Alternative Pain Management via Endocannabinoids in the Time of the Opioid Epidemic: Peripheral Neuromodulation and Pharmacological Interventions

Abstract

The use of opioids in pain management is hampered by the emergence of analgesic tolerance, which leads to increased dosing and side effects, both of which have contributed to the opioid epidemic. One promising potential approach to limit opioid analgesic tolerance is activating the CNS endocannabinoid system, via activation of CB₁ receptors (CB₁Rs) in the descending pain inhibitory pathway. In this review, we first discuss preclinical and clinical evidence revealing the potential of pharmacological activation of CB₁Rs in modulating opioid tolerance, including activation by phytocannabinoids, synthetic CB₁R agonists, endocannabinoid degradation enzyme inhibitors, and recently discovered CB₁R positive allosteric modulators. On the other hand, as nonpharmacological pain relief is advocated by the US-NIH to combat the opioid epidemic, we also discuss contributions of peripheral neuromodulation, involving the electrostimulation of peripheral nerves, in addressing chronic pain and opioid tolerance. The involvement of supraspinal endocannabinoid systems in peripheral neuromodulation-induced analgesia is also discussed.

1. Opioid epidemic

The opioid epidemic is an issue of great concern in Northern America and many European countries (OECD, 2019) due to a continuously increasing death rate from misuse or overuse of opioids. In 2018, opioids contributed to more than two-third of overdose deaths in the US (US-CDC, 2020), exerting an enormous economic burden to the healthcare system (Hagemeier, 2018). Opioids are commonly prescribed for relieving cancer pain, postoperative pain or severe neuropathic pain. However, patients often continued to consume opioids even though opioids are categorized as a second or third-line treatment for chronic neuropathic pain (Urits et al., 2020). This is possibly due to persistent painful conditions or postsurgical opioid over-prescription (Neuman et al., 2019). Besides prescription opioids, improper disposal of leftover opioids and the availability of over-the-counter opioids have also substantially contributed to opioid misuse (Sobczak and Gorynski, 2020). Repeated opioid use leads to analgesic tolerance, dependence, and abuse (Skolnick, 2018). Increasing dosing of opioids for better pain control also increases the risk of constipation and respiratory depression, due to the excessive activation of opioid receptors in gastrointestinal and respiratory systems. The latter leads to opioid overdose death. In fact, the consumption of opioids for relieving non-cancer pain, as compared with other analgesics, positively correlates with an increased risk of all-cause mortality (Ray et al., 2016). Thus, unresolved chronic pain and potentially mismanaged opioid administration are significant contributors to the ongoing opioid crisis (Skolnick, 2018). Although there are several non-opioid analgesic options, opioids retain their central role in the approach of many practitioners to pain management, despite inconclusive or negative findings in benefit to risk analyses.

2. Opioids, endorphins, and descending pain inhibition

2.1. Opioid-induced analgesia

There are four members in the opioid receptor family, μ- (MOR), κ- (KOR) and δ- (DOR) opioid receptors as well as NOP receptors (NOR) (Alexander et al., 2019). The analgesic effect of clinically used opioids is mainly mediated by MORs, which are abundantly expressed in neuronal tissues involved in the descending pain inhibitory pathway, including the periaqueductal gray (PAG), rostral ventromedial medulla (RVM) and spinal cord, as well as in sensory nerve endings (Stein, 2016). MOR activation leads to neuronal inhibition by activating G protein-coupled-inward rectifying K⁺ (GIRK) channels and/or inhibiting Ca²⁺ channels, both mechanisms can contribute to opioid-induced analgesia. GIRK channel activation induces neuronal membrane hyperpolarization and thus directly inhibits neuronal excitability, which probably contributes to the analgesic effect of opioids at peripheral and spinal levels (Nockemann et al., 2013). Ca²⁺ channel inhibition at nerve terminals either leads to reduced glutamate release at the spinal dorsal horn or produces disinhibition in the PAG by reducing GABA release onto the glutamatergic neurons in the PAG, which project to the RVM, and subsequently suppress nociceptive transmission at the spinal cord (Bagley and Ingram, 2020).

2.2. Opioid tolerance and the descending pain inhibitory pathway

Repeated application of MOR agonists typically elicits MOR desensitization (Williams et al., 2013) and internalization (Koch and Hollt, 2008) via β-arrestin-2- and/or PKA-dependent pathways, both of which may contribute to analgesic tolerance of opioids. The descending pain inhibitory pathway, where MORs are densely expressed, plays an important role in the generation of analgesic tolerance, especially the PAG. Direct microinjection of morphine into the PAG induced a stronger analgesic effect and quickly elicited analgesia tolerance, compared to injection at the dorsal raphe nucleus (Campion et al., 2016). Similar susceptibility of the PAG to morphine analgesic tolerance was observed in another comparative study between PAG and RVM (Morgan et al., 2005). These data demonstrated that opioid-sensitive neurons in the PAG play a prominent role in the development of opioid analgesic tolerance. Indeed, electrophysiological studies in PAG slices of morphine tolerant mice, found reduced responsiveness of the PAG neurons towards an MOR agonist, suggesting down-regulation of MOR number and/or dysfunctional MOR activity (Bagley et al., 2005).

2.3. Endogenous opioids in PAG stimulation-induced analgesia and stress-induced analgesia

Electrical stimulation of the PAG was able to produce analgesia, which is mediated by endogenous opioids in rodents (Cannon et al., 1982) and humans (Young and Chambi, 1987). This PAG stimulation-produced analgesia (PAG-SPA) was naloxone-sensitive (Akil et al., 1976), mediated by a mechanism similar to morphine analgesia, and exhibited analgesic tolerance after repeated administrations (Mayer and Hayes, 1975). PAG-SPA is known to exhibit crosstalk with the neuronal circuit involved in the phenomenon of stress-induced analgesia (SIA) (Terman et al., 1985), an evolutionarily conserved, self-protecting mechanism in mammals engaged during life-threatening situations. During stress, endogenous opioids, e.g., β-endorphin (Rubinstein et al., 1996), are released to induce analgesia via activating the descending pain inhibitory pathway.

