Adenosine Receptors as Potential Therapeutic Analgesic Targets

Pain represents an international burden and a major socio-economic public health problem. New findings, detailed in this review, suggest that adenosine plays a significant role in neuropathic and inflammatory pain, by acting on its metabotropic adenosine receptors (A1AR, A2AAR, A2BAR, A3AR). Adenosine receptor ligands have a practical translational potential based on the favorable efficacy and safety profiles that emerged from clinical research on various agonists and antagonists for different pathologies. The present review collects the latest studies on selected adenosine receptor ligands in different pain models. Here, we also covered the many hypothesized pathways and the role of newly synthesized allosteric adenosine receptor modulators. This review aims to present a summary of recent research on adenosine receptors as prospective therapeutic targets for a range of pain-related disorders.


Pain
Chronic pain is a global burden and a major socio-economic public health problem, profoundly limiting and adversely impacting the quality of life and promoting disability and unnecessary suffering [1]. An estimated 10-30% of people worldwide are affected, with a high societal impact [2,3]. Despite extensive investigational efforts, current treatment approaches (nonsteroidal anti-inflammatory drugs, opioids, antidepressants, or anticonvulsants) for chronic pain are ineffective, inadequate, and/or have intolerable side effects [4][5][6][7]. New drugs and therapies are required because a safe and effective treatment for chronic pain has not yet been developed. Pharmaceutical research focuses on effective analgesic drugs, with G-protein-coupled receptors (GPCRs) being a key focus in pain-relieving medication development [8][9][10]. Among them, adenosine metabotropic receptors (AR) modulators are increasingly being implicated in the antinociceptive and/or antihyperalgesic effects seen in various chronic pain models [2,[11][12][13]. This review aims to describe the most recent developments in our knowledge of the function of ARs in pain. We will review and analyze the preclinical experimental studies that examined the function and mode of action of each receptor subtype in the modulation of acute and chronic pain.

Role of Adenosine in Inflammation and Pain
New findings suggest that adenosinergic modulators have substantial therapeutic potential in chronic diseases, particularly, in inflammation dysregulation situations [14]. In neuropathic and inflammatory pain models, adenosine controls neuronal and non-neuronal cell functions [15], modulates primary afferent transmission, and prevents pain behavior. Furthermore, adenosine exerts multiple effects on the transmission of pain in peripheral and spinal areas and plays an important role in CNS pain processing by regulating excitatory neurotransmission, sustaining neuronal signal transmission, and regulating glial activation and proliferation [7,[15][16][17][18]. Adenosine is an endogenous inflammation modulator that influences almost all physiological and pathophysiological functions [19,20]. Indeed, the adenosinergic system represents an outstanding pharmacological target for pain management, in which inflammation represents a pathogenetic mechanism [11,20]. Indeed, ARs mainly serve as anti-inflammatory receptors since they have an important role in modulating cytokine release and expression as well as in oxidation damage [11]. Adenosine signaling also stimulates the release of neurotransmitters in the brain [21] and contributes to the antinociceptive features of norepinephrine, opioids, caffeine, tricyclic antidepressants, 5-hydroxytryptamine, and transcutaneous electrical nerve stimulation [16]. Adenosine receptor modulators administered spinally or systemically cause antinociception in a large number of studies [12,[22][23][24][25], both in animal models [21,26] and in clinical studies [27,28].
The role of AR activation in inflammatory pain models has been recently investigated [13]. Inflammation can lead to altered pain perception and is mediated by proinflammatory chemokines, reactive oxygen species, and various other mediators and secondary messengers [29]. During this event, also extracellular adenosine concentrations are heightened from multiple sources, including vesicular release and its precursor ATP. Thus, the nonspecific or mechanism-based release of nucleotides, which are later metabolized to adenosine, can indirectly increase the concentrations of adenosine [15]. Adenosine can modulate intrinsic neuronal transmission by increasing K + conductivity and, presynaptically, at sensory nerve endings to prevent the release of substance P and, potentially, glutamate [30,31]. The effects of adenosine on inflammation regulation support the therapeutic application of selective agonists or antagonists capable of activating or inactivating ARs in numerous diseases and conditions, notoriously inflammatory pain.
An initial investigation highlighted the effectiveness of intrathecal (i.t.) adenosine analogs in lowering hypersensitive symptoms [32]. Further research verified their substantial action in nerve damage models, and it has been underlined that adenosine systems may serve as a target in neuropathic pain syndromes. The spinal infusion of adenosine relieved mechanical allodynia after nerve damage, but not in healthy animals [21]; this effectiveness was remarkably long-lasting (i.e., to 24 h; [33]). In addition, when adenosine is administered in an inflamed area, leukocyte rise, activation, and adhesion are reduced [34] ( Table 1).  [72].

