**1. Introduction**

Among different types of chronic pain, neuropathic pain is defined by the International Association for the Study of Pain (IASP) as pain caused by a lesion or disease of the somatosensory nervous system (IASP 2012). There are many available treatment approaches for the management of neuropathic pain. Yet, despite these advances, it remains an unmet medical need because most of the treatment approaches intended to halt this pain condition are not effective enough or sometimes effective but limited by side effects. Thus, finding new targets and innovative future strategies that might help to improve neuropathic pain control are of clinical need.

µ-Opioid receptor (MOR) agonists are the mainstay treatment for different forms of chronic pain [1–4]. However, their efficacy in the management of neuropathic pain is a longstanding question of debate. Yet, international guidelines restrict opioids to second- or third-

**Citation:** Király, K.; Karádi, D.Á.; Zádor, F.; Mohammadzadeh, A.; Galambos, A.R.; Balogh, M.; Riba, P.; Tábi, T.; Zádori, Z.S.; Szök˝o, É.; et al. Shedding Light on the Pharmacological Interactions between µ-Opioid Analgesics and Angiotensin Receptor Modulators: A New Option for Treating Chronic Pain. *Molecules* **2021**, *26*, 6168. https://doi.org/10.3390/ molecules26206168

Academic Editor: Lorenzo Di Cesare Mannelli

Received: 15 September 2021 Accepted: 8 October 2021 Published: 13 October 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

line therapy, with no clear consensus on their effect [5–7]. MOR agonists with significantly higher intrinsic efficacy than morphine produced acceptable analgesia in preclinical models of neuropathic pain [8,9]; however, this has not been successfully utilized clinically because clinical trials showed controversial results related to their efficacy and liability for side effects [10–13]. In response to this argument, many studies have been conducted to increase the efficacy and decrease the side effects of opioids when used in the management of neuropathic pain. Some of the encouraging strategies that aim to improve the analgesic effect and decrease the side effects of currently used analgesics, such as opioids, are based on combining two or more different agents. However, so far, clinical research data that is based on combination strategies have not met expectations [14]. Chaparro et al., reviewed clinical trials on the efficacy and safety of various agent combinations for neuropathic pain [14]. Their analysis revealed that the combination of opioids with gabapentin was significantly better than gabapentin alone in reducing the symptoms. However, the number of treated patients that was required for a single patient to benefit was still 9.5, and significantly more participants experienced side effects and thus dropped out of the studies with opioids plus gabapentin than with gabapentin alone [14]. On the other hand, studies assessing the effects of opioids in combination with other sensory-sensitization blocking agents could be of high clinical value. Thus, continuing preclinical research based on the application of multi-target drugs or combination strategies that involve implementing different agents might bring a new treatment option for neuropathic pain. In the former case, for instance, applying opioid receptor ligands that display agonist and non-opioid effects, such as tapentadol, display both the MOR agonist and norepinephrine reuptake inhibitory effects in the same molecule [15]. Recently, our group reported on the promising effect of the combination of glycine transporter 1 and 2 inhibitors in the management of neuropathic pain evoked by sciatic nerve ligation [16]. In such a strategy, we need to consider how the individual drugs affect pain transmission.

Accumulating evidence has proven that drugs affecting the renin–angiotensin system can modulate pain transmission [17–34]. Recent studies have also shown that drugs mimic or antagonize angiotensin type 1 and 2 (AT1R and AT2R) receptor-mediated actions do produce a beneficial analgesic effect in rodent models of chronic pain types [17,20,22,28,29,35–38]. The analgesic effect of ligands affecting angiotensin receptors in neuropathic pain is explained by the contribution of these receptors to neuroregeneration and neuroprotection partially by reducing neural inflammatory processes [18,24,37,39–41]. Nevertheless, much remains unclear regarding the role and clinical utility of these receptors in analgesia.

This review briefly highlights how the effect of MOR agonist-induced analgesia is altered under neuropathic pain conditions, showing the advantages and drawbacks, as well as principal factors that negatively impact the analgesic effect of MOR analgesics in this pain entity. The next sections review the implication of angiotensin and its receptors in chronic pain, particularly that associated with neuropathy, and also the neuroanatomical overlap between MORs and angiotensin receptors in relation to pain. Finally, according to the reviewed data, perspectives on the future drug combination-based research strategy to treat neuropathic pain are provided. With respect to angiotensin IV and its receptor, the presence of the peptide has been reported in human dorsal root ganglia (DRG) and trigeminal nucleus (TG) [42,43]. However, there are little data related to their analgesic effect. Thus, they will not be discussed in the present review.

