Effects of cannabinoids on ligand-gated ion channels

Phytocannabinoids such as Δ9-tetrahydrocannabinol and cannabidiol, endocannabinoids such as N-arachidonoylethanolamine (anandamide) and 2-arachidonoylglycerol, and synthetic cannabinoids such as CP47,497 and JWH-018 constitute major groups of structurally diverse cannabinoids. Along with these cannabinoids, CB1 and CB2 cannabinoid receptors and enzymes involved in synthesis and degradation of endocannabinoids comprise the major components of the cannabinoid system. Although, cannabinoid receptors are known to be involved in anti-convulsant, anti-nociceptive, anti-psychotic, anti-emetic, and anti-oxidant effects of cannabinoids, in recent years, an increasing number of studies suggest that, at pharmacologically relevant concentrations, these compounds interact with several molecular targets including G-protein coupled receptors, ion channels, and enzymes in a cannabinoid-receptor independent manner. In this report, the direct actions of endo-, phyto-, and synthetic cannabinoids on the functional properties of ligand-gated ion channels and the plausible mechanisms mediating these effects were reviewed and discussed.


Introduction
Cannabis sativa plant (marijuana) consists of more than 500 chemical compounds, and about 120 of them are structurally related phytocannabinoids which are terpenophenolic compounds chemically related to the terpenes with their ring structure derived from a geranyl pyrophosphate (ElSohly et al., 2017). Among these phytocannabinoids, Δ 9tetrahydrocannabinol (THC) and cannabidiol (CBD) are major psychoactive and nonpsychoactive components of marijuana, respectively (Izzo et al., 2009).
Following structural characterization of series of phytocannabinoids in 1960s, CB1 and CB2 cannabinoid (CB) receptors, and endocannabinoids anandamide (N-arachidonoylethanolamine; AEA) and 2-arachidonoylglycerol (2-AG) were identified and shown to activate CB1 and CB2 with high affinity and efficacy (Devane et al., 1992;Mechoulam et al., 1995;Cristino et al., 2020). Subsequently, enzymes involved in synthesis and inactivation of endocannabinoids were identified. Synthesis of AEA and OPEN ACCESS EDITED BY other N-acylethanolamines is catalyzed by N-acylphosphatidylethanolamine (NAPE)-specific phospholipase D-like hydrolase (NAPE-PLD), and biosynthesis of 2-AG and other monoacylglycerols are catalyzed by Diacylglycerol lipase α (DAGLα) and DAGLβ. On the other hand, while anandamide and other N-acylethanolamines are hydrolyzed by fatty acid amide hydrolase (FAAH), the hydrolysis of 2-AG and other monoacylglcerols catalyzed by monoacylglycerol lipase (MAGL) (Fowler et al., 2017;Cristino et al., 2020). This system of endogenous ligands, receptors and metabolic enzymes became known as the endocannabinoid system. In addition to phyto-and endocannabinoids, intensive structureactivity studies to synthetize novel high-affinity ligands for cannabinoid receptors have resulted in emergence of several compounds such as WIN55,212-2, HU210, and CP55,940 with significantly diverse chemical structures that constitute third group of cannabinoids collectively coined synthetic cannabinoids (Huffman and Padgett, 2005;Wiley et al., 2014;Le Boisselier et al., 2017;Howlett et al., 2021).
In recent years, lack of psychoactive effects of some phytocannabinoids such as CBD, increased access to cannabis products due to decriminalization of medical use of cannabis in several western countries, and new approvals for clinical use of cannabinoid-based compounds have revitalized the interest in pharmacological effects of phytocannabinoids, specifically the CBD. In clinical practice, Nabilone (Cesamet), a synthetic cannabinoid, and Dronabinol (Marinol), synthetic (-) enantiomer of THC, have been used for treatment of anorexia and chemotherapy-induced nausea and vomiting in the United States, Canada, and United Kingdom (Wright and Guy, 2014;Pagano et al., 2022). Nabiximols (Sativex), an oral spray containing CBD and THC in a 1: 1 ratio, has been approved in several countries including United Kingdom, European Union, and Canada for the treatment of multiple sclerosisassociated spasticity (Ortiz et al., 2022;Pagano et al., 2022). CBD has been recently approved in the United States and the European Union as an add-on antiepileptic drug (Epidiolex) for the treatment of patients affected by refractory epilepsy such as Dravet syndrome and Lennox-Gastaut syndrome (Ortiz et al., 2022;Pagano et al., 2022) and caused resurgence of interest in the pharmacology of cannabinoids in general and phytocannabinoids in particular.
Since CBD does not activate CB1 and CB2 receptors, some of the attention in cannabinoid research has recently been focused on CB1 and CB2-independent cellular and molecular targets for CBD and other phytocannabinoids such as cannabigerol and cannabidivarin (Ghovanloo et al., 2022;Isaev et al., 2022;Mirlohi et al., 2022). In fact, in earlier studies, cannabinoid-receptor independent effects have been described for both endogenous and synthetic cannabinoids as well (Johnson et al., 1993;Fan, 1995;Oz et al., 2000). AEA stimulates GTPγS binding in brain membranes isolated from mice lacking CB1 receptors, and this effect is not altered by CB1 and CB2 antagonists (Di Marzo et al., 2000). THC and AEA-induced analgesic effects in hot-plate (for AEA) and tail flick tests (for THC) remain intact in CB1 knockout (Zimmer et al., 1999;Di Marzo et al., 2000) or both CB1 and CB2 knock-out mice (Rácz et al., 2008). Similarly, AEA still induces catalepsy and analgesia and decreases spontaneous activity in CB1-deficient mice (Di Marzo et al., 1994;Baskfield et al., 2004) and exert CB1 antagonist-resistant behavioral effects in the open-field test (Järbe et al., 2003a;Järbe et al., 2003b).