Adenosine Metabolism
Adenosine is a ubiquitous but short-lived endogenous modulator [14] whose effects are mostly exerted by the activation of its four G protein-coupled P1 receptor subtypes: A 1 , A 2A , A 2B , and A 3 ARs [47,66,73,74]. The effects of each AR depend on the extracellular adenosine concentrations, which are controlled by different intra-and extra-cellular enzymes involved in its synthesis and degradation, as well as by uptake processes at the membrane level. Using this sophisticated system, the cells can adjust purinergic signaling in response to changes in the health state of a tissue [18]. The ectonucleotidase CD39 converts ATP/ADP to AMP, and subsequently, CD73 converts AMP to adenosine. This sequence is the primary source of adenosine in the extracellular space [75]. Adenosine, in turn, is removed by enzymes that are specifically designed to metabolize it, such as adenosine deaminase (ADA), which converts extracellular adenosine to inosine, and adenosine kinase, which phosphorylates adenosine to form AMP [76]. These latter two intracellular enzymes reduce extracellular adenosine to attenuate AR signaling because these two pools are linked through equilibrative nucleoside transporters (ENTs) and concentrative nucleoside transporters (CNTs) on the cell membrane [77]. Typical extracellular endogenous adenosine concentrations during normal conditions in tissues and organs are in the nanomolar range [78]. However, adenosine is dramatically elevated in the vicinity of damaged cells following trauma or injury, particularly under hypoxic/ischemic insults [79,80], where it can reach micromolar concentrations and consequently activate all AR subtypes [18].

Adenosine Receptors
Various studies probed which ARs, when either activated or blocked, can contribute to an antinociceptive effect. The receptor subtypes differ in the G subunit to which they bind: A 1 AR and A 3 AR are preferentially coupled to Gi proteins, which means that they inhibit the activity of adenylyl cyclase (AC); while A 2A AR and A 2B AR stimulate AC via Gs proteins, resulting in increased cyclic adenosine monophosphate (cAMP) levels [19,47,73]. Adenosine can have various impacts on peripheral pain signals (i.e., normal tissue, inflammation, or following nerve injury), depending on the implicated receptor subtype, its location, and the tissue circumstances ( Figure 1).

Adenosine Receptors
Various studies probed which ARs, when either activated or blocked, can contribute to an antinociceptive effect. The receptor subtypes differ in the G subunit to which they bind: A1AR and A3AR are preferentially coupled to Gi proteins, which means that they inhibit the activity of adenylyl cyclase (AC); while A2AAR and A2BAR stimulate AC via Gs proteins, resulting in increased cyclic adenosine monophosphate (cAMP) levels [19,47,73]. Adenosine can have various impacts on peripheral pain signals (i.e., normal tissue, inflammation, or following nerve injury), depending on the implicated receptor subtype, its location, and the tissue circumstances ( Figure 1). Figure 1. Adenosine receptors and their main signaling pathways. Schematic representation of Gcoupled A1A, A2A, A2B, and A3A receptor subtypes (A1AR, A2AAR, A2BAR, and A3AR) activated by extracellular adenosine and the main intracellular pathways involved. A2AR and A2BAR receptors are coupled to Gs protein, which leads to adenylyl cyclase (AC) activation and cyclic AMP (cAMP) increase. On the other hand, A1AR and A3AR are coupled to Gi protein, which inhibits AC and reduces cAMP. In some districts, A2BAR and A3AR receptors are also coupled to Gq protein, and A1AR is coupled to Go, which stimulates Ca 2+ release from intracellular stores. Moreover, all receptors are coupled to mitogen-activated protein kinase (MAPK) pathways, including extracellular signal-regulated kinase 1 (ERK1), ERK2, p38, and MAPK.
Increases in cAMP at the sensory nerve terminal and the activation of PKA have been shown to induce hyperalgesia [81,82]. In this context, a recent study shows that, following partial sciatic nerve ligation in rats, cAMP inhibition reduces the behavioral signs of neuropathic pain and the phosphorylation of the cAMP response element binding (CREB) protein in the spinal cord [83]. Researchers have been able to be er understand the role of these receptors in pain signaling by monitoring how gene deletions alter thresholds to various types of pain stimulation since the development of A1, A2A, A2B, and A3AR-deficient mouse strains [47].

Adenosine A1 Receptors
The Gi/o protein family includes A1AR, and its activation causes a majority of the biological effects mostly by inhibiting the second messenger cAMP [17,84,85]. In addition, when activated, the G protein beta and gamma subunits stimulate phospholipase C [86], which cleaves phosphatidylinositol-4,5-bisphosphate [87] into diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (IP3), thereby increasing the Ca 2+ levels. Additionally, elevated intracellular Ca 2+ levels can trigger the activity of several enzymes, including protein AR, and A 3 AR) activated by extracellular adenosine and the main intracellular pathways involved. A 2 AR and A 2B AR receptors are coupled to Gs protein, which leads to adenylyl cyclase (AC) activation and cyclic AMP (cAMP) increase. On the other hand, A 1 AR and A 3 AR are coupled to Gi protein, which inhibits AC and reduces cAMP. In some districts, A 2B AR and A 3 AR receptors are also coupled to Gq protein, and A 1 AR is coupled to Go, which stimulates Ca 2+ release from intracellular stores. Moreover, all receptors are coupled to mitogen-activated protein kinase (MAPK) pathways, including extracellular signal-regulated kinase 1 (ERK1), ERK2, p38, and MAPK.
Increases in cAMP at the sensory nerve terminal and the activation of PKA have been shown to induce hyperalgesia [81,82]. In this context, a recent study shows that, following partial sciatic nerve ligation in rats, cAMP inhibition reduces the behavioral signs of neuropathic pain and the phosphorylation of the cAMP response element binding (CREB) protein in the spinal cord [83]. Researchers have been able to better understand the role of these receptors in pain signaling by monitoring how gene deletions alter thresholds to various types of pain stimulation since the development of A 1 , A 2A , A 2B , and A 3 ARdeficient mouse strains [47].

