The Endocannabinoid System and Related Lipids as Potential Targets for the Treatment of Migraine-related Pain

Rosaria Greco PhD; Chiara Demartini PhD; Anna Maria Zanaboni BSc; Miriam Francavilla MSc; Roberto De Icco MD, PhD; Lara Ahmad MD; Cristina Tassorelli MD, PhD


Headache. 2022;62(3):227-240. 

In This Article



Of the three main strains of cannabis (Cannabis sativa, C. ruderalis, and C. indica), C. sativa has been used as an anesthetic, analgesic, and anxiolytic in different conditions;[11] its broad application is owing to its anti-inflammatory, anticancer, neuroprotective, and antinociceptive effects.[11] Phytocannabinoid characterization is still in its early stages, with most research to date focusing on the psychoactive compound Δ9-tetrahydrocannabinol (THC) and on cannabidiol (CBD), which is a nonpsychotropic cannabinoid. In the late 20th century, the first medical cannabis-derived compound was approved for clinical use, for multiple sclerosis.[12] Regarding the headache and migraine field, preclinical data suggest that cannabinoids could play a role in the control of migraine-related pain. In cultured platelets, cannabinoids have been shown to block serotonin release that was induced by incubation with a patient's plasma obtained during a migraine attack. THC, acting at the CB1 receptor,[13–15] may operate in the same way as endocannabinoids by inhibiting both vasorelaxation and CGRP release. In rats, CB receptor agonists, including THC, induce a state of latent sensitization characterized by increased sensitivity to stress (a well-known migraine trigger); this suggests that cannabinoid/cannabis overuse in vulnerable individuals may increase their risk of developing medication overuse headache (MOH).[16] In the clinical field, medical marijuana has been used in the acute and preventive treatment of migraine; although preliminary data suggest a reduction of attack frequency (Table 1), the evidence is weak because of a lack of controlled clinical studies. Archival data from a large database showed greater improvements in men than in women and suggested that concentrated preparations were more effective than flower consumption.[23] In a small, single-center trial, Pini et al.[18] tested a synthetic cannabinoid (nabilone) in patients with MOH and found reductions in pain duration and intensity and in daily intake of analgesics. In a pilot study, Nicolodi et al.[19] tested a combination of THC with CBD in the acute and preventive treatment of patients with chronic migraine (CM) and reported a reduction in pain intensity. Methodologically sound studies are now needed to investigate the possible effects of cannabis in migraine treatment and to define strains, formulations, and dosage.

The ECS and Related Lipids

The discovery of the ECS[24] was a major advance in the cannabinoid research. Thus, it is not surprising that recent decades have seen multiple attempts to develop novel analgesic compounds for chronic pain by targeting it.[25,26] Recent studies suggest that compounds interacting with the ECS are also able to modulate the pathways involved in migraine-related pain.[27–29]

The ECS is a complex cell-signaling system widely present in peripheral tissues and the nervous system. The ECS is traditionally understood to consist of the CB1 and CB2 receptors; the endogenous ligands that bind the CB1/CB2 receptors, of which N-arachidonoylethanolamine (anandamide) and 2-arachidonoylglycerol (2-AG) are the most studied; and synthetic and degradative enzymes. The ECS is functionally connected with other signaling pathways that use fatty acids, esters, and amides, such as palmitoylethanolamide (PEA).[30–32] These lipids are not classified as "endocannabinoids" because they do not bind to CB receptors, but they may be synthesized and hydrolyzed by the endocannabinoid metabolic enzymes (see the "Synthesis and Catabolism" section for more details). Endocannabinoid tone enhancement has been proposed as an alternative modality of activation of CB receptors and is possibly devoid of the psychotropic effects reported with CB receptor agonists. From these considerations, it is evident that the complexity of the ECS poses a challenge for the development of selective endocannabinoid-based drugs but, at the same time, offers interesting therapeutic opportunities.

