Goods and Bads of the Endocannabinoid System as a Therapeutic Target: Lessons Learned after 30 Years

Abstract

The cannabis derivative marijuana is the most widely used recreational drug in the Western world and is consumed by an estimated 83 million individuals (∼3% of the world population). In recent years, there has been a marked transformation in society regarding the risk perception of cannabis, driven by its legalization and medical use in many states in the United States and worldwide. Compelling research evidence and the Food and Drug Administration cannabis-derived cannabidiol approval for severe childhood epilepsy have confirmed the large therapeutic potential of cannabidiol itself, Δ⁹-tetrahydrocannabinol and other plant-derived cannabinoids (phytocannabinoids). Of note, our body has a complex endocannabinoid system (ECS)-made of receptors, metabolic enzymes, and transporters-that is also regulated by phytocannabinoids. The first endocannabinoid to be discovered 30 years ago was anandamide (N-arachidonoyl-ethanolamine); since then, distinct elements of the ECS have been the target of drug design programs aimed at curing (or at least slowing down) a number of human diseases, both in the central nervous system and at the periphery. Here a critical review of our knowledge of the goods and bads of the ECS as a therapeutic target is presented to define the benefits of ECS-active phytocannabinoids and ECS-oriented synthetic drugs for human health. SIGNIFICANCE STATEMENT: The endocannabinoid system plays important roles virtually everywhere in our body and is either involved in mediating key processes of central and peripheral diseases or represents a therapeutic target for treatment. Therefore, understanding the structure, function, and pharmacology of the components of this complex system, and in particular of key receptors (like cannabinoid receptors 1 and 2) and metabolic enzymes (like fatty acid amide hydrolase and monoacylglycerol lipase), will advance our understanding of endocannabinoid signaling and activity at molecular, cellular, and system levels, providing new opportunities to treat patients.

Graphical abstract

Fig. 1

Endocannabinoid binding receptors. The two major endocannabinoids anandamide and 2-arachidonoylglycerol bind to and activate metabotropic and ionotropic membrane receptors (with either an intracellular or an extracellular binding site) and nuclear receptors.

Fig. 2

Endocannabinoid signaling pathways. Receptor binding by anandamide and 2-arachidonoylglycerol triggers various signal transduction pathways, which activate G proteins, ion channels, as well as gene transcription.

Fig. 3

Biosynthetic pathways of anandamide. AEA can be synthesized from membrane phospholipid precursors via different routes. The Ca²⁺-dependent hydrolysis of NArPE by NAPE-PLD is considered the most relevant among these biosynthetic pathways.

Fig. 4

Catabolic pathways of anandamide. AEA can be cleaved to arachidonic acid and ethanolamine by different hydrolytic routes. FAAH-1 is considered the most relevant among these catabolic pathways. Alternatively to hydrolytic routes, AEA can be oxidized by LOXs, COX-2, or cytochrome P450 to generate various eicosanoid-like PG-ethanolamides or hydroxy-AEAs.

Fig. 5

Biosynthetic pathways of 2-arachidonoylglycerol. 2-AG can be synthesized from membrane phospholipid precursors via different routes. The Ca²⁺- and glutathione-dependent hydrolysis of DAG by DAGLα/β is considered the most relevant among these biosynthetic pathways.

Fig. 6

Catabolic pathways of 2-arachidonoylglycerol. 2-AG can be cleaved into arachidonic acid and glycerol by different hydrolytic routes. MAGL is considered the most relevant among these catabolic pathways. Alternatively to hydrolytic routes, 2-AG can be oxidized by LOXs or COX-2 to generate various eicosanoid-like PG-glyceryl esters or hydroxy-2-AGs.

Fig. 7

Transport of endocannabinoids. (A) Anandamide and 2-arachidonoylglycerol can cross the plasma membrane via different mechanisms, which include passive diffusion, exocytosis of microvesicles and a putative membrane transporter. (B) Intracellular trafficking of anandamide and 2-arachidonoylglycerol is driven by various carriers that include structurally unrelated proteins like albumin, RBP2, HSP70, FABPs, SCP2, and FLAT.

