The neuropathic pain state, which can occur after injury of a nerve or be a concurrent symptom of various clinical disorders, is considered to be one of the most important problems in medicine today (Burma et al., 2016; Burke et al., 2017). The limited understanding of its pathomechanism prevents the development of an effective treatment. Opioid drugs, used commonly in clinics for inflammatory pain therapy, lose their efficacy in neuropathic pain (Furlan et al., 2006). In recent years, a rising number of reports have confirmed the crucial role of microglial cells in neuropathic pain development (Tsuda et al., 2003; Beggs and Salter, 2013). Those cells have been shown to be strongly activated after peripheral nerve injury on the ipsilateral side of the spinal cord (Tsuda et al., 2003; Beggs and Salter, 2013). This is directly connected with the morphological changes of microglia into an amoeboid form and the enhanced expression of nociceptive factors and microglial receptors levels (Beggs and Salter, 2013). Matrix metalloproteinase 9 (MMP-9) appears to induce the development of neuropathic pain by active cleavage of IL-1β and activating p38MAPK in spinal microglia (Schonbeck et al., 1998; Kawasaki et al., 2008). While activated, microglial cells are known to release pronociceptive cytokines (for instance IL-1β, IL-6, and IL-18), which cause pain and counteract opioid analgesia (DeLeo and Yezierski, 2001; Mika et al., 2008; Narita et al., 2008). Recently, many studies have highlighted the crucial role of antinociceptive cytokines (IL-1Ra, IL-10, and IL-18BP) against pain (Ledeboer et al., 2007; Mika et al., 2013; Pilat et al., 2015, 2016). We, as well as other laboratories, have previously reported that inhibiting microglia activation with minocycline significantly reduces pain behavior and the level of pronociceptive interleukins and enhances the effectiveness of morphine in neuropathy (Mika et al., 2007; Cui et al., 2008; Zychowska et al., 2016).

Certain subtypes of purinergic ionotropic (P2X) receptors are reported to contribute to neuropathic pain development (Tsuda et al., 2003; Grygorowicz et al., 2016). We decided to take a close look at microglial subtype 4 (P2X4R) because of the increasing number of studies indicating its importance in neuropathy pathogenesis when activated by its endogenous agonist, adenosine, 5′-triphosphate (ATP) (Sim and North, 2010). Elevated levels of this nucleotide are observed after injuries, which induce the depolarization of sensory neurons (North and Khakh, 2006). The role of ATP was examined in various contexts, e.g., in knock-out mouse models or pharmacomodulation (Cockayne et al., 2000; Tsuda et al., 2003). Ulmann et al. (2008) examined P2X4 knockout mice, which showed a remarkable reduction in mechanical hypersensitivity after spinal nerve injury in comparison with wild-type animals. P2X4R might thus represent a potential therapeutic target to limit microglia-mediated inflammatory responses associated with nervous system injury (Ulmann et al., 2008). There is abundant evidence that extracellular ATP plays an important role in pain signaling in both the peripheral and central nervous system. It has been demonstrated that extracellular ATP and/or its metabolites are factors possibly involved in the transduction of mechanical stimuli into enhanced excitability of nociceptive nerve fibers in the trunk of a peripheral nerve, and it is well-documented that ATP is released by different types of mechanical stimuli (Irnich et al., 2002; Lazarowski et al., 2003). Gemes et al. (2012) indicated that nerve injury increases plasma membrane Ca²⁺-ATPase activity in sensory neurons after spinal nerve ligation. It has also been shown that an antagonist of all P2X receptors subtypes (TNP-ATP) but not an antagonist specific to all subtypes excluding P2X4R (PPADS) has a strong analgesic potency and ability to block P2X4R upregulation on microglial cell cultures, pointing to the crucial role of this specific P2X receptor subtype in pain sensation (Tsuda et al., 2003). Tricarbonyldichlororuthenium (II) dimer (CORM-2) has been reported to be an effective P2X4R antagonist (Wilkinson and Kemp, 2011). We have also recently reported that repeated intraperitoneal (ip.) administration of CORM-2 significantly reduced mechanical and thermal hypersensitivity in a rat model of neuropathy, and as preliminary data, that repeated intrathecal (ith.) administration attenuate thermal hypersensitivity. Moreover, a single ip. administration of CORM-2 was able to enhance opioid analgesia (Jurga et al., 2016a).