In addition to opioid-dependent PAG-SPA, a naloxone-insensitive PAG-SPA was also reported (Yaksh et al., 1976), but its mechanism(s) remained unclear for decades. Similarly, opioid-independent SIA was also reported (Terman et al., 1986), which, unlike opioid-dependent SIA, did not develop tolerance after repeated induction (Suplita et al., 2008). The opioid-independent form of SIA was, very much later, revealed to be mediated, at least in part, via the endocannabinoid (eCB) system (Hohmann et al., 2005; Lee et al., 2016; Lee et al., 2020). Thus, it is not surprising that the (endo)cannabinoid system has been regarded as a viable alternative pharmacological target for analgesia, although the efficacy of its clinical application remains inconclusive. In some countries, cannabinoids have been suggested as second- or third-line treatments for chronic pain (Urits et al., 2020).

3. Cannabinoid-induced analgesia

3.1. Cannabinoid pharmacology

Cannabinoids include phytocannabinoids, synthetic cannabinoids and eCBs, as reviewed by Schurman et al. (2020). Phytocannabinoids are isolated from marijuana and hemp (Cannabis sativa), where Δ9-tetrahydrocannabinol (THC) is the major psychomimetic constituent. Synthetic cannabinoids, such as CP-55,940, WIN-55,212, levonantradol, etc, have been synthesized for research or drug development purposes and have higher affinity and often higher efficacy at cannabinoid receptors, as compared to THC. eCBs are endogenous ligands of cannabinoid receptors. They are notable for being synthesized on demand, unlike classical neurotransmitters that are pre-synthesized, stored in vesicles of nerve terminals, and rapidly released following increases in nerve terminal calcium. Anandamide (AEA) and 2-arachydononylglycerol (2-AG) are two main eCBs, and both are well documented for their roles in analgesia. 2-AG, in particular, appears to be synthesized in postsynaptic neurons and then travels retrogradely to activate presynaptic cannabinoid CB₁ receptors (CB₁Rs) to inhibit neurotransmitter release. The synthesis and degradation of these two eCBs are enzymatically regulated, which will be discussed in the later section.

Cannabinoids can produce several biological activities via CB₁Rs and CB₂Rs, such as analgesia, anti-emesis, hyperthermia, appetite stimulation, anti-spasticity, anti-epilepsy, anti-inflammation and anti-tumorigenesis (Alexander and Molina-Holgado, 2019). Among these, the analgesic effect mediated by CB₁Rs is noteworthy, although CB₂Rs also have a role in pain regulation (Iyer et al., 2020). CB₁Rs are expressed in neuronal tissues along the descending pain inhibitory pathway, including the PAG (Bouchet and Ingram, 2020; Lau and Vaughan, 2014), RVM and spinal cord (Kelly and Chapman, 2001), as well as in some peripheral nerves endings (Kelly et al., 2003). At the cellular level, CB₁R activation, via G_i/o protein activation, leads to inhibition of adenylyl cyclase, activation of K⁺ channels, and inhibition of Ca²⁺ channels, which can contribute to the analgesic effects of cannabinoids (Alexander and Molina-Holgado, 2019).

Interestingly, opioids and cannabinoids share the same cellular action mechanisms and a similar regional distribution in the descending pain inhibitory pathway (Wilson-Poe et al., 2012), while both exert their analgesic effects through their respective receptors (Lau and Vaughan, 2014). Cannabinoids can also activate the descending pain inhibitory pathway, which encompasses the midbrain PAG and its downstream RVM, and ultimately inhibit pain transmission at the spinal cord (Bouchet and Ingram, 2020). Through its distinct pharmacological receptors and profiling, the cannabinoid system is regarded as a main non-opioid analgesic mechanism to complement opioid analgesia in the descending pain inhibitory pathway, especially in the PAG (Lau and Vaughan, 2014). Thus, cannabinoids are well-positioned to be a non-opioid alternative in pain management.

3.2. Cannabinoids and pain relief

Through their evaluation in multiple preclinical pain models, both phytocannabinoids and synthetic cannabinoids have been reported to suppress acute and chronic pain, mainly, via CB₁Rs (Narouze, 2020). However, like opioids, the tolerance development is also a concern when using cannabinoids as an alternative pain-relieving agent. Like MORs, prolonged activation of CB₁Rs may result in β-arrestin-dependent desensitization (Daigle et al., 2008; Nguyen et al., 2012), which may contribute to analgesic tolerance of cannabinoids. The analgesic effect of repeated low doses of THC, although less susceptible to tolerance development (McKinney et al., 2008), was found to be associated with cognitive dysfunction in rodents (Sarne et al., 2011). Clinically, substantial human studies unfortunately also indicated a limited efficacy of using CB₁R agonists in relieving chronic non-cancer pain (Stockings et al., 2018). Often, the therapeutic benefits of CB₁R agonists are limited by the risk or emergence of adverse psychoactive effects.

3.3. Endocannabinoids and pain relief

Besides exogenous agonists, pharmacological activation of CB₁Rs can be achieved via increasing eCB levels through inhibiting their respective degradation enzymes, i.e. fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL) for AEA and 2-AG, respectively (Donvito et al., 2018). At CB₁Rs, AEA is a low efficacy agonist, whereas 2-AG is a high efficacy agonist, although AEA has a higher affinity than 2-AG (Narouze, 2020). An electrophysiological study has shown that inhibiting either FAAH or MAGL enhanced GABAergic transmission via CB₁Rs in rat PAG slices (Lau et al., 2014), which can lead to PAG disinhibition, subsequently activating the descending pain inhibitory pathway and inducing analgesia (Guindon et al., 2013; Lau and Vaughan, 2014). Several in vivo preclinical studies also confirmed the analgesic effect of eCB degradation inhibitors in various animal models of inflammatory and neuropathic pain, as reviewed previously (Donvito et al., 2018; Hossain et al., 2020). It is noteworthy that majority of these preclinical studies demonstrated that eCB degradation inhibitors possess antiallodynic and/or anti-hyperalgesia effect in inflammatory and neuropathic pain models (Donvito et al., 2018; Hossain et al., 2020), insufficient finding are available in the literature to demonstrate whether both classes of eCB degradation inhibitors (FAAH and MAGL inhibitors), can enhance pain threshold in naïve animals, similar to opioids.