Synthesis and Catabolism

Unlike classic neurotransmitters, endocannabinoids are not stored in vesicles but are enzymatically produced on demand from membrane glycerophospholipid precursors; this occurs in response to cellular depolarization leading to Ca2+ influx. The specific enzymes responsible for endocannabinoid biosynthesis are N-acyl-phosphatidyl-ethanolamine-selective phospholipase D for anandamide, and diacylglycerol lipases α and β for 2-AG. Fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL) are the catabolic enzymes involved in the degradation of anandamide and 2-AG, respectively. Other lipids, such as PEA, are acylethanolamines formed from N-acylated phosphatidylethanolamine via numerous enzymatic pathways.[33] There are also several other acylethanolamines, including oleoylethanolamide (OEA) and stearoylethanolamide. Acylethanolamines can be hydrolyzed by FAAH in the endoplasmic reticulum[34] and by N-acylethanolamine-hydrolyzing acid amidase (NAAA) in lysosomes.[35] By contrast, MAGL activity has been described in both cytosolic and membrane-bound locations.[36]

Inhibition of endocannabinoid catabolic enzymes increases tissue levels of endocannabinoids and other lipid mediators. This issue is discussed in depth in "ECS Modulation As a Potential Therapeutic Target."


As mentioned above, endocannabinoids interact with CB receptors and CB-like receptors.[37] In particular, anandamide is a partial agonist of CB1 receptors, whereas 2-AG is a full agonist of both CB1 and CB2 receptors.[38] CB1 receptors are present to different degrees in peripheral tissues, such as the heart, liver, and immune cells;[39,40] activation of CB1 receptors in the brain, where they are mainly expressed, induces effects on mood, sensitivity, cognition, and locomotor activity.[41] By contrast, CB2 receptors are primarily located in immune cells,[42] although they can also be detected in the CNS, specifically in neurons and glial cells, particularly microglia.[43,44] CB2 receptors are expressed at lower levels than CB1 receptors in the CNS, which suggests that they might not be involved in cannabis-related effects under physiological conditions.[45] However, CB2 expression is dynamic and in some pathological conditions (e.g., addiction, inflammation, anxiety, and epilepsy), it can be induced in specific brain areas, thus suggesting CB2 involvement in various psychiatric and neurological diseases.[45]

Through CB receptor activation, anandamide and 2-AG may affect numerous physiological processes, including pain modulation[46–48] and inhibition of pro-inflammatory mediators.[49] As reported above, endocannabinoids may, in part depending on their concentration, also act on other targets besides CB receptors. For instance, anandamide is an endogenous agonist of CB1 receptors but may also interact with transient receptor potential (TRP) cation channels and peroxisome proliferator-activated receptors (PPARs), specifically TRPV1 and PPARγ, respectively. Anandamide activates TRPA1 and TRPV2 at high micromolar concentrations[50,51] and behaves as an inhibitor of TRPM8 and Cav3 Ca2+ channels.[51,52] Instead, 2-AG binds to and activates TRPV1 channels and GABAA receptors, while it inhibits Cav3 channels.[53]

Other atypical CB receptors include some of the orphan G protein–coupled receptors,[53] such as GPR55, which binds to PEA, phytocannabinoids, and different cannabinoid ligands of synthetic origin, including HU210, JWH-015, and R(+)-methanandamide.[54,55] Cannabinoids have the ability to bind allosterically to opioid receptors,[56] which are located on peripheral nerves and involved in the modulation of neurogenic inflammation.

In the brain, OEA and PEA can have biological effects of their own, independently of CB receptors and, like anandamide, they can activate TRPV1 receptors.[57] In vivo studies have demonstrated that OEA can induce either pain or analgesia via TRPV1 activation, depending on the experimental setting.[58,59] Furthermore, the anti-inflammatory and neuroprotective effects of PEA and OEA may also be mediated by activation of PPARα and PPARγ,[60] which are ubiquitously distributed in the CNS[61] and expressed in neurons, astrocytes, and microglia.[62] Of note, the rapid effects of OEA and PEA may be mediated by nongenomic activation of protein kinases.[60]

Endocannabinoid-like Mediators

Other long-chain N-acyl-amides, including N-acyl-taurines, N-acyl-serotonins, N-acyl-dopamines, fatty acid primary amides, and N-acyl-amino acids, have heterogeneous targets and have been included in the ECS pathways, showing why this intricate signaling system has been dubbed the "endocannabinoidome."[63,64] These lipids exert a variety of biological activities in the CNS, as they control biological processes in neurons and glial cells. These endocannabinoid-like mediators have an affinity for CB1 or CB2 receptors, but that affinity is much lower than that of classical endocannabinoids.