Despite the many studies on the topic, the involvement and changes of the P2X4R under neuropathic pain remain to be elucidated, but the P2X4R is known to be localized mainly on microglia/macrophage cells. Therefore, we started our experiments with an evaluation of the time course changes in P2X4R levels in a rat neuropathic pain model [chronic constriction injury of the sciatic nerve (CCI)]. Next, we decided to use behavioral tests to examine if or how CORM-2 injected intrathecally (ith.) can influence nociceptive transmission under neuropathic pain. Moreover, using the Western blot technique, we have measured the influence of CORM-2 on activated glia and, in parallel, analyzed the protein levels of both the pronociceptive (IL-1β, IL-18, IL-6) and antinociceptive (IL-1Ra, IL-18BP, IL-10) factors in the dorsal horn of the lumbar spinal cord and DRG in the CCI-exposed rats. Knowing that the mature forms of IL-1β and IL-18 are cleaved by MMP-9 (Schonbeck et al., 1998; Kawasaki et al., 2008) and/or nucleotide-binding oligomerization domain, leucine rich repeat and pyrin domain containing 3/caspase 1 (NLRP3/Casp-1) inflammasome complex (Franchi et al., 2009), we decided to examine the influence of CORM-2 on the activation of those factors in the neuropathic pain model. Additionally, we investigated the impact of this P2X4R antagonist on the activation of the p38MAPK and ERK1/2 pathways and changes in NFκB activation, which are known to be crucial for the production of nociceptive factors during neuropathic pain development (Tsuda et al., 2004; Hains and Waxman, 2006). Recent publications report also contribution of phosphoinositide 3-kinase PI3K/Akt pathway in development of hypersensitivity in different pain models (Tsuda et al., 2009; Pereira et al., 2011; Xu et al., 2014), so we have included it in our experiments. Moreover, we investigated how a single intrathecal administration of CORM-2 would influence opioid (morphine and buprenorphine) analgesia in a rat model of neuropathic pain.

Our studies provide new evidence that spinal microglial P2X4R contribute to neuropathy development, since we observed spinal upregulation of P2X4R levels in every measured time point in parallel with activation of IBA-1-positive cells but no changes in the DRG. Here, we report that chronic, as well as single pharmacological blockade of P2XR with CORM-2 (ith. administration, once daily for 9 days) diminished both mechanical and thermal hypersensitivity in a rat model of neuropathic pain on day 7 after CCI, which is a continuation of our previously published results suggesting an important role of P2X4R, however, we have measured only its influence on thermal hypersensitivity at that time (Jurga et al., 2016a). Our results indicate, for the first time, that blockade of P2X4R with chronic ith. administration of CORM-2 prevents elevated levels (on day 7 after CCI) of glial markers (IBA-1 and GFAP) in the spinal cord and/or DRG bringing it closer to the level observed in the INTACT animals. A certain amount of pronociceptive factors, e.g., cytokines, are known to be upregulated in neuropathic pain and in the CCI model of neuropathy. We have examined the protein levels of IL-1β, IL-18, and IL-6 released by activated glial cells in the spinal cord and DRG. The results we report suggest that blockade of microglial P2X4R transmission brings those factors back to physiological levels. The antinociceptive factors we examined (IL-1Ra, IL-18BP, and IL-10) remained at the same level after CCI as those in INTACT animals; however, we observed a strong upregulation of IL-1Ra in CORM-2-treated CCI-exposed rats. The observation that even a single ith. administration of CORM-2 has additive analgesic with morphine and buprenorphine on day 7 after CCI supplements our previous results on chronic ip. administration of CORM-2 and suggests that P2X4R is an interesting molecular target for the treatment of neuropathic pain.

As we have reported previously (Jurga et al., 2016a), activated spinal microglia upregulation starts as early as day 2 after CCI, with the biggest change on day 7, and lasts until day 14, when it is still significant. Similarly, upregulation of microglia protein markers was reported by others in the rat and mouse sciatic nerve injury model on day 7 (Bishay et al., 2010) and day 14 (Kang et al., 2015). Astrocyte upregulation was reported in CD1 mice in the CCI model on days 10 and 21 after surgery (Chen et al., 2014), which, in combination with our results, indicate that astroglia are activated later than microglia. In relation to those results, we have observed spinal upregulation of the P2X4R protein level parallel to microglial activation. This suggests a relation between P2X4R and microglial activation in neuropathy development in the Wistar rat CCI model, especially considering that P2X4R was reported to co-localize only with microglia and not with astrocytes or neurons in the central nervous system (Tsuda et al., 2003). Results from the Wistar rat SNL model showed upregulation in P2X4R protein levels on days 1, 3, 7, and 14 (Tsuda et al., 2003) and in the Sprague-Dawley rat CCI model on days 3, 7, and 14 (Chen X.M. et al., 2015). DRG tissue analysis has not been reported yet; however, here we show that the P2X4R level remains unchanged. This supports the hypothesis on the contribution of microglial P2X4R to neuropathy development because IBA-1-positive cells responsible for the upregulation in the DRG are macrophages and not microglia.