Nonetheless, similar to exogenous CB₁R agonists, repeated CB₁R-activation by eCBs, especially by 2-AG, can also cause analgesic tolerance. This was used to explain the underlying cause of analgesic tolerance observed in repeated administrations of MAGL inhibitors, such as JZL184 and KML29 (Crowe et al., 2017; Schlosburg et al., 2010), due to the eCB overload phenomenon (Lichtman et al., 2010). Repeated administrations of a low dose of JZL184, which was expected to modestly increase level of 2-AG and activate CB₁Rs in a manner similar to low doses of THC, seems to induce analgesia without tolerance in mice (Kinsey et al., 2013). On the other hand, although repeated administration of exogenous AEA caused analgesic tolerance (Welch, 1997), repeated administration of its degradation enzyme (FAAH) inhibitor generally does not elicit analgesic tolerance (Kiso et al., 2020; Slivicki et al., 2018a; Slivicki et al., 2019), but see Okine et al. (2012).

To date, the successful application of FAAH and MAGL inhibitors in pain management is at the preclinical stage only. An earlier clinical study with an FAAH inhibitor, PF-04457845, did not show significant analgesia in phase II trial in patients with osteoarthritic pain (Huggins et al., 2012). The phase I clinical trial of another FAAH inhibitor, BIA 10–2474, was terminated early due to flaws in research design and a likely off target effect (Kaur et al., 2016). A clinical study of an MAGL inhibitor, ABX-1431, for experimental pain and peripheral neuropathy is completed, but the results have not been revealed (Deng and Li, 2020). Further clinical studies may be planned to investigate the efficacy of FAAH and/or MAGL inhibitors in humans.

Recently, Hohmann’s group identified a novel pharmacological approach to enhance eCB-CB₁R-mediated analgesia without tolerance by using a positive allosteric modulator (PAM) of CB₁Rs. This PAM retained its analgesic efficacy after chronic administration without tolerance in animal models of inflammatory pain and chemotherapy-induced neuropathic pain (Slivicki et al., 2018b). This pharmacological manipulation has the potential to enhance eCB-CB₁R transmission at low or normal levels of eCBs and thus minimize the occurrence of CB₁R desensitization due to eCB overload. However, further studies will be needed to determine its role in managing chronic pathological pain in humans.

4. Opioid-sparing effects by cannabinoids in opioid-tolerant subjects

4.1. Synergistic analgesia between opioids and cannabinoids

Concurrent activation of MORs and CB₁Rs was found to synergistically induce analgesia in preclinical studies, where co-administration of THC with opioids, including morphine and codeine, was found to increase the potency of opioids by 1.2–24.8 folds, as compared with the group co-administered with the vehicle of THC (Cichewicz, 2004). Thus, co-administration of opioids with cannabinoids is considered to be a potential strategy to reduce the amount of opioid needed to produce analgesia, i.e., provide an opioid-sparing effect, in chronic opioid users. If co-administration of a cannabinoid reduces the amount of opioid needed for satisfactory analgesia, this may be a way to decrease the development of opioid analgesic tolerance. In this section, we will discuss the preclinical and clinical evidence of the opioid-sparing effect of cannabinoids.

In rodents, systemic administration of a CB₁R agonist together with morphine can induce synergistic analgesia in neuropathic (Kazantzis et al., 2016) and inflammatory (Chen et al., 2019) pain models. Interestingly, several studies also showed synergistic analgesia when morphine was co-administered with an ineffective dose of cannabinoids (Alsalem et al., 2019), or when both morphine and cannabinoids were at sub-effective doses (Smith et al., 2007). The PAG is an important site of action for this synergistic analgesia. Repeated intra-PAG (i.pag.) microinjections of a CB₁R agonist enhanced the analgesic efficacy of subsequent i.pag. morphine injection without producing analgesic tolerance (Wilson et al., 2008). Thus, the synergistic analgesic effect of cannabinoids with low doses of opioids may be utilized to reduce opioid tolerance. Interestingly, a continuation study by the same group showed that the analgesic synergism between CB₁R and MOR agonists in rat PAG is bidirectional (Wilson-Poe et al., 2013). Nevertheless, an asymmetrical analgesic synergism was reported when CB₁R and MOR agonists were systemically administered (Vigano et al., 2005). In morphine (i.p.)-tolerant rats, i.p. injection of a CB₁R agonist can induce analgesia, but not vice versa. The synergistic analgesia produced by opioids and cannabinoids was also reported in nonhuman primates (Nilges et al., 2019), a preclinical model known to closely mimic human responses.

However, the outcomes of clinical studies examining cannabinoid/opioid interactions are not as consistently promising as expected from preclinical studies (Nielsen et al., 2017). In healthy volunteers, cannabis enhances the analgesic effects of sub-threshold oxycodone (Cooper et al., 2018). However, coadministration of dronabinol with oxycodone not only did not enhance oxycodone-induced analgesic effect, but it also increased abuse- and impairment-related subjective responses in healthy volunteers (Babalonis et al., 2019). In patients suffering from chronic non-cancer pain, indeed some studies showed a positive correlation of enhanced opioid analgesia with cannabis co-administration (Degenhardt et al., 2015) and some patients consider using cannabinoids as a substitution for opioids (Takakuwa and Sulak, 2020). However, in cancer pain patients with optimized opioid therapy, add-on administration with a cannabinoid did not show superior pain relief as compared with placebo (Lichtman et al., 2018). Meta-analysis studies also indicated that no significant opioid reduction was achieved by medical cannabis or CB₁R agonists in patients with postoperative pain (Abdallah et al., 2020), non-cancer chronic pain (Okusanya et al., 2020) and cancer-related chronic pain (Boland et al., 2020). Furthermore, some clinical studies indicated that cannabinoid treatment is associated with higher pain intensity (Liu et al., 2019), increased opioid consumption (Bhashyam et al., 2018), and enhanced opioid abuse potential (Babalonis et al., 2019) in patients with acute or postoperative pain. Furthermore, a history of cannabis use can increase the tendency of inpatient opioid use (Dalal et al., 2020).