#### **2. The Opioid System and the** µ**-Opioid Receptor in Different Pain Entities**

The opioid system is a physiological system for controlling pain, but it also participates in addictive behaviors and immune defense, among others. Mammalian endogenous opioid peptides and exogenous natural, semisynthetic and synthetic opioid agonists can produce their effects through the activation of opioid receptors, namely µ-(MOR), δ-(DOR), and κ-(KOR) opioid receptors. Opioid receptors belong to the class A G-proteins of the pertussis toxin-sensitive Gi/Go family. Their effectors include adenylyl cyclase, N- and L-type Ca2+ channels, and inwardly rectifying K<sup>+</sup> channels. Upon activation, adenylyl

cyclase and Ca2+ channels are inhibited, whereas K<sup>+</sup> channels are activated. Thus, both the limitation of Ca2+ entry and the hyperpolarization of the cells may give a tenable explanation for the inhibition of transmitter release at pain traffic points [44,45]. With respect to pain, central MORs are the principal target for mediating the analgesic effects of opioids. As in MOR-knockout mice, selective MOR agonists failed to produce analgesia as well as MOR-induced opioid side effects, such as respiratory depression, gastrointestinal transit inhibition, and addiction liability [46,47]. Since the identification of functional peripheral MORs, it has become obvious that the analgesic effects of opioids do not solely depend on MORs at the central nervous system (CNS) [48]. It is worth noting that achieving peripheral analgesia requires prerequisite factors that are related both to the physicochemical properties of opioid analgesics (limited CNS penetration) and pain entity. In the case of the latter, the pathological state of pain largely reflects the effects of opioid analgesics. In inflammatory or acute non-inflammatory pain, MORs number is increased or maintained at normal level, respectively [9,48–50]. Several opioid researchers have proven that functional MORs in the periphery are targetable, particularly in inflammatory pain types [51–54]. However, under neuropathic pain conditions, several studies have demonstrated the downregulation of MORs in the dorsal spinal cord and DRG [9,55]. The efficacy of currently available MOR agonists in neuropathic pain is a question of debate. Taken together, in cases of acute or inflammatory pain types, opioid analgesics can provide adequate pain control, which is somewhat hampered by above mentioned unwanted effects. However, in the case of neuropathic pain, the desired analgesia itself is often unachievable, consequently demanding dose-escalation, therefore causing more pronounced side effects (Figure 1A) (Karádi and Al-Khrasani, unpublished data) and (Figure 1B) (adopted from our previous work [16]).

**Figure 1.** (**A**) The analgesic effect of morphine measured on a dynamic plantar aesthesiometer (DPA) test at 30 min, after s.c. administration to mononeuropathic animals. Columns represent the paw withdrawal threshold of the animals in grams ± S.E.M. Asterisks indicate the significant differences between treatment groups or operated (R) and non-operated (L) hind paws (\* *p* < 0.05; \*\*\* *p* < 0.001 and \*\*\*\* *p* < 0.0001). Statistical differences were determined with one-way ANOVA and Tukey post-hoc test. Data represent means ± S.E.M (n = 5–12 per group). (Karádi, D.Á.; Al-Khrasani, M.; unpublished data). (**B**) Effect of the systemic administration of morphine to the motor function of rats. Columns represent the time latency of the animals in sec ± S.E.M. at 30 min post-treatment in the rotarod test. Asterisks indicate the significant differences compared to the saline group (one-way ANOVA, Newman–Keuls post-hoc test; \*\*\* *p* < 0.001). In each treatment group, 4–7 animals were used. These results were adopted from our previous work [16].

For restoring the effect of opioids in neuropathic pain, many attempts have been focused on the mechanisms related to changes in the number of functional MORs on sensory neurons in subjects with painful neuropathy. In our and other studies carried out in rats with neuropathic pain induced either by streptozotocin (STZ) or chronic constriction injury (CCI), the number of MORs was found to be decreased in DRG and spinal tissue [9,56,57]. This reduction in MOR number was accompanied by a decrease in the analgesic effects of opioids.

