Microglial Cannabinoid CB2 Receptors in Pain Modulation

Pain, especially chronic pain, can strongly affect patients’ quality of life. Cannabinoids ponhave been reported to produce potent analgesic effects in different preclinical pain models, where they primarily function as agonists of Gi/o protein-coupled cannabinoid CB1 and CB2 receptors. The CB1 receptors are abundantly expressed in both the peripheral and central nervous systems. The central activation of CB1 receptors is strongly associated with psychotropic adverse effects, thus largely limiting its therapeutic potential. However, the CB2 receptors are promising targets for pain treatment without psychotropic adverse effects, as they are primarily expressed in immune cells. Additionally, as the resident immune cells in the central nervous system, microglia are increasingly recognized as critical players in chronic pain. Accumulating evidence has demonstrated that the expression of CB2 receptors is significantly increased in activated microglia in the spinal cord, which exerts protective consequences within the surrounding neural circuitry by regulating the activity and function of microglia. In this review, we focused on recent advances in understanding the role of microglial CB2 receptors in spinal nociceptive circuitry, highlighting the mechanism of CB2 receptors in modulating microglia function and its implications for CB2 receptor- selective agonist-mediated analgesia.


Introduction
Chronic pain is a complex web of emotional experiences and subjective senses that brings patients enormous physical and psychological burdens. Due to the current unmet demand for pain relief, attention has been highly focused on effective analgesics without major central adverse effects. The endocannabinoid system serves as an important neuromodulator signaling, pathway that is deeply involved in the complex modulation of endogenous homeostasis and a variety of pathological processes. Currently, in parallel to the great interest in the endocannabinoid system, cannabinoid compounds, including endocannabinoids, plant-derived, and synthetic cannabinoids, are increasingly recognized as potential therapeutic alternatives for pathological pain, because of their vital roles in modulating nociceptive information processing [1,2].
The biological effects of these cannabinoids are mainly mediated by two endocannabinoid receptors, cannabinoid receptors 1 (CB 1 R) and 2 (CB 2 R). However, currently available cannabis-based medicines are largely targeting CB 1 R, which is detected abundantly in both the peripheral and central nervous systems (PNS and CNS, respectively). Their activation can result in a spectrum of adverse effects, such as the development of tolerance [3], addiction, and psychotomimetic effects [4], thus limiting their therapeutic potential. Subsequently, several strategies have been developed for reducing the central side effects of cannabinoid compounds acting as analgesics, including peripherally restricted CB 1 R agonists, topically applied cannabinoids, cannabinoid metabolic enzyme inhibitors, bifunctional cannabinoids ligands, as well as selective CB 2 R agonists. For example, previously Figure 1. Components of the endocannabinoid system are involved in the main routes of biosynthesis, action, and degradation of endocannabinoids in the nervous system. 2-AG is mainly produced from the hydrolysis of DAG, mediated by two diacylglycerol lipases DAGLα/β. DAG is derived from phosphatidylinositol trisphosphate (PIP 2 ), hydrolyzed by PLC. Most AEA appears to be derived from its membrane precursor, NAPE, which is produced by N-acyltransferase (NAT) using phosphatidylethanolamine (PE) and phosphatidylcholine (PC). NAPE can be hydrolyzed by a specific phospholipase D (NAPE-PLD). Microglia may be the primary cellular source of 2-AG and AEA in neuroinflammatory conditions, as they are capable of producing 20 times more endocannabinoids than other glial cells and neurons. AEA and 2-AG benefit from their strong lipid solubility and can be released into the intercellular space through the cell membrane soon after production. AEA mainly plays a role by activating CB 1 R expressed on the presynaptic membrane and postsynaptic membrane. 2-AG can not only activate CB 1 R, but also activate CB 2 R expressed on microglia. After performing their functions, endocannabinoids undergo re-uptake into the neurons and microglia by membrane transporters and are hydrolyzed by different enzymes. 2-AG is degraded by MAGL, ABHD-6, ABHD-12, or COX-2 into arachidonic acid, ethanolamine, and glycerol, while AEA is mainly metabolized by FAAH or COX-2 into arachidonic acid and ethanolamine.