Adenosine A 1 Receptors
The G i/o protein family includes A 1 AR, and its activation causes a majority of the biological effects mostly by inhibiting the second messenger cAMP [17,84,85]. In addition, when activated, the G protein beta and gamma subunits stimulate phospholipase C [86], which cleaves phosphatidylinositol-4,5-bisphosphate [87] into diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (IP3), thereby increasing the Ca 2+ levels. Additionally, elevated intracellular Ca 2+ levels can trigger the activity of several enzymes, including protein kinase C [88], phospholipase D (PLD), phospholipase A2 (PLA2), and others. The PI3K and MAPK signaling pathways, specifically, ERK1/2 and MEK, can be activated by A 1 AR activation, which induces alterations in gene expression and the activation of glial cells [51,88,89].
A 1 AR activation can elicit both pro-inflammatory and anti-inflammatory responses. According to the literature, adenosine can serve as an analgesic by binding to A 1 ARs [41] and as an inducer of an anti-inflammatory response by binding to A 2A ARs [90] in the early phases of inflammation. Activated spinal A 1 AR prevents sensory transmission by inhibiting the slow ventral root potential that is connected to nociception through C-fibers.
Additionally, i.t. adenosine enhances noradrenaline release in in vivo models through A 1 AR activation, but not under normal conditions [40,41], and its antiallodynic effects rely probably on spinal adrenergic processes [21]. Substances that cause an increase in endogenous adenosine concentrations, such as inhibitors of adenosine metabolism, can activate spinal adenosine A 1 ARs. As a result, increased antinociception is produced by the intravenous administration of prototype nucleoside inhibitors of adenosine kinase [30,90]). After systemic administration, a group of novel nucleoside and non-nucleoside adenosine kinase inhibitors (A-134974, A-286501, ABT-702) were shown to reduce inflammatory and neuropathic pain [31,[91][92][93]. The adenosine concentration in the spinal cord was raised as a result of adenosine metabolism inhibition, which resulted in the activation of A 1 ARs [94,95]. While the intracerebroventricular (supraspinal) and intraplantar (peripheral) activities of adenosine kinase inhibitors were less effective compared to i.t. administration in neuropathic pain models, the antinociceptive effects of these drugs largely occurred at spinal locations [93]. Accordingly, peripheral locations were not involved in analgesic effects in an inflammatory condition [91].
A 1 AR activation is necessary for the majority of antinociceptive effects of adenosine [35,36]. A 1 ARs have recently been discovered to be present on intrinsic neurons and in laminae I and II of the dorsal horn of the spinal cord [17,37,38,96,97]. At the II lamina level, numerous afferent sensory nerves interact with postsynaptic neurons. In addition, A 1 ARs are localized in the descending projection within the posterior horn [38]. Presynaptic receptors could be present, since dorsal root ligation increased A 1 AR immunoreactivity on the side next to the dorsal root ganglion (DRG), despite rhizotomy failing to detect a significant presynaptic A 1 AR population on sensory neuron terminals [37,98]. Neuronal activity in the spinal cord and the DRGs is predominantly prevented by A 1 AR activation [25,42,43,99]. The ability of A 1 AR to regulate presynaptic excitatory transmission to substantia gelatinosa neurons in the spinal cord was connected to its ability to reduce pain [100]. According to DeLander et al. [101], the spinal cord is likely where adenosine-induced NMDA receptor inhibition also takes place, which lessens the plasticity and central sensitization mechanisms associated with chronic pain.
Additionally, multiple animal models of acute pain patterns showed that systemic administration of different A 1 AR agonists might have an analgesic effect [17,21,26,30,102,103]. The intracellular NO/cGMP/PKG/KATP signaling pathway is activated by peripheral adenosine A 1 AR, leading to reduced pain [35]. The A 1 AR agonists alleviate thermal, but not mechanical, allodynia induced by sciatic nerve damage. Since thermal hyperalgesia is mediated by C-fibers and mechanical allodynia by A-fibers, these results demonstrate that A 1 ARs are present in C-but not in A-fibers [15]. A 1 AR agonists, such as R-phenyl-isopropyladenosine (R-PIA) [104] and GR79236 [16], inhibit the mechanical allodynia induced by spinal nerve ligation in rats. GR79236 inhibits also carrageenan-induced inflammatory hyperalgesia [16], while R-PIA decreases the thermal pain threshold in rats with spinal cord damage [105]. The latter prevents formalin-induced pain and reduces hyperalgesia induced by PGE2 by activating A 1 AR [44,53]; however, it is to be noted that R-PIA is poorly selective for A 1 AR and can induce A 3 AR effects as well [106,107]. Another A 1 AR agonist, N 6 -cyclopentyladenosine (CPA), when administered to rats, prevented arthritis-induced pain and neuropathy-induced pain [46]. In particular, reports suggested that the i.t. administration of A 1 AR agonists caused analgesia in multiple animal models of acute pain, comprising tail flicking, tail immersion, hot plate, formalin, acetic acid, capsaicin models, and others [104,108,109]. Moreover, the analgesic effect induced by i.t. adenosine and the increase in thermal hyperalgesia in A 1 AR knockout mice were reversed [110]. In these animals, a lower pain threshold was found in hyperalgesia tests [12], and they also showed moderate hyperalgesia to heat stimulation [110].