#### **3. Angiotensin Receptor Mimetics and Antagonists in Relation to Pain**

#### *3.1. Endogenous Angiotensin Ligands and Angiotensin Receptors*

Components of the renin–angiotensin system (RAS) have been previously reviewed or discussed extensively [19,24,58–64]. Nevertheless, the main findings are briefly summarized here for an overview. Among the endogenous peptides of the RAS, neuronal angiotensin II (Ang II) is the most significant in relation to pain. Ang II is an octapeptide derived from the inactive precursor angiotensinogen, which is initially cleaved by renin, resulting in the inactive intermediate angiotensin I (Ang I). Ang II is cleaved from Ang I by the angiotensin-converting enzyme 1 (ACE1). Ang II equally binds to and activates the AT1R and AT2R (see later on). Another relevant endogenous peptide of the RAS to this review is angiotensin 1-7 (Ang (1-7)), which is cleaved by the angiotensin-converting enzyme 2 (ACE2) from Ang II or by ACE1 from Ang I via the intermediate angiotensin 1-9. Ang (1-7) activates the Ang (1-7) receptor or MAS receptor, but it can also bind with lower affinity to AT2R.

There are four angiotensin receptor types known so far within the RAS; namely angiotensin II type 1 and 2 receptors, the angiotensin IV receptor, and the Ang (1-7) receptor or MAS receptor (abbreviated as AT1R, AT2R, AT4R, and AT7R or MASR, respectively). Additionally, in mice and rats, two AT1R isoforms have been identified, namely AT1aR and AT1bR [65,66]. In relation to the RAS, this review will focus on data of AT1R, AT2R, and MASR, with respect to pain, particularly from preclinical studies. They all belong to the rhodopsin-like G-protein coupled receptor family (GPCR); however, they differ significantly in terms of activation of signaling pathways and cellular and tissue distribution patterns. The latter will be discussed in detail in a separate section. The AT1R is a prime example of a GPCR that upon activation can be dependent and independent from heterotrimeric G-proteins, allowing the receptor to have a wide range of signaling responses to Ang II. In terms of G-protein dependent signaling pathways, the AT1R couples to multiple types of Gα, (Gq/11, Gi, G12, and G13), but it also includes the activation of small G-proteins. G-protein independent signaling of AT1R involves β-arrestin 1 and 2, tyrosine kinaserelated signaling, reactive oxygen species signaling, receptor-interacting scaffold proteins, or heterodimerization with AT2R or MASR. In the case of AT2R, signaling pathways are still not fully elucidated, in spite of the intensive research. In fact, it is one of the least understood areas of the renin–angiotensin system. Most interestingly, it fails to demonstrate classic GPCR signaling features, such as affecting second messengers (e.g., cAMP, diacylglycerol) or the lack of phosphorylation-induced receptor desensitization, or internalization in most tissue types. However, it has been proven that AT2R is sensitive to GTPγS and pertussis toxin in rat locus coeruleus, indicating Gi/o coupling [67]. AT2R can also stimulate protein phosphatases and nitric oxide production. In addition, AT2R mediates the inactivation of mitogen-activated protein kinase (MAPK) inhibition which is important in the induction of apoptosis [60,67]. The AT2R and Ang II interaction leads to neurite formation and growth via the modulation of polymerized β-tubulin, microtubule-associated proteins (MAP), the activation of the p42/p44 MAPK phosphorylation of trkA. MASR, similar to AT1R and AT2R, can couple to many downstream signaling pathways via Ang (1-7) activation. These include the activation of phospholipase C and A2, arachidonic acid release, or calcium-independent nitric oxide synthase activation. MASR also modulates several kinase-related pathways/effectors, such as the p38 MAPK, ERK1/2, phosphatidylinositol 3-kinase/Akt, RhoA, and cAMP/PKA, in different cell lines. MASR was also demonstrated

to constitutively couple to Gαi, Gαq, and Gα12/13 [63]. On the other hand, similar to AT2R, in most cases, MASR fails to induce the conventional G-protein mediated signaling response, defined by the levels of classical second messengers, such Ca2+, or inositol trisphosphate (IP3), despite belonging to the GPCR family.