Notably, the spinal cord distribution of the CB 2 R (mRNA or protein) has been identified primarily in microglia, especially activated microglia, and poorly expressed in neurons by the preponderance of evidence using quantitative real-time PCR [6,9], in situ hybridization [28], western blotting [92,93], immunohistochemical staining [17], or single-cell sequencing [60]. Intriguingly, similar to peripheral CB 2 R, which is mainly detected in immune cells, central CB 2 R is predominantly expressed in microglia, which indicates that CB 2 R is critically involved in the modulation of immune-related responses, including immune-neuron interactions and the occurrence of neuroinflammatory and pathological pain [94].

Microglia Express the CB 2 R-Related Functional Endocannabinoid System
In acting as the central custodians protected by the blood-brain barrier, microglia are derived from primitive macrophages that emanate from the embryonic yolk sac and invade the CNS via the circulatory system during development [95,96]. Over recent decades, a large amount of strong evidence has supported the hypothesis that microglia-dependent phagocytosis/degradation and neuroimmune response critically contributes to the alterations in synaptic remodeling and signaling pathways for chronic pain development [33,34,97,98]. In order to maintain stability in an ever-changing environment, microglia adopt different phenotype states characterized by distinct morphological types to perform their functions.
In addition, after this M1 response, aimed at eliminating noxious stimuli and restraining the initial inflammation from the area, the resolution of the inflammatory process is essential to bypass neurotoxicity and chronic inflammation [103]. For this resolving phase to occur, a shift in microglia phenotype along a spectrum of activation states from proinflammatory to phagocytic anti-inflammatory (M2, alternative activation state) is typically observed, which could increase the production of anti-inflammatory mediators, including transforming growth factor (TGF-β), interleukin 10 (IL-10), IL-13, and IL-4 [104,105]. This process can dampen the pro-inflammatory cytokine levels and promote the resolution of inflammation and the recovery of tissue homeostasis [105,106].
Previous studies in CB 2 R −/− mice have demonstrated that an exacerbated microgliosis occurs in the spinal cord after peripheral nerve injury [29,107], which indicates the relevant role of CB 2 R in controlling microglial proliferation and reactivity. Indeed, accumulating reports have shown that microglia express functional endocannabinoid signaling, which could modulate the activity of microglia to restore CNS homeostasis during pathological neuroinflammatory conditions [108,109]. Firstly, classic studies have shown that cultured microglia from humans [103,110], rats [77,111], or mouse tissues [9,85] and the BV-2 microglia cell line [85], express large amounts of CB 2 R. As previously mentioned in Section 2, CB 2 R is predominantly expressed in activated microglia within the CNS, which is consistent with CB 2 R upregulation in activated macrophages in the peripheral [77,112]. By contrast, the expression of CB 1 R in microglia is ambiguous. Preliminary evidence from early studies suggested that CB 1 R is expressed in culture microglia prepared from mollusks and rodents [111,113], but not in humans [110]. However, it has recently been shown that, unlike Cnr2 (CB 2 R) transcripts, Cnr1 (CB 1 R) expression was not detected in naïve microglia by microfluidic RT-qPCR [114]. These results require further functional and anatomical evidence to investigate whether CB 1 R can regulate microglial function. To date, Agnès Nadjar et al. have provided the only evidence of CB 1 R functional expression on microglia [115].Their results showed that neither male nor female animals exhibited any peculiar behavioral phenotype when CB 1 R was selectively deleted in CX3CR1-positive cells, including motor activity, anxiety levels, learning, and memory. These results suggested that there is a possibility that low expression of CB 1 R in microglia has no clear role on microglia function. Futhermore, Maciej Pietr et al. found that GPR55 is significantly expressed in both the BV-2 cells and primary mouse microglia. Their expression pattern response to the state of microglia is very similar to that of CB 2 R, suggesting it might also be involved in the neuroinflammatory process [116].
