoring metabolic activity in vivo (Pertwee et al., 2010). PPARγ expression predominates in adipose tissue, but is also observed in the liver, large intestine, and spleen (Lehrke & Lazar, 2005; Vidal-Puig et al., 1996; Villapol, 2018). Within the CNS, PPARγ expression has been detected in the piriform cortex, ventral pallidum, caudate putamen (Moreno, Farioli-Vecchioli, & Cerù, 2004), and to a lesser extent the prefrontal cortex (PFC), nucleus accumbens (NAc), and amygdala (Warden et al., 2016). Immunohistochemical images showed colocalization of PPARγ in neurons, some staining in astrocytes, but not in microglia (Warden et al., 2016).

2.5. Transient receptor potential vanilloid 1 (TRPV1) channel

Six families of transient receptor potential channels (TRP) have been identified: canonical, vanilloid (TRPV), melastatin (TRPM), polycystin, mucolipin and ankyrin (TRPA). TRPs are ion channels with a nonselective cation pore and six transmembrane domains that are involved in sensory transduction. Δ⁹-THC has no effect on TRPV1 functional activity at 100 μM, while AEA is a potent TRPV1 agonist with a EC₅₀ value of 0.16–1.15 μM (Table 1). Δ⁹-THC is a mild agonist at TRPV2 (EC₅₀: ~0.65μM) and has intermediate effects at TRPA1 (EC₅₀: ~0.23μM) and TRPM8 (IC50: ~0.16μM; De Petrocellis et al., 2011). TRPV2 is activated by elevations in temperature and inflammation (De Petrocellis, Nabissi, Santoni, & Ligresti, 2017) and distributed in the paraventricular nucleus, arcuate nucleus, nucleus of the solitary tract, locus coeruleus as well as a number of other regions in the rat forebrain and hindbrain (Nedungadi, Dutta, Bathina, Caterina, & Cunningham, 2012). TRPV2 is colocalized with neurons and to a lesser extent, astrocytes (Nedungadi et al., 2012; Shibasaki, Ishizaki, & Mandadi, 2013).

2.6. Other targets

Beyond the five main receptor systems described above, cannabinoids may also interact with other targets possibly by forming heterodimers or functioning as opioid receptor allosteric modulators. For the purposes of this book chapter, we will only discuss the few receptors implicated in cannabinoid action in later Sections 3–6.

Opioid receptors:

A significant amount of work indicates cross-talk between the endocannabinoid and endogenous opioid system. Opioid receptors are inhibitory GPCRs. There are four receptor subtypes: μ, δ, κ, and nociception, with endorphins, enkephalins, dynorphins, and nociceptin as the endogenous ligands, respectively. CB₁R was reported to form heterodimers with μ, δ, and κ opioid receptors and signaling at μ opioid receptors (MORs) is reduced by CB₁R agonism (Rios, Gomes, & Devi, 2006). Colocalization of CB₁Rs and MORs has also been detected in striatal medium-spiny neurons and the dorsal horn of the spinal cord (Rodriguez, Mackie, & Pickel, 2001; Salio et al., 2001). In addition, Δ⁹-THC was reported to increase the rate of dissociation of MOR and δ opioid receptor (DOR) ligands from their orthosteric binding sites designating THC as an allosteric modulator at these receptors (Kathmann, Flau, Redmer, Tränkle, & Schlicker, 2006).

Adenosine receptors:

Cannabinoids also have activity at the adenosine receptors, which are divided into four subtypes: A₁, A_2A, A_2B and A₃. Prior work has demonstrated that CB₁R antagonism prevents A₁R activation (Savinainen, Saario, Niemi, Järvinen, & Laitinen, 2003). Heteromeric complexes between CB₁Rs and A_2ARs have been detected in the striatum (Ferre et al., 2011; Ferreira et al., 2015) and hippocampus (Aso et al., 2019). The endocannabinoids (2-AG, AEA), but not synthetic cannabinoids (WIN55,212-2, CP55940), were reported to function as negative allosteric modulators at the A₃ receptor (Lane, Beukers, Mulder-Krieger, & Ijzerman, 2010).

Serotonin receptors:

Additionally, a subset of serotonergic receptors is targeted by cannabinoids. There are seven families of 5-HT receptors (5-HT_1–7) and further subcategories within these classes. The majority of 5-HT receptors are GPCRs, not including the 5HT₃R. CB₁-5HT_2A heterodimers have been identified in the hippocampus, caudate putamen, and somatosensory cortex (Viñals et al., 2015). The Δ⁹-THC metabolites, 11-hydroxy-Δ⁸-THC and 11-oxo-Δ⁸-THC, attenuated serotonin binding to 5-HT_1A, 5-HT_1B, 5-HT_1D, 5-HT_1E, and 5-HT_2C receptors (Kimura, Ohta, Watanabe, Yoshimura, & Yamamoto, 1998; Kimura, Yamamoto, Ohta, Yoshida, & Watanabe, 1996). Similarly, AEA weakened radioligand binding of 5-HT to 5-HT_2A, 5-HT_2B, and 5-HT_2C (Kimura et al., 1998). On the other hand, HU210 increased 5-HT binding to 5-HT₂Rs on rat cortical membranes (Cheer, Cadogan, Marsden, Fone, & Kendall, 1999). In addition, CB₁R may also form heterodimers with dopamine D₂Rs in the striatum (Marcellino et al., 2008)

3. Cannabinoid tetrad effects

High doses of cannabinoids such as Δ⁹-THC or WIN55,212-2 (a potent CB₁R and CB₂R agonist) produce classical tetrad effects—analgesia, hypothermia, catalepsy, and hypolocomotion, which are often used to determine whether a novel compound is cannabimimetic in nature. The receptor mechanisms mediating cannabinoid effects in the tetrad are not fully understood. However, transgenic mice with conditional knockouts of different cannabinoid receptors have been widely used to identify the neuronal populations that mediate Δ⁹-THC or other cannabinoid effects. Significant progress has been made. Here, we discuss each assay within the tetrad and the current knowledge regarding the neural underpinnings of Δ⁹-THC- or WIN55,212-2-induced changes.

3.1. Analgesia

In pre-clinical work, Δ⁹-THC induces a strong antinociceptive effect across multiple behavioral tests (hot plate, tail flick test, formalin test) and models of chronic pain (inflammatory pain, neuropathic pain; Casey, Atwal, & Vaughan, 2017; Craft, Haas, Wiley, Yu, & Clowers, 2017; Finn et al., 2004; Wang et al., 2020).

CB₁R mechanisms:

Early work demonstrated that pharmacological blockade or genetic deletion of CB₁Rs in CB₁-KO mice blocked Δ⁹-THC- or WIN55,212-2-induced analgesia (Compton, Aceto, Lowe, & Martin, 1996; Ledent et al., 1999; Rinaldi-Carmona et al., 1994; Varvel et al., 2005; Wiley & Martin, 2003; Zimmer, Zimmer, Hohmann, Herkenham, & Bonner, 1999). To address the anatomical locus and the cell type-specific receptor mechanisms underlying cannabinoid modulation of pain, multiple cell-type specific CB₁-KO mice have been developed with CB₁R deleted from a restricted neuronal population. As stated above, CB₁Rs are highly expressed on GABA and glutamatergic neurons. However, selective deletion of CB₁Rs on cortical glutamatergic neurons (Glu-CB₁-KO, generated by crossing CB₁-floxed mice with NEX-Cre mice), forebrain GABAergic interneurons (GABA-CB₁-KO, generated by crossing CB₁-floxed mice with Dlx5/6-Cre mice), or dopamine D₁ receptor-expressing neurons (Drd₁-CB₁-KO, generated by crossing CB₁-floxed mice with D₁-Cre mice) failed to alter Δ⁹-THC antinociception (Table 2; Monory et al., 2007). These findings indicate that CB₁Rs on cortical GABAergic or glutamatergic neurons as well as D₁-expressing neurons do not mediate cannabinoid antinociception.

Table 2.

Summary of cannabinoid (THC, WIN55,212-2, JWH133)-induced tetrad effects in transgenic animals.