#### *3.2. AT1 and AT2 Receptor Agonists*

Following the discovery of the neuronal RAS, numerous studies have reported on the implication of AT1R/AT2R agonists on nociception [27,30,33,68–76]. In spite of the high number of studies conducted, literature data remain highly controversial. Some publications describe the analgesic activity of AngII, AngIII, or renin on acute pain tests following central (intracerebroventricular [27,69,71,72,76] or intrathecal [33]) administration. These reports proposed different possible mechanisms of action behind the observed effects. Many of them indicate the role of the endogenous opioid system as the analgesic activity of test compounds was naloxone-sensitive [27,33,69,71,72]. Next, Shimamura et al., suggested a kinetic interaction between AngIII and met-enkephalin, namely the inhibition of cleavage of the latter [71]. Georgieva et al., found that AngII administered intracerebroventricularly (icv.) produced an antinociceptive effect in the acetic-acid writhing pain model, yet the AngII-induced antinociception was blocked by PD123319, an AT2R selective antagonist but not by losartan, an AT1R antagonist [75]. In this study, the authors concluded that AT2Rs but not AT1Rs are involved in the mechanism behind the analgesic action in acute inflammatory pain. Since then, studies assessing the effects of RAS peptides (angiotensinogen, AngI, AngII, or AngIII) microinjected into different regions of the periaqueductal gray (PAG) were conducted in rats. In these studies, all test peptides were proven to be analgesic on the tail-flick assay, and their effect was AT1R or AT2R antagonist reversible [77]. Another observation is that spontaneously hypertensive rats show longer latency on the hot plate but not on the tail-flick test, when compared to wild-type animals. Moreover, this increase in latency can be reversed by orally administered captopril or losartan, but not by antihypertensive agents which are acting on targets other than the RAS [73]. In contrast to the above-mentioned studies, Cridland et al., reported that AngII failed to show either antior pronociceptive effect [72]. However, at present, we cannot judge this issue because, to the best of our knowledge, there is no other study that supports Cridland's observations. It is also worth considering the article of Pavel et al., which examined the effect of AngII and losartan in rats undergoing CCI. In these animals, intraperitoneal AngII was found to be pronociceptive in the von Frey test (mechanical stimuli), constant hot- and cold-plate tests and decremental cold plate test (thermal stimuli). Losartan fully reversed the effect of AngII in case of mechanical stimuli, partially reversed it in case of constant cold-plate test, but further aggravated it in the decremental cold plate test. In the incremental hot plate test, the pain threshold was unchanged both following AngII or AngII + losartan administration [78]. The differences observed in this study between the effect of angiotensin in response to constant or decremental/incremental thermal stimuli is difficult to explain.

Further on, the direct pronociceptive activity of AngII and AngIII was described as spontaneous painful behavior (scratching) was observed following intrathecal administration [40,41]. It is worth noting that the study of Cridland et al., showed neither anti- nor pronociceptive action of AngII, whereas Nemoto and coworkers reported a pronociceptive action. Despite the similar administration route, the phenotype of the animals, as well as the dose applied, was different in these studies [40,41,73]. Therefore, further studies are needed to elucidate the effect of AngII at the spinal level. Indirectly supporting the pronociceptive action of AngII, Kaneko et al., reported icv. administered AngII to attenuate the analgesic activity of morphine in a dose-dependent manner in hot plate and tail pinch tests [69]. Similarly, Yamada et al., found that icv. administrated AngII or the AT2R agonist novokin decreased the antinociceptive effect of morphine in the tail-pinch test [79]. Shepherd et al., also reported an increased mechanical but not thermal allodynia following intraplantar AngII administration in mice after spared nerve injury (SNI) [80].