In neuronal systems, fast signal transmission between pre and post synaptic structures is carried out by the conformational changes induced by the signaling molecules, neurotransmitters in integral membrane proteins coined ligand-gated ion channels (LGICs). The LGICs are divided into three super families: the Cys-loop superfamily, the glutamate receptors [NMDA (N-methyl-D-aspartate), AMPA (α-amino-3-hydroxy-5methyl-4-isoxazolepropionic acid) and kainate], and the ATPgated channels (Alexander et al., 2017). The Cys-loop superfamily comprises both cationic receptors such as nicotinic acetylcholine (nACh) and 5-HT 3 receptors and anionic receptors such as GABA (γ -aminobutyric acid) type A (GABA A ), and glycine receptors. Typical LGIC is composed of multiple subunits in homo or hetero pentameric structure with a pore forming α subunit that allows the regulated flow of ions across the plasma membrane in response to binding of endogenous or exogenous molecules. Structurally, an αsubunit of LGIC contains a large extracellular N-terminus, four transmembrane (TM) domains, a short extracellular C-terminus, and a large cytoplasmic domain between TM3 and TM4 (Plested, 2016). While this review focuses on the direct cannabinoid receptor-independent effects of cannabinoids on the LGICs, the effects of these compounds on other ion channels, various G-protein coupled receptors, enzymes, and neurotransmitter transporters have been reviewed in earlier reports (Oz, 2006;Pertwee et al., 2010;Soderstrom et al., 2017;Cifelli et al., 2020;Senn et al., 2020;Vitale et al., 2021).

Effects of cannabinoids on nicotinic acetylcholine receptors
Nicotinic acetylcholine receptors (nAChRs) are cationic channels belonging to the Cys-loop superfamily of ligandgated ion channels and their opening is controlled by the endogenous neurotransmitter ACh or exogenous ligands such as nicotine (Albuquerque et al., 2009;Dineley et al., 2015). They are homo or hetero-pentameric complexes comprised of α1-10, β1-6, δ, and γ subunit combinations and mediate fast cholinergic neurotransmission in both central and peripheral nervous systems and are expressed in non-neuronal cells as well (Albuquerque et al., 2009;Zoli et al., 2015). The homomeric α7-nAChR subtype plays an important role in synaptic plasticity and various disease pathologies including pain, neuroinflammation, and neurodegenerative diseases such as Alzheimer and Parkinson diseases (Oz et al., 2013;Lorke et al., 2016;Oz et al., 2016;Borroni and Barrantes, 2021;Mizrachi et al., 2021;Papke and Horenstein, 2021).
In vivo studies, WIN55,212-2, non-selective cannabinoid receptor agonist, has been shown to impair memory-related effects of nicotine in male rats (Robinson et al., 2010) and mice (Biala and Kruk, 2008). In Xenopus oocytes expressing α7-nACh receptors, AEA and 2-AG noncompetitively inhibited α7-nACh receptor-mediated currents with IC 50 values of 229 and 168 nM, respectively (Oz et al., 2003;Oz et al., 2004b). In line with these findings, radioligand binding studies indicated that AEA, at concentration range of 30-300 μM, does not alter specific binding of [ 3 H]nicotine in human frontal cortex (Lagalwar et al., 1999) and rat thalamic membranes (Butt et al., 2008), further suggesting a noncompetitive nature of AEA effect on nACh receptors. Inhibition of α7-nACh receptor by AEA was not reversed by either SR 141716A (Rimonabant), specific CB1 antagonist, or SR 144528, CB2 antagonist and was not sensitive to pertussis toxin treatment (Oz et al., 2003). In line with earlier findings, arachidonic acid (AA), a fatty acid moiety of AEA (Vijayaraghavan et al., 1995), but not ethanolamine or glycerol, inhibited the function of α7-nACh receptors suggesting that it was the intact endocannabinoid and not the metabolite AA that altered the function of α7-nACh receptor. In an in vivo study, Baranowska et al., demonstrated that, in the presence of AM2521, a CB1 antagonist, methanandamide, a metabolically stable analog of AEA inhibited nicotine-induced tachycardiac responses mediated by the activation of α7-nACh receptors on the cardiac postganglionic sympathetic neurons of urethane anesthetized rats (Baranowska et al., 2008). In another in vivo study, increased endocannabinoid levels modulate the somatic signs of nicotine withdrawal by a mechanism that does not involve CB1 receptors (Merritt et al., 2008).
Effects of synthetic cannabinoids on the functions of nACh receptors have been tested in a few studies. In Xenopus oocytes, WIN55,212-2 was ineffective up to concentrations of 10 μM, but CP55940 inhibited α7-nACh receptor with an IC 50 value of 2.7 µM (Oz et al., 2004b). On the other hand, in cultured rat trigeminal ganglion neurons, patch-clamp studies indicated that native nicotinic receptor was inhibited by WIN55,212-2 with an IC 50 value about 3 µM in a cannabinoid receptor independent manner (Lu et al., 2011).