Transgenic mouse line	Analgesia	Hypothermia	Catalepsy	Hypoactivity	References
Wildtype	Normal THC or WIN effect	Normal effect	Normal effect	Normal effect	Monory et al. (2007) and Wang et al. (2020)
Global CB1-KO	No effect (to THC or WIN)	No effect	No effect	No effect	Ledent et al. (1999), Monory et al. (2007), Wang et al. (2020), and Zimmer, Zimmer, Hohmann, Herkenham, and Bonner (1999)
CaMK-CB1-KO (${\text{CB}}_1^{\text{flox}}$ xCaMK-Cre)	↓ THC effect	No effect	No effect	No effect	Monory et al. (2007)
Glu-CB1-KO (${\text{CB}}_1^{\text{flox}}$ xNEX-Cre)	Normal THC effect	↓ THC effect	Normal effect	↓ THC effect	Monory et al. (2007)
GABA-CB1-KO (${\text{CB}}_1^{\text{flox}}$ xDlx5/6-Cre)	Normal THC effect	Normal effect	Normal effect	Normal effect	Monory et al. (2007)
D1-CB1-KO (${\text{CB}}_1^{\text{flox}}$ xDrd1-Cre)	Normal THC effect	↓ THC effect	↓ THC effect	Normal effect	Monory et al. (2007)
SNS-CB1-KO (${\text{CB}}_1^{\text{flox}}$ xNav1.8-Cre)	↓ WIN effect	N/D	Normal effect	N/D	Agarwal et al. (2007)
Global CB2-KO	↓ THC or WIN effect	Normal effect	↓ THC or WIN effect	Normal effect (rotarod) ↓ THC effect (open-field)	Wang et al. (2020) and Li et al. (2021)
DA-CB2-KO (${\text{CB}}_2^{\text{flox}}$ xDAT-Cre)	Normal effect (to WIN)	Normal effect	↓ WIN effect	Normal effect	Liu et al. (2017))
CX3CR1-CB2-KO (${\text{CB}}_2^{\text{flox}}$ xCX3CR1-Cre)	Normal effect (to WIN, ACEA)	Normal effect	Normal effect	Normal effect	Liu, Canseco-Alba, Liang, Ishiguro, and Onaivi (2020)
Immune-CB2-KO (CB2flox xLysM-Cre)	Normal JWH133 effect	N/D	N/D	N/D	Cabañero et al. (2020)
Syn-CB2-KO (${\text{CB}}_2^{\text{flox}}$ xSyn-Cre)	↓ JWH133 effect	N/D	N/D	N/D	Cabañero et al. (2020)
SNS-CB2-KO (${\text{CB}}_2^{\text{flox}}$ xNav1.8-Cre)	↓ JWH133 effect in thermal pain	N/D	N/D	N/D	Cabañero et al. (2020)
GPR55-KO	Normal THC effect ↑ WIN effect	↑ THC effect	↑ THC effect	↑ THC effect	Wang et al. (2020)

Abbreviations: CaMK—Ca⁺⁺/calmodulin-dependent protein kinase II; CB1^flox—CB1-floxed mice; NEX—a gene encodes neuronal helix-loop-helix protein (a transcription factor), expressed in pyramidal neurons of the dorsal telencephalon during embryonic development; Dlx5/6—distal-less homeobox genes 5/6, expressed in forebrain GABAergic neurons during embryotic development; SNS—sensory neurons expressing Nav1.8 channel; DA—dopamine; LysM—lymphocytes/monocytes; Syn—Synapsin (a neuronal marker).

In a more recent report, De Giacomo and colleagues (2020) used a conditional rescue model in which CB₁Rs are restored only in distinct neuronal subpopulations in full CB₁-KO mice and compared to full CB₁R-KO mice to determine whether CB₁Rs in a given region can reproduce Δ⁹-THC antinociception. In line with the findings from conditional CB₁-KO mice, the rescue of CB₁R expression in dorsal telencephalic glutamate neurons (Glu-CB₁-RS) or forebrain GABA neurons (GABA-CB₁-RS) did not re-establish Δ⁹-THC-induced analgesia (De Giacomo et al., 2020), suggesting that activation of CB₁Rs in forebrain GABA or glutamate neurons is insufficient to produce analgesic effects. In contrast, mice lacking CB₁Rs on CaMKIIα-positive neurons (CaMK-CB₁-KO, generated by CB₁-floxed mice with CaMKIIα-Cre mice) demonstrated attenuated (but not abolished) Δ⁹-THC-induced analgesia (Monory et al., 2007), suggesting that CB₁Rs in CaMKIIα-expressing neurons partially mediate Δ⁹-THC-induced analgesia. In addition, CB₁Rs may still be expressed by GABAergic neurons in other brain regions in the forebrain of GABA-CB₁-KO mice since the Dlx5/6 (distal-less homeobox 5 and 6) genes are expressed in progenitors of GABAergic interneurons only in developing forebrain and their expression strongly diminishes after birth (Dimidschstein et al., 2016). Moreover, in the striatum these genes are not GABA specific. Similarly, CB₁Rs may still be expressed in glutamatergic neurons in other brain regions in the forebrain Glu-CB₁-KO mice since the NEX gene is mainly expressed in pyramidal neurons of the dorsal telencephalon during embryonic development (Schwab et al., 2000). Thus, more studies are required to determine the role of CB₁Rs in GABA or glutamate neurons of other non-forebrain regions in cannabinoid antinociception.

The major site of pain perception is the sensory nervous system, to be more precise, the sensory neurons in the dorsal root ganglia (DRG) as well as the dorsal horn neurons in the spinal cord, which also contain high densities of CB1Rs (Ahluwalia, Urban, Capogna, Bevan, & Nagy, 2000; Farquhar-Smith et al., 2000). To determine the role of CB₁Rs in primary sensory neurons, a peripheral CB₁-KO mouse line was developed, in which CB1Rs in DRG nociceptive (Na_v1.8-expressing) sensory neurons were deleted (SNS-CB1-KO, generated by crossing CB₁-floxed mice with SNS-Cre mice; Agarwal et al., 2007). The nociceptor-specific loss of CB₁Rs substantially reduced analgesia produced by local and systemic, but not intrathecal, delivery of WIN55,212-2 (Table 2; Agarwal et al., 2007). This suggests that CB₁Rs expressed on the peripheral terminals of nociceptors (DRG sensory neurons) are critical in cannabinoid-induced analgesia (Fig. 5). These findings are consistent with work demonstrating that systemic administration of a novel peripherally acting CB₁R agonist, AZ11713908, produced robust analgesia (Yu et al., 2010). Interestingly, CaMKIIα is also highly expressed in DRG neurons (Carlton & Hargett, 2002), suggesting that a peripheral CB₁R mechanism could also contribute to the reduction of Δ⁹-THC-induced analgesia observed in CaMK-CB₁-KO mice. Thus, CB1R mechanisms in peripheral sensory neurons appear to be the primary mechanism underlying cannabinoid analgesic effects (Fig. 5).

Fig. 5.

A summary of the major neural mechanisms underlying THC-induced tetrad effects.

CB₂R mechanisms:

Although CB₁R activation appears to be the primary mechanism underlying Δ⁹-THC’s analgesic effects, other targets have been discovered. For instance, CB₂R agonists have been demonstrated to produce potent analgesic effects in animal models of chronic inflammatory and neuropathic pain (Maldonado, Baños, & Cabañero, 2016; Shang & Tang, 2017). Further, we have recently reported that deletion of CB₂Rs in CB₂-KO mice significantly reduces Δ⁹-THC- or WIN55,212-2-induced analgesia, implicating CB₂Rs in cannabinoid antinociception (Wang et al., 2020). However, the anatomical substrates underlying CB₂R-mediated analgesia are still unclear. Recently, it was reported that mice can learn to self-administer the CB₂R agonist JWH133 to inhibit neuropathic pain (Cabañero et al., 2020). This behavior was blocked by global CB₂-KO mice, suggesting a CB₂R-mediated effect. Interestingly, selective deletion of CB₂R from neurons in neuronal CB₂-KO mice (CB₂-floxed X Syn-Cre) caused an increase in JWH133 self-administration, while selective deletion of CB₂R from immune cells in monocyte-specific CB₂-KO mice (CB₂-floxed X LysM-Cre) did not alter JWH133 self-administration, suggesting that increased spontaneous pain occurs in neuronal CB₂-KO mice and high doses of JWH133 are required to relieve neuropathic pain (Cabañero et al., 2020) (Table 2). These findings provide clear evidence supporting a neuronal CB₂R mechanism underlying CB2R-induced analgesia. However, selective deletion of CB₂R from DGR neurons in peripheral neuronal CB₂-KO mice (CB₂-loxed X Nav1.8-Cre) did not significantly alter JWH133-produced analgesic effects, suggesting that a neuronal CB₂R mechanism in the brain play a dominant role in JWH133- or other cannabinoid-induced analgesia (Fig. 5).