There is large literature data on neural regeneration and differentiation mediated by the AT2R, which were recently reviewed by Danigo et al. [24]. From this aspect, activating the AT2R induces positive changes in terms of neural injury. This neuroprotective action linked to the AT2R has been associated with an increase in neuronal BDNF expression by several reports. The AT2R agonist "compound 21" (C21) has been reported to increase neurite growth following spinal nerve injury [81] and to improve survival while attenuating post-stroke neurological deficit in mice [82]. Under these conditions, the common feature was an increase in neuronal BDNF expression. In contrast, increasing BDNF level is not necessarily beneficial in cases of peripheral nerve injury from the aspect of pathological pain, since Madara et al., showed that BDNF could induce glutamate release by enhancing the action of presynaptic NMDA receptors [83]. BDNF release governs the spinal long-term potentiation of C-fibers [84]. Long-term potentiation and a consequently increased glutamatergic tone, involving the increased activity of spinal NMDA receptors, are hallmarks of neuropathic pain or other chronic pain states [85,86]. Furthermore, Chen et al., proved that spinal NMDA receptor-potentiation on primary afferents in neuropathic pain could be blocked either by the BDNF scavenger trkB-Fc or by the trkB receptor antagonist ANA-12 [87]. The contribution of BDNF to pain was validated by Sikandar et al., where they demonstrated that the conditional knockout of BDNF from mouse sensory neurons results in unchanged response to most acute pain types and displayed hypoalgesia in chronic inflammatory or neuropathic pain [88].

#### *3.3. MAS Receptor Agonists*

Primarily the Ang (1-7)-MASR branch of RAS acts as an antagonist of the AngII-AT1R activity. The activity linked to AT2Rs is similar in general; however, with respect to pain transmission, this is not the case. The possible analgesic effect of Ang (1-7) was investigated following mostly local (intraplantar [21,23] or intrathecal [34,89–93]) administration. Studies using intraplantar administration reported that Ang (1-7) attenuated PGE2 [21,23,90,91] or carrageenan [23] induced inflammatory mechanical hyperalgesia. The antihyperalgesic effect of Ang (1-7) was lost in MASR KO mice [23] and was reversible by MASR, nNOS, guanylyl cyclase, or ATP-sensitive potassium channel blockers [94] as well as by different adrenergic antagonists [21], but not by naloxone [95].

Intrathecal administration of Ang (1-7) resulted in a decrease in spontaneous nociceptive behavior induced by intrathecal AngII [91], AngIII [92], substance P or NMDA [34]. Furthermore, intrathecal Ang (1-7) showed an antiallodynic and antihyperalgesic effect in neuropathic pain induced by CCI [89], STZ [90], or genetic model of diabetes (ob/ob mice) [93]. Moreover, several authors reported that Ang (1-7) effectively decreased the pathological increased p38 phosphorylation in the spinal cord [90–92,96]. Similar results were reported following intrathecal administration of ACE2 activator DIZE, namely reduced nociceptive behavior in the formalin test and decreased spinal p38 phosphorylation [96]. On the other hand, intraplantar Ang (1-7) was ineffective in the treatment of CCI induced neuropathic pain [23].

The effect of systemic (ip.) administration of Ang (1-7) on bone cancer pain was investigated by Forte et al., In this model, Ang (1-7) reduced spontaneous pain reactions, increased von Frey threshold and tail immersion latency following acute or chronic administration. The authors reported no anti-tumor activity [97].

#### *3.4. AT1 and AT2 Receptor Antagonists*

A growing body of literature data supports that antagonists of the AT1R, such as losartan, candesartan, or telmisartan, among others, display analgesic action in different pain models, including acute thermal, inflammatory, or neuropathic pain [17,23,30,35,36,39–41]. With respect to the analgesic effect of telmisartan, our unpublished results also support such findings because it could reduce the partial sciatic nerve CCI-induced allodynia after systemic administration in rats (Figure 2) (Karádi and Al-Khrasani, unpublished data)).

**Figure 2.** The analgesic effect of telmisartan measured on a dynamic plantar aesthesiometer (DPA) test at 120 min, after p.o. administration to mononeuropathic animals induced by partial sciatic nerve ligation rat model described by Seltzer et al. [98]. Columns represent the paw withdrawal threshold (PWT) of the animals in grams ± S.E.M. Asterisk indicates the significant differences between treatment groups or operated (R) and non-operated (L) hind paws (\*\* *p* < 0.01 and \*\*\*\* *p* < 0.0001). Statistical differences were determined with one-way ANOVA and Tukey post-hoc test. Data represent means ± S.E.M (n = 5 per group). (Karádi, D.Á.; Al-Khrasani, M.; unpublished data).