Microglia not only express CB 2 R responding to cannabinoid ligands but also produce and inactivate the endocannabinoids AEA and 2-AG by the complex cellular machinery [111,117,118]. The endocannabinoids are highly lipophilic molecules, that are naturally synthesized from glia and neuron membrane phospholipids during physiological events in the CNS. As a result, endocannabinoids are released immediately after their production to stimulate CB 1 R and CB 2 R, and in turn, they are inactivated by re-uptake and the ensuing hydrolysis [119,120]. Notably, under neuroinflammatory conditions in vitro, microglia are capable of producing 20 times more endocannabinoids than other glial cells and neurons. Therefore, microglia may be the primary cellular source of endocannabinoids in vivo [35,85]. Note that the expression of AEA and 2-AG is significantly upregulated when microglia are activated and switch to a protective phenotype (M2) under pathological conditions [68,103,121]. Further studies have demonstrated that the production of 2-AG and AEA is dependent on sustained rises in intracellular Ca 2+ concentration, which may be mediated by the activation of ionotropic P 2 X purinergic receptors P 2 X 4 and P 2 X 7 [118,122]. The molecular mechanism underlying this process involves the fact that an increase in intracellular Ca 2+ can directly increase NAPE-PLD and DAGL activity while inhibiting MAGL activity in microglia [85,123,124]. The inverse sensitivity of DAGL and MAGL to Ca 2+ constitutes a primitive and efficient modality for microglia to continuously accumulate 2-AG [118,125]. Additionally, microglia also express the metabolic enzyme FAAH, which is responsible for the degradation of 2-AG and AEA [123]. Several reports have detailed that under neuropathological conditions, activated spinal microglia may produce significantly more 2-AG and AEA by upregulating DAGL and NAPE-PLD, while downregulating FAAH [103,126,127]. Meanwhile, it has been suggested that the exposure of human or rat microglia to 2-AG and AEA at low concentrations increases the expression of the M2 microglial marker arginase 1 (Arg-1), together with other markers such as suppressor of cytokine signaling 3 (SOCS 3 ) [103]. These results support that the production and actions of endocannabinoids are closely related to the phenotype of microglia, and the endocannabinoid system also plays an important role in microglial immunomodulation by inducing an M2 phenotype. It also explains that some of the neuroprotective effects of the endocannabinoid system in diverse pathological states of inflammatory models may be mediated by their immunomodulatory actions, such as disrupting pro-inflammatory processes [128][129][130].
The increased production of endocannabinoids may also upregulate the expression of microglial CB 2 R, which in turn may activate more CB 2 R to dampen nociceptive signaling cascades, amplifying the anti-inflammatory responses [103]. Notably, the activation of CB 2 R has been reported to transform microglia from the M1 to the M2 phenotype [131,132], suggesting that CB 2 R signaling is very important for microglia to polarize towards the M2 phenotype with phagocytic capacity by morphology alterations [103]. These findings are consistent with previous studies that showed that upregulation of CB 2 R in activated microglia has been associated with improvement of the disease consequences in specific neuroinflammatory conditions [126,133,134].
In summary, the overall consequence of CB 2 R activation on microglia by endocannabinoids AEA and 2-AG or by exogenous cannabinoids appears to be to exert beneficial properties of microglia, such as the release of anti-inflammatory mediators, by promoting the generation of neuroprotective microglia phenotype (M2). The M2 phenotype could reduce neuronal hyperexcitation causally involved in central sensitization, with the capacity of phagocytosis and reduction of releasing detrimental factors like pro-inflammatory cytokines and free radicals [15,124,127,[135][136][137][138][139][140][141]. The expression profiles of CB 2 R and endocannabinoids in homeostatic and activated microglia are summarized in Figure 2.
of microglial CB2R, which in turn may activate more CB2R to dampen nociceptive signal-ing cascades, amplifying the anti-inflammatory responses [103]. Notably, the activation of CB2R has been reported to transform microglia from the M1 to the M2 phenotype [131,132], suggesting that CB2R signaling is very important for microglia to polarize towards the M2 phenotype with phagocytic capacity by morphology alterations [103]. These findings are consistent with previous studies that showed that upregulation of CB2R in activated microglia has been associated with improvement of the disease consequences in specific neuroinflammatory conditions [126,133,134].