4.2. Synergistic analgesia between opioids and eCBs

As for the potential synergistic effect of eCBs on opioid analgesia, either through inhibiting eCB degradation or CB₁R PAMs, so far, only preclinical studies are available. When administered alone, an FAAH inhibitor or MAGL inhibitor, which was known to increase AEA or 2-AG levels, significantly inhibited chemotherapy-induced (Guindon et al., 2013) and chronic constriction injury (CCI)-induced (Kinsey et al., 2009) neuropathic pain in mice, via CB₁Rs as analgesia was prevented by pretreatment with a CB₁R antagonist. Table 1 compiles the available reports on the opioid-sparing and analgesic tolerance-preventive effects of eCB degradation enzyme inhibitors. Co-administration of a selective MAGL inhibitor with a sub-effective dose of morphine was synergistic in eliciting analgesia in formalin-induced inflammatory pain in rats (Clapper et al., 2018) or in CCI-induced neuropathic pain in mice (Wilkerson et al., 2016). Interestingly, repeated co-administrations of these agents did not show analgesic tolerance (Wilkerson et al., 2016). A similar opioid-sparing effect was observed in rodents co-treated with morphine and a selective FAAH inhibitor (Hasanein and Ghafari-Vahed, 2016; Slivicki et al., 2018a) or a dual FAAH/MAGL inhibitor (Wilkerson et al., 2017). Repeated co-administrations of an FAAH inhibitor (Fotio et al., 2020; Hasanein and Ghafari-Vahed, 2016) or a dual inhibitor (Wilkerson et al., 2017) with morphine also prevented or attenuated morphine analgesic tolerance.

Table 1.

Synergistic effects on opioid analgesia induced by endocannabinoids.

Pain Model	Animal	eCB enhancer		Interaction with morphine		Reference
		AEA	2-AG
		FAAH inhibitor	MAGL inhibitor	Acute morphinea (Dose)	Repeated morphineb (Dose, days)
Inflammatory pain (formalin test)	SD rat		ABD-1970	Synergistic analgesia (2.49 mg kg−1, s.c.)		Clapper et al., 2018
Neuropathic pain (CCI)	C57BL/6J mouse		MJN110	Synergistic analgesia (0.82 mg kg−1, i.p.)	Reduced tolerance (0.82 mg kg−1, 2× day−1, 6 days)	Wilkerson et al., 2016
Acute pain (tail-flick)	Wistar rat	URB597			Reduce tolerance (10 mg kg−1, 2× day−1, 7 days)	Hasanein and Ghafari-Vahed, 2016
Neuropathic pain (chemotherapy)	C57BL/6J mouse	URB597 URB937		Reduction of morphine ED50		Slivicki et al., 2018a
Neuropathic pain (CCI)	C57BL/6J mouse	SA-57 (dual inhibitor)		Reduction of morphine ED50	Without tolerance (1.12 mg kg−1, 2× day−1, 5 days)	Wilkerson et al., 2017
Acute pain (tail-immersion)	CD1 mouse	URB597		No interaction (15 mg kg−1, s.c.)	Reduced tolerance (15–30 mg kg−1, 2× day−1, 7 days)	Fotio et al., 2020
		URB937		No interaction (15 mg kg−1, s.c.)	No interaction (15–30 mg kg−1, 2× day−1, 7 days)
Neuropathic pain (chemotherapy)	C57BL/6J mouse	GAT211 (CB1R PAM)		Reduction of morphine ED50	Without tolerance (10 mg kg−1, 1× day−1, 20 days)	Slivicki et al., 2020

^aThe acute dose of morphine when co-administrated with the eCB enhancer.

^bThe repeated doses of morphine and treatment duration when co-administrated with the eCB enhancer. URB597 is a BBB-permeable and URB937 is a BBB-impermeable FAAH inhibitors. SA-57 is an FAAH and MAGL dual inhibitor. 2-AG: 2-arachidonoylglycerol; AEA: anandamide; CB1R PAM: CB1 receptor positive allosteric modulator; CCI, chronic constriction injury; eCB, endocannabinoid; FAAH, fatty acid amide hydrolase; i.p., intraperitoneal injection; MAGL, monoacylglycerol lipase; s.c., subcutaneous injection; SD rat, Sprague Dawley rat.

Fig. 1 (A & B) depicted a working model of the possible signaling pathway in the PAG explaining how eCB degradation enzymes restore analgesia in opioid-tolerant conditions. Briefly, downregulation of MORs in GABA interneurons of the PAG following repeated exposure to exogenous opioids, causes analgesic tolerance as it impairs opioid-induced disinhibition signaling onto the projection neuron that is required to activate the descending inhibitory pathway (Bagley and Ingram, 2020; Bouchet and Ingram, 2020). Pharmacological inhibition of MAGL and/or FAAH, via their respective inhibitors, causes an accumulation of eCBs that can activate the CB₁Rs on GABA interneurons, thus restoring the disinhibition signaling in the PAG (Fig. 1A). On the other hand, a recent study demonstrated that a CB₁R PAM co-administered with morphine augmented the analgesic effect of morphine and prevented the development of morphine tolerance (Slivicki et al., 2020). This may be attributed to an enhancement of the effect of eCBs on presynaptic CB₁Rs on GABAergic terminals of PAG neurons, leading to the restoration of PAG disinhibition, which was impaired during opioid-tolerant conditions (Fig. 1B). Taken together, similar to preclinical studies of CB₁R agonists, eCB degrading enzyme inhibitors and a CB₁R PAM showed convincing analgesia and attenuated opioid tolerance, but further clinical studies will need to be conducted to provide translational validity of their pharmacotherapeutic potentials.

Fig. 1. A working model of possible interactions between opioids and endocannabinoids in the periaqueductal gray (PAG) in control (A) and opioid-tolerant conditions during pharmacological (B) and peripheral neuromodulation (C) interventions.