Other mechanisms:

Additionally, the TRPA1 channel is a receptor critically involved in thermal pain perception. Evidence has shown that TRPA1 is implicated in cannabinoid antinociception. Specifically, Akopian et al. (2008) found that WIN55,212-2 induces analgesia in a peripheral capsaicin pain model, which is absent in TRPA1-KO mice. JWH133 produced a significant reduction in either mechanical or thermal hypernociception in a neuropathic pain model (Cabañero et al., 2020). Genetic deletion of TRPA1 blocked JWH133-induced reduction in thermal, but not, mechanical, pain, suggesting possible involvement of TRPA1 in cannabinoid analgesia. In addition, the opioid system was reported to be involved in cannabinoid analgesia. Specifically, κ-opioid receptors (KORs) are involved in Δ⁹-THC-induced analgesia as the κ agonist dynorphin was increased by Δ⁹-THC administration and mice with a genetic deletion of dynorphin showed attenuated Δ⁹-THC-induced analgesia (Houser, Eads, Embrey, & Welch, 2000; Zimmer et al., 2001). However, KOR-KO mice displayed no change in Δ⁹-THC antinociception (Ghozland et al., 2002). Given that Δ⁹-THC does not bind to KORs, dynorphin may indirectly alter Δ⁹-THC’s action against pain. Unexpectedly, genetic deletion of GPR55 produced an opposite enhancement in WIN55,212-2-induced analgesia (Wang et al., 2020), suggesting involvement of GPR55 mechanism in cannabinoid analgesia.

3.2. Hypothermia

Intermediate to high doses of Δ⁹-THC produce a drop in body temperature across different routes of administration and species (Hayakawa et al., 2007; McMahon, Amin, & France, 2005; Taffe, Kevin, Creehan, & Vandewater, 2015; Taffe, Creehan, Vandewater, Kerr, & Cole, 2021; Varvel et al., 2006).

CB₁R mechanism:

CB₁Rs mediate the hypothermic effects of Δ⁹-THC as pharmacological antagonism or genetic deletion of CB₁Rs blocks Δ⁹-THC-induced decreases in temperature (Hayakawa et al., 2007; Ledent et al., 1999; McMahon et al., 2005; Varvel et al., 2005; Zimmer et al., 1999). In contrast to the analgesic effects, hypothermic responses to Δ⁹-THC were significantly attenuated in CaMK-CB₁-KO and forebrain Glu-CB₁-KO mice, but not forebrain GABA-CB₁-KO mice, implicating cortical glutamatergic neurons in Δ⁹-THC-induced hypothermia (Monory et al., 2007). The hypothermic effects of Δ⁹-THC are likely mediated mainly by CB₁Rs expressed in the preoptic anterior hypothalamus (POAH), a major thermoregulatory brain area (Fitton & Pertwee, 1982; Rawls, Cabassa, Geller, & Adler, 2002). In forebrain Glu-CB₁-RS and GABA-CB₁-RS mice, CB₁R expression is rescued in the hypothalamus relative to full CB₁-KO mice (Gutierrez-Rodríguez et al., 2017; Remmers et al., 2017; Ruehle et al., 2013). Consistent with the data from conditional knockout mice, Glu-CB₁-RS, but not GABA-CB₁-RS, mice showed a partial rescue of Δ⁹-THC hypothermia. As such, glutamatergic CB₁Rs may play a dominant role in Δ⁹-THC-mediated hypothermia. Glutamate tonically increases body temperature by binding to NMDA receptors in the preoptic hypothalamus (Sengupta, Jaryal, & Mallick, 2016). One report found that the hypothermic response to WIN55,212-2 is synergistically enhanced by NMDA receptor antagonism (Rawls, Cowan, Tallarida, Geller, & Adler, 2002). Microinjections of WIN55,212-2 into the POAH induced hypothermia (Rawls, Cabassa, et al., 2002). These findings suggest that cannabinoids may act on a glutamatergic pathway in the POAH to produce hypothermia (Fig. 5).

Non-CB₁R mechanisms:

Additional non-CB₁R receptor mechanisms have been implicated in Δ⁹-THC-induced temperature shifts. We found that blockade or deletion of CB₂Rs in global CB₂-KO mice or selective deletion of CB2Rs in midbrain DA neurons failed to alter Δ⁹-THC- or WIN55,121-2-induced hypothermia (Liu et al., 2017; Wang et al., 2020). In contrast, selective antagonism or genetic deletion of GPR55 receptors augmented hypothermia in response to Δ⁹-THC or WIN55,212-2 (Wang et al., 2020), suggesting that activation of GPR55 has a suppressive effect on Δ⁹-THC-induced hypothermia. No prior work has established a role for GPR55 in temperature control. However, knowledge of this receptor is limited, and future work should investigate this possibility. In addition, serotonergic 5-HT_1A receptors and dopamine D₂ receptors also regulate Δ⁹-THC-induced hypothermia in an opposing manner such that D₂ receptor antagonists attenuate and 5-HT_1A receptor antagonists potentiate Δ⁹-THC’s hypothermic effects and vice versa with their respective agonists (Malone & Taylor, 2001; Nava, Carta, & Gessa, 2000). Although Δ⁹-THC has no direct binding affinity at 5-HT_1A and D2 receptors, these effects could be mediated indirectly via Δ⁹-THC metabolite activity at the 5-HT_1AR and CB₁-D₂ heterodimer interactions.

3.3. Catalepsy

In rodents, Δ⁹-THC induces catalepsy at high doses (10mg/kg and above; Long et al., 2010; Metna-Laurent, Mondésir, Grel, Vallée, & Piazza, 2017). The most common behavioral assay of catalepsy is the bar test in which an animals forepaws are placed on a horizontal bar and the amount of time it takes them to move out of this unusual conformation and put both paws on the ground is recorded (Sanberg, Bunsey, Giordano, & Norman, 1988).

CB₁R mechanism:

As with Δ⁹-THC induced hypothermia, blockage or deletion of CB₁Rs effectively abolishes Δ⁹-THC’s cataleptic effects (Ledent et al., 1999; Lichtman & Martin, 1997; Tseng & Craft, 2004; Varvel et al., 2005; Zimmer, Zimmer, Hohmann, Herkenham, & Bonner, 1999). An early study found that deletion of CB₁Rs on forebrain GABA or glutamate neurons failed to alter the cataleptic effects of Δ⁹-THC (Monory et al., 2007), demonstrating that CB₁Rs on both neuronal cell types in the cortex do not mediate Δ⁹-THC-induced catalepsy. This is consistent with findings from conditional rescue mice in which neither dorsal telencephalic glutamatergic CB₁Rs nor forebrain GABAergic CB₁Rs were sufficient to rescue the cataleptic effect of Δ⁹-THC (De Giacomo et al., 2020).

Interestingly, deletion of CB₁Rs from CaMKIIα (CaMK-CB₁-KO) or D₁-expressing neurons (Drd1-CB₁-KO) abolished Δ⁹-THC-induced catalepsy (Monory et al., 2007), suggesting that CB1Rs on both types of neurons play a critically important role in catalepsy produced by cannabinoids. CaMKIIα is expressed in numerous neuronal cell types that project to a myriad of brain regions. Therefore, it is unknown exactly how CB₁Rs in CaMKIIα-expressing neurons underlie cannabinoid-induced catalepsy. In contrast, D1Rs are mainly distributed in one population of GABAergic medium-spiny neurons (D₁-MSNs) in the striatum and glutamatergic neurons in the cortex. It is well known that D₁-MSNs regulate voluntary motor movements (van der Stelt & Di Marzo, 2003). Activation of CB₁Rs on D₁-MSNs likely inhibits GABAergic MSNs in the striatum, producing motor impairment. This is supported by the finding that microinjections of Δ⁹-THC into the nucleus accumbens (NAc) produced catalepsy (Sano et al., 2008) that was inhibited by both serotoninergic agonists and NMDA receptor antagonists (Nobuaki Egashira et al., 2006; Kinoshita et al., 1994). These findings suggest that CB₁R expression in the striatum may be a primary brain region underlying Δ⁹-THC-induced catalepsy (Fig. 5).

CB₂R mechanism:

In addition to CB₁R, CB₂R also plays a role in the cataleptic effects of Δ⁹-THC or WIN55212-2. Indeed, deletion and pharmacological antagonism of CB₂Rs attenuated cataleptic behavior following Δ⁹-THC or WIN55,212-2 administration (Wang et al., 2020). Furthermore, selective deletion of CB₂Rs from midbrain DA neurons attenuated WIN55,212-2-induced catalepsy (Liu et al., 2017), while deletion of CB₂R from microglia (CB₂-floxed X CX3CR1-Cre) had no effect on WIN55,212-2-induced catalepsy (Liu, Canseco-Alba, Liang, Ishiguro, & Onaivi, 2020), suggesting that a neuronal, not microglial, CB₂R mechanism underlies cannabinoid-induced catalepsy. In addition, it was recently reported that CB₂Rs are highly expressed in glutamate neurons in the red nucleus of the midbrain and modulate locomotor activity (Zhang, Shen, et al., 2021). These findings together suggest that activation of CB₂Rs in the mesolimbic DA neurons and the motor circuit glutamate neurons at least in part underlies cannabinoid-induced catalepsy (Fig. 5).