Other mechanisms, such as the inhibition of glutamate release, are involved in A1ARassociated analgesic processes in chronic pain. As mentioned above, the antinociceptive impact of A 1 AR agonists has been ascribed to AC inhibition and the reduced generation of cAMP in sensory nerve endings [44,53,106,111]. Moreover, A 1 ARs inhibit calcitonin gene-related peptide release [16] and Ca 2+ entry in sensory neurons when studied in vitro [106,107]. The primary mechanisms for spinal A 1 AR-mediated antinociception are described as: (1) increased K + conductance and hyperpolarization of intrinsic neurons in the dorsal horn, (2) inhibition of the release of peptides, such as substance P and CGRP, and (3) inhibition of glutamate release [25,39,100,112,113]. According to [97], A 1 AR mechanisms also cause interneurons in the dorsal horn to release less γ-aminobutyric acid (GABA); however, the overall effect of this process is likely to be stimulatory, making it difficult to determine how much this action contributes to the suppression of pain.
It has been hypothesized that peripheral A 1 ARs physically associate with µ-opioid and α 2 -adrenergic receptors and generate tolerance, cross-tolerance, and cross-dependence with other drugs, i.e., heterologous desensitization, when an A 1 AR agonist is administered repeatedly [106]. The effectiveness of adenosine in such situations may be influenced by the activation of the α 2 -adrenergic receptor. Since those A 1 ARs have pronounced inhibitory activities, it is unclear what mechanism underlies this fundamentally excitatory function. This activity, meanwhile, highlights the possible complexity of A 1 AR responses to nerve damage, where a variety of spinal alterations in nociceptive transmission take place (e.g., central sensitization, disinhibition, phenotype switch).
Accordingly, using peripheral microdialysis, it was found that capsaicin (which activates TRPV1 receptors specifically expressed on C-fibers [114]), formalin (which causes neurogenic and tissue inflammation [115]), and glutamate [116] increased the peripheral extracellular levels of tissue adenosine. In each case, these actions were inhibited by capsaicin pretreatment, which indicated an interaction between A 1 ARs and TRPV1 receptors (see below).
Collectively, these findings suggest that directly acting agonists, as well as substances that boost the local tissue levels of extracellular adenosine, can activate inhibitory A 1 ARs on sensory afferents to suppress pain under specific circumstances (such as in the presence of a mild degree of stimulation, specific mediators, and nerve injury).
Unfortunately, despite their capacity to inhibit pain, the therapeutic advantage of A 1 AR agonists is constrained by their cardiovascular side effects [71]. Therefore, they are not currently being pursued for pain.

Adenosine A 2A and A 2B Receptors
The role of A 2A AR and A 2B AR in pain is more controversial, as their activation has both pro-and antinociceptive effects [45,52,117,118]. Furthermore, the peripheral activation of A 2A AR and A 2B AR, which have pronociceptive and vasodilatory properties, can result in the side effects of AR agonists. [44,49]. Regarding the potential mechanism of action of these receptors, A 2A AR is coupled to Gs and Golf proteins, the latter mostly in the striatum [119]. Following the activation of these proteins, the stimulation of AC and increased cAMP synthesis are the highly significant intracellular events [15,49] that could stimulate cAMP-dependent kinase [120]. As a result, PKA can activate several pathways through the binding of PKC, influencing Ca 2+ and K + concentrations, cAMP-responsive elements [83], MAPK, and PLC activation [51,84,85,103]. Furthermore, the A 2B AR receptor, which has a lesser affinity for adenosine compared to other ARs, promotes signaling by activating Gs and Gq. By stimulating PLC, A 2B AR can increase IP3 and intracellular Ca 2+ and stimulates the arachidonic acid pathway [49,121]. A 2B AR may occasionally couple to Gq and Golf proteins as well [47]. It is known that both A 2A ARs and A 2B ARs, by increasing cAMP accumulation and activating downstream signal pathways, can lead to alterations in gene transcription relevant to the nociception effect. Increases in cAMP in sensory neurons cause PKA activation, Na + channel phosphorylation, and activation of sensory afferents, all of which are mediators of hyperalgesia [81].
A 2A ARs are highly expressed in the basal ganglia and the olfactory bulb of the CNS but are expressed at low levels in other brain areas, whereas A 2B ARs and A 3 ARs are expressed at extremely low levels [73]. Moreover, the DRG sensory neurons contain adenosine A 2A ARs [48] and A 2B ARs [57]. The local peripheral injection of A 2A AR agonists in the rat hind paw caused an increase in mechanical hyperalgesia in functional tests, as well as an increase in the flinching response to formalin [52].