In summary, the overall consequence of CB2R activation on microglia by endocannabinoids AEA and 2-AG or by exogenous cannabinoids appears to be to exert beneficial properties of microglia, such as the release of anti-inflammatory mediators, by promoting the generation of neuroprotective microglia phenotype (M2). The M2 phenotype could reduce neuronal hyperexcitation causally involved in central sensitization, with the capacity of phagocytosis and reduction of releasing detrimental factors like pro-inflammatory cytokines and free radicals [15,124,127,[135][136][137][138][139][140][141]. The expression profiles of CB2R and endocannabinoids in homeostatic and activated microglia are summarized in Figure 2. The expression profiles and possible molecular mechanisms of CB2R-related functional endocannabinoid system in homeostatsis and activated microglia in pain processing. When the primary afferent nerve is injured or in a state of chronic pain, the resting microglia will be activated by the mediator released from the central terminal of the primary afferent and transform into pro-inflammatory (M1) microglia. When ATP activates the increased expression of P2X4 and P2X7 on microglia, Ca 2+ enters microglia and regulates the activities of MAGL, DAGL, and NAPE-PLD, which Figure 2. The expression profiles and possible molecular mechanisms of CB 2 R-related functional endocannabinoid system in homeostatsis and activated microglia in pain processing. When the primary afferent nerve is injured or in a state of chronic pain, the resting microglia will be activated by the mediator released from the central terminal of the primary afferent and transform into pro-inflammatory (M1) microglia. When ATP activates the increased expression of P 2 X 4 and P 2 X 7 on microglia, Ca 2+ enters microglia and regulates the activities of MAGL, DAGL, and NAPE-PLD, which lead to increased production and relation of endocannabinoids such as AEA and 2-AG and pro-inflammatory mediators including IL-1β, IL-6, IL-12, IFN-γ, and TNF-α in reactive microglia. This transition was also accompanied by a distinct morphological change in the microglia, from a small soma with long, branched processes to a more amoeba-like shape. At the same time, endocannabinoid such as 2-AG or AEA and exogenous cannabinoids such as AM1241 can act on the increased expression of CB 2 R on microglia. Activation of CB 2 R can inhibit adenylate cyclase (AC), which results in a reduction of intracellular cAMP levels. Diminished cAMP level intracellularly suppresses the activity of PKA and changes the expression of respective ion channels such as P 2 X 4 and P 2 X 7 on microglia, leading to decreased cytosolic Ca 2+ concentration. Changes in Ca 2+ distribution upon CB 2 R stimulation can also regulate the activities and expressions of MAGL, DAGL, FAAH, and NAPE-PLD. Meanwhile, CB 2 R activation is also accompanied by downstream PLC activation through secondary messengers to regulate the activity of the members of the MAPK family, such as ERK 1/2 and p38. As a final consequence, these processes can down-regulate the release of pro-inflammatory cytokines and up-regulate the release of anti-inflammatory cytokines such as IL-4, IL-10, and TGF-β by regulating the activity of different transcription factors, leading to a switch of microglia to an anti-inflammatory phenotype (M2).

Antinociceptive Effects of Well-Characterized CB 2 R Selective Agonists
Currently, intense interest has been focused on the use of cannabinoid compounds typically acting upon the CB 1 R and CB 2 R for the treatment of pathological pain. These widely researched compounds include endocannabinoids like AEA and 2-AG, phytocannabinoids like ∆-9-tetrahydrocannabinol and cannabidiol, as well as a large number of synthetic cannabinoids [20,[142][143][144][145][146]. However, in parallel with a deeper understanding of the endocannabinoid system's underlying expression profiles and the physiological and pharmacological properties of cannabinoid receptors, CB 2 R selective agonist compounds are increasingly recognized as safer novel therapeutic candidates, with their properties to bypass certain centrally mediated unwanted effects associated with the activation of CB 1 R.
The following section summarizes the antinociceptive effects of various wellcharacterized CB 2 R selective agonists, including HU308, JWH-015, JWH-133, GW405833, AM1241, and MDA7, in different pain states to provide direct support for the hypothesis that CB 2 R can serve as a promising therapeutic target for pain relief ( Table 1).
The HU308 is the first CB 2 R selective synthetic compound (K i = 22.7 ± 3.9 nM) that exhibits low affinity for CB 1 R (K i > 10 µM), which exerted anti-inflammatory and peripheral antinociceptive activities in an arachidonic acid-induced mouse inflammatory pain model and the late phase of the mouse formalin pain model. These activities were significantly inhibited by the use of selective CB 2 R antagonist SR144528 [147]. In the rat postoperative pain model, surgical incision-induced tactile allodynia was significantly suppressed by HU308 [148]. The topical HU308 also has been found to reduce corneal hyperalgesia and inflammation in wild-type mice, but not in CB 2 R −/− mice, further validating that CB 2 R is a drug target of this compound [149].