GPR55 mechanism:

On the other hand, it was recently reported that GPR55 receptors are densely distributed in the striatum and administration of a GPR55 agonist (abnormal-cannabidiol) has been shown to block catalepsy produced by haloperidol (Marichal-Cancino, Fajardo-Valdez, E. Ruiz-Contreras, Mendez-Díaz, & Prospero-García, 2017; Celorrio et al., 2017), while pharmacological blockade of GPR55 potentiate Δ⁹-THC-induced catalepsy (Wang et al., 2020). Similarly, mice lacking GPR55 demonstrated enhanced Δ⁹-THC- or WIN55,212-2-induced catalepsy (Wang et al., 2020). These findings suggest that GPR55 activation may produce an anti-cataleptic effect. As such, the final behavioral expression of cannabinoid-induced catalepsy may depend on the respective contributions of CB₁R, CB₂R, and GPR55.

3.4. Hypolocomotion

Δ⁹-THC suppresses locomotor activity at doses of 3mg/kg and above. The open field locomotion test is most often utilized to measure changes in movement. Mice are placed in a large, empty container and the distance they travel is monitored. The rotarod test is another measure of locomotor performance, particularly motor coordination, in which mice are placed on an elevated revolving rod and the time it takes them to fall is recorded.

CB₁R mechanism:

Like the other assays within the tetrad, Δ⁹-THC or WIN55,212-2 alters locomotor activity via activation of CB₁Rs as deletion of CB₁Rs abolished Δ⁹-THC and other cannabinoids-induced locomotor impairment (Ledent et al., 1999; Nguyen et al., 2016; Taffe, Creehan, & Vandewater, 2015; Zimmer et al., 1999). Findings from three different conditional CB₁-KO mice strains (Glu-CB₁-KO, CaMK-CB₁-KO and VgluT2-CB₁-KO) implicate glutamatergic neurons in Δ⁹-THC’s effects on locomotion (Monory et al., 2007; Han et al., 2017). These findings parallel work with conditional rescue mice in which restoration of CB1R expression in dorsal telencephalic glutamatergic neurons (Glu-CB₁-RS mice) reestablished Δ⁹-THC-induced locomotor suppression (De Giacomo et al., 2020). In contrast, deletion of CB₁Rs in forebrain GABA neurons failed to alter Δ⁹-THC-induced locomotor impairment (Monory et al., 2007). Similarly, GABA-CB₁-RS mice with CB₁R expression rescued in forebrain GABAergic neurons showed no evidence of Δ⁹-THC locomotor inhibition (De Giacomo et al., 2020). Previous work has demonstrated that cannabinoids attenuate excitatory glutamatergic input in the striatum (Brown, Brotchie, & Fitzjohn, 2003). Thus, CB₁Rs on corticostriatal glutamatergic projection neurons likely mediate hypolocomotion produced by Δ⁹-THC (Monory et al., 2007) (Fig. 5).

The basal ganglia contains two major GABAergic neuronal populations—D₁-MSNs and D₂-MSNs. Both populations of neurons express CB₁Rs (Hermann, Marsicano, & Lutz, 2002) and control basal ganglia motoric output (Graybiel, 2000). Activation of D₁-MSNs enhances, while activation of D₂-MSNs inhibits locomotion (Calabresi, Picconi, Tozzi, Ghiglieri, & Di Filippo, 2014; Kravitz et al., 2010). The D₁-expressing MSNs have become a locus of interest since Δ⁹-THC treatment may directly inhibit this locomotor-enhancing population of neurons. However, CB₁R deletion from D₁-MSNs had no effect on Δ⁹-THC-induced hypolocomotion (Monory et al., 2007). Interestingly, the effects of Δ⁹-THC on overall locomotor activity (open field test) vs. motor coordination (rotarod) may have distinct neural underpinnings. Indeed, Blazquez and colleagues (2020) found that Glu-CB₁-KO and WT mice had comparable deficits in motor coordination following Δ⁹-THC in direct contrast to the findings described above using the open-field. Further, the motor dyscoordinating effects of Δ⁹-THC were absent in Drd1-CB₁-KO mice, indicating that D₁-MSNs are critical for Δ⁹-THC-induced deficits in motor coordination (Blázquez et al., 2020) (Fig. 5).

CB₂R mechanism:

In addition to CB₁R mechanisms, dopaminergic CB₂Rs may also underlie Δ⁹-THC-induced locomotor depression. We have previously reported that CB₁ and CB₂ receptors modulate locomotor activity in opposite directions (Li et al., 2021; Li et al., 2009; Wang et al., 2020; Xi et al., 2011). Specifically, genetic deletion of CB₁Rs decreased basal locomotor activity, while genetic deletion of CB₂Rs produced a moderate increase, indicating that activation of CB₂Rs inhibits locomotor behavior (Li et al., 2021; Wang et al., 2020). Systemic or intra-NAc administration of JWH133, a selective CB₂R agonist, inhibits basal level locomotion and decreases cocaine’s locomotor activating effects in a dose-dependent manner (Xi et al., 2011). Further, genetic deletion of CB₂Rs blocked the Δ⁹-THC-induced reduction in open-field locomotion (Li et al., 2021; Wang et al., 2020), implicating CB2Rs in Δ⁹-THC’s locomotor suppressant effects. When CB₂Rs are selectively deleted from midbrain DA neurons, mice show an increase in basal locomotor activity (Canseco-Alba et al., 2019). These findings suggest that dopaminergic CB₂Rs contribute to Δ⁹-THC-induced hypolocomotion (Galaj & Xi, 2019; Jordan & Xi, 2019) (Fig. 5).

GPR55 mechanism:

Finally, GPR55-KO mice showed heightened Δ⁹-THC or WIN55,212-2-induced deficits in motor coordination on the rotarod (Wang et al., 2020). These findings complement prior work in which administration of a GPR55 agonist improved performance on the rotarod following selective lesions of striatal DA neurons (Fatemi, Abdollahi, Shamsizadeh, Allahtavakoli, & Roohbakhsh, 2021). This work demonstrates that GPR55 agonism is involved in motor coordination and has an obverse effect on Δ⁹-THC hypomotility in the tetrad, although the cell types and brain regions responsible are unknown.

In summary, Δ⁹-THC and other cannabinoids produce classical tetrad effects through multiple receptor mechanisms, including CB₁R, CB₂R and GPR55 with CB₁R predominant (Table 2). Technical advances in detecting low level gene expression and the development of conditional transgenic animals have begun to uncover the region and cell type-specific subpopulations that underlie Δ⁹-THC’s effects in the tetrad. In brief, CB₁R in peripheral primary sensory neurons of the DRG and CB₂R in super-spinal neurons appear to be the major targets underlying THC-induced analgesia, while glutamatergic CB₁Rs in the preoptic anterior hypothalamus are not only necessary, but also sufficient for Δ⁹-THC-induced hypothermia. The cataleptic and locomotor suppressant effects of Δ⁹-THC are likely mediated mainly by activation of CB₁Rs on corticostriatal glutamatergic projection neurons and CB₂Rs on midbrain DA neurons and red nucleus glutamate neurons (Fig. 5).

4. Cannabinoid subjective effects

The subjective experience of cannabis varies on the affective spectrum from person to person. The majority of human users report enjoyment, relaxation and laughter, while others describe paranoia, anxiety and depression (Green, Kavanagh, & Young, 2003). In preclinical work, negative or aversive effects of Δ⁹-THC are most commonly observed, particularly at high doses, whereas reward is rarer, difficult to replicate and only observed with low doses. A number of preclinical models of drug reward are utilized in addiction research. The gold standard is intravenous self-administration where animals are implanted with a jugular catheter and trained to make operant responses for drug infusions. Another commonly used model is place conditioning in which the amount of time spent in a context formerly associated with drug exposure is used as a measure of reward. Aversion can also be assessed in this model if time in the drug paired context drops considerably after conditioning. Lastly, intracranial self-stimulation (ICSS) is a behavioral test that assesses how drugs of abuse alter operant responding for electrical stimulation of the median forebrain bundle or optical stimulation of a specific phenotype of neurons such as DA neurons or glutamate neurons. A decrease or increase in brain-stimulation reward (BSR) thresholds denotes a rewarding or aversive drug effect, respectively. The following section will walk through studies investigating the neural mechanisms underlying Δ⁹-THC reward versus aversion using these behavioral models.