Deregulation of the ECS and PEA Changes in Migraine

Clinical observations support the hypothesis that the ECS may be dysfunctional in migraine and may interact with numerous migraine-related pathways. Exploration of ECS modulation in animal models of migraine has also given interesting findings. Table 2 and Table 3 summarize preclinical and clinical studies on the relationship between migraine and the ECS or related lipids.

Clinical Evidence. Endocannabinoid deficiency was first postulated to have a role in migraine pathophysiology in the early 2000s.[87] This hypothesis was supported by subsequent studies that showed higher platelet activity of both FAAH and anandamide transporter in female patients with migraine as compared with controls.[65] A positron emission tomography (PET) study showed increased CB1 binding in specific brain areas implicated in pain processing in subjects with episodic migraine (EM).[66] This upregulation, which is associated with alterations in the gene expression of metabolic enzymes, was also observed in peripheral blood mononuclear cells (PBMCs), indicating that these cells may mirror cerebral changes.[71] Through haplotype classification, migraine has been found to be related to the CRN1 gene; the significant effect of this gene on migraine headaches was suggested to be related to abnormal trigeminovascular activation.[88] More recently, Gouveia-Figueira et al.[67] failed to detect any differences in peripheral levels of anandamide in patients with EM as compared with controls. This finding indicates that the ECS is not completely dysregulated in EM, or alternatively that efficient compensatory mechanisms are invoked.

Similarly, we did not detect differences in plasma levels of anandamide and PEA in patients with EM versus healthy patients.[68] Nevertheless, PEA plasma levels were found to be increased during the ictal phase of spontaneous migraine-like attacks triggered experimentally by administration of nitroglycerin (NTG).[68] This phenomenon suggests that PEA release is a compensatory mechanism serving to counteract the neurovascular changes that occur during the early phase of NTG-induced attacks. Preliminary evidence suggests that PEA may be useful in the prevention of migraine attacks in adult and pediatric populations and could potentially be a new therapeutic tool for management of this condition[20–22] (Table 1). PEA may have indirect effects on CB receptors, FAAH expression, and TRPV1 channels.[89] Through FAAH inhibition, PEA might modulate anandamide levels,[90] which in turn would stimulate CB2 or CB1 receptors or desensitize TRPV1 channels, thus mediating analgesic and anti-inflammatory effects.

Dysregulation of the ECS is more pronounced in CM than in EM. Compared with controls and with individuals with EM, subjects with CM (with and without MOH) showed lower levels of FAAH and endocannabinoid membrane transporter activity in platelets.[69]

Our group showed that withdrawal of overused drugs in subjects with MOH induced a reduction in platelet FAAH activity associated with normalization of spinal nociception.[70] In addition, in PBMCs from patients with CM, we detected changes in gene expression of enzymes involved in 2-AG metabolism.[71] These findings are in agreement with previous studies reporting decreased levels of 2-AG and anandamide in the cerebrospinal fluid (CSF) or platelets of people with CM[72,73] and thus confirm that there is chronic deregulation of the ECS in migraine. An increase in PEA levels was also described in the CSF of patients with both CM and MOH,[72] which can probably be interpreted as a compensatory response to the decreased levels of anandamide. Elsewhere, 2-AG and anandamide platelet levels were also significantly lower in patients with MOH and in ones with CM compared with controls, without significant differences between the two patient groups. On the other hand, chronic medication overuse can reduce endocannabinoid levels in different brain areas, such as the midbrain, hippocampus, striatum, and cerebral cortex,[91] contributing to the decreased levels of anandamide and 2-AG found in patients with MOH. Serotonin levels were strongly reduced in the two patient groups and were correlated with 2-AG levels, with higher values for patients with MOH,[73] suggesting an interaction between 2-AG and serotonin in modulating different signaling pathways.

Preclinical Evidence. Changes in the ECS have been reported in experimental models of migraine as well, suggesting that targeting this system or related lipids may open the way for future therapeutic strategies.[28] In the NTG-based animal model of migraine, we found increased FAAH and MAGL activity and an increased number of CB receptor binding sites in the mesencephalon 4 h after NTG administration.[92] FAAH activity was also increased in the hypothalamus and medulla;[92] these findings point to ECS deregulation as a pathophysiological mechanism of migraine, or alternatively indicate an adaptive response to a disease state.