The JWH-015, one of the earliest discovered compounds, was found to have improved selectivity for the CB 2 R (K i = 13.8 nM at CB 2 R and K i = 383 nM at CB 1 R) from the aminoalkylindole classification of CB 2 R agonists, which is effective in alleviating painrelated behaviors and reducing inflammatory responses without inducing psychotropic effects even when intrathecally applied [21,52,[150][151][152]. The intrathecal administration of JWH015 reduced the paw incision and caused postoperative hypersensitivity and microglial activation in the spinal cord without inducing behavioral side effects. This effect was prevented by intrathecal injection of the CB 2 R selective antagonist AM630 [10,153], indicating a CB 2 R-dependent mechanism of action. In spinal nerve ligation (SNL) or lumbar 5 nerve transection (L5NT) neuropathic pain models, intrathecal JWH015 treatment significantly reduced nerve injury-induced hypersensitivity, which can also be blocked by intrathecal AM630 [13,17]. Additionally, it has also been reported that intrathecal or intraperitoneal injection of JWH-015 displayed an analgesic effect to attenuate bone cancer-induced spontaneous pain and mechanical allodynia [15,92,154].
The JWH-133 is also a well-characterized CB 2 R agonist, one of the most highly selective ligands for the CB 2 R (K i = 677 ± 132 nM at CB 1 R and K i = 3.4 ± 1.0 nM at CB 2 R) [155,156], which inhibits both inflammatory and neuropathic hyperalgesia through a CB 2 R-selective mechanism. For example, spinal (i.t.) or local (i.pl.) administration of JWH-133 reduced noxious mechanical stimulation evoked responses in wide dynamic range neurons recorded in SNL neuropathic pain, carrageenan-induced inflammatory pain and osteoarthritis pain, in a manner that was prevented by SR144528 [157][158][159]. The JWH-133, administered systemically (s.c.), can increase weight bearing and decrease peripheral edema or allodynia in the carrageenan-inflamed paw and osteoarthritis pain [160]. However, the JWH-133 has been shown to functionally interact with opioids to modulate antinociception in the formalin test without inducing tolerance, and can also attenuate cross-tolerance with morphine [161]. Furthermore, recent studies using CB 2 R constitutive knockout and tissuespecific genetic deletion mice suggested that self-administration of JWH-133 not only attenuated spontaneous pain and anxiety-associated behavior in the partial sciatic nerve ligation (PSNL) induced neuropathic pain model but also void of reinforcing effects in animals without pain, indicating the absence of abuse liability [14].
The GW405833, another highly selective CB 2 R ligand, is also classified as an aminoalkylindole [162]. In addition, the GW405833 was determined to be a selective human CB 2 R agonist in a recombinant binding assay (K i = 2043 ± 183 nM at CB 1 R and K i = 14 ± 6 nM at CB 2 R), while its selectivity appeared to be lower for the rat CB 2 R (K i = 273 ± 42.6 nM at CB 1 R and K i = 3.6 ± 1.1 nM at CB 2 R) [163,164]. In both rats and mice, pharmacological characterization of GW405833 has been previously shown to elicit efficacious antihyperalgesic and anti-inflammatory effects in several pain models, including PSNL, hind paw incision, and complete Freund's adjuvant (CFA)-induced inflammatory pain, without eliciting the centrally CB 1 R-mediated side effects [163,165].
Similarly, systemic administration of GW405833 reduced the late phase of formalin pain and allodynia elicited by SNL in a dose-dependent manner [9]. Additionally, hind paw incision, chronic constriction injury (CCI)-induced tactile allodynia, and carrageenanevoked peripheral edema or weight bearing were relieved by GW405833 [148,166,167]. These above-mentioned effects were demonstrated to be dependent upon CB 2 R activation rather than the activation of CB 1 R or opioid receptors by the experiments performed in CB 2 R −/− mice or utilizing CB 2 R and opioid receptor selective antagonists. Interestingly, analgesic effects of high-dose GW405833 (i.p. 100 mg/kg) were also evident in the tail flick and hot plate tests in CB 2 R −/− mice, which might be attributed to both moderate affinities for CB 1 R and significant CNS penetration [163,165]. Moreover, in stark contrast to treatment with WIN55,212-22, a mixed CB 1 R/CB 2 R agonist, chronic repeated injection of GW405833 was able to provide sustained reversal of allodynia following SNL without tolerance development [168].