4.1. Cannabinoid reward

Self-administration of Δ⁹-THC has been demonstrated in squirrel monkeys at low doses (4μg/kg/infusion), but not in rhesus monkeys, an effect that can be blocked by rimonabant, a selective CB1R blocker (John et al., 2018; Justinova, Tanda, Redhi, & Goldberg, 2003; Mansbach, Nicholson, Martin, & Balster, 1994; Tanda, Munzar, & Goldberg, 2000). However, in rodents (rats and mice), Δ⁹-THC or WIN55,212-2 alone cannot maintain reliable self-administration possibly due to the limited reinforcing efficacy and anxiogenic effects of cannabinoids (Lefever, Marusich, Antonazzo, & Wiley, 2014; Takahashi & Singer, 1979). Interestingly, it was recently reported that passive Δ⁹-THC pre-exposure or co-administration of Δ⁹-THC with cannabidiol (CBD), a phytocannabinoid devoid of psychotomimetic effects, improved cannabinoid self-administration in rats (Spencer et al., 2018). A small number of studies have demonstrated Δ⁹-THC-induced place preferences and decreases in BSR thresholds at the low end of the dose range (0.075–1mg/kg), which are absent in the presence of a CB₁R antagonist or in CB₁-KO mice (Braida, Iosue, Pegorini, & Sala, 2004; Foll, Wiggins, & Goldberg, 2006; Gardner et al., 1988; Ghozland et al., 2002; Katsidoni, Kastellakis, & Panagis, 2013; Lepore, Liu, Savage, Matalon, & Gardner, 1996; Lepore, Vorel, Lowinson, & Gardner, 1995; Li et al., 2021; Soria et al., 2004; Valjent & Maldonado, 2000). Microinjections of Δ⁹-THC directly into the NAc shell and posterior ventral tegmental area (VTA) also support place preferences in rats, implicating these brain regions in Δ⁹-THC-induced reward (Zangen, Solinas, Ikemoto, Goldberg, & Wise, 2006).

GABAergic CB₁R mechanism:

Δ⁹-THC administration produces a rise in DA concentration in the NAc (Tanda, Pontieri, & Chiara, 1997) and increases the firing rate of dopaminergic neurons in the VTA (French, Dillon, & Wu, 1997). Thus, it was hypothesized that CB₁Rs on GABA neurons mediate the rewarding effects of Δ⁹-THC via disinhibition of VTA DA neurons (Fig. 6). This hypothesis is supported by electrophysiological data demonstrating decreased GABA activity in midbrain slices in the presence of Δ⁹-THC and WIN55,212-2 (Friend et al., 2017; Szabo et al., 2002). Additionally, transgenic FAAH^C/A knock-in mice, which recapitulate the FAAH (fatty acid amide hydrolase) polymorphism and display decreased FAAH expression and elevated circulating AEA, produced an enhanced place preference in adolescent female FAAH^C/A mice relative to controls (Burgdorf et al., 2020). Importantly, this increase in cannabinoid reward was accompanied by greater expression of GABAergic CB₁Rs and lower expression of glutamatergic CB₁Rs in the VTA (Burgdorf et al., 2020).

Fig. 6.

Neural mechanisms underlying cannabinoid reward vs. aversion. CB₁Rs are expressed in VTA GABAergic neurons and glutamatergic neurons as well as their afferents projected from other brain regions to VTA DA neurons (data shown). CB₂Rs are found in VTA DA neurons. Cannabinoids modulate the mesolimbic DA system via activation of brain CB₁Rs and CB₂Rs. Cannabinoids such as Δ⁹-THC or WIN55,212-2 produce rewarding effects by binding to CB₁Rs on VTA GABAergic interneurons and/or their afferents, thereby reducing GABA-mediated disinhibition of VTA DA neurons and cannabinoid reward. Conversely, Δ⁹-THC or WIN55,212-2 may also produce aversive effects by activating CB₁Rs on glutamatergic neurons in the VTA or glutamatergic afferents, and CB₂Rs on midbrain DA neurons, thereby inhibiting VTA DA release to the NAc. The subjective effects of cannabinoids may thus depend on the balance of opposing CB1R and CB₂R effects and individual differences in expression of cannabinoid receptors. DA, dopamine; GABA, γ-aminobutyric acid; NAc, nucleus accumbens; VTA, ventral tegmental area.

Non-CB₁R mechanisms:

Other work has evaluated whether non-cannabinoid receptor systems are involved in the rewarding effects of Δ⁹-THC. In one report, the MOR antagonist, naltrexone, decreased Δ⁹-THC self-administration (Justinova, Tanda, Munzar, & Goldberg, 2004) and Δ⁹-THC place preferences were absent in MOR-KO mice (Ghozland et al., 2002). Δ⁹-THC increases β-endorphin release in the VTA, which could explain the lack of rewarding effects when MORs are antagonized or deleted (Solinas, Zangen, Thiriet, & Goldberg, 2004). Another series of studies implicated A_2A adenosine receptors (A_2ARs) in the rewarding effects of Δ⁹-THC. CB₁Rs form heterodimers with A_2ARs on presynaptic cells and activation of A_2ARs counteracts the inhibitory effects of CB₁Rs on glutamate release in corticostriatal terminals (Ferreira et al., 2015; Köfalvi et al., 2020). Antagonism of presynaptic A_2ARs was shown to reduce Δ⁹-THC self-administration in squirrel monkeys, indicating that a decrease in cortical striatal glutamate attenuates Δ⁹-THC reward (Justinová et al., 2011). Similarly, inhibition of postsynaptic A_2ARs potentiated Δ⁹-THC self-administration (Justinová, Redhi, Goldberg, & Ferre, 2014). These findings suggest that excess glutamate in corticostriatal brain regions, perhaps the NAc, may also contribute to the rewarding effects of Δ⁹-THC likely by stimulating dopamine release in the striatum. Finally, α7 nicotinic acetylcholine receptors (α7nAChRs) also play a role in Δ⁹-THC reward. Kynurenic acid (KYNA) acts as a negative allosteric modulator of α7nAChRs and increased levels of KYNA inhibit Δ⁹-THC-induced increases in NAc dopamine and Δ⁹-THC self-administration in squirrel monkeys (Justinova et al., 2013). α7nAChRs are located on glutamatergic neurons that project to the NAc shell (Dani & Bertrand, 2007; Secci et al., 2019). As such, suppression of glutamatergic activity by KYNA and subsequent decreases in NAc DA likely underlie the decrease in Δ⁹-THC reward. This work by Justinova and colleagues provides further evidence in support of a glutamatergic accumbal mechanism meditating Δ⁹-THC’s rewarding properties perhaps in conjunction with disinhibition of GABAergic tone in the VTA.

4.2. Cannabinoid aversion

The aversive effects of Δ⁹-THC are well documented in preclinical work. Subjects develop robust place aversions and increases in BSR thresholds in ICSS particularly at intermediate to high doses (3–20mg/kg; Braida et al., 2004; Hempel, Clasen, Nelson, Woloshchuk, & Riley, 2018; Katsidoni et al., 2013; Li et al., 2021; Mallet & Beninger, 1998a, b; Schramm-Sapyta et al., 2007; Spiller et al., 2019; Valjent & Maldonado, 2000; Vann et al., 2008; Wiebelhaus et al., 2015). We have recently tested cannabinoids in a new assay, optogenetic ICSS (oICSS), to further evaluate the rewarding (or reward-enhancing) versus aversive (or reward-attenuating) effects of a given drug (Han et al., 2017; Jordan et al., 2019; Newman et al., 2019). In this procedure, an adeno-associated virus (AAV) carrying a Cre-dependent channelrhodopsin 2 (ChR2) gene is microinjected into the VTA to express light-sensitive ChR2 in DA neurons of transgenic DA transporter (DAT)-Cre mice or glutamatergic neurons in VgluT2-Cre mice. Using this assay, we found that systemic administration of Δ⁹-THC or WIN55,212-2 dose-dependently inhibited oICSS maintained by optical stimulation of VTA DA neurons and shifted the stimulation-response curve rightward or downward (Humburg et al., 2021), suggesting that cannabinoids are aversive in mice.