Among the phytocannabinoids tested, THC and cannabinol up to 10 µM were ineffective (10-20% inhibition) on α7-nACh receptors expressed in Xenopus oocytes (Oz et al., 2004b;Mahgoub et al., 2013). Similarly, THC (1 µM) did not affect the amplitudes of currents through α4β2-nACh receptors (Spivak et al., 2007). CBD, on the other hand, at relatively high concentration modulates the function of nACh receptors. In earlier studies, CBD was reported to decrease the amplitudes of miniature end-plate potentials in frog neuromuscular junction suggesting an effect on the postsynaptic nACh receptors (Turkanis and Karler, 1986). Later studies indicated that CBD inhibits human α7-nACh receptors with an IC 50 of 11.3 µM (Mahgoub et al., 2013;Oz et al., 2014) and suppresses cholineinduced inward currents in rat hippocampal interneurons. CBD has also been shown to inhibit nicotine-induced [ 3 H]norepinephrine release in rat hippocampal slices (Mahgoub et al., 2013), and reduce withdrawal symptoms in nicotinedependent rats (Smith et al., 2021) further suggesting an interaction between CBD and nicotinic receptors. Notably, Inhibitory effect of CBD was time-dependent with 50% inhibition occurring in 2.1 min sustained CBD application, voltage-independent, and non-competitive, i.e., while the EC 50 of ACh remained unaltered, maximal ACh responses were significantly decreased (Mahgoub et al., 2013). In line with these findings, specific binding of [ 125 I] α-bungarotoxin binding was not affected by CBD.
In addition to α7-nicotinic receptor, the α4β2 is the major nACh receptor subtype in the CNS and has been implicated in mediating both the positive-reinforcing and cognitive effects of nicotine (Tapper et al., 2004). The presence of the α4β2-nACh receptor subunit appears to be necessary and sufficient for the development of nicotine-induced tolerance and sensitization in vivo (Tapper et al., 2004). Using the wholecell patch-clamp technique, Spivak et al., showed that AEA, in the concentration range of 200 nM to 2 μM, reduced the maximal amplitudes and increased the desensitization of acetylcholine-induced currents mediated by human α4β2 nACh receptors expressed in SH-EP1 cells (Spivak et al., 2007). The effects of AEA could be neither reversed by the SR141716A (1 μM) nor replicated by the THC (1 μM). Interestingly, AEA exerted inhibitory effect when applied extracellularly but not during intracellular dialysis. Kinetic analysis of the effect of AEA on α4β2 nACh receptor-mediated currents demonstrated that the first forward rate constant leading to desensitization increased nearly 30-fold as a linear function of AEA concentration and the energy levels of the activated state were raised by AEA (Spivak et al., 2007). Another study on α4β2 nACh receptors utilized rat thalamic synaptosomes where the ACh-induced 86 Rb + effluxes carried mainly through native α4β2 nACh receptors were reversibly inhibited by AEA with an IC 50 of 0.9 µM in a noncompetitive manner (Butt et al., 2008). Inhibitory effect of AEA was not altered by pretreatments with the CB1 receptor antagonist SR141716A (1 μM), the CB2 receptor antagonist SR144528 (1 μM), and PTX. Although α7 and α4β2 receptor subtypes are affected by AEA, the function of muscle type nicotinic receptor is not altered by AEA (Oz, 2006). The results of key studies investigating the effects of cannabinoids on nACh receptors are summarized in Table 1.

Effects of cannabinoids on serotonin type 3 (5-HT 3 ) receptors.
There are seven classes of 5-HT receptors (5-HT1-5-HT7). 5-HT 3 receptor is the only ionotropic cation-selective ion channel which belongs to the superfamily of Cys-loop ligandgated ion channels (Barnes et al., 2009;Lummis, 2012), remaining 5-HT receptor classes are G-protein-coupled receptors. To date, there are five 5-HT 3 receptor genes termed A-E. Only 5-HT 3A subunit can assemble as a functional homopentameric channel and 5-HT 3A is an obligate participant in all other 5-HT 3 receptor complexes (Lummis, 2012). Activation of 5-HT 3 receptors depolarizes neurons and mediates fast, excitatory synaptic transmission in the central and peripheral nervous systems. The highest number of 5-HT 3 receptor binding sites occurs in the area postrema and solitary tract nucleus (Morales and Wang, 2002). In addition, large numbers of 5-HT 3 receptors are found in the gastrointestinal system where activation of 5-HT 3 receptors are shown to alter gastrointestinal motility and to regulate the vomiting reflex (Sanger and Andrews, 2006), and they also appear to play a role in irritable bowel syndrome, visceral pain and inflammation (Faerber et al., 2007;Thompson and Lummis, 2007;Machu, 2011).
Interestingly, potencies of AEA actions on 5-HT 3 receptors differ significantly between cell lines such as HEK-293 (Barann et al., 2002) and Xenopus oocytes (Oz et al., 2002a). The mechanisms underlying these seemingly discrepant results were studied in a recent study (Xiong et al., 2008). It was found that differences in the potencies of AEA inhibition of 5-HT 3 receptors between Xenopus laevis oocytes and HEK-  Frontiers in Physiology frontiersin.org 04 293 cells were mainly due to different levels of steady-state receptor density at the cell surface. The magnitude of AEA inhibition was inversely correlated with the expression level of receptor protein; i.e., increasing surface receptor expression decreased the magnitude of AEA inhibition. In line with these findings, pretreatment with actinomycin D, an inhibitor of transcription process, decreased the amplitude of current activated by maximal 5-HT concentrations and increased the magnitude of AEA inhibition. AEA accelerated 5-HT 3 receptor desensitization time in a concentration-dependent manner without significantly changing receptor activation and deactivation time. The desensitization time was correlated with the AEA-induced inhibition and mean 5-HT current density. Applications of 5-hydroxyindole and nocodazole, a microtubule disruptor, significantly slowed 5-HT 3 receptor desensitization and reduced the magnitude of AEA inhibition. Collectively, these findings suggested that 5-HT 3 receptor density at the steady state regulates receptor desensitization kinetics and the potency of AEA-induced inhibiting effect on the receptors.