In the same migraine model, anandamide administration reduced NTG-induced hyperalgesia during the plantar formalin test and neuronal activation in the trigeminal nucleus caudalis (TNC).[74] This action was likely mediated by CB1 and CB2 receptors,[74,75] although anandamide can modulate central sensitization of activated B cells expression as well via TRPV1, cyclooxygenase-2 (COX-2), and the nuclear factor κ light chain enhancer.[75] Of note, treatment with methanandamide, an anandamide synthetic analogue, was found to attenuate NTG-induced CGRP increases in plasma, trigeminal ganglia, and the brainstem, and it inhibited dural mast cell degranulation.[77]

Anandamide has significantly reduced neurogenic inflammation in other animal models of migraine, too.[76,79,93] Anandamide inhibited neurogenic, as well as CGRP-induced and NO-induced, dural vasodilation.[76] Moreover, hyperexcitability of cortical neurons, one of the key mechanisms in migraine pathophysiology, was inhibited by WIN55,212–2 (a CB1 receptor agonist), whereas JWH 133 (a CB2 receptor agonist) was not effective.[78] CB1 receptor activation diminished dural stimulation-evoked activity in A-fiber neurons and basal spontaneous activity in the trigeminocervical complex of rats, suggesting that anandamide may be involved in the modulation of descending pain control.[94] Accordingly, endocannabinoids may lead to analgesia by suppressing GABAergic inhibition of periaqueductal gray (PAG) output neurons, which project along a descending analgesic pathway.[95] In contrast, WIN55,212–2 showed conflicting results in its action on trigeminal sensory afferent neurons, as it may have both excitatory and inhibitory effects[93] depending on the dose used.

ECS Modulation as a Potential Therapeutic Target

Figure 1 schematically illustrates the potential therapeutic action of ECS modulation in migraine-related pain and the putative crosstalk with other pathways.

Figure 1.

Schematic representation of the possible therapeutic approaches to modulate the activity of the endocannabinoid system. Reduction of pain (including migraine-related pain) can be achieved via the: (i) enhancement (+) or inhibition (−) of different receptors through either orthosteric or allosteric ligands; (ii) inhibition (−) of degrading enzymes, resulting in an increased endocannabinoids and related lipids tone. The putative crosstalk between the endocannabinoid and opioid systems is also represented. 2-AG, 2-arachidonoylglycerol; AA, arachidonic acid; AEA, anandamide; Allosteric, allosteric ligands of the cannabinoid receptors; CBs, cannabinoid receptors; COX-2, cyclooxygenase-2; EA, ethanolamine; ECS, endocannabinoid system; FAAH, fatty acid amide hydrolase; Gly, glycerol; GPCRs, G-protein-coupled receptors; MAGL, monoacylglycerol lipase; NAAA, N-acylethanolamine acid amide hydrolase; OA, oleic acid; OEA, oleoylethanolamide; ORs, opioid receptors; OS, opioid system; PA, palmitic acid; PEA, palmitoylethanolamide; PG, prostaglandin; PPARs, peroxisome proliferator-activated receptors; TRPV1, transient receptor potential cation channel subfamily V member 1 [Color figure can be viewed at]

Endocannabinoid Catabolic Enzyme Inhibitors. Inhibition of endocannabinoid catabolic enzymes is a new strategy designed to circumvent the adverse effects associated with direct and generalized activation of CB receptors. Indeed, inhibition of FAAH and MAGL induces increased endocannabinoid tone at the site where the lipids are needed and then released, thus resulting in tissue-selective CB receptor activation. Such endocannabinoid tone enhancement may be beneficial in the treatment of several pathological conditions, including migraine. Increasing knowledge of the pathophysiology of migraine has allowed new research into disease treatment options that may modulate the underlying trigeminocervical complex activation. In this context, distinct profiles of MAGL and FAAH inhibitors in the PNS and CNS have recently been reported.[83,85,96] In recent years, we have also shown that FAAH and MAGL inhibitor effects at the peripheral and central levels are involved in the modulation of migraine-related pain. In particular, evaluation of the activity of these inhibitors in animal models of EM and CM demonstrated that systemic administration of the peripherally restricted FAAH inhibitor URB937 attenuated NTG-induced trigeminal and spinal hyperalgesia in rats.[84,85] This effect was associated with increases in anandamide and PEA levels together with decreases in mRNA expression of CGRP, substance P, and pro-inflammatory cytokines (tumor necrosis factor α and interleukin-6) in the trigeminal ganglion. Similar effects were also observed in the acute NTG model following pretreatment with URB597, a global FAAH inhibitor. More specifically, we reported that URB597 modulated NTG-induced neuronal activation in structures involved in migraine-related pain, such as the TNC.[83] In line with this, Nozaki et al.[82] reported that NTG-induced mechanical allodynia and c-Fos expression in the TNC in mice were abolished by URB597 when it was administered 2 h before NTG.