Glutamatergic CB₁R mechanism:

The receptor mechanisms underlying Δ⁹-THC’s aversive effects are not fully understood. In prior work, cannabinoid-induced reductions in BSR thresholds were prevented by CB₁R antagonism (SR141716A or AM251; Katsidoni, Kastellakis, & Panagis, 2013; Spiller et al., 2019), pointing to a CB₁R mechanism in cannabinoid-induced aversion. To determine the specific cell types involved, we have recently used optogenetics to stimulate VTA glutamate neurons in VgluT2-Cre mice (Han et al., 2017). Unexpectedly, Δ⁹-THC significantly inhibited oICSS maintained by optical stimulation of VTA glutamatergic neurons, but this effect was absent in VgluT2-CB₁-KO mice. Similarly, deletion of CB₁Rs on glutamate neurons prevented the expression of Δ⁹-THC place aversions (Han et al., 2017). These findings suggest that CB₁Rs on VTA glutamate neurons are involved in Δ⁹-THC-induced aversion (Fig. 6). This fits with the model of Δ⁹-THC’s affective properties described above wherein increased expression of CB₁Rs on GABA neurons and lower expression of CB₁Rs on glutamate neurons shifts the affective properties of Δ⁹-THC towards reward (Burgdorf et al., 2020).

Dopaminergic CB₂R mechanism:

Beyond CB₁Rs, Δ⁹-THC’s aversive effects were also blocked by CB₂R antagonism, as assessed by an increase in BSR thresholds in rats (Spiller et al., 2019). Additionally, genetic deletion of CB₂Rs shifted Δ⁹-THC place conditioning from a place aversion in wildtype mice to a place preference in CB₂-KO mice (Li et al., 2021), indicating that CB₂Rs play a role in the initial aversive effects of Δ⁹-THC. CB₂Rs are expressed on VTA dopaminergic neurons (Zhang et al., 2014; Zhang et al., 2017; Zhang et al., 2019) and can decrease the firing rate of these cells as well as diminish DA release in the NAc (Ma et al., 2019; Zhang et al., 2014; Zhang et al., 2017). These findings implicate CB₂Rs on mesolimbic DA neurons in THC-induced aversion (Fig. 6).

PPAR mechanisms:

Little is known regarding the role of PPARs in cannabinoid action. However, we have recently explored the function of PPARs in Δ⁹-THC induced aversion using oICSS (Hempel, Bi, Klein, & Xi, 2021). Administration of Δ⁹-THC decreased responding for optical stimulation of VTA DA neurons in DAT-cre mice and this effect was attenuated by administration of a PPARα or PPARγ antagonist. As previously stated, Δ⁹-THC is a potent PPARγ agonist (Table 1), which delineates how PPARγ antagonism reduced Δ⁹-THC action on oICSS. PPARα receptors have been detected in the VTA and NAc at low levels (Warden et al., 2016). However, the mechanism through which PPARα modulates Δ⁹-THC’s aversive effects is still unclear. While Δ⁹-THC administration does produce changes in PPARα gene transcription, it does not bind to the α isoform.

Kappa opioid receptor mechanism:

Finally, κ-opioid receptors (KORs) have been investigated as a potential receptor target mediating Δ⁹-THC aversion. Mice with elevated expression of the opioid encoding gene prodynorphin, a precursor of the KOR agonist dynorphin, demonstrated enhanced Δ⁹-THC place aversions relative to controls (Cheng, Laviolette, van der Kooy, & Penninger, 2004). In the same vein, dynorphin-deficient and KOR-KO mice developed attenuated Δ⁹-THC place aversions as did mice administered a KOR antagonist prior to assessments of Δ⁹-THC aversion (Clasen et al., 2017; Ghozland et al., 2002; Zimmer et al., 2001). Microinjections of Δ⁹-THC into the posterior NAc shell produced a significant place aversion that was attenuated by KOR antagonism, implicating NAc KORs in the expression of Δ⁹-THC aversion (Norris, Szkudlarek, Pereira, Rushlow, & Laviolette, 2019).

5. Cannabinoid effects on anxiety

One of the most commonly cited reasons for cannabis use in humans is the relaxing effects of the drug (Ewusi Boisvert et al., 2020; Green, Kavanagh, & Young, 2003); however, some individuals experience anxiety under the influence (Spindle et al., 2018). In preclinical models, Δ⁹-THC has a biphasic effect on anxiety: anxiolytic at low doses and anxiogenic at high doses. The elevated plus maze (EPM) is a frequently used animal model in which rodents are placed on a raised apparatus containing two crossed arms – one of which is enclosed by walls and the other is open. Greater time spent in the open arms is a measure of decreased anxiety. The light dark test is another assay of anxiety that takes advantage of rodents’ preference for dark, enclosed spaces. Subjects have access to two compartments separated by a door – one compartment is open and well-lit and the other is dark and enclosed. Anxiolytic drugs increase the proportion of time spent in the light compartment. The following subsections will update the current understanding of the receptor mechanisms underlying the anxiolytic and anxiogenic effects of Δ⁹-THC.

5.1. Anxiolytic

As with the majority of cannabinoid effects on the central nervous system, administration of a CB₁ antagonist blocks Δ⁹-THC-induced anxiolytic properties (Berrendero & Maldonado, 2002; Rubino et al., 2007). CB₁Rs in the amygdala and prefrontal cortex (PFC) underlie this effect (Tiziana Rubino et al., 2007). CP55,940 also produces an anxiolytic-like response at a low dose (1μg/kg), which is absent in Glu-CB₁-KO, but not GABA-CB₁-KO mice (Rey, Purrio, Viveros, & Lutz, 2012), suggesting that glutamatergic CB₁Rs in the PFC and amygdala mediate the initial anxiolytic effects of Δ⁹-THC. However, further work is needed to confirm this.

CB₁R and CB₂R mechanisms:

One study found that brief exposure to a predator odor produced an anxiety-like state in rats that was blocked by the selective monoacylglycerol lipase (MGL) inhibitors, KML29 and JZL184 (Ivy et al., 2020). MGL inhibitors prevent 2-AG degradation and elevate brain 2-AG concentrations, indicating that increased levels of 2-AG were anxiolytic in this model. Unexpectedly, the behavioral response to JZL184 was abolished by a CB₂R, but not CB₁R, antagonist. Specifically, the selective CB₂R agonist, JWH133, produced anxiolytic-like effects in rats exposed to a predator odor stressor and this response was blocked by the CB₂R antagonist AM630 (Ivy et al., 2020). However, early studies demonstrated that JZL184 produced marked anti-anxiety effects that were prevented by administration of the CB₁R antagonist SR141716A (Kinsey, O’Neal, Long, Cravatt, & Lichtman, 2011; Sciolino, Zhou, & Hohmann, 2011). Thus, 2-AG relieves anxiety potentially through activation of both CB₁Rs and CB₂Rs.

Other receptor mechanisms:

Outside of the cannabinoid receptor family, other systems also modulate Δ⁹-THC’s anti-anxiety effect (Table 3). For instance, antagonism of the MOR and DOR suppressed the anxiolytic response to Δ⁹-THC, although no additional work has investigated this neurobiological mechanism (Berrendero & Maldonado, 2002). Additionally, serotonergic receptors are known to regulate anxiety and have been investigated in this context. Administration of a 5-HT_1A receptor antagonist abolished the anxiolytic effect of Δ⁹-THC (Braida, Limonta, Malabarba, Zani, & Sala, 2007). Similarly, genetic deletion of 5-HT_2A receptors (5-HT_2ARs) blocked the anxiolytic response to Δ⁹-THC (Viñals et al., 2015). Viñals et al. (2015) further demonstrated that CB1Rs form heterodimeric complexes with 5-HT_2ARs and perturbation of this relationship via transmembrane helix interference peptides suppressed the anti-anxiety effect of Δ⁹-THC. These heterodimers were found in the cortex, hippocampus and striatum, although it’s unknown exactly which regions mediate the anxiolytic effects of Δ⁹-THC.

Table 3.

Receptor mechanisms underlying Δ⁹-THC’s affective properties.