Thus, it appears that the differences in IC 50 values for AEA inhibition of 5-HT 3 receptor in Xenopus oocytes (3.7 μM; (Oz et al., 2002a); and HEK-293 cells (130 nM; (Barann et al., 2002)) are mainly due to the different receptor expression levels among various cell types.
In vivo studies, WIN55,212-2 and CP55,940 inhibited the responses mediated by 5-HT 3 receptors located on the terminals of cardiopulmonary afferent C-fibers in anesthetized and SR141716A pretreated rats (Godlewski et al., 2003). Similarly, CBD inhibited the reflex bradycardia induced by the 5-HT 3 receptor agonist phenylbiguanide in spontaneously hypertensive rats (Kossakowski et al., 2019). In another in vivo study, the inhibitory effect of WIN55,212-2 on cocaineinduced locomotor hyperactivity was reported to be mediated by the inhibition of 5-HT 3 receptors (Przegaliński et al., 2005). In a recent study, mice deficient in CB1 and CB2 receptors were treated with THC and AEA, in the presence of the 5-HT 3 antagonist ondansetron (Rácz et al., 2008). AEA induced analgesia, but not catalepsy, is completely blocked by ondansetron, suggesting that 5-HT 3 receptors are involved in cannabinoid-induced analgesia in a manner independent of known cannabinoid receptors (Rácz et al., 2008).
Both THC and CBD were shown to directly interact with 5-HT 3 receptors expressed in HEK-293 cells (Barann et al., 2002;Xiong et al., 2011b), Xenopus oocytes, and NGNs (Yang et al., 2010a;Yang et al., 2010b;Xiong et al., 2011b). Depending on the expression levels of the receptor, THC and CBD inhibited human 5-HT 3A receptors with respective IC 50 values of 1.2 µM and 0.6 µM in Xenopus oocytes (Yang et al., 2010a;Yang et al., 2010b) and 119 nM and 329 nM in HEK-293 cells (Xiong et al., 2011b). Inhibition by these cannabinoids was voltage-independent, noncompetitive, and increased with sustained application (preincubation) time. In line with non-competitive nature of their effect, THC and CBD did not alter the specific [ 3 H] GR65630 binding on 5-HT 3A receptors (Yang et al., 2010a;Yang et al., 2010b). Importantly, slowing the receptor desensitization by pharmacological agents such as 5hydroxyindole and nocodazole or mutations such as R427L markedly decreased the extent of 5-HT 3 receptor inhibition by CBD (Xiong et al., 2011b). Similarly, increasing the expression level and density of 5-HT 3 receptors on the cell membrane caused a noticeable decrease of receptor desensitization and the potency of THC on 5-HT 3 receptor (Yang et al., 2010b). Finally, 5-HT3 and CB1 receptors colocalize in hippocampal and dentate gyrus interneurons (Morales and Bäckman, 2002) and in other brain regions as well (Morales et al., 2004). Therefore, it is likely that the actions of cannabinoids including AEA and THC within the same cell can be mediated by simultaneous dual actions of cannabinoids on cannabinoid and 5-HT 3 receptors. The results of key studies investigating the effects of cannabinoids on 5-HT 3 receptors are summarized in Table 2.

Effects of cannabinoids on ionotropic glutamate receptors
Glutamate mediates most excitatory neurotransmission in the mammalian central nervous system through activation of metabotropic, G protein-coupled glutamate receptors, and ionotropic glutamate receptors, which are cation-selective ligand-gated ion channels. Ionotropic glutamate receptors are divided into different functional classes, namely α-amino-3hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, kainate receptors, N-methyl-D-aspartate (NMDA) receptors, and GluD (Δ) receptors (Hansen et al., 2021). In vivo studies, synthetic cannabinoid HU211 protected against tremorogenic, convulsive, and lethal effects of NMDA in mice and acted as a functional antagonist in radioligand binding studies (Feigenbaum et al., 1989). Similarly, NMDA-induced toxicity was reversed by THC, but not WIN55,212-2 in SR141716A insensitive manner in AF5 cells (Chen et al., 2005). AEA (1-10 µM), but not THC, potentiated NR1/NR2A NMDA receptor-mediated currents in Xenopus oocytes and increased NMDA-induced intracellular Ca 2+ transients in the presence of SR141716A in rat cortical, cerebellar, and hippocampal slices (Hampson et al., 1998). 2-AG (5 µM) enhanced NMDA-evoked currents in a CB1 and TRPV1 receptor-independent manner through activation of PKC/Src signaling pathway (Yang et al., 2014). In studies of blood pressure control in rats, it was found that AEA produces a pressor effect in the presence of SR141716A and this increase in pressure was partially reduced by the NMDA receptor antagonist MK-801, suggesting a possible direct interaction between NMDA receptors and AEA with relevance to central control of blood pressure (Malinowska et al., 2010).
CBD inhibited AMPA-glutamate receptor-mediated evoked excitatory postsynaptic currents and suppressed the frequency Frontiers in Physiology frontiersin.org 05 and amplitudes of miniature excitatory postsynaptic currents in mice hippocampal neurons (Yu et al., 2020). Furthermore, in HEK-293 cells, CBD was shown to inhibit recombinant GluA1 currents with IC 50 of 22.5 µM and facilitate the deactivations of GluA1 and GluA2 receptors and slow the recovery of GluA1 receptor from desensitization (Yu et al., 2020). In line with these findings, CBD alleviate behavioral hyperactivity and hippocampal c-Fos expression in Gria1−/− mice which is a model sensitive to drugs that downregulate glutamatergic transmission such AMPA antagonists (Aitta-Aho et al., 2019). AEA, at high concentrations, inhibits homomeric GluR1 and GluR3 subunit-mediated currents with IC 50 of 161 µM and 143 μM, respectively (Akinshola et al., 1999). AEA inhibited heteromeric GluR1/3 and GluR2/3 receptor subunits with similar IC 50 value of 148 µM and 241 μM, respectively.