ATP, acting via P2 purinergic receptors, has been suggested to mediate neuropathic pain development. The ATP-gated P2X4R is considered to be involved in the modulation of spinal microglia activity and peripheral macrophages and in the maintenance of hypersensitivity caused by peripheral nerve injury (Trang and Salter, 2012). Here, we report that pharmacological blockade of P2X4R with CORM-2 (repeated and single ith. administration) has the potential to diminish both mechanical and thermal hypersensitivity in the rat CCI model of neuropathic pain as measured on day 7 after the operation. This complements our previously published results regarding ip. administration (Jurga et al., 2016a) and supports the hypothesis on the important role of P2X4R in neuropathic pain development. Moreover, treatment with other available P2XR antagonists showed results also favoring this supposition. For instance, experiments on mice where the effects of two antagonists of P2XR were compared (one selective to P2X4R and one to all other P2XRs) clearly indicated that P2X4R is the most important among the P2X receptor family (Tsuda et al., 2003). Until now, it has been reported that ip. administration of CORM-2 has an analgesic potency in a mouse SNL model (Liu et al., 2015), a mouse inflammatory pain CFA model (Negrete et al., 2014) and in a Sprague-Dawley rat CCI model 14 days after CCI (Chen Y. et al., 2015). Blockade of receptors on microglia cells is a promising target for neuropathic pain pharmacotherapy, as reports regarding antagonism of Toll-like receptors 4 (Lewis et al., 2013; Jurga et al., 2016b), histamine H4 receptors (Medhurst et al., 2008; Hsieh et al., 2010; Dong et al., 2014), or chemokine receptors (e.g., CCR5, CXCR4) (Bhangoo et al., 2007; Hu et al., 2015; Bai et al., 2016; Kwiatkowski et al., 2016; Luo et al., 2016; Piotrowska et al., 2016) are increasingly common. We decided to investigate the changes caused by P2X4R antagonism with CORM-2 (repeated ith. administration) in some pathways related to the activity of those receptors whose role in neuropathic pain has already been proposed. First, we examined the p38MAPK pathway. It is known that p38MAPK is activated in microglia/macrophages cells in neuropathic pain models (Jin et al., 2003; Tsuda et al., 2004). Interestingly, it is reported that TNP-ATP, a specific P2X4R antagonist, has the ability to downregulate this receptor in microglial cultures when its level was previously elevated after ATP stimulation (Tsuda et al., 2003). This draws attention to the possible relationship between activation of P2X4R and the p38MAPK pathway, which leads to the enhanced production of pronociceptive factors, including NO and cytokines (DeLeo and Yezierski, 2001), and to the activation of transcription factors responsible for regulation of genes involved in nociception (Potucek et al., 2006). Pharmacological inhibition of p38MAPK and activation of microglia/macrophages has analgesic potency in neuropathic pain models, as has been reported with minocycline (Tikka et al., 2001; Popiolek-Barczyk et al., 2015) or SB203580 (Tsuda et al., 2004) treatment. In our experiments, we observed a downregulation of p38MAPK by CORM-2 in the DRG but not in the spinal cord in the neuropathic pain model. Another member of the MAPK family that we examined was ERK1/2, which is strongly activated within neurons and glia during neuropathic pain. Our results are in agreement with previously published results that show that ERK1/2 phosphorylation is increased on day 7 after CCI (Rojewska et al., 2015), which is correlated with enhanced glial cell activation (Popiolek-Barczyk et al., 2015). Previously, published results suggested that ERK1/2 is essential for the intracellular signaling that leads to the production of various nociceptive