(A) Under control condition, the glutamatergic projection neuron (grey neuron) in the PAG receives GABA (green circles) inhibitory transmission via GABA_A receptors (green receptors). The inhibition of GABA release from the presynaptic GABA interneuron (green neuron) (disinhibition) can be achieved via presynaptic activation of either μ-opioid receptors (MORs, yellow receptors) by endorphins (yellow circles) or CB1 receptors (CB₁Rs, pinkish-purple receptors) by endocannabinoids (eCBs), i.e., 2-arachidonoylglycerol (2-AG, purple circles) or anandamide (AEA, pink circles). The levels of 2-AG and AEA are regulated by their respective degradation enzymes, monoacylglycerol lipase (MAGL, purple circular sectors in the presynaptic GABAergic terminal) or fatty acid amide hydrolase (FAAH, pink circular sectors in the postsynaptic neuron). (B) During opioid-tolerant conditions, the expression of MORs on GABA interneurons is downregulated due to repeated treatments with exogenous opioids, e.g., morphine (yellow triangles), leading to a reduced inhibition of GABAergic transmission by opioids, i.e., reduced disinhibition. This reduction of disinhibition on the postsynaptic glutamatergic projection neuron can possibly be restored by pharmacological inactivation of eCB degradation enzymes with an MAGL inhibitor (red cross) and/or an FAAH inhibitor (black cross), or by a CB₁R positive allosteric modulator (PAM). (C) On the other hand, the reduction of disinhibition on the postsynaptic projection neuron during opioid-tolerant conditions can also be achieved via peripheral neuromodulation through an opioid-independent and cannabinoid-dependent mechanism. That is, electrostimulation at peripheral nerves, e.g., the median nerve, can lead to activation of the Gq protein-coupled receptors (G_qPCRs) by increasing the release of G_qPCR ligands (e.g., orexins or glutamate) in the PAG. Via the phospholipase C (PLC)-diacylglycerol lipase (DAGL) pathway, 2-AG can be synthesized and produce retrograde inhibition of GABA release via presynaptic CB₁Rs. The images of neurons, ligands and receptors are adapted from Illustration Toolkit Neuroscience by Motifolio.

It is noteworthy that the abovementioned preclinical studies on the analgesic effect of cannabinoids and eCBs use acute neurogenic pain (thermal or mechanical), inflammatory pain or chronic neuropathic pain (surgical-, chemotherapy-, or diabetic induced) models. Although they are often assumed to be translationally relevant for clinical pain, they may not adequately mimic the multidimensionality of clinical pain conditions, e.g., the motor, neurological and psychological complications (Finn et al., 2021) associated with chronic pain in humans. Further studies are required to collectively examine the potential effects of cannabinoids and eCBs on the full multidimensionality of clinical pain conditions in humans.

5. Peripheral neuromodulation and chronic pain relief

5.1. Peripheral neuromodulation

The search for pharmacological replacements for opioid analgesics is regarded as an important component to resolve the opioid crisis (Yaksh et al., 2018). However, the financial and time costs of novel drug development are considerable barriers to a rapid solution. The US National Institute of Health has recognized that nonpharmacological treatments of pain may be an alternative approach to address chronic pain as well as the opioid crisis (Abbasi, 2018). Peripheral neuromodulation may be one of the feasible options.

The main principle of peripheral neuromodulation involves the application of electrical stimulation on the peripheral nerves to achieve therapeutic effects, including analgesia (Slavin et al., 2015). In an early implementation, peripheral modulation was achieved via peripheral nerve stimulation (PNS), where large electrodes were implanted next to the target nerves via open surgery (Slavin et al., 2015). Gradually, several less invasive procedures of peripheral neuromodulation were introduced, including transcutaneous electrical nerve stimulation (TENS), percutaneous electrical nerve stimulation (PENS), and ultrasound-guided percutaneous PNS (pPNS). TENS involves delivering transcutaneous electrical stimulation through intact skin to activate the underlying nerve fibers (Teoli and An, 2021). When TENS is applied to an acupoint is referred as transcutaneous acupoint electrical stimulation (TEAS) (Liang et al., 2019). PENS is conducted via inserting acupuncture-like needles into the dermatome(s) of the target peripheral nerve (Ghoname et al., 1999). The later ultrasound-guided pPNS was developed by percutaneously inserting stimulating electrodes to the target nerve guided under ultrasound, utilizing similar principles as conventional PNS while avoiding an invasive open surgery (Gabriel and Ilfeld, 2021).

5.2. Analgesic mechanisms of peripheral neuromodulation

5.2.1. Opioid-dependent

Since the first report of peripheral nerve stimulation for pain management in 1967 (Wall and Sweet, 1967), the Gate Theory of Pain or Gate Control Theory is attributed as the underlying mechanism (Melzack and Wall, 1965). This theory posits that nociceptive transmission via Aδ and C fibers can be interfered with by non-painful stimuli on Aβ fibers, with the spinal dorsal horn as the “gate” to suppress pain transmission to the brain. However, the gate theory cannot explain the systemic analgesia or the pain relief at the location distal to the stimulating point of peripheral neuromodulation (Gozani, 2019). Thus, mechanisms other than the gate theory may be involved.

Endogenous opioids are one of the candidates that contribute to peripheral neuromodulation-induced analgesia. PNS can increase β-endorphin levels in human cerebrospinal fluid (CSF) (Clement-Jones et al., 1980; Salar et al., 1981). Direct PAG electrostimulation can also induce β-endorphin elevation in the CSF (Amano et al., 1980), which contributes to SPA in humans (Hosobuchi et al., 1979). Thus, it is likely that PNS can activate the opioid system in the brain, releasing β-endorphin that subsequently triggers the PAG-RVM-spinal descending inhibitory pathway. Indeed, Sluka’s group reported that peripheral neuromodulation-induced analgesia can be reversed by blocking opioid receptors at PAG (DeSantana et al., 2009), RVM (Kalra et al., 2001) and spinal (Sluka et al., 1999) levels in rats with inflammatory pain. Preclinical reports on naloxone-sensitive peripheral neuromodulation are also available in the literature (Chen et al., 1996; Jorum and Shyu, 1988).

5.2.2. Opioid-independent

The opioid-dependent analgesic effects induced by peripheral neuromodulation or SPA, similar to morphine-induced analgesia, also develop tolerance after repeated stimulations, and displayed cross-tolerance to morphine in humans (Leonard et al., 2011; Young and Chambi, 1987) and rats (Mayer and Hayes, 1975). Nonetheless, numerous longitudinal studies have found that peripheral neuromodulation can surprisingly induce long-term analgesic effects, ranging from months to decades (Cohen et al., 2019; Johnson and Goebel, 2016; Kupers et al., 2011). Thus, tolerance-prone opioid signaling is unlikely to be the underlying mechanism. Indeed, the SPA induced by PAG stimulation in rats was only partially reversed by naloxone (Akil et al., 1976; Morozova and Zvartau, 1986) at the dose that completely reversed morphine-induced analgesia, suggesting an involvement of a non-opioid mechanism in the PAG, which is supported by a later clinical study (Young and Chambi, 1987).