Fig. 8

Chemical structure, CB₂R binding affinity and selectivity of relevant nonclassic cannabinoids. ^aConsensus human CB₂R binding affinity values from a multicentric collaborative profiling effort between multiple independent academic laboratories and industry (Soethoudt et al., 2017). ^bCB₂R selectivity (10^∧(pKi CB₂R-pKi CB₁R).

Fig. 9

Chemical structure, CB₂R binding affinity and selectivity of representative aminoalkylindole CB₂R ligands. ^aConsensus human CB₂R binding affinity values from a multicentric collaborative profiling effort between multiple independent academic laboratories and industry (Soethoudt et al., 2017). ^bCB₂R selectivity (10^∧(pKi CB₂R-pKi CB₁R).

Fig. 10

(A) X-ray structure of CB₁R (blue) bound to the antagonist AM6538. (B) Cryo-EM structure of CB₁R (green) in complex with G proteins (α subunit in yellow, β subunit in blue, γ subunit in purple) and the classic cannabinoid agonist AM841. (C) Chemical structures of AM6538 and AM841.

Fig. 11

Comparison of ligand binding modes in CB₁R and CB₂R. (A) The binding pocket of AM10257 in CB₂R crystal structure (PDB code 5ZTY). AM10257 and the key residues are shown in sticks as the following color code: CB₂R, brown; AM10257, light coral. (B–D) Binding pose comparison of AM6538 in CB₁R (PDB code 5TGZ), and AM10257 in CB₂R, using color code as follows: CB₁R, slate blue; AM6538, dodger blue; CB₂R, brown; AM10257, light coral. (E–F) The binding pocket of AM12033 in CB₂R (PDB code 6KPF) and WIN55,212-2 in CB₂R (PDB code 6TP0). Ligands and the key residues are shown in sticks as the following color code: AM12033, brown; CB₂R (6KPF), dark green; WIN55,212-2, royal blue; CB₂R (6TP0), dark salmon. (G) The conformational comparison of “toggle switch” residues Trp258^6.48 between AM12033- and WIN55,212-2-bound CB₂R. (H–J) Binding pose comparison of THC-like agonist in CB₁R (PDB code 6KPG) and CB₂R (PDB code 6KPF). THC-like agonists are shown as sticks (H) and surface (I–J), the key residues are shown in sticks as the following color code: CB₂R, dark green; AM12033, brown; CB₁R, maroon; AM841, dark khaki. (K-M) Binding pose comparison of agonist FUB in CB₁R (PDB code 6N4B) and agonist WIN55,212-2 in CB₂R (PDB code 6TP0). FUB and WIN55,212-2 are shown as sticks (K) and surface (L–M), the key residues are shown in sticks as the following color code: CB₂R, dark salmon; WIN55,212-2, royal blue; CB₁R, dark cyan; FUB, orange.

Fig. 12

Conformational changes during CB₂R activation. (A–C) The conformational change of key residues between inactive- and active-CB₂R. “Toggle switch residue” (A), D^3.49R^3.50Y^3.51 motif (B), and N^7.49P^7.50xxY^7.53 motif (C). (D–F) The overall structure (D), the extracellular region (E), and intracellular region (F) comparison of inactive- (brown) and active-state (dark green) CB₂R structures.

Fig. 13

Structures of the clinically tested cannabinoid agonists (A) and the selective CB₁R antagonists (B).