factors, including cytokines, that participate in the intensification of the neuropathic pain sensations (Popiolek-Barczyk et al., 2015; Rojewska et al., 2015). In our experiments, we measured the spinal upregulation of p-ERK1/2 by CORM-2 in the neuropathic pain model and, in parallel, the level of antinociceptive IL-1Ra. However, whether this or other (unknown yet) relationship is important for the observed effect requires clarification and further studies. In the DRG we did not observe any impact of CORM-2 administration on IL-1Ra levels, which can be explain by ith. administration. Excessive production of the above-mentioned pronociceptive factors, e.g., interleukins, are known to be upregulated in neuropathy models (Verri et al., 2007; Rojewska et al., 2014, 2015; Popiolek-Barczyk et al., 2015). IL-1β and IL-18 promotes pain behavior development when injected intrathecally (Malcangio et al., 1996; Miyoshi et al., 2008). IL-1β, IL-18 and IL-6 are spinally upregulated in neuropathic pain models, including our reports (Okamoto et al., 2001; Pilat et al., 2015, 2016; Rojewska et al., 2016). Interestingly, the results we observed suggest that blockade of microglial P2X4R transmission restores the elevated levels of pronociceptive factors in neuropathy. Moreover, we have shown that the MMP-9 level is elevated in rats after CCI in the spinal cord and DRG (Schonbeck et al., 1998; Kawasaki et al., 2008; Rojewska et al., 2014), but here, we are the first to show that preemptive and repeated P2X4R blockade has the ability to prevent this upregulation. This observation is important, since Kawasaki et al. (2008) has proven that MMP-9 produced in the injured DRG neurons serves as one of the triggers for spinal microglia activation and neuropathic pain development and that MMP-9-induced pathophysiology involves IL-1β cleavage and microglia p38MAPK activation. These observations, along with the identical regulation of pronociceptive spinal IL-1β in the spinal cord, are extremely important. Our results are also in agreement with Vaccari et al. (2012), who reported a decrease in IL-1β release in P2X4R^-/- mice after spinal cord injury, which resulted in an improvement in the functional outcome. The results of our studies provided evidence that blocking P2X4Rs prevents microglial activation and, consequently, the release of pronociceptive molecules. It is claimed that under neuropathic pain, the physiological balance between pronociceptive and antinociceptive factors is disturbed (Mika et al., 2008; Rojewska et al., 2014, 2015). The corresponding antinociceptive factors that we examined were expressed at the same level as those in the INTACT group on day 7 after CCI. However, we observed a strong upregulation of spinal IL-1Ra in CORM-2-treated CCI-exposed rats. IL-1Ra is a member of the IL-1 family that binds to IL-1 receptors but does not induce any intracellular response. The IL-1Ra blocks the pronociceptive effects of IL-1β and therefore has a strong analgesic effect under neuropathy (Hook et al., 2011; Pilat et al., 2015). Chronic ith. administration of CORM-2 inhibited microglial activation in neuropathy and enhanced the level of IL-1Ra, an important antinociceptive factor. We also examined the level of IL-18BP, which is known to inactivate pronociceptive IL-18 (Kim et al., 2000; Pilat et al., 2016); however CORM-2 administration did not influence its level. Another important antinociceptive cytokine is IL-10 (Bluthé et al., 1999; Heyen et al., 2000), but like IL-18BP, its level was not regulated by CORM-2 administration. This may suggest that the mechanism of IL-1Ra activation relies on the P2X4R-mediated pathway and that this molecule, along with pronociceptive IL-1β (IL-1Ra blocks its activation), plays a key role in neuropathy development.