A 2A AR activation by extracellular adenosine resulted in an anti-inflammatory effect through its expression on diverse cell types [122]. In addition, the anti-inflammatory actions of the prototypical A 2A AR agonist CGS21680 also included an increase in the serum interleukin (IL)-10 levels in complete Freund's adjuvant (CFA)-injected rats, reflecting the data reported in lymphocytes from patients. In addition, treatment with CGS21680 induced a reduction in the secretion of proinflammatory cytokines in a murine calvary model of bone resorption induced by abrasion particles, while the IL-10 levels in the bone was considerably improved [54]. The local administration of CGS21680 in the hind paw led to mechanical hyperalgesia in rats [53], and the systemic administration of the A 2A AR antagonist SCH 58261 attenuated the nociceptive responses in both acute and inflammatory tests in mice [27,63], which signifies a related role of A 2A ARs in peripheral nociceptive signaling pathways. Previous studies demonstrated that LASSBio-1359, a novel A 2A AR agonist, reversed the hyperalgesic response in mice, caused by the stimulation of the inflammatory process, through peripheral A 2A AR activation [55]. Taiwo and Levine (1990) showed distinctly different effects of A 1 AR and A 2A AR activation on peripheral pain [44]. They found that A 1 AR mediated analgesia in peripheral locations, while A 2A AR enabled painful perception, as reported in other investigations [44,52,56,123,124]. However, the role of A 2A AR is not clear. Indeed, other researchers showed a decrease in pain when A 2A AR is activated [125][126][127]. These controversial results could be connected to the intracellular signaling of A 2A AR. Its activation increases cAMP production, which could cause pain, but also opens the K + channels, which could prevent pain [17,49,50].
Furthermore, A 2A AR expressed on glial cells increased inflammatory mediator secretion, which caused and sustained chronic pain [90]. Therefore, A 2A AR blocking may be a novel strategy for the treatment of neuropathic and chronic pain. A 2A AR knockout animals are less sensitive to pain, suggesting that A 2A AR antagonism is a pain reliever for acute [56] and chronic pain. The A 2A AR agonist CGS21680, which was administered spinally [26,30] and supraspinally [50], was reported to produce antinociception. However, it was less active than A 1 AR agonists, and lacking the effects of selective A 2A AR antagonists, it is unclear whether this receptor was responsible for such actions. Moreover, the A 2A AR effects on spinal electrophysiology are complicated [100,112], and it is unclear which receptors cause these effects. The local administration of CGS21680 heightened the pain response [45,52], an effect that was likely mediated by the Gs-AC-PKA pathway. Intriguingly, one study found that forskolin and dibutyryl cAMP, which are PKA activators, increased the sensitivity to heat [128].
Contrarily, the activation of the adenosine receptor A 2A AR increased nociception by sensitizing peripheral afferent axons that transmit to the spinal cord [56]. As a result, mice lacking the A 2A AR responded less strongly to nociceptive thermal stimuli [117,118], whereas animals lacking the A 1 AR responded more strongly to nociceptive thermal stimuli [12,117]. Other research found no change in heat thresholds or antinociceptive reactions upon the systemic administration of morphine, a µ-opioid receptor agonist [129]. It was suggested that these functional variations in opioid actions reflected altered opioid receptor patterns in the spinal cord and, maybe, altered A 2A ARs in sensory neurons. Because of the involvement of A 2A ARs in inflammation [130] and the growing understanding of how inflammation contributes to neuropathic pain, responses to inflammation or nerve damage have not yet been described in A 2A AR-null mice [131]. It has been shown that mice lacking A 2A AR are hypoalgesic, which may indicate a peripheral pronociceptive role. [117].
The majority of inflammatory cells express A 2B AR, and the activation of this receptor has both pro-and anti-inflammatory effects [58]. A 2B AR has pro-inflammatory effects by promoting the secretion by mast cells and macrophages of IL-1, IL-13, IL-3, IL-8, IL-4, and VEGF [61]. Moreover, A 2B AR induced the generation of IL-19 and TNF-α from bronchial epithelial cells [62]. The inhibition or deletion of the A 2B ARs in mice reduced intestinal inflammation and slowed the progression of murine colitis/inflammatory bowel disease [132]. A 2B ARs could be activated in pathological circumstances like hypoxia/ischemia and inflammation, where adenosine is considerably raised, as a larger quantity of adenosine is required for A 2B AR activation compared to other AR subtypes [47,133]. A 2B ARs are expressed both centrally and peripherally, in a wide variety of tissues, most notably in immune system cells. It has been suggested that these receptors, which have the lowest adenosine affinity compared to the other AR subtype, are particularly important in immunological processes like inflammation that are characterized by high adenosine concentrations. Several studies showed the peripheral pronociceptive impact of A 2B AR activation [64,134], while the injection of A 2B AR antagonists showed an analgesic effect. Other authors found that a prolonged enhancement in plasma adenosine activated A 2B ARs on myeloid cells, resulting in the increase in circulating IL-6 and sIL-6R. Data on the advantages of using A 2B AR antagonists in inflammatory processes, due to their anti-inflammatory actions [58,65] in addition to their analgesic effects [64], are available in the literature. However, there are limited data available on PSB-603, which is commercially available as a highly powerful and selective A 2B AR antagonist in multiple species, including humans and rodents [135][136][137]. The tests carried out demonstrated that PSB-603 (i.p. 5 mg/kg) significantly decreased inflammation both in local and in systemic inflammation models. In particular, treatment with PSB-603 significantly decreased the quantities of both TNF-α and IL-6 and the levels of ROS in the paws of mice with inflammation generated according to the local inflammation model. The A 2B AR antagonist did not affect the plasma levels of TNF-α and IL-6 in the systemic inflammation model. However, A 2B AR activation was demonstrated to be excitotoxic even with heightened adenosine concentrations and would consequently initiate an inflammatory process [138]. The A 2B AR antagonist was able to enhance the analgesic effects of morphine and paracetamol while decreasing thermal hyperalgesia [63,64]. A variety of A 2B AR antagonists were also demonstrated to elicit antinociception when administered systemically; this effect was considered peripheral, since it was also seen in a structure that did not penetrate the central nervous system [64]. In the mouse tail immersion and hot plate tests, blocking A 2B ARs enhanced the effects of paracetamol, and blocking A 2A ARs had an antinociceptive effect even in the absence of paracetamol [63].