The AM1241, another agonist possessing a high affinity for the CB 2 R (K i = 3.4 ± 0.5 nM at the CB 2 R and K i = 280 ± 41 nM at CB 1 R), belong to the aminoalkylindole class [169]. There is growing evidence supporting the hypothesis that AM1241 produced antinociceptive effects in preclinical inflammatory and neuropathic pain models lacking CNS side effects in a tetrad of behavioral tests that is used to assess cardinal signs of central CB 1 R [169]. The systemic (i.p.) and local (i.pl.) administration of AM1241 exhibited a thermal antinociceptive effect in the acute pain model of rats, which was significantly blocked by the CB 2 R selective antagonist AM630 [19]. These actions of CB 2 R have been confirmed by further studies, in which acute nociception of AM1241 was lost in the tail flick and hot plate tests of CB 2 R −/− mice [22]. In the rat postoperative pain model, AM1241 also obviously suppressed tactile allodynia [148]. Moreover, systemic (i.p.) or local (i.pl.) administration of AM1241 suppressed allodynia, hyperalgesia, and peripheral edema in the carrageenan-evoked rat inflammatory pain model in a CB 2 R-dependent manner because SR144528 or AM630 specifically blocked these effects [170][171][172][173]. Similarly, intravenously administered AM1241 reduced the late phase of formalin pain, which also depends upon CB 2 R activation [9]. In the SNL or CCI of sciatic nerve-induced neuropathic pain models and the CFA-induced chronic inflammatory pain model, AM1241 (i.p., i.DRG., or i.t.) produced a significant reversal of established mechanical and thermal hypersensitivity in rats or CB 1 R −/− mice [6,9,169,174]. The AM1241 could also reduce pain symptoms in a CB 2 R dependent manner in the vincristine-induced neuropathic pain model and bone the cancer-induced pain model [93,175,176]. The MDA7, one of the acylhydrazone derivatives, is a more promising CB 2 R selective agonist (hCB 1 R K i > 10,000 nM; hCB 2 R K i = 422 nM; rCB 1 R K i = 2565 nM; rCB 2 R K i = 238 nM) for the treatment of pain [24,177]. It has been shown that systemic administration of MDA7 exhibits an attenuated SNL-induced tactile allodynia in rats in a dose-dependent way. The target specificity of MDA7 was confirmed by pretreatment with selective antagonists, while attenuation of the antiallodynic effects was mediated by AM630 but not either the CB 1 R selective antagonist AM251 or the opioid antagonist naloxone [24]. This molecule has also been shown to effectively suppress mechanical allodynia rats and mice in paclitaxel (PTX)-induced neuropathic pain models. In addition, MDA7 can produce a modest thermal antinociceptive effect in naive rats without affecting locomotor activity. These effects were blocked after pretreatment with AM630 in wild type mice or were absent in CB 2 R −/− mice, which indicates that the action of MDA7 directly involves the activation of CB 2 R [24][25][26].
As a result of the great potential of targeting CB 2 R, new potential drugs are constantly being developed. Some of them are being tested in clinical trials. For example, Olorinab, an oral and highly selective full agonist of CB 2 R, reached phase II trials for abdominal pain in Crohn's disease and for irritable bowel syndrome [178]. However, there is still no CB 2 R selective agonist on the market as a new analgesic drug. This situation resulted from many reasons. Firstly, most CB 2 R ligands were highly lipophilic and, as such, not optimal for clinical application due to unfavorable physicochemical properties, which potentially contributed to modest or lack of clinical efficacy. Secondly, these compounds will be required to have high affinity and selectivity for CB 2 R to avoid the adverse effects of activating CB 1 R, while the target engagement of current CB 2 R ligands is poor. Aside from the development and optimization of CB 2 R ligands, the fact that human and rodent CB 2 R sequences have relatively low homology should be considered, which may give rise to differences in ligand engagement and efficacy [179]. Furthermore, preclinical pain models in animals might not fully and accurately reflect human pathological mechanisms, which may also affect the clinical translation of CB 2 R agonists. Nonetheless, substantial efforts to better optimize CB 2 R ligands for clinical application are ongoing, and many existing ligands have reached the most advanced phases, such as JBT-101 [180,181]. There is no doubt that specifically activating CB 2 R is considered a good strategy for developing new analgesic agents with fewer side effects.