URB602, an inhibitor of MAGL, inhibited NTG-induced reduction in tail-flick latency,[86] counteracted NTG-induced hyperalgesia during the orofacial and plantar formalin tests, and attenuated c-Fos expression in the PAG and TNC.[86] Of note, when MAGL inhibitors (URB602 or JZL184) were administered with vehicle (instead of NTG), hyperalgesic behavior during the orofacial formalin test was found to be increased, which indirectly suggests that higher 2-AG levels may have a paradoxical analgesic action at the trigeminal level.[86] In agreement with this observation, Spradley et al.[97] described differences in peripheral endocannabinoid modulation after MAGL inhibition; they showed that increases in peripheral levels of 2-AG were associated with a pro-nociceptive effect in the facial skin but not in the hind paw skin. On the other hand, MAGL inhibition had no effect on the antinociceptive activity of 2-AG in a model of trigeminal pain.[98] Given that FAAH and MAGL are localized in different neuronal regions and regulate the degradation of different endocannabinoids, they may have dissimilar roles in regulating physiological functions, including nociception. Thus, although the action of FAAH inhibitors in the trigeminal system seems straightforward, this topographically segregated effect of MAGL inhibition in trigeminal/extra-trigeminal areas warrants further research. Of interest, on the basis of our findings using the dual FAAH/MAGL inhibitor JZL195 in the NTG model,[99] we recently suggested a synergistic role for anandamide and 2-AG in trigeminal pain modulation.[100] This suggestion is in line with the report by Zubrzycki et al.,[98] which found that JZL195 modulated analgesia in an animal model of orofacial pain.

The now great interest in FAAH inhibitors for the treatment of pain and CNS disorders is documented by the numerous molecules under development and tested in clinical trials.[101] However, the development of potent and safe FAAH inhibitors is delayed by their possible off-target mediated adverse effects, like the unfortunate event that occurred in the context of the BIA10-2474 trial.[102] Thus, extensive preliminary studies, including in silico methods, could be useful to identify possible serious off-target effects, at the cellular level, of FAAH inhibitors.[102]

Allosteric Ligands of CB1 and CB2 Receptors. Further to the above-mentioned strategies to potentiate the ECS without directly stimulating CB receptors, a different approach has recently gained attention. It is based on allosteric cannabinoid ligands that bind to sites other than the orthosteric binding sites (i.e., the receptor sites where the endogenous ligand binds). These compounds act as allosteric modulators by modifying the cannabinoid protein conformation and by modulating the affinity and/or efficacy of orthosteric ligands. This interaction positively or negatively affects the orthosteric ligand-driven signaling. This approach may result in pharmacological advantages (e.g., higher specificity) and thus reduce adverse effects.[103] Of interest, CBD is a negative allosteric modulator of CB1 and a CB2 partial agonist.[104,105] On the other hand, CBD appears to have numerous molecular targets and several effects outside the ECS.[106] For instance, it acts as a partial agonist also on PPARγ[107] and 5-HT1A receptors[108] and on TRPV1.[107] Therefore, its exploration as a putative treatment option in relation to migraine pathology might prove interesting. Recent studies reported that a positive allosteric modulator of CB1 and CB2 receptors displayed antinociceptive activity in animal models of neuropathic pain.[109,110] Moreover, a CB1 allosteric ligand showed synergistic antiallodynic effects with FAAH and MAGL inhibitors.[108]

These observations, together with the lack of intrinsic agonism of allosteric modulators, suggest that this approach deserves further investigation; in addition, it might provide interesting leads for clinical development, given that it may have a favorable side effect profile[111] with limited psychomimetic and depressant effects.