Mast cell mediators such as histamine and 5-hydroxytryptamine are released more often when A 2B AR and A 3 AR are activated, which might augment pain signaling [17,24]. A 2A AR agonists improved plasma extravasation when injected into the knee joint, although a local treatment with A 2A AR, A 2B AR, and A 3 AR agonists also caused edema that included mast cell degranulation [139][140][141]. These effects on the skin and joints are thought to be the result of pro-inflammatory activity.
Different methods revealed other factors that affect nociception. Although such action cannot be explained by A 1 AR inhibition, the analgesic properties of A 2A AR and A 2B AR antagonists have attracted interest.

Adenosine A 3 Receptors
The exploration of the role of A 3 AR in pain and of the mechanisms and sites of action of A 3 AR agonists has begun recently. In many pain models, A 3 AR exhibited preclinical antinociceptive efficacy [5][6][7]71,142] and efficacy and safety in studies on non-painful states [7,143]. A 3 AR, like A 1 AR, is coupled to G i/o proteins. Its coupling to the G q/11 protein is less well established. The inhibition of AC and the activation of PLC, IP3, DAG, PKC, and PLD, are the primary mechanisms involved in A 3 AR signaling. These receptors activate the MAPK pathway, namely, ERK1/2, as other ARs [67,68,72,143]. It was demonstrated that A 3 AR agonists reduced the pro-nociceptive N-type Ca 2+ channels in rat DRG neurons [67,68] and by this mechanism inhibited post-visceral hypersensitivity [72]. More recent research showed that activating A 3 AR had an antihyperalgesic impact in many mouse models of neuropathic pain, which is in contrast with previous reports. Numerous preclinical pain models showed that A 3 AR ligands have significant anti-nociceptive effects [49,125]. Previous studies showed that the anti-nociceptive effects of endogenous adenosine depend on A 3 AR activation at spinal and supraspinal locations, suggesting that a selective A 3 AR agonist might be an effective analgesic [73]. The highly selective A 3 AR agonist MRS5698 decreased mechanical allodynia in numerous rat neuropathic pain models, including models of chronic constriction injury (CCI) and chemotherapy-induced neuropathic pain [49]. Previous research showed that the antinociceptive effects mediated by endogenous adenosine depend on A 3 AR activation at spinal and supraspinal locations, and pharmacological data identified a selective A 3 AR agonist as a potent non-narcotic drug [6].
Many cells are involved in the dual impact of A 3 AR on inflammatory processes, and this receptor subtype can have overlapping and opposing functions [64,144]. A 3 AR activation can also produce a pro-inflammatory effect by inducing the release of histamines and other allergic mediators [145], preventing eosinophilic chemotaxis and inhibiting apoptosis [69]. A 3 AR also boosted the fast infiltration of inflammatory cells by attracting eosinophils and macrophages in the lungs [146] and promoting Ca 2+ signaling [147]. Concerning its role in pain, several previous publications described a pronociceptive profile of A 3 AR, since animals with its gene deletion demonstrated less hyperalgesia with inflammation and hypoalgesia in a nociceptive threshold test compared to control mice [14], but its effect in reducing chronic pain was revealed in subsequent studies.
The A 3 AR agonist IB-MECA was effective in increasing the mechanical threshold in a chronic inflammation model. However, the adenosine A 3 AR seems to have intricate effects on the CNS, with pro-inflammatory and anti-inflammatory functions, notably in healthy conditions. In agreement with this, previous studies indicated that IB-MECA reduced the IL-1 and IL-10 levels in the control group. Although A 3 AR is not highly expressed in the nervous system, adenosine has anti-inflammatory properties that can help relieve pain associated with peripheral inflammation [70]. A 3 AR agonists, such as IB-MECA, were previously tested in phase II and phase III clinical research for disorders other than pain and, so far, have demonstrated good safety profiles, in contrast to A 1 AR agonists, which have limited therapeutic application [148]. According to hot plate test results, mice lacking A 3 AR showed diminished nociception [149]. This was most likely due to a decline in the supraspinal processing and in the ability to "detect" painful stimuli. This is in line with the expression of A 3 AR in thalamic nuclei [150], where it may be involved in the processing of nociceptive inputs.
It is important to note that many of the earlier studies with supposedly A 1 AR-selective agonists, like R-PIA, have to be reinterpreted in light of the evidence that most of these agonists can also active A 3 AR in mouse models [151,152]. In addition to species differences in AR ligand affinity and efficacy, there are also species differences regarding the role of the receptor in inflammatory processes.

Allosteric Modulators
Using natural adenosine concentrations to activate the receptor in conjunction with an allosteric enhancer of the A 1 AR is an alternative method of promoting A 1 AR agonism. As this method depends on adenosine being created at the target area, it should have few side effects. Adenosine release is a normal compensatory process that occurs in some illness situations to assist tissues in regaining equilibrium.