Molecular Mechanisms Involved in the Action of Microglial CB 2 R in Pain Processing
The dorsal horn of the spinal cord is the vital site for controlling pain intensity because it is where efficient transmission of nociceptive information occurs between the central terminals of primary afferents and second-order interneurons. Additionally, the crosstalk between spinal neurons, astrocytes, oligodendrocytes, and microglia is indispensable to mediating central pain sensitization of neuronal circuits. A considerable amount of evidence has implicated the crucial role of selective agonists of CB 2 R in treating pathologic pain symptoms through the modulation of microglia in a CB 2 R dependent manner [17,25,26,29,177,[182][183][184][185]. Therefore, microglial CB 2 R can serve as a promising therapeutic target for pain relief because of the important role of spinal microglia in regulating central sensitization. Later in the remainder of this review, we will focus on the molecular mechanism of targeting spinal cord CB 2 R to inhibit neuroinflammatory signaling pathways for pain relief with a microglial-centric view.
As described above, reactive microglia express CB 2 R [17,26,184]. The protein and mRNA levels of spinal cord CB 2 R were both significantly upregulated in chronic pain conditions, including CFA-induced inflammatory pain [106], SNI, SNL, CCI, and chemotherapyinduced neuropathic pain [29,30]. The above studies indicate that CB 2 R over-expression in activated microglia at the dorsal horn of the spinal cord under pathological pain conditions may occur as a result of specific neuroinflammatory responses. After that, modulation targeting CB 2 R may result in a neuroprotective effect [186]. Accordingly, the preponderance of evidence has implicated that activation of the CB 2 R system via spinal administration of CB 2 R agonists produces significant control over inflammatory and neuropathic pain in multiple models. For example, intrathecal administration of AM1241 attenuates allodynia or thermal hyperalgesia induced by CFA, SNL [6], CCI, or bone cancer models [93]. The therapeutic utility of other CB 2 R selective agonists that exhibit analgesic effects to treat different pains has been thoroughly described in Section 4.1.
Below, we will summarize our current understanding of the cellular and molecular mechanisms involved in the action of microglial CB 2 R in pain processing. However, several studies have demonstrated that spinal CB 2 R activation limits microglia activity to the ipsilateral dorsal horn, because constitutive knockout of CB 2 R results in a spread of microgliosis to the contralateral dorsal horn in an arthritis model or after sciatic nerve injury [187][188][189]. The possible mechanisms involved in this process are discussed in detail. Firstly, the activation of CB 2 R can inhibit the activities of adenylyl cyclase [190] and MAPK [174,191]. The CB 2 R selective agonists exerted anti-allodynic effects in rats by reducing MAPK (p38 and ERK 1/2 ) phosphorylation and inducing MAPK phosphatases (MKP-1 and MKP-3, the major regulators of MAPKs) expression in the spinal dorsal horn [13]. The downregulation of the p38 MAPK pathway can lead to a reduction of the cytokines IL-1β, TNF-α, and brain-derived neurotrophic factor (BDNF) [174,192,193], and the suppression of ERK 1/2 can decrease microglia proliferation [26,194,195]. Furthermore, the AMPK pathway is upregulated after CB 2 R activation, which can downregulate the synthesis of nitric oxide (NO) [196]. Actually, the activation of CB 2 R in microglial cells has been found to lead to spinal decreased iNOS, IL-6, BDNF, CCR2, and TNFα receptor expression during neuropathic pain [25,197], and increased release of anti-inflammatory cytokines, such as IL-10 [174,198,199]. Specifically, the activation of spinal CB 2 R by exerciseinduced AEA release also reduces the production of IL-1β and TNFα in mice within a carrageenan-induced pain model [136]. This may also be closely related to the reduction of endocannabinoid degradative enzymes at the spinal cord level. It has been shown that intrathecal AM1241 not only modulates critical glial factors but also reduces the expression levels of MAGL, while not altering FAAH [174]. In addition, CB 2 R agonist treatment can reduce microglial purinergic receptor P 2 X 4 upregulation [25], which may be another mechanism by which CB 2 R activation reduces microglial contributions to pain. The P 2 X 4 was identified as a microglia-specific molecule that was activated and upregulated after peripheral nerve injury and also plays critical roles in processing nociceptive information and contributing to microglial-dependent central pain sensitization [200].