ATP, an endogenous agonist of purinergic receptors, has the ability to activate p38MAPK with the release of cytokines and other microglial factors engaged in nociception. According to the literature, the second pathway important to investigate was the NLRP3/Caspase-1 inflammasome, because this complex is responsible for the cleavage of the preforms of IL-1β and IL-18 into active molecules (Rathinam et al., 2012). Moreover, Jiang L. et al. (2016) reported that in mice, CORM-2 administration in an acute lung injury model downregulates the level of elevated NLRP3/Caspase-1 in the lung bronchoalveolar lavage fluid, suggesting engagement of P2X4R in the activation of the inflammasome. However, we did not observe activation of this complex in the CCI model, which we found was in agreement with the studies of Curto-Reyes et al. (2015) who recently reported that pain development is not dependent on NLRP3/Caspase-1 in a mouse SNL model (Table 1).

The recent literature data suggest that activation of PI3K/Akt pathway is strongly involved in release of interleukins and other microglial factors involved in nociception, which is in agreement with our results showing the activation of PI3K/Akt pathway after CCI in rats on day 7. Similarly, it has been presented, that this pathway is upregulated undergo activation, e.g., in Oxaliplatin chemotherapy-induced neuropathic pain in mice (Jiang S.P. et al., 2016). Significant upregulation of p-Akt in DRG neurons was also reported after L5 spinal nerve ligation rat model, with attenuation of behavioral pain symptoms after PI3K inhibitors administration (wortmannin or LY294002) (Xu et al., 2014). It was proven that AS605240 (PI3K inhibitor) produced a significant relief in tripsine-induced nociception in mice (Pereira et al., 2011). Not to neglect remains also the fact, that administration of PI3K inhibitor (LY294002) into primary microglial culture suppresses the P2X4R upregulation after fibronectin introduction (Tsuda et al., 2009). Moreover, it has been shown, using chemoattractant protein-1 (MCP1)-mediated bone cancer pain model, that there is a significant dependence between PI3K/Akt and microglia activation (Jin et al., 2015). Here we show that repeated ith. administration of CORM-2 lowers levels of both PI3K and p-Akt protein in spinal cord and DRG suggesting its significant contribution to P2X4R-mediated pain development.

The P2X4R antagonist has potency to not only diminish mechanical and thermal hypersensitivity in a rat CCI model of neuropathic pain but also enhance opioid analgesia. The observation that a single ith. administration of CORM-2 enhances the analgesic effects of single-injected morphine and buprenorphine on day 7 after CCI supplements our previous results obtained after ip. administration of CORM-2 (Jurga et al., 2016a). It is known that opioids, commonly used in the clinic for inflammatory or acute pain therapy, lose their efficacy in neuropathic pain. This may be connected with the development of tolerance (Narita et al., 2008), downregulation of opioid receptors (Mika et al., 2014) and activation of p38MAPK along with the enhanced release of pronociceptive factors (Raghavendra et al., 2004; Hutchinson et al., 2008a). Inhibition of microglia activation and secretion of nociceptive factors by certain pharmacological tools has been shown to improve opioid analgesia, for example treatment with an inhibitor of microglia – minocycline (Hutchinson et al., 2008b), pronociceptive interleukin – propentofylline (Miki and Miki, 1991; Raghavendra et al., 2004) or p38MAPK inhibitor - SB203580 (Cui et al., 2006). Moreover, in 2015 authors suggested that the activation of PI3K/Akt pathway can influence the opioid effectiveness, which stays in agreement with our pharmacological results showing both morphine and buprenorphine analgesic properties are enhanced by CORM-2, inhibitor of this pathway (Lutz et al., 2015). According to those results, pharmacological blockade of P2X4R may be a promising target for neuropathic pain therapy because, as we report here, CORM-2 administration leads to the attenuation of pronociceptive cytokine levels and significant analgesia with the ability to enhance both morphine and buprenorphine effectiveness. Together, this indicates that P2X4R may play an important role in nociception and opioid effectiveness in neuropathic pain and is, therefore, a promising target for effective therapy construction.

Our results support the hypothesis that spinal microglial P2X4R contribute to neuropathy development, since we have observed spinal upregulation of P2X4R levels at every time point in parallel with activation of IBA-1-positive and GFAP-positive cells. Our results confirmed that preemptive and repeated pharmacological spinal blockade of P2X4R can attenuate pain as well as improve opioid analgesia on day 7 after CCI. Interestingly, we observed (Scheme 1) that CORM-2 inhibited pronociceptive factors (IL-1β, IL-18, IL-6) in the spinal cord and/or DRG along with the MMP-9 factor, which is a well-known contributor to IL-1β activation and spinal microglia activation. Moreover, CORM-2 treatment led to the elevation of antinociceptive IL-1Ra, a molecule whose action opposes that of pronociceptive IL-1β. With strong support from our results summarized in Scheme 1, we are proposing a hypothesis that MMP-9, p38MAPK, ERK1/2 and PI3K/Akt, not TIMP-1 and the NLRP3/Caspase-1 inflammasome complex, are involved in the analgesic effects of CORM-2 in rat neuropathic pain. Taking this into consideration, it is highly probable that microglial P2X4R may be involved in nerve-injury caused nociception because our results showed that treatment with CORM-2 restores the physiological level of pronociceptive factors in neuropathy and attenuates pain. This suggests that P2XR is a promising target for pharmacomodulation in future trials.

Only the minimal essential number of animals was used, and all of the procedures were performed according to the recommendations of IASP (Zimmermann, 1983) and the NIH Guide for the Care and Use of Laboratory Animals. This study was carried out in accordance with the recommendations of local Ethics Committee (Krakow, Poland), permission number: 1055.

AJ, AP, WM, BP, and JM performed the experiments, AJ and JM designed the study, AJ, AP, JM analyzed and interpreted the results, AJ, AP, WM, BP, and JM drafted the manuscript and accepted the finalized version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer LVD and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.