Effects of cannabinoids on glycine receptors
Glycine receptors are inhibitory, anion-selective Cys-loop ligand-gated ion channels that mediate fast inhibitory neurotransmission in the spinal cord, brainstem and retina (Lynch, 2009). There are four known glycine receptor α subunits (α1-α4) and a single β subunit in vertebrates. Functional receptors assemble as homo-pentamers of α subunits, or as hetero-pentamers of α and β subunits (Burgos et al., 2016). Malfunctions of glycine receptors have been associated with a range of neurological disorders including hyperekplexia, temporal lobe epilepsy, autism, breathing disorders, and chronic inflammatory pain (Lynch et al., 2017).

Cannabinoid tested Effect and conclusion References
Frontiers in Physiology frontiersin.org 2011). In another study, AEA (1-30 µM) substantially increased the amplitude of glycine-activated currents in cultured rat spinal neurons (Xiong et al., 2012b). The effect reached a steady-state within 5 min continuous AEA application and markedly diminished by increasing glycine concentrations. In HEK-293 cells, AEA potentiation of homomeric α1 and α3 receptors was significantly higher than α2 glycine receptors. While S296A and S307A mutations in transmembrane three domains of α1 and α3 receptors markedly decreased the potentiation by sustained (continuous), but not simultaneous co-application of AEA, A303S mutation increased the potentiation of α2 glycine receptors by AEA. The S296A mutation also decreased the extent of α1 receptor potentiation by other cannabinoids such as CBD, THC, and HU-210 (Xiong et al., 2012b). Importantly, the authors of this study also showed in control studies that these mutations have no significant effect on the pharmacological properties of glycine receptors. Furthermore, removal of hydroxy or oxygen groups in AEA and THC significantly reduced the potentiating effects of these compounds on α1 glycine receptor suggesting that these groups are important for their effects on glycine receptor. In outside-out patches in CHO cells expressing human α1 glycine receptor, continuous application of 2-AG (1 µM) inhibited the peak amplitudes, decreased the rise time, accelerated the desensitization rate, and slowed down the deactivation of the receptor activated by glycine at the level of EC 50 (40-70 µM). Notably, inhibitory effect of AEA was significantly less at low (20 µM) and high (1 mM) concentrations of glycine (Lozovaya et al., 2011). In further studies mimicking synaptic stimulation in outside-out patches from CHO cells by applying high concentration of glycine (1 mM) for 2 ms duration at 10 Hz frequency, application of 2-AG markedly reduced the amplitudes of glycine-induced currents. Similarly, in hypoglossal motoneurons of CB1 knock-out mice, 2-AG abolished the facilitation of synaptic glycinergic currents induced by repetitive (10-20 Hz) stimulations indicating that endocannabinoids can modulate synaptic transmission under physiological conditions in a cannabinoid-receptor-independent manner (Lozovaya et al., 2011). CBD and THC have been shown to potentiate glycine receptor-mediated currents in a cannabinoid receptorindependent manner in Xenopus oocytes expressing homomeric α1 and heteromeric α1β glycine receptors, and in acutely isolated CNS neurons (Hejazi et al., 2006;Xiong et al., 2011a;Wells et al., 2015). CBD was reported to act as positive allosteric modulator of α1 and α1β1 glycine receptors expressed in HEK-293 cells with EC 50 values of 12.3 µM and 18.1 µM, respectively (Ahrens et al., 2009a). Notably, at higher concentration range, CBD directly acts as agonist of α1 and α1β1 glycine receptors with respective EC 50 values of 132.4 µM and 144.3 µM. The mutation of the α1 subunit transmembrane two serine 267) residue to isoleucine abolished both co-activation and direct activation of glycine receptor by CBD (Foadi et al., 2010). However, THC-induced potentiation was not affected when the S267Q mutant α1 glycine receptors were expressed in Xenopus oocytes (Hejazi et al., 2006). In subsequent studies, it was shown that continuous application, but not co-application, of THC (1 µM) for about 5 min caused a gradually developing marked potentiation (up to 1,200%) of glycine activated currents in HEK-293 cells expressing α1, α3, and α1β1, but not α2, subunits of glycine receptor (Xiong et al., 2011a). Notably, S296 residue of the α1 and the S307 of the α3 subunits have been shown to be critical for the potentiation of glycine receptors by continuous, but not simultaneous, THC application (Xiong et al., 2011a;Xiong et al., 2012b). Conversely, substitution of the corresponding residue, Ala303 of the α2 subunit with a serine converted the α2 subunit from a low to high THC sensitivity. Similar to THC, the CBD, dehydroxyl-CBD, and dehydroxyl-THC have also been shown to potentiate glycine-activated currents of α1 and α3 subunits (Xiong et al., 2011a;Xiong et al., 2012a;Xiong et al., 2012b). Further studies indicated that CBD and dehydroxyl-CBD suppress persistent inflammatory and neuropathic pain by potentiating α3 glycine receptor-mediated currents in HEK-293 cells and cultured spinal neurons and the CBD suppression of neuropathic pain was reversed in α3-knock-out mouse (Xiong et al., 2012a). In subsequent studies, dehydroxyl-CBD was reported to suppress inflammatory pain and potentiate α1-glycine receptors by interacting with S296 amino acid residue. Accordingly, analgesic effect of dehydroxyl-CBD was reversed in the transgenic mice carrying S296A mutation, which also blocks DH-CBD-induced potentiation of glycine-activated currents in neurons of the spinal cord dorsal horn in mutant mouse (Lu et al., 2018). In another study, dehydroxyl-CBD rescued functional deficiency of glycine receptor and exaggerated acoustic and tactile startle responses in mice bearing point mutations in α1 glycine receptors that are responsible for a hereditary startle-hyperekplexia disease (Xiong et al., 2014). Recently, cannabinoids such as THC and CBD were reported to rescue cocaine-induced seizures by restoring brain glycine receptor dysfunction in a cannabinoid-receptor independent manner (Zou et al., 2020b). The results of key studies investigating the effects of cannabinoids on glycine receptors are summarized in Table 3.