Similarly, the evidence supporting that a naloxone-insensitive mechanism can contribute to peripheral neuromodulation-induced analgesia has been reported in humans since 1981. In patients with primary dysmenorrhea, Walker and Katz (1981a) found that repeated electrostimulation at the radial, median and/or saphenous nerves exerted a prolonged systemic analgesic effect that was naloxone-insensitive and did not show cross-tolerance with morphine. Their next publication in the same year emphasized that the same mode of PNS did not develop analgesic tolerance after repeated treatments (Walker and Katz, 1981b). The authors claimed that this was the first clinical evidence that described a “non-opioid pathway can produce lasting pain relief in patients with severe clinical symptoms” (Walker and Katz, 1981a).

The studies conducted in nonhuman primates by Willis’s group also demonstrated the involvement of a non-opioid mechanism in the analgesia induced by peripheral neuromodulation. For instance, electrostimulation at the tibial, sciatic or median nerve significantly suppressed neuronal activity in the spinothalamic tract (STT) activated by painful sural nerve stimulation, suggesting a central participation in peripheral neuromodulation (Chung et al., 1984b). Naloxone slightly but significantly reversed the suppression induced by tibial nerve stimulation on the elicited STT neuronal activity (Chung et al., 1984a). Furthermore, TENS at the hindlimb nerves also inhibited STT neuronal activity and induced analgesia (Foreman et al., 1975) in a manner not reversed by naloxone (Lee et al., 1985). These data suggested an opioid-independent mechanism contributes to peripheral neuromodulation-induced analgesia.

5.2.3. Involvement of eCBs

Among reported analgesic mechanisms other than opioids, the eCB system in the PAG is the most likely candidate involved in peripheral neuromodulation-induced analgesia. The first report indicating that the PAG eCB system can be activated by peripheral neuromodulation was demonstrated by Longhurst’s group in a study investigating the cardiovascular suppressive effect induced by peripheral neuromodulation in rats. They found that electrical stimulation at PC5 (Jianshi) and PC6 (Neiguan) acupoints, i.e. electroacupuncture at PC5-PC6, reduced GABA levels in the PAG in a manner reversed by a CB₁R antagonist (Fu and Longhurst, 2009). PC5 and PC6 acupoints are known to overlie the median nerve (Joo Oh et al., 2012). Chiou’s group subsequently substantiated that median nerve stimulation (MNS) can trigger a CB₁R-mediated inhibition of GABA release, i.e. disinhibition, in the PAG (Chen et al., 2018). That is, MNS at the PC6 acupoint (MNS-PC6) significantly suppressed the hot-plate nociceptive response in normal mice and mechanical allodynia in CCI-mice, via a CB₁R-mediated disinhibition in the PAG through an endogenous orexin-initiated eCB signaling (Chen et al., 2018). This will be discussed in section 6. The same study also indicated that MNS-PC6-induced analgesia is naloxone-insensitive, in agreement with earlier studies in humans (Walker and Katz, 1981a; Walker and Katz, 1981b) and nonhuman primates (Lee et al., 1985) that found an opioid-independent mechanism underlying this form of peripheral neuromodulation-induced analgesia.

Besides MNS, stimulation of other peripheral nerves by the electroacupuncture procedure can also induce analgesia via the CB₁R-mediated disinhibition mechanism in the PAG. A study reported that electroacupuncture at acupoints GB30 (Huantiao) and GB34 (Yanglingquan), a procedure similar to percutaneous sciatic nerve stimulation (Shao et al., 2015), significantly suppressed inflammatory and neuropathic pain responses in mice via CB₁R-mediated inhibition of GABA neurons in the PAG (Zhu et al., 2019). In addition to the PAG, eCB-CB₁R transmission in the periphery and spinal cord are also involved in peripheral neuromodulation-induced analgesia. TENS of the hindpaw of mice was found to induce analgesia, accompanied by increased anandamide and CB₁R expression in paw, spinal, and PAG tissues (de Oliveira et al., 2020).

5.3. Opioid-sparing effect of peripheral neuromodulation providing pain relief in opioid-tolerant subjects

In clinical practice, the treatment duration of prescription opioids in an opioid-naïve postoperative patient strongly correlates with the likelihood to develop opioid dependence, which may lead to unintentional overdose or misuse of opioids. In addition to pharmacological intervention with cannabinoids, peripheral neuromodulation has been demonstrated to have opioid-sparing effects in patients with postoperative pain (Gabriel and Ilfeld, 2021). Table 2 summarizes the available literature on the opioid-sparing effect of peripheral neuromodulation. In patients with varied chronic pain conditions, direct PNS of the forearm nerves (ulnar, median or radial nerves) by implanted electrodes has been reported to remarkably reduce opioid consumption in 23 out of 24 (Strege et al., 1994) and 8 out of 9 (Deer et al., 2010) patients. Recently, clinical reports indicated that percutaneous PNS at the femoral and sciatic regions provided adjunct analgesic effect in patients receiving total knee arthroplasty surgery and led to an earlier cessation of opioid consumption (Ilfeld et al., 2019). Several placebo-controlled clinical studies also demonstrated that the consumption of opioids to relieve postoperative pain was significantly reduced by TENS treatment at the dermatome of the skin incision sites in patients receiving major spinal surgery (Unterrainer et al., 2010) or major gynecological procedures (Hamza et al., 1999; Wang et al., 1997). Interestingly, transcutaneous acupoint stimulation (TEAS) treatments also significantly reduced post-operative opioid consumption with the stimulating acupoints including the LI4 (Hegu) acupoint (Lan et al., 2012; Wang et al., 1997) targeting the radial nerve (Umemoto et al., 2019), the ST36 (Zusanli) acupoint (Chen et al., 1998; Lan et al., 2012) targeting the sciatic nerve (Jung et al., 2018), and the PC6 acupoint (Lan et al., 2012) targeting the median nerve (Chen et al., 1998) (Table 2). Besides acute postoperative pain, the requirement for oral opioids in patients with lower back pain was reported to be relieved by regional PENS or pPNS of the lower back (Kapural et al., 2018). Interestingly, a reduction in opioid consumption by TENS was accompanied by decreased cortisol levels 24 h after surgery (Szmit et al., 2021), suggesting that the TENS procedure reduces stress.