In addition to modulating microglial immune function by reducing the production of pro-inflammatory cytokines and increasing the release of anti-inflammatory cytokines involved in neuroinflammatory signaling pathways, activation of CB 2 R also can switch microglia into a more anti-inflammatory state by limiting migration and promoting phagocytic function. For example, it has been found that 2-AG can induce the recruitment of microglia partly by stimulating CB 2 R in BV-2 cells [85]. However, another study showed that CB 2 R activation in microglia stimulates MKP phosphatases, which can inhibit the ERK pathway, thus decreasing microglial chemotaxis/migration mediated by ADP [201]. Moreover, it has been found that the activation of CB 2 R is capable of inducing the removal of native beta-amyloid both in situ and in vitro by promoting the phagocytic function of macrophages [138]. On the contrary, some results also show that the activation of CB 2 R can inhibit the phagocytosis of microglia by activating ERK 1/2 /AKT-Nurr1 signal pathways [202]. These different results of microglial migration or phagocytosis mediated by CB 2 R activation may be due to the different stages of inflammation development. However, as described in Section 3 of this review, the CB 2 R activation can promote the shift of M1 to M2 microglia. All of this may enhance the beneficial properties of microglia and be associate with the restoration of microglial activity.
Overall, accumulating evidence reveals that activation of CB 2 R vitally regulates microglial immune function by blocking the normal inflammatory response with increased production of anti-inflammatory mediators and decreased production of proinflammatory mediators causally involved in central sensitization [29,31,127,135,203,204].

Conclusions
In this review article, we summarize the analgesic effects mediated by CB 2 R and the mechanisms involved in pain regulation. Firstly, it is well known that the endocannabinoid system exerts an important role in neuronal regulation. Within the CNS, CB 2 R mainly expresses in homeostatic microglia, while there is a unique feature that their expression is rapidly upregulated in activated microglia under certain pathological conditions. The CB 2 R might serve as an intriguing target for the development of drugs for the management of pain because of its ability to mediate analgesia with few psychoactive effects. Indeed, accumulating data have demonstrated that the CB 2 R agonists exert analgesic effects in various preclinical pain models, such as inflammatory and neuropathic pain. Additionally, spinal microglia can modulate the activity of spinal cord neurons and have a critical role in the development and maintenance of chronic pain. The activation of CB 2 R can reduce pain signaling by regulating the activity of spinal microglia and inhibiting neuroinflammation. Specifically, the CB 2 R activation has been reported to transform microglia from the pro-inflammatory M1 to the neuroprotective M2 phenotype by promoting the beneficial properties of microglia, such as the releasing of anti-inflammatory mediators, or the induction of phagocytosis, and reducing their ability to release pro-inflammatory cytokines involved in central sensitization. Overall, we provided an improved understanding of the underlying mechanisms involved in the action of microglial CB 2 R in pain processing. However, further studies are needed to dissect the specific role of CB 2 R expressed in different phenotype microglia to provide a better alternative to controlling pain by regulating CB 2 R.
Author Contributions: K.X.: writing-original draft preparation, visualization. and conceptualization; Y.W., Z.T., and Y.X.: writing-review and editing; C.W.: writing-review and editing, supervision, and project administration. Z.W.: writing-review and editing, conceptualization, visualization, supervision, and project administration. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this manuscript. Acknowledgments: Figures 1 and 2 were created with Available online: BioRender.com.

Conflicts of Interest:
The authors declare no conflict of interest. Suppressor of cytokine signaling 3 TGF-β Transforming growth factor TNF-α Tumor necrosis factor α 2-AG 2-arachidonoyl-glycerol