Effects of cannabinoids on GABA A receptors
γ-Aminobutyric acid (GABA), the major inhibitory neurotransmitter in the brain, exerts its action via ionotropic GABA A and metabotropic GABA B receptors. GABA A receptors, like glycine receptors, are anion-selective Cys-loop ligand-gated ion channels activated by GABA and the selective agonist muscimol, blocked by bicuculline and picrotoxin, and Frontiers in Physiology frontiersin.org modulated by benzodiazepines, barbiturates, and some other classes of depressants (Olsen and Sieghart, 2008;Sigel and Steinmann, 2012). Consequently, drugs modulating the activity of GABA A receptors play a critical role in the treatment of neuropsychiatric disorders such as epilepsy, anxiety, and insomnia, as well as in anesthesia (Olsen and Sieghart, 2008;Olsen, 2018;Ghit et al., 2021). A total of 19 GABA A receptor subunit genes have been identified in humans that code for six α1-6, three β1-3, three γ1-3, three ρ1-3, and one each of the δ, ε, π, and θ (Sigel and Steinmann, 2012). They are usually constructed with two copies of an α subunit, two copies of a β subunit, and one copy of either a γ subunit, or another such as Δ, to form several combinations of hetero-pentameric GABA A receptor subtypes (Sigel and Steinmann, 2012;Olsen, 2018).
In early radioligand binding studies, synthetic cannabinoids such as levonantradol, nabilone, CP-47,497 (-)-CP-55,940 and (-)-CP-55,244 enhanced the specific binding of [ 3 H]-flunitrazepam to mouse brain in vivo, potentiated the anticonvulsant effects of diazepam against pentylenetetrazol, and elicited analgesic effects that correlated with the potency of their facilitatory effects on [ 3 H]-flunitrazepam binding (Koe and Weissman, 1981;Koe et al., 1985) suggesting that cannabinoids can interact with GABA A receptors. Furthermore, benzodiazepines such as diazepam partially substitutes for THC in drug discrimination studies (Mokler et al., 1986) and this effect was not antagonized by CB1 receptor antagonist SR141716 (Wiley and Martin, 1999). More recently, effects of cannabinoids on the functional properties of GABA A receptors has been shown in heterologous expression systems and native neurons. In Xenopus oocytes expressing α1-6β2γ2 subunits of GABA A receptors, 2-AG potentiated currents activated low GABA concentrations (EC 5 = 0.1-10 µM) with EC 50 values ranging from 1.5 µM to 15.7 µM and significantly shifted GABA concentration-response curves to the left (Sigel et al., 2011;   Frontiers in Physiology frontiersin.org Bakas et al., 2017). 2-AG also potentiated extrasynaptic δ-subunit containing α4β2δ GABA A receptors (Bakas et al., 2017). Potentiation of GABA-evoked currents by 2-AG was significantly greater at GABA A -receptors containing a β2 or β3 subunit and markedly reduced by α2β2(V436T)γ2L and α2β2(VF439L)γ2L mutations in transmembrane domain 4 (Sigel et al., 2011). Subsequent studies have identified four amino acid residues in transmembrane domain 4 of the β2 subunit, β2W428, β2S429, β2F432, and β2V443 and in transmembrane domain 3, β2V302, in addition to the residues previously described, β2M294 and β2L301 (Baur et al., 2013b). In addition to cannabinoid receptor agonists, antagonists of these receptors have also been shown to directly modulate GABA A receptors. SR141716 (rimonabant) and AM251 allosterically potentiated low (0.5 µM) GABA activated currents to maximal potentiations of 3,381% and 881% of controls with EC 50 values of 7.3 µM and 0.4 µM, respectively in Xenopus oocytes expressing α1β2γ2 GABA A receptor (Baur et al., 2012). Another study on dissociated basolateral amygdala neurons, reported that AM251 (1 µM) caused 30% reduction on the peak amplitudes of GABA (5 µM)-activated currents (Zhu and Lovinger, 2005).