In a rat model of plantar surgical damage, T-62, an A 1 AR positive allosteric modulator (PAM), reduced pain hypersensitivity in a dose-dependent manner (0.3-1 mcg i.t.). When clonidine was also provided, the amount of T-62 needed to produce an antihyperalgesic effect was cut in half, achieving only 40% of the maximum effect. Clinical testing of T-62 was conducted on patients with postherpetic neuralgia who were in pain [153][154][155]. Unfortunately, a clinical trial of the lead drug T-62 in post-herpetic neuropathic pain was discontinued [156]. However, future clinical trials of other A 1 AR allosteric enhancers might still be worthwhile.
The past 30 years witnessed a significant amount of research in the area of A 1 AR, which led to the development of clinical candidates for A 1 AR agonism, antagonism, and allosteric alteration. Pharmacologically speaking, the development of A 1 AR antagonists should be simpler than that of A 1 AR agonists because of the complications of developing A 1 AR agonists, such as receptor desensitization and the potential for severe cardiovascular and CNS side effects. The A 1 AR allosteric enhancer T-62 was previously evaluated in the clinic after showing encouraging effects in neuropathic pain animal models. Therapeutics that target A 1 AR (A 1 AR antagonists, A 1 AR agonists, and allosteric enhancers) may soon achieve long-awaited clinical success as a result of these significant scientific and clinical advancements. In addition to directly acting agonists and antagonists of A 1 AR and A 3 AR, allosteric modulators of these receptors have also been developed [69]. Allosteric modulators have the advantage of increasing the effects of endogenous adenosine released in response to stress at a specific location or tissue, causing a biological impact in the absence of an agonist. Numerous allosteric enhancers of the activation of ARs by agonists are being studied as therapeutic prospects.
The first known AR allosteric modulators were 2-aminothiophenes, exemplified by PD-81,723. Clinical trials are now examining the A 1 AR allosteric enhancer T-62 as a potential therapy for neuropathic pain. The difficulties in creating A 1 AR agonists, antagonists, or allosteric enhancers for therapeutic intervention in humans have now been well defined.
TRR469 was recently found to be one of the most potent A 1 AR positive allosteric modulators ever reported, with the ability to induce a 33-fold increase in adenosine affinity [155,157,158]. In the formalin and writhing tests, TRR469 successfully reduced nociceptive responses with results comparable to morphine. Additionally, unlike the orthosteric CCPA, it did not cause locomotion or cataleptic side effects in the diabetic neuropathic pain paradigm generated by streptozotocin (STZ) [158].

Clinical Studies
In several situations, it has been reported that the systemic administration of adenosine through i.v. infusion in humans (50-70 g/kg per minute over 45-60 min) resulted in some degree of pain relief. Adenosine infusions (i.v.) produced analgesia in healthy volunteers when used to treat experimental pain caused by cutaneous heat thresholds [159], ischemic pain [160,161], allodynia brought on by mustard oil [162], and cutaneous inflammatory pain [163]. In clinical studies including peripheral neuropathic pain [164,165] and mixedetiology neuropathic pain, with postsurgical/posttraumatic neuropathic pain as the most common diagnosis [166], i.v. adenosine decreased spontaneous and evoked pain. In the latter experiment, which used an enhanced enrollment strategy, 40/66 patients (61%) were found to respond to adenosine during the initial open phase of the trial. In situations of neuropathic pain, one infusion of adenosine can occasionally result in long-term pain relief lasting months (estimated at 5-10% in [167], reported in 2/26 subjects in [165], and observed in 3/62 subjects in [166]). It is important to note that a comparable long-lasting pain relief from postherpetic neuralgia has been observed following a single [168] or repeated infusions of ATP, which is rapidly converted to adenosine [169]. When adenosine is administered during surgery, it also has analgesic effects on the patient. Segerdahl and colleagues found that adenosine (i.v.; 80 g/kg per minute) during surgery might decrease the need for anesthesia and/or postoperative analgesics [170][171][172]. Both medications stabilized the cardiovascular system during surgery, according to studies comparing intravenous adenosine infusions to intravenous remifentanil (a short-acting opioid) infusions [173,174]. However, the use of adenosine reduced postoperative pain and the need for opioid analgesics. The most striking effects (degree, duration of 48 h) and the lowest incidence of postoperative nausea were recorded by Fukunaga et al. [174]. They utilized a dose of 82 g/kg per minute of adenosine, with a cumulative dose of 2.5 g, compared to the dose of 17 g/kg per minute used by Zárate et al. [173], with a mean total dose of 1.4 mg.
It has been interesting to examine the possible effectiveness of adenosine administered spinally in humans in light of preclinical studies showing adenosine (i.t.) and A 1 AR agonists elicit antinociception [175]. Adenosine was shown to lessen experimentally produced allodynia from mustard oil [176] as well as hyperalgesia and allodynia after intradermal capsaicin injection in human volunteers when given intravenously (i.v) [177,178]. Adenosine (i.t.) decreased the area of allodynia 2-24 h post injection in a double-blind trial on individuals with neuropathic pain, supporting the observations of prior case reports [27,179]. Contrarily, adenosine (i.t.) had no impact on chronic pain following hysterectomy or on acute pain following the application of noxious heat [176,177]. According to a recent study, adenosine (i.t.) was unlikely to be helpful as an analgesic monotherapy due to its low efficacy (25% decrease in allodynia, no effect on continuous pain), the induction of back pain, and its inability to treat other types of pain [180].