Fig. 14

Structures of the CB₂R in different states. (A) Crystal structure of antagonist AM10257-bound CB₂R (PDB code 5ZTY). (B) Crystal structure of agonist AM12033-bound CB₂R (PDB code 6KPC). (C) Cryo-EM structure of AM12033-bound CB₂R-G_i complex (PDB code 6KPF). (D) Cryo-EM structure of WIN55,212-2-bound CB₂R-G_i complex (PDB code 6TP0), using color code as follows: CB₂R-AM10257, brown; CB₂R-AM12033 (PDB code 6KPC), sky blue; CB₂R-AM12033 (PDB code 6KPF), green; CB₂R-WIN55,212-2, dark salmon; Gα_i in CB₂R-AM12033, purple; Gβ in CB₂R-AM12033, teal; Gγ in CB₂R-AM12033, orchid; scFv16 in CB₂R-AM12033, cornflower blue; Gα_i in CB₂R-WIN55,212-2, medium purple; Gβ in CB₂R-WIN55,212-2, turquoise; Gγ in CB₂R-WIN55,212-2, plum; scFv16 in CB₂R-WIN55,212-2, light blue.

Fig. 15

Chemical structure and CB₂R binding affinity of THC, N-arachidonoylethanolamine, and 2-arachidonoyl glycerol. ^aConsensus human CB₂R binding affinity values from a multicentric collaborative profiling effort between multiple independent academic laboratories and industry (Soethoudt et al., 2017). ^bCB₂R selectivity (10^∧(pKi CB₂R-pKi CB₁R).

Fig. 16

Chemical structure and CB₂R binding affinity of noladin ether and synthetic eCB analogs. ^aBinding to spleen cannabinoid receptor. ^bWith phenylmethanesulfonyl fluoride.

Fig. 17

Chemical structure, CB₂R binding affinity and selectivity of representative classic cannabinoids. ^aConsensus human CB₂R binding affinity values from a multicentric collaborative profiling effort between multiple independent academic laboratories and industry (Soethoudt et al., 2017). ^bCB₂R selectivity (10^∧(pKi CB₂R-pKi CB₁R).

Fig. 18

Chemical structure, CB₂R binding affinity or functional activity, and selectivity of clinically evaluated bicyclic (het)aryl derived CB₂R ligands.

Fig. 19

Chemical structure, CB₂R binding affinity or functional activity, and selectivity of clinically evaluated bicyclic aliphatic (het)aryl arrays.

Fig. 20

Chemical structure, CB₂R binding affinity or functional activity, and selectivity of clinically evaluated CB₂R agonists and CB₂R inverse agonists SR144528 containing five- and six-membered central cores. ^aConsensus human CB₂R binding affinity values from a multicentric collaborative profiling effort between multiple independent academic laboratories and industry (Soethoudt et al., 2017). ^bCB₂R selectivity (10^∧(pKi CB₂R-pKi CB₁R).

Fig. 21

Chemical structure of validated CB₂R allosteric modulators.

Fig. 22

Chemical structure, CB₂R binding affinity, and selectivity of CB₂R radioligands, PET tracers, fluorescent and pAfBPP probes.

Fig. 23

Chemical structures of representative inhibitors of FAAH.

Fig. 24

Chemical structures of representative inhibitors of NAAA.

Fig. 25

(A) Structured part of the AlphaFold model for human DAGLα, residues 1-681; red: transmembrane domain, blue: catalytic domain, green: regulatory loop. (B) Unstructured tail region from the AlphaFold model, residues 682–1042 highlighting potential phosphorylation sites, as discussed in the text, and Homer binding domain. (C) Schematic representation with highlighted regions and relevant serines shown.

Fig. 26

Model of eCB membrane transport and trafficking showing druggable targets. Molecular pharmacology and possible implications for therapeutic intervention of using diverse eCB transport inhibitors are shown. See text for details and abbreviations.

Fig. 27

The eCBome receptors as a pharmacological substrate for plant-derived cannabinoids and host or commensal bacteria-derived eCBs and eCB-like molecules. The elements of the ECS as part of the eCBome are shown squared in red. The chemical structures of commendamide, N-miristoyl-alanine, and N-oleoyl-serinol are shown from the top right and down. CBD, cannabidiol; THC, Δ⁹-tetrahydrocannabinol; THCV, Δ⁹-tetrahydrocannabidivarin.

Fig. 28

Reciprocal modulation of the ECS and the gut microbiome. The emerging microbendocannabinoidome (μbeCBome), also summarized in Table 14, enlarges the span of microbe-host communication and expands it to the eCBome.