Table 2.

Opioid-sparing effect of peripheral neuromodulation in chronic pain or postoperative pain relief in clinical setting.

Pain types	Subjecta	Target nerve (Acupoint)	Stimulation Mode	Opioid	Opioid Reduction (PostOP time)b	Reference
Chronic peripheral nerve pain	American (24 vs 0)	Ulnar/ Median/ Radial	PNS	Narcotics (Meperidine)	Opioid cessation in 23/24 patients.	Strege et al. (1994)
Chronic pain (Carpal tunnel syndrome)	American (9 vs 0)	Median	PNS	Oral narcotics	Opioid reduction in 8/9 patients.	Deer et al. (2010)
Postoperative (total knee arthroplasty)	American (7 vs 0)	Femoral/ sciatic	pPNS	Oxycodone	Opioid cessation: 45–60 to 6 days.	Ilfeld et al. (2019)
Postoperative (major spinal surgery)	Austrian (14 vs 11)	Dermatome of incision site	TENS	Piritramide	62.1%.	Unterrainer et al. (2010)
Postoperative (major gynaecological procedures)	American♀ (25 vs 25)	Dermatome of incision site	TENS	Morphine	50% (24 hr) 53% (overall).	Hamza et al. (1999)
Postoperative (lower abdominal surgery)	American♀ (25 vs 25)	Radial (LI4)	TEAS	Hydromorphone	34% (24 hr) 46% (overall).	Wang et al. (1997)
Postoperative (total abdominal hysterectomy/ myomectomy surgery)	American♀ (25 vs 25)	Sciatic (ST36)	TEAS /TENS	Hydromorphone	39% (24 hr) 38% (overall).	Chen et al. (1998)
Postoperative (total hip arthroplasty)	Chinese; (30 vs 30)	Median (PC6) /Radial (LI4) /Sciatic (GB31-ST36)	TEAS	Fentanyl	37% (24 hr) 31% (48hr).	Lan et al. (2012)
Postoperative (inguinal hernia repair)	Polish (24 vs 23)	Radial (LI4)/ Dermatome of incision site	TEAS /TENS	Morphine	51.6% (24 hr).	Szmit et al. (2021)

^aThe number of subjects in the treatment and placebo groups, respectively. Both genders were recruited, unless specified.

^bThe opioid reduction percentage 24 hours post-operation or overall. PNS: direct peripheral nerve stimulation; PostOP: post-operation; pPNS: ultrasound-guided percutaneous peripheral nerve stimulation; TEAS: transcutaneous acupoint electrical stimulation at the GB31 (Fengshi), LI4 (Hegu), PC6 (Neiguan) or ST36 (Zusanli) acupoint. TENS: transcutaneous electrical nerve stimulation.

To the best of our knowledge, until now, only one preclinical study has investigated the potential beneficial effect of peripheral neuromodulation in opioid-tolerant animals. Taking advantage of the naloxone-insensitive analgesic effect induced by MNS-PC6 (Chen et al., 2018), Chiou’s group revealed that analgesic tolerance did not develop after repeated MNS-PC6 treatments in mice with neuropathic pain. They also found that MNS-PC6 was able to provide significant analgesia in neuropathic mice that had developed tolerance to escalating doses of morphine (Lee et al., 2021). MNS-PC6-induced analgesia is mediated by an eCB (2-AG)-mediated disinhibition of the PAG, a sequence after activation of orexin 1 receptors (OX₁Rs), a type of G_q protein-coupled receptors (G_qPCRs) (Fig. 1C). This study indirectly supports the notion that the eCBs, which are synthesized on demand and released under optimal spatial (in the synaptic cleft) and temporal (upon OX₁R activation) conditions can induce analgesia without causing eCB overload and tolerance, an important distinction to the effect of pharmacological elevation of eCBs (Lichtman et al., 2010), thus avoiding CB₁R desensitization. Although the mechanistic findings from the preclinical model are somewhat in line with prior clinical reports of peripheral neuromodulation-induced analgesia, it should be noted that these studies may not fully reflect other neurological and psychological aspects of chronic pain and/or opioid use disorders, thus further studies should be conducted.

6. Involvement of the orexin-eCB signaling in the PAG in analgesia induced by peripheral neuromodulation and stress

The orexin system consists of two hypothalamic neuropeptides, orexin-A and orexin-B, and two receptors, OX₁ and OX₂. Orexins are well-known to be involved in arousal, hormonal, metabolic and cardiovascular functions (Li and de Lecea, 2020), and also in pain regulation (Chiou et al., 2010). Using electrophysiological and behavioral approaches, Chiou’s group has reported that the analgesic effect induced by MNS-PC6 is opioid-independent (Chen et al., 2018), as observed in the opioid-independent form of SIA (Lee et al., 2016; Lee et al., 2020). Both modes of analgesia share the same disinhibition mechanism mediated by an endogenous orexin-initiated eCB cascade, as first revealed in the PAG (Ho et al., 2011). This mechanism can lead to analgesia via activating the descending pain inhibitory pathway. As depicted in Fig. 2, when orexin neurons in the lateral hypothalamus are activated by acute stress (Lee et al., 2016) or MNS-PC6 (Chen et al., 2018), orexins are released in the PAG to activate postsynaptic OX₁ receptors (OX₁Rs), a G_qPCR, resulting in the synthesis of 2-AG via a phospholipase C (PLC)–diacylglycerol lipase (DAGL) enzymatic pathway. 2-AG then produces retrograde inhibition of GABA release by activating CB₁Rs on GABAergic terminals, leading to disinhibition of the PAG excitatory neurons that project to the RVM that in turn send inhibitory inputs to the spinal cord, culminating in the activation of the descending pain inhibitory pathway that is constituted by the PAG–RVM–spinal cord circuit and ultimately leading to analgesia (Ho et al., 2011). It is noteworthy that activation of other G_qPCRs in the PAG, e.g. mGlu₅ receptors was shown to cause GABA disinhibition via a similar downstream signaling pathway (Drew et al., 2008) and mediate SIA in mice (Lee et al., 2020). Further studies will need to be carried out to discern their possible involvement in peripheral neuromodulation-induced analgesia.