CBD has been shown to act as a positive allosteric modulator, with EC 50 values ranging from 0.9 µM to 16.1 µM and magnitudes of potentiation in the range of 72%-332% at αβγ2 receptor combinations, with higher level of potentiation on α2 containing subtype combination (Bakas et al., 2017). The greatest levels of enhancements by CBD were reported as 332% on α2β2γ2L and 752% on α4β2δ subunit combinations and the classical benzodiazepine binding site located at α-γ2L interface does not seem to be involved in CBD interaction with GABA A receptors. Furthermore, CBD potentiation of α2β2γ2L was significantly decreased by α2β2(V436T)γ2L mutation on transmembrane domain 4. Notably, CBD potentiation of GABA A receptors was significantly decreased with increasing GABA concentrations, resulting in decreased GABA EC 50 values with no apparent change in the E max . In a recent study, CBD (2 µM) induced 30-50% potentiation of GABA (1-10 µM) activated currents in Xenopus oocytes expressing human α1β2γ2, α1β2, and α2β2γ2, α2β2 subunit combinations of GABA A receptors (Ruffolo et al., 2018). In another investigation, CBD (10 µM) caused 106% potentiation, with EC 50 of 2.4 µM, of GABA (15 µM)-activated currents in Xenopus oocytes expressing α1β2γ2 subunit combination of GABA A receptors (Anderson et al., 2019). In addition to CBD, THC (3 µM) has also been shown to potentiate (about 100%) α1β2γ2 GABA A receptors activated by low concentrations (EC 2 ) of GABA (Yao et al., 2020b). In another study, the rundown of GABA (500 µM) activated currents in Xenopus oocytes transplanted with hippocampal membranes from epileptic patients was significantly reversed by 2 h pretreatment with 50 nM cannabidivarin, a non-psychoactive homolog of CBD (Morano et al., 2016). In a recent study, cannabigerolic acid, the biosynthetic precursor to both THC and CBD, has been shown to increase GABA (15 µM) activated currents to 271% of controls with EC 50 of 910 nM in Xenopus oocytes expressing human α1β2γ2 GABA A receptors (Anderson et al., 2021). In addition, this study showed that cannabigerolic acid potentiated the anticonvulsant effects of clobazam against hyperthermiainduced and spontaneous seizures and displayed anticonvulsant effects in maximal electroshock model. Recently, dehydroxylcannabidiol, a synthetic nonpsychoactive cannabinoid, has been shown to restore the GABA-and glycine-activated currents in HEK-293 cells co-expressing a major GABA A receptor isoform (α1β2γ2) and α1 glycine receptor carrying a human hyperekplexia-associated mutation (R271Q), suggesting that cannabinoids can be potentially valuable candidate drugs to manage hyperekplexia (Zou et al., 2020a). Collectively, these results indicate that both CBD and THC act as positive allosteric modulators of glycine and GABA A receptors, especially at low agonist concentrations and subunit specific manner. The results of key studies investigating the effects of cannabinoids on GABA A receptors are summarized in Table 4.

Discussion
Cannabinoids are highly lipophilic compounds with a LogP (octanol-water partition coefficient) values ranging between 4 and 9. Thus, it is likely that these lipophilic molecules first dissolve into the lipid membrane and then diffuse into a nonannular lipid space to allosterically inhibit the ion channels. Consistent with this idea, the effect of cannabinoids on ion channels usually reaches a maximal level within several minutes (5-10 min) of sustained applications.
Notably, functional properties of ligand-gated ion channels have been shown to be affected by the activation of second messenger pathways (Siara et al., 1990;Hoffman et al., 1994;Zhang et al., 1995;Nishizaki and Sumikawa, 1998). However, agents modulating cAMP and protein kinase C pathways and chelation of intracellular Ca 2+ do not alter the effects of endocannabinoids on LGICs (Oz et al., 2002a;Oz et al., 2003). Beside cannabinoids, actions of several lipophilic modulators, such as capsaicin (Lundbaek et al., 2005;Alzaabi et al., 2019;Nebrisi et al., 2020), endocannabinoids (Oz et al., Frontiers in Physiology frontiersin.org 2004a; Spivak et al., 2007), general anesthetics (Zhang et al., 1997;Jackson et al., 2008), and steroids (Oz et al., 2002b) on various ion channels require 5-10 min continuous application times to reach their maxima, suggesting that their binding site(s) is/are located inside the lipid membrane and require a relatively slow equilibrium time to exert their effect. From this aspect, it appears that alone or the combination of two mechanisms can describe the lipophilic actions of cannabinoids (Oz, 2006;Oz et al., 2014;Ghovanloo and Ruben, 2021). First, cannabinoids, like other lipophilic molecules, partition into the lipid bilayer and alter the biophysical properties of the membrane by reducing membrane electrical resistance (Bach et al., 1976), increasing membrane fluidity (Hillard et al., 1985;Mavromoustakos et al., 2001;Dainese et al., 2012), changing membrane order (Bloom et al., 1997), increasing membrane stiffness Ghovanloo and Ruben, 2021), increasing membrane elasticity (Medeiros et al., 2017;James et al., 2022), and changing physicochemical and structural properties of bilayer membranes (Makriyannis et al., 1990;Yang et al., 1992;Ambrosi et al., 2005;Tiburu et al., 2007;Tian et al., 2011). Secondly, cannabinoids can bind directly to transmembrane domains of ion channels embedded in the cell membrane (Xiong et al., 2011a;Xiong et al., 2012b;Lu et al., 2018;. In fact, residue S296 in transmembrane domain of α1 glycine receptor has been reported to interact with cannabinoids and mediate their potentiating effect on these receptors (Xiong et al., 2011a;Xiong et al., 2012b;Lu et al., 2018). In favor of specific binding site on ion channels, WIN55,212-3, cannabinoid receptor-inactive enantiomer of WIN55,212-2, did not affect the 5-HT 3 receptor-mediated currents (Barann et al., 2002), confirming earlier results with CP56667 on nodose ganglion neurons (Fan, 1995) and indicating that although their effects are not mediated by the activation of cannabinoid receptors, these synthetic cannabinoids inhibit 5-HT 3 receptors in an enantiomer-specific manner possibly through a hydrophobic bindings site within the bilayer membrane. As a result of these mechanisms, it is likely that cannabinoids affect the energy requirements for gating-related conformational changes and allosterically modulate the functional properties of ion channels (Spivak et al., 2007). Interaction between allosteric modulators and cannabinoids has also been investigated in a