Dual Blockers. At high concentrations, anandamide can activate TRPV1 channels, prompting them to stimulate pain pathways via CGRP release.[112] We have already mentioned that FAAH inhibition may potentiate the biological activity of anandamide, thereby mediating an antihyperalgesic response. This is not always observed in animal models of neuropathic pain,[113] probably owing to the existence of alternative anandamide biotransformation pathways that favor TRPV1-mediated pro-nociceptive signaling.[114] Hence, the development of combined FAAH/TRPV1 blockers may be a promising strategy.[115] For instance, the endogenous lipid-signaling molecule N-arachidonoyl-5-hydroxytryptamine (AA-5-HT) has been shown to act as a dual FAAH/TRPV1 blocker and to possess anti-inflammatory effects.[116] AA-5-HT increased anandamide at central and peripheral levels and antagonized human and rat recombinant TRPV1 channels.[116] This blocker induced analgesic effects in different pain models.[115] Another FAAH/TRPV1 blocker is OMDM198, which reduced pain in a rat model of osteoarthritis[117,118] and in a formalin-induced inflammatory pain model.[119] A similar effect was also reported with the dual FAAH/TRPA1 agonist OMDM202.[119] Both compounds increased anandamide levels in the spinal cord.

There is evidence suggesting a potential role for dual COX/FAAH inhibitors in pain management. A synergistic effect of common NSAIDs and anandamide or FAAH inhibitors was demonstrated in different experimental models of acute and chronic pain.[115] This led to the synthesis of dual-acting compounds, such as ARN2508, with promising analgesic actions.[120] Although inhibiting both COX and FAAH individually has obvious beneficial effects, it should be noted that when FAAH is inhibited, the COX-2 pathway will be responsible for anandamide biotransformation, leading to a decrease in endocannabinoid tone and to the generation of inflammatory metabolites.[121]

The experimental data obtained with the dual-acting compounds suggest that, by operating simultaneously at interconnected systems, they exert a potential beneficial effect also on migraine and may avoid some of the adverse effects of the single-compound treatments.[122]

NAAA Inhibitors. Given the interesting properties and low toxicity shown by PEA, this lipid has been studied in clinical trials to evaluate its anti-inflammatory effects.[123,124] PEA levels may increase during inflammation, producing local anti-inflammatory and analgesic actions.[125] PEA modulation of mast cell activation and degranulation[126] is another modality through which this lipid interferes with inflammation and chronic pain. Additionally, PEA is produced and hydrolyzed by microglia in response to inflammatory events,[127] and its activity on the PPARα receptors promotes the expression of anti-inflammatory molecules.[128] By using inhibitors of NAAA, the enzyme that preferentially hydrolyzes PEA, it would thus seem possible to treat inflammatory pain. Local interventions, through PEA/PPAR-α signaling pathway modulation by NAAA inhibitors solely at the site of inflammation, may be an option for minimizing potential adverse effects. Although these findings seem encouraging, only a limited number of NAAA inhibitors are available, and they possess suboptimal druggability profiles.[129]

The Opioid System in Endocannabinoid Antinociception. In recent years, the interconnection between the endogenous antinociceptive systems has been the focus of growing attention.[130] A role for the opioid system in cannabinoid antinociception has been demonstrated, as has the reverse.[131,132] This is not surprising because the endogenous elements composing these systems share the same pattern of anatomic distribution and have comparable functional properties.[133,134] Activation of opioid or CB receptors may result in similar behavioral effects, including analgesia,[135] because they display comparable signal transduction properties.[136,137] Of interest, especially regarding migraine-related pain, is the fact that the pathways of both systems are involved in the modulation of brainstem descending control of trigeminal nociception.[130] For example, blockade of both opioid and CB1 receptors led to a significant reduction in anandamide-mediated antinociception.[98] Moreover, CB2 receptor activation caused release of β-endorphin from keratinocytes, thereby inducing antinociception.[138] That said, combining cannabinoids and opioid drugs could be a promising approach to generate long-lasting analgesia, while avoiding or minimizing the tolerance issues associated with opioid-based drugs.[139]