Assay	Target	Antagonist	Knockout mouse	Results	References
Reward
IVSA	CB1Rs	AM251 SR141716A	N/A	Blocked	Freels et al. (2020, Justinova, Tanda, Redhi, and Goldberg (2003), and Spencer et al. (2018)
CPP	CB1Rs	SR141716A	CB1-KO	Blocked	Braida, Iosue, Pegorini, and Sala (2004), Li et al. (2021)
ICSS	CB1Rs	SR141716A	N/A	Blocked	Katsidoni, Kastellakis, and Panagis (2013)
IVSA	MORs	Naltrexone	N/A	Attenuated	Justinova, Tanda, Munzar, & Goldberg, 2004
CPP	MORs	N/A	MOR-KO	Blocked	Ghozland et al. (2002)
IVSA	A2ARs	MSX-3	N/A	Attenuated	Justinová et al. (2011)
IVSA	Presynaptic A2ARs Postsynaptic A2ARs	SCH-442416 KW-6002	N/A	Attenuated Enhanced	Justinová et al. (2011) and Justinová, Redhi, Goldberg, and Ferré (2014)
IVSA	α7nAChRs	Ro 61-8048	N/A	Attenuated	Justinova et al. (2013)
Aversion
ICSS	CB1Rs	SR141716A	N/A	Blocked	Katsidoni, Kastellakis, and Panagis (2013)
ICSS	CB2Rs	AM630	N/A	Blocked	Spiller et al. (2019)
oICSS	PPARα/PPARγ	GW6471 GW9662	N/A	Attenuated	Hempel et al. (2021)
oICSS CPA	Subcortical glutamate-CB1Rs	N/A	VgluT2-CB1-KO	Attenuated Blocked	Han et al. (2017)
CPA	CB2Rs	N/A	CB2-KO	Blocked	Li et al. (2021)
CPA	Prodynorphin gene	N/A	Dream-KO	Enhanced	Cheng, Laviolette, van der Kooy, and Penninger (2004)
CPA	Prodynorphin gene	N/A	Pdyn-KO	Attenuated	Zimmer et al. (2001))
CPA	KORs	Nor-BNI	KOR-KO	Blocked	Clasen et al. (2017), Ghozland et al. (2002) and Norris, Szkudlarek, Pereira, Rushlow, and Laviolette (2019)
Anxiolytic effect
LDB EPM	CB1Rs	SR141716A	N/A	Blocked	Berrendero and Maldonado (2002) and Rubino et al. (2007)
LDB	MORs DORs	β-funaltrexamine Naltrindole	N/A	Blocked	Berrendero and Maldonado (2002)
EPM	5-HT1ARs	WAY100635	N/A	Blocked	Braida, Limonta, Malabarba, Zani, and Sala (2007)
EPM	5-HT2ARs	N/A	5-HT2AR-KO	Blocked	Viñals et al. (2015)
EPM	CB2Rs	AM630	N/A	Blocked	Ivy et al. (2020)
Anxiogenic effect
EPM	CB1Rs	AM251	N/A	Blocked	Rubino et al. (2008)
EPM	PPARα	GW6471	N/A	Attenuated	Hempel et al. (2021)
EPM Open field	Unknown	CBD	N/A	Attenuated	Liu et al. (2017), Murphy et al. (2017), and Todd and Arnold (2016)

Abbreviations: IVSA—intravenous self-administration, CPP—conditioned place preference, CPA—conditioned place aversion, ICSS—intracranial self-stimulation, oICSS—optical ICSS, LDB—Light dark box, EPM—elevated plus maze, DREAM—downstream regulatory element antagonistic modulator, Pdyn, prodynorphin.

5.2. Anxiogenic

CB₁Rs in the basolateral amygdala mediate the anxiogenic effects of Δ⁹-THC, as prior work has demonstrated that microinjections of Δ⁹-THC into this brain region produced anxiety that was attenuated by pretreatment with the CB₁R antagonist AM251 (Rubino et al., 2008). Interestingly, mice lacking CB₁Rs on GABAergic, but not glutamatergic neurons, failed to demonstrate anxiety in response to CP55,940 (50μg/kg; Rey, Purrio, Viveros, & Lutz, 2012). A GABA_B receptor-related mechanism may also contribute to the anxiogenic effect as positive allosteric modulation of GABA_B receptors blocked CP55,940-induced anxiety. Whether these findings extend to Δ⁹-THC’s anxiogenic response is unknown. In addition, we have recently discovered that inhibition of PPARα can attenuate the anxiety produced by 5mg/kg Δ⁹-THC (Hempel et al., 2021). PPARα is expressed brain regions that regulate anxiety including the amygdala and hippocampus (Kainu, Wikström, Gustafsson, & Pelto-Huikko, 1994; Warden et al., 2016). However, the function of PPARα in the CNS is still being explored and very little is currently known.

A number of studies have demonstrated that co-administration of CBD can inhibit the anxiogenic effects of Δ⁹-THC (Liu, Scott, & Burnham, 2021; Murphy et al., 2017; Szkudlarek et al., 2019; Todd & Arnold, 2016; Zuardi, Shirakawa, Finkelfarb, & Karniol, 1982). CBD targets a number of receptors, but could suppress the effects of Δ⁹-THC on anxiety via negative allosteric modulation of the CB1R or 5-HT_1AR activation (Campos & Guimarães, 2008; Galaj & Xi, 2021; Laprairie, Bagher, Kelly, & Denovan-Wright, 2015; Russo, Burnett, Hall, & Parker, 2005). Prior work has demonstrated that CBD can block anxiety caused by restraint stress via activity at the 5-HT_1AR (Resstel et al., 2009). Along these lines, Szkudlarek et al. (2019) found that intra-PFC injections of CBD suppressed Δ⁹-THC-induced anxiety in the EPM and 5-HT_1ARs are highly expressed in PFC neurons. Further work is needed to ascertain if direct 5-HT_1AR activation mediates CBD’s inhibitory effect on Δ⁹-THC in preclinical models of anxiety (for a summary see Table 3).

6. Cannabinoid cognitive effects

Acute exposure to cannabis impairs executive functions in human users across a number of domains such as attention, inhibitory control, psychomotor control, short term episodic memory, working memory and spatial memory (Crane, Schuster, Fusar-Poli, & Gonzalez, 2013; Crean, Crane, & Mason, 2011; Ranganathan & D’Souza, 2006). In animals, deficits in learning and memory are observed in spatial learning and memory, short term memory, repeated acquisition, habit formation, and fear conditioning (Goodman & Packard, 2015; Kangas et al., 2016; Prini et al., 2020; Resstel, Moreira, & Guimarães, 2009). A comprehensive review of the literature regarding the neurobiology of cannabinoids and cognition is beyond the scope of the present work. Here, we will focus on the neural mechanisms underlying acute effects of the most consistent Δ⁹-THC-induced neurocognitive impairments, namely spatial learning and memory and short-term memory. A number of different animal models have been used in this context. Two commonly employed tests are the Morris water maze (MWM) and the novel object recognition task (NOR). The MWM assesses spatial learning and memory and in this task, animals are placed in a tank filled with cloudy liquid and must find a hidden platform to escape. In the NOR, subjects are initially exposed to two objects and in a subsequent session they are presented with one familiar and one novel object. Time spent exploring the new object is a measure of short-term memory impairment (for a summary see Table 4).

Table 4.

Receptor mechanisms underlying Δ⁹-THC-induced learning and memory impairments.

Assay	Target	Antagonist	Knockout mouse	Results	References
Spatial learning and memory
RAM	CB1R	SR141716A	N/A	Blocked	Lichtman and Martin (1996)
MWM	CB1R	SR141716A	CB1-KO	Blocked	Varvel and Lichtman (2002)
MWM	Cell-type specific CB1Rs	N/A	GABA-CB1-KO Glu-CB1-KO GFAP-CB1-KO	No effect No effect Blocked	Han et al. (2012)
MWM	NDMARs	AP-5	N/A	Blocked	Han et al. (2012)
MWM	COX-2	NS-398	COX-2-KO−	Blocked	Chen et al. (2013)
MWM	A1R	Caffeine	N/A	Enhanced	Sousa et al. (2011)
RAM	5-HT	Clomipramine 5-MeODMT	N/A	Attenuated Attenuated	Egashira et al. (2002)
RAM	5-HT2	DOI	N/A	Attenuated	Egashira et al. (2002)
RAM	5-HT1A	8-OHDPAT	N/A	Reversed	Inui et al. (2004)
RAM	AchE	Physostigmine Tetrahydroaminoacridine	N/A	Attenuated Attenuated	Mishima, Egashira, Matsumoto, Iwasaki, and Fujiwara (2002)
Short-term memory
DNMS	CB1R	SR141617A	N/A	Blocked	Hampson and Deadwyler (2000)
DNMP	CB1R	SR141617A	N/A	Attenuated	Hampson and Deadwyler (2000) and Mallet and Beninger (1998a, b)
NOR	Mitochondria-specific CB1R	N/A	DN22-CB1	Blocked	Hebert-Chatelain et al. (2016)
NOR	PKC	NPC CHE	N/A	Blocked Blocked	Busquets-Garcia et al. (2018)
DNMP	A1R	Caffeine CPT	N/A	Enhanced Enhanced	Panlilio et al. (2012)
NOR	A2AR	SCH442416 KW-6002	N/A	Attenuated No effect	Aso et al. (2019)
NOR	5-HT2AR	MDL 100907	N/A	Reversed	Viñals et al. (2015)
NOR	5-HT2AR	N/A	5-HT2AR-K0	Attenuated	Viñals et al. (2015)
DNMS	ACHe	Rivastigmine	N/A	Blocked	Goonawardena, Robinson, Hampson, and Riedel (2010)

Abbreviations: RAM—radial arm maze, MWM—Morris water maze, DNMS—delayed nonmatch to sample, DNMP—delayed nonmatch to position, NOR—novel object recognition.