few studies. While, general anesthetics (Jackson et al., 2008), ethanol (Oz et al., 2005), and neurosteroids (Sigel et al., 2011) exert additive effects with endocannabinoids, fatty acid amides such as docosatetraenylethanolamide (Baur et al., 2013a;Baur et al., 2013b) antagonize the positive allosteric effects of endocannabinoids suggesting distinct binding sites for these modulators. In conclusion, both membrane disturbing effects and a hydrophobic bindings site(s) within the transmembrane regions of the LGIC can mediate the modulatory actions of cannabinoids on these channels (Figure 1). Cannabinoids also appear to interact with other membrane lipids such as cholesterol to exert their effects (Martin et al., (Bruggemann and Melchior, 1983). In another study, cholesterol was shown to stimulate both the insertion of AEA into bilayer membranes, and its transport across these membranes (Di Pasquale et al., 2009). Interestingly, cholesterol seems to provide structural support to CBD binding site on the glycine receptors. Removal of membrane cholesterol by cyclodextrin (Yao et al., 2020b) or anti-cholesterol drug such as simvastatin (Yao et al., 2020a) markedly reversed THC or dehydroxyl-CBD potentiation of α1 and α3 glycine receptors by interacting with the S296 residue suggesting that cholesterol directly interact with these cannabinoids on this binding pocket. Another recent study indicates that membrane orientation of CBD and its effect on water permeability were significantly altered in the presence of cholesterol suggesting an interaction between CBD and cholesterol in lipid membranes (Perez et al., 2022). Among more than 120 phytocannabinoids found in cannabis, to date, beside THC, some other cannabinoids such as cannabinol, (−)-trans-Δ 8 -tetrahydrocannabinol, (−)-trans-Δ 9tetrahydrocannabivarin, cannabidivarin, cannabigerol, cannabichromene, and the sesquiterpene (E)-β-caryophyllene have also been shown to bind and fully or partially activate cannabinoid receptors (Husni et al., 2014;Pertwee and Cascio, 2014;Zagzoog et al., 2020). Notably, majority of phytocannabinoids that do not interact with cannabinoids receptors and several phenolic terpenes found in the cannabis plant also act as allosteric modulators of various ion channels including several members of LGICs (Nurulain et al., 2015;Oz et al., 2015;Hoffmann et al., 2016;Lozon et al., 2016;Al Kury et al., 2018), suggesting that these compounds can also potentially contribute to allosteric effects of cannabis through cannabinoid receptor-independent mechanisms. Importantly, in addition to LGICs, cannabinoids act through cannabinoidindependent mechanisms to modulate the functions of TRP channels (Muller et al., 2018;Storozhuk and Zholos, 2018), PPARs Lago-Fernandez et al., 2021), 5-HT1A receptors (Russo et al., 2005;Rodrigues da Silva et al., 2020;Yano et al., 2020), and other receptors (Ibeas Bih et al., 2015;Morales et al., 2017;Turner et al., 2017). Lipophilic features of cannabinoids markedly affect their pharmacokinetic properties as well. In biochemical studies, tissue AEA and 2-AG concentrations range between 10-214 pmol/g (9 nM-195 nM) and 1-21 nmol/g (1 µM-19 µM), respectively

FIGURE 1
Cannabinoids interact with the lipid membrane and influence the functional properties of ion channels and other integral membrane proteins. Cannabinoids enter the lipid membrane and binds to cannabinoid receptors through lipid membrane (Reggio and Traore, 2000;Barnett-Norris et al., 2005;Makriyannis et al., 2005). In addition, cannabinoids directly affect channel function by changing the biophysical properties of the lipid membrane or binding to a hydrophobic binding site(s) located on the transmembrane regions of ligand-gated ion channels (represented with large dashed red arrow). Cannabinoids, like other lipophilic molecules, partition into the lipid bilayer and alter the biophysical properties of the membrane by reducing membrane electrical resistance, increasing membrane fluidity, changing membrane order, increasing membrane stiffness, increasing membrane elasticity, and changing physicochemical and structural properties of bilayer membranes (represented with thinner diagonal red arrows). Secondly, cannabinoids can bind directly to transmembrane domains of ion channels embedded in the cell membrane (see discussion).
Frontiers in Physiology frontiersin.org (Buczynski and Parsons, 2010). However, tissue endocannabinoid concentrations display significant spatiotemporal differences among brain regions and within the cell as well. Presumably, de novo synthesized endocannabinoids are preferentially incorporated into lipid membranes and give rise to spatially localized signaling within the cell. Similarly, phytocannabinoids, due to their high lipophilicity, are expected to attain membrane concentrations that are considerably higher than blood levels (Deiana et al., 2012). In an earlier study, it has been shown that perfusion of isolated rat hearts with buffer containing [ 3 H]-CBD results in strong accumulation of radioactivity in the tissue (Smiley et al., 1976). Recently, chronic (28 days) administration of CBD (230 mg/kg) was shown to reach CBD concentrations of 2.7 µM and 3.3 µM in muscle and liver tissue, respectively (Child and Tallon, 2022). Therefore, the membrane concentrations of cannabinoids are likely to reach to pharmacologically relevant ranges to exert their effects on the functional properties of LGICs described in this review. However, the results of studies using concentrations of cannabinoids above 20-30 µM may have no pharmacological relevance.

Conclusion
Cannabinoids have appeared as important modulators of diverse pathological and physiological processes and intensive research efforts have examined the efficacy of cannabinoid antagonist and agonist as therapeutic agents. Although cannabinoids primarily exert their cellular and organ system effects by interacting with CB1 and CB2 cannabinoid receptors, in recent years, it has been shown that not all effects of these agents are mediated by cannabinoid receptors. Several lines of evidence show that cannabinoids exert their effects by modulating activities of ion channels; transporters; enzymes, and other G-proteincoupled receptors in a cannabinoid receptor-independent manner. Among these diverse array of cellular macromolecules, ligand-gated ion channels constitute an important target that can effectively modulate neurotransmission in both central and peripheral nervous system.

Author contributions
All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.