Humans who received a peripheral adenosine injection experienced pain reactions similar to those brought on by ischemia circumstances [17]. Substance P [181] and nicotine [182] can increase the pain-initiating effects of adenosine, which are typically a limiting factor in the use of adenosine-related drugs for the treatment of chronic pain. Histamine and 5-hydroxytryptamine are released from mast cells upon activation of A 3 ARs, and these chemicals operate on the sensory nerve terminals, causing pain [17].
Peripheral adenosine injection has an inhibitory overall effect on pain processing, because increasing extracellular adenosine concentrations has an antinociceptive effect. Since adenosine has dual A 1 AR-and A 2A AR-mediated anti-inflammatory and antinociceptive actions, there is growing interest in developing medicines that may have analgesic effects by altering the extracellular adenosine concentrations. Inhibitors of AK, whose spinally mediated antinociceptive effects were recognized more than ten years ago [23], are examples of effective medications. Orally administered AK inhibitors are more efficient at reducing inflammatory pain than neuropathic or acute pain, most likely because of the anti-inflammatory properties of adenosine [92]. According to studies comparing the antinociceptive and anti-inflammatory properties of AK inhibitors administered at the ipsilateral or contralateral sides of the injury, the antinociceptive action of AK inhibitors is systemically mediated at the level of the spinal dorsal horn [134]. The peripheral injection of an inflammatory agent causes enhanced c-Fos expression in the spinal dorsal horn, which could be decreased by AK inhibitors [134].
Antidepressants are frequently used to treat neuropathic pain; however, their analgesic effectiveness appears to occur without regard to any impact on one's mood and may even include a rise in extracellular adenosine concentrations. This was demonstrated in rat models of neuropathic pain following either acute [183] or chronic [183] amitriptyline administration. In a rat model of excruciating diabetic neuropathy, endogenous adenosine also appeared to play a role in the antiallodynic activity of amitriptyline [184]. Esser and   [183] noted that amitriptyline reduced pain by modulating endogenous adenosine levels.
The analgesic effects of opioids may be influenced by rises in adenosine concentrations. Humans receiving i.t. morphine showed an increase in adenosine concentrations in their cerebrospinal fluid [179]. It is interesting to note that morphine-induced adenosine release is diminished in neuropathic rats [185], which may help to explain why opioids are less effective and potent at treating neuropathic pain. A 2A AR-deficient animals were also found to have changes in the expression of various opioid receptor types, suggesting a functional interaction between A 2A AR and opioid receptors in the modulation of pain [129].
A critical analysis of the uses of adenosine and ATP in pain control, which summarizes the majority of human investigations, indicated that adenosine compounds have a high potential to reduce pain [168]. This study emphasized that it is essential to consider medication dosages, routes, timing, and tissue penetration and that additional basic research is required to fully understand several issues. Caffeine can modulate pain through its antagonistic effects on ARs, but as recently highlighted [186], the type of effect (such as headache creation or relief) relies on the site of action, the dosage, and the timing of exposure. The interaction between paracetamol and coffee in pain control likely involves both A 2A ARs and A 2B ARs. Moreover, theophylline reduces chest discomfort in people with a hypersensitive esophageal mucosa, perhaps through modifying adenosine-mediated nociception [187]. Humans who received peripheral adenosine injection experienced pain reactions similar to those brought on by ischemia circumstances [17].
The use of A 1 AR agonists could be another strategy for treating migraine headaches by blocking the trigeminovascular system and CGRP. By acting on the trigeminal nucleus and preventing the release of CGRP into the bloodstream, A 1 AR activation limits trigeminovascular activation. This second effect is due to A 1 AR activation on peripheral trigeminal nerve terminals [16]. It was demonstrated that tonic A 1 AR activation prevents the facilitatory effects of CGRP in the hippocampus [188]. Interestingly, CGRP actions are promoted by A 2A AR activation [188,189]; however, it is unclear how this information relates to using A 2A AR antagonists as a migraine therapy strategy.

Conclusions
Adenosine receptors have been suggested as a possible target for the development of new analgesics due to the widespread capacity of A 1 AR agonists to attenuate pain responses in a variety of pain models after systemic administration. Nevertheless, in addition to producing antinociception, A 1 AR agonists also caused hypothermia, a reduction in locomotor activity, and cardiovascular alterations, and these effects are elicited by doses similar to those inducing antinociception. Non-nucleoside AK inhibitors are more effective in promoting antinociception, which has encouraged the development of indirectly acting drugs rather than directly acting agonists in the setting of a systemic administration [94]. Moreover, in several preclinical models of chronic pain, the regulation of A 3 AR results in strong anti-hypersensitive effects [70]. It is interesting to note that authors discovered that the A 1 AR-mediated antiallodynic effect was much shorter-lived than the A 2A AR-or A 2B ARmediated pain control (lasting up to 6 weeks) [66]. As discussed above, these unorthodox pharmacological approaches have shown promising results in preclinical animal models of pain and could offer a solution to the challenges AR ligands have previously encountered in the clinic. Overall, the results reported in this review highlight the therapeutic potential of ARs as drug targets for the treatment of acute and chronic pain, as well as the need for the creation of novel, more effective approaches.