6.1. Spatial memory

CB₁Rs mediate the impairment in spatial memory after acute administration of Δ⁹-THC as demonstrated by work in which CB₁Rs are antagonized or genetically deleted (Lichtman & Martin, 1996; Varvel & Lichtman, 2002). Intracerebral microinjection studies have established that CB1Rs in the hippocampus and medial PFC (mPFC) underlie this effect (Egashira, Mishima, Iwasaki, & Fujiwara, 2002). Moreover, the synthetic cannabinoid, HU210, produced deficits in spatial memory that were present in the forebrain GABA-CB₁-KO and Glu-CB₁-KO mice, but absent in mice with a conditional deletion of CB₁Rs in astrocytes (Han et al., 2012). In this report, antagonism of glutamatergic NMDA receptors (NMDARs) blocked the effects of HU210 on spatial memory and long-term synaptic depression in the hippocampus (Han et al., 2012). These findings indicate that cannabinoid-induced changes in spatial memory rely on stimulation of CB₁Rs on astrocytes and subsequent release of glutamate and changes in NMDAR signaling that induce LTD. It’s unknown whether a similar pathway underlies the effects of Δ⁹-THC in the MWM. However, antagonism and genetic ablation of cyclooxygenase-2 (COX-2), an enzyme responsible for the conversion of arachidonic acid to prostanoids, blocked Δ⁹-THC impairments in spatial memory (Chen et al., 2013). These findings tie in with the astrocytic CB₁R mechanism as COX-2 enhances prostaglandin E₂ (PGE₂) release from astrocytes (Chen et al., 2013) and PGE₂ signaling can facilitate glutamatergic gliotransmission (Cali, Lopatar, Petrelli, Pucci, & Bezzi, 2014). Overall, astrocytic glutamate release appears to play a pivotal role in the expression of cannabinoid memory impairment.

A number of other pathways are involved in the effects of Δ⁹-THC on spatial memory. For instance, pretreatment with an adenosine A₁ receptor (A₁R) agonist (caffeine) significantly elevated the Δ⁹-THC-induced deficit in spatial memory (Sousa et al., 2011). A₁Rs were detected on glutamatergic terminals in the hippocampus and stimulation of A₁Rs disinhibited CB₁R suppression of hippocampal glutamatergic signaling (Sousa et al., 2011). Activation of the serotonergic system had the opposite effect on Δ⁹-THC spatial memory impairments, i.e. the deficit was reversed by 5-HT_1A and 5HT₂ receptor agonism (Egashira et al., 2002; Inui et al., 2004). Δ⁹-THC also attenuated acetylcholine release in the dorsal hippocampus, which was blocked by administration of a 5-HT_1AR agonist in agreement with prior work in which facilitation of acetylcholine release rescued Δ⁹-THC-induced memory deficits (Inui et al., 2004; Mishima, Egashira, Matsumoto, Iwasaki, & Fujiwara, 2002). These systems may work in concert with astrocyte mediated glutamate release to mediate Δ⁹-THC’s effects on spatial memory.

6.2. Short-term memory

Unsurprisingly, activation of CB₁Rs mediates the Δ⁹-THC-induced deficits in short-term memory as confirmed by CB₁R antagonism studies (Hampson & Deadwyler, 2000; Mallet & Beninger, 1998a, b). The hippocampus appears to be a locus of interest in this respect as in vivo recording from groups of hippocampal neurons during performance of a short-term memory task revealed attenuated firing strength following Δ⁹-THC administration (Hampson & Deadwyler, 2000). One group has demonstrated that mice lacking CB₁Rs in hippocampal mitochondria fail to demonstrate a disruption in short-term memory following WIN55,212-2 administration (Hebert-Chatelain et al., 2016). CB₁Rs on mitochondria inhibit soluble-adenylyl cyclase, which decreases protein kinase A phosphorylation and has downstream impacts on mitochondrial energetic activity (Hebert-Chatelain et al., 2016). However, recent work has shown that intra-hippocampal injections of a protein kinase C inhibitor blocked Δ⁹-THC’s effects on short-term memory (Busquets-Garcia et al., 2018). It’s not clear whether the differential activity at these two protein kinase families works in conjunction to produce cannabinoid memory deficits or if the cannabinoid agent selected drives activity at one or the other group of enzymes.

The adenosine system has also been implicated in Δ⁹-THC’s inhibitory effect on short-term memory. Administration of caffeine or a selective A₁R antagonist potentiated Δ⁹-THC-induced short-term memory impairments (Panlilio et al., 2012). On the other hand, A_2AR antagonism mediated the therapeutic effect of CBD on Δ⁹-THC short-term memory deficits (Aso et al., 2019). A₁ and A_2AR antagonists have recently been shown to differentially regulate synaptic plasticity in the hippocampus, which could underlie their opposing actions on cannabinoid memory changes (Reis et al., 2019). As described earlier, CB₁Rs form heterodimers with 5HT_2ARs and evidence of these structures has been detected in the hippocampus (Viñals et al., 2015). Both genetic deletion and pharmacological antagonism of 5HT_2ARs reversed Δ⁹-THC-induced changes in short-term memory (Viñals et al., 2015). Finally, cholinergic hypofunction has been hypothesized to play a role in the cognitive deficits produced by cannabinoids. Goonawardena, Robinson, Hampson, and Riedel (2010) found that an acetylcholinesterase inhibitor blocked WIN55,212-2-induced memory impairing properties and restored the normal firing patterns of hippocampal neurons. Whether these findings can be applied to Δ⁹-THC is unknown.

7. Conclusion

The majority of Δ⁹-THC’s pharmacological and behavioral effects are mediated by activation of CB₁Rs (Tables 2–4). However, the cell types and brain regions expressing these receptors varies depending on the behavior assessed. Despite the widespread expression of CB₁Rs on GABAergic neurons, glutamatergic CB₁Rs preferentially underlie many of Δ⁹-THC’s effects including hypothermia, hypolocomotion, aversion, and anxiety relief. New research has discovered functional roles for CB₁Rs expressed on astrocytes and mitochondria in cannabinoid-induced cognitive impairments. This represents an exciting and promising area for future work, as non-neuronal CB₁R expression has been mostly ignored. Similarly, Δ⁹-THC binds to a number of other targets such as CB₂R, GPR55, PPARs. The data from our lab has demonstrated that activation of these non-CB₁ receptors also modulate Δ⁹-THC-induced behavior (Hempel et al., 2021; Li et al., 2021; Spiller et al., 2019; Wang et al., 2020). A host of additional non-cannabinoid receptor systems have been implicated in Δ⁹-THC’s behavioral effects primarily serotonergic (5-HT_1ARs & 5-HT_2ARs), opioidergic (MORs & KORs), and adenosinergic (A₁Rs & A_2ARs) signaling. Activity at these receptors may stem from downstream activity following CB₁R or CB₂R stimulation. Given the increasing use of cannabis for both recreational and medicinal purposes, understanding the neurobiology of Δ⁹-THC’s CNS effects is of vital concern. Moreover, this research provides a basis for the design of pharmacotherapeutics for substance use disorders including cannabis use disorder as well as non-psychoactive alternatives for medical marijuana users.

Abbreviations

Δ⁹-THC

Δ⁹-tetrahydrocannabinol

AEA

Anandamide

2-AG

2-arachidonoyl glycerol

ChR2

channelrhodopsin 2

GABA

gamma-aminobutyric acid

GPR55

G protein-coupled receptor 55

GPCR

G protein-coupled receptor

ISH

in situ hybridization

NAc

nucleus accumbens

oICSS

optical intracranial self-stimulation

PPARα

peroxisome proliferator-activated nuclear receptor alpha

PPARγ

peroxisome proliferator-activated nuclear receptor gamma

TRPV1

transient receptor potential vanilloid 1 channel

VTA

ventral tegmental area