Fig. 2. A proposed working model demonstrating the involvement of orexin-initiated endocannabinoid signaling in stress- and peripheral neuromodulation-induced analgesia.

The periaqueductal gray (PAG) is known to be a pivotal brain region mediating the supraspinal analgesic effect of orexins (red circles) as intra-PAG (i.pag.) microinjection of orexin was shown to produce analgesia in rodents via activation of OX₁ receptors (OX₁Rs, red receptor) in the projection neuron (grey neuron). Through the phospholipase C (PLC)–diacylglycerol lipase (DAGL) pathway, activation of OX₁Rs leads to the production of 2-arachidonoylglycerol (2-AG, purple circles), an endocannabinoid (eCB) that travel retrogradely to activate presynaptic CB₁ receptors (CB₁Rs, pinkish-purple receptor) on the terminal of the GABA neuron (green neuron), resulting in decreased GABA (green circles) release, and thus disinhibition of the projection neuron, leading to analgesia. This orexin-initiated eCB signaling in the PAG was reported to mediate opioid-independent stress- and peripheral neuromodulation-induced analgesia, via activating orexin neurons (red neuron) in the lateral hypothalamus (LH). The numbers in square brackets refer to the numbers assigned to the supporting studies depicted below the schemas. The images of neurons, ligands and receptors are adapted from Illustration Toolkit Neuroscience by Motifolio.

7. Conclusions and future perspectives

Opioid analgesics are the “gold standard” in pain management and remain irreplaceable, although their chronic clinical use is limited by several unwanted side effects, especially the analgesic tolerance that leads to a dose escalation and ultimately increases the risk of respiratory depression. Therapeutic interventions to delay the emergence of opioid tolerance or maintain the analgesic efficacy after repeated opioid dosing may be achieved by activating the cannabinoid system in the descending pain inhibitory pathway. Mechanistic studies in laboratory animals support this notion, including the finding that exogenous cannabinoids (Chen et al., 2019) or inhibitors of eCB degradation (Wilkerson et al., 2017; Wilkerson et al., 2016) have an opioid-sparing effect by preventing opioid analgesic tolerance. However, inconclusive findings were reported in the clinical setting (Le Foll, 2021), and often the analgesic benefits of cannabinoids were masked by their neurocognitive side effects (Yanes et al., 2019). As direct CB₁R activation is subject to tolerance as well, the cannabis-use history and the amount of cannabinoid intake may need to be established to ensure a higher success rate in clinical trials. Recent advancement in cannabinoid pharmacology is the introduction of CB₁R PAMs, which have been reported to be devoid of analgesic tolerance and cannabimimetic side effects, such as hypothermia and catalepsy in rodents (Ignatowska-Jankowska et al., 2015; Slivicki et al., 2018b). Thus, CB₁R PAMs seem to be a promising therapeutic agent for opioid-independent chronic pain control, provided that the efficacy is translated in clinical studies. Furthermore, their safety pharmacology and toxicology profiles should be established to ensure an adequate therapeutic window when used for chronic pain management.

On the other hand, peripheral neuromodulation via activating the eCB system in the descending pain pathway may be a potential alternative nonpharmacological option for chronic pain management. Since its clinical introduction in the 1960s, peripheral nerve stimulating devices have evolved, due to the advancement of biomedical technology, from surgically implanted electrodes to minimally invasive miniature implants (Banks and Winfree, 2019), and even more recently simplified to non-invasive wearable devices (Kong and Gozani, 2018). Thus, peripheral neuromodulation has become easily accessible. Extensive mechanistic studies (Chen et al., 2018; Lee et al., 2021) in animals have supported the involvement of endocannabinoids in the analgesic mechanism of peripheral neuromodulation by MNS. It can be speculated that this mode of opioid-independent analgesic management may be easily achieved by wearable devices in patients with opioid tolerance, providing satisfactory clinical outcomes, and thus may be of great benefit in palliative care or reducing opioid use and lessening the opioid crisis. Although clinical case studies and animal studies both showed similar efficacy in pain suppression, larger-scale randomized control trials for peripheral neuromodulation in opioid-tolerant patients should be conducted to discern the efficacy in this unique patient population.

Clinically, prescription opioids and chronic pain are both reported to be associated with depression and anxiety (Rosoff et al., 2021). The opioid and endocannabinoid systems in periaqueductal gray are known to, at least in part, be involved in these neuropsychiatric disorders. Furthermore, other domains of opioid use disorders, such as dependence, withdrawal, rewards, etc. have been demonstrated to be subject to cannabinoid-opioid interactions (Mohammadkhani and Borgland, 2020; Norris et al., 2019). Nonetheless, the neuropsychiatric interactions between opioid and cannabinoid systems would be another interesting and crucial topic of review.

Abbreviations:

2-AG

2-arachidonoylglycerol

BBB

blood brain bar

CB1R

CB1 receptor

CB2R

CB2 receptor

CCI

chronic constriction injury

CSF

cerebrospinal fluid

DAGL

diacylglycerol lipase

DOR

δ-opioid receptor

eCB

endocannabinoid

i.p.

intraperitoneal injection

FAAH

fatty acid amide hydrolase

GIRK

G protein-coupled-inward rectifying K+ channels

G_qPCRs

G_q protein-coupled receptors

KOR

κ- opioid receptor

LH

lateral hypothalamus

MAGL

monoacylglycerol lipase

MOR

μ-opioid receptor

MNS

median nerve stimulation

MNS-PC6

median nerve stimulation at PC6 acupoint

NOR

NOP receptor

OX₁R

OX₁ receptor

PAG

periaqueductal gray

PAM

positive allosteric modulator

PLC

phospholipase C

PNS

direct peripheral nerve stimulation

PostOP

post-operation

pPNS

ultrasound-guided percutaneous peripheral nerve stimulation

RVM

rostral ventromedial medulla

s.c.

subcutaneous injection

SIA

stress-induced analgesia

SPA

stimulation-produced analgesia

STT

spinothalamic tract

TEAS

transcutaneous acupoint electrical stimulation at acupoint

TENS

transcutaneous electrical nerve stimulation

THC

Δ9-tetrahydrocannabinol