In addition, intrathecal administration of losartan has been reported to block AngIIinduced spontaneous pain [39], both phases of formalin test [41], and STZ-induced allodynia [99]. On the other hand, microinjection of AT1R and AT2R antagonists into the PAG has been reported to aggravate incisional allodynia [26,77]. Local administration of losartan was also investigated by Costa et al., In this study, intraplantar (ipl.) losartan effectively reversed prostaglandin E2 (PGE2) and carrageenan-induced mechanical hyperalgesia but was ineffective in CCI induced neuropathic pain [23]. In contrast, numerous publications have reported that systemic administration of AT1R antagonists to be beneficial [17,20,35,36,79]. Most of these reports suggest that blocking AT1R could also attenuate the inflammatory reaction in DRG [35,36] or the sciatic nerve [17] and elevate the decreased BDNF level in the sciatic nerve [17] following neuronal damage.

Bessaguet et al., investigated the effect of candesartan on resiniferatoxin-induced neurotoxic thermal hypoalgesia in mice and proved that intraperitoneal candesartan was able to reverse the evoked hypoalgesia in this assay, yet the same effect was achieved following the treatment with AT2R antagonist, EMA200 (PD123319). The authors proposed that candesartan may increase the AT2R binding of endogenous AngII, thus lowering the thermal threshold of animals. This proposal is further supported by the lack of efficacy of candesartan in AT2R KO mice [20]. In agreement with these results, Hashikawa-Hobara et al., reported that hypoesthesia caused by fructose induced diabetes was reversible by orally administered candesartan [100]. Obagata et al., showed that intrathecal losartan can attenuate the allodynia evoked by STZ in mice. In addition, they found that Ang II, as well as ACE expression, were increased, indicating the involvement of AngII in neuropathic pain conditions. It has also been reported that candesartan is capable of inducing neuroprotective, anti-inflammatory, and pro-angiogenetic effects accompanied by an increase in BDNF expression [101,102]. In these studies, the beneficial effects of AT1R antagonism were reversible by the AT2 receptor antagonist, EMA200 [101,102]. Similar to the above-mentioned studies, the authors hypothesized that AT1R antagonism causes a shift in endogenous AngII binding from the AT1R to the AT2R, thus indirectly causing AT2R activation.

There are numerous studies indicating that AT2R antagonism can be beneficial in treating different pain entities. In case of inflammatory pain types, the proposal that reduction in hyperinnervation can attenuate pain is in agreement with literature data [103,104]. Chakrabarty et al., reported that EMA200 reduced thermal hyperalgesia, mechanical allodynia, and pathological hyperinnervation of inflamed tissue in a model of inflammatory pain induced by complete Freund's Adjuvant (CFA) [18,22]. The same compound was also effective in the treatment of cancer-induced bone pain, which is mostly an inflam-

matory pain type, strongly depending on local inflammatory mediators [105]. The most clinically promising results, however, came from the investigation of the analgesic effect of EMA200 and its analogs in neuropathic pain, partially contradicting the above-mentioned data [28,29,37,38,80,106–108]. These include rodent models of mononeuropathic pain and even human clinical trials. AT2R antagonists were shown to be able to attenuate mechanical [37,38,81,107,108] and cold [107] allodynia in different mononeuropathic models, such as CCI or SNI. Moreover, the effect of EMA200 was validated on complex behavioral pain assays as well [109]. The most clinically relevant result, however, is that the analgesic effect of EMA401, the orally available analog of EMA200, was tested in clinical trials for postherpetic neuralgia [28,29] and diabetic neuropathy [28]. The efficacy in attenuating symptoms of the patients enrolled was acceptable in both conditions; however, two of the three studies were prematurely terminated because of preclinical data on the possible hepatotoxic effect of the test compound upon long-term administration [28]. There is no clear consensus whether AT2Rs are expressed on sensory neurons creating a direct pharmacological target for analgesia [18,37,38,106,107,110], or the observed beneficial effect is mediated by immune cells infiltrating injured nerves [80,107]. The neuro-immune cross-talk proposed by the latter studies was recently reviewed by Balogh et al. [19].
