Abstract

Neuropathic pain is a pathological symptom experienced worldwide by patients suffering with nervous system dysfunction caused by various diseases. Treatment of neuropathic pain is always accompanied by a poor response and undesired adverse effects. Therefore, developing a novel “pain-kill” drug design strategy is critical in this field. Recent evidence demonstrates that neuroinflammation and immune response contributes to the development of neuropathic pain. Nerve damage can initiate inflammatory and immune responses, as evidenced by the upregulation of cytokines and chemokines. In this paper, we demonstrated that different chemokines and chemokine receptors (e.g., CX3CL1/CX3CR1, CCL2/CCR2, CCL3/CCR1, CCL4/CCR5 and CCL5/CCR5) serve as mediators for neuron–glia communication subsequently modulate nociceptive signal transmission. By extensively understanding the role of chemokines in neurons and glial cells in nociceptive signal transmission, a novel strategy for a target-specific drug design could be developed.

Keywords

chemokines; cytokines; pain: neuropathic;

1. Introduction

Neuropathic pain is a common pathological symptom of multiple neural diseases. It has been extensively studied. Growing evidence demonstrates that the recruitment and infiltration of inflammatory cells such as neutrophils and macrophages (which are activated during nerve injury) lead to microenvironmental changes that subsequently initiate hyperalgesia. In previous studies, a relationship has been identified between mediators released by the activation of inflammatory cells and the development of hyperalgesia in neuropathic pain. Proinflammatory cytokines such as tumor necrosis factor-α (TNF-α), interleukin 1 β (IL-1β), and interleukin 6 (IL-6) induce acute or short-term hyperalgesia and are implicated directly in chronic hyperalgesia and allodynia. Besides inflammatory cytokines, recent evidence supports that chemokines and chemokine receptors are involved in the development of hyperalgesia and in pain transmission mechanisms. The review in this paper therefore will (1) briefly focus on elucidating the relationship between inflammation and neuropathic pain and (2) focus on the role of each major chemokine and chemokine receptor that is being studied at this moment.

2. Inflammation and neuropathic pain

It is widely accepted that the loss of myelinated or unmyelinated sensory axons in sciatic nerve injury is a major factor in neuropathic pain. Since 1990, other chemical stimuli such as inflammatory mediators have been demonstrated to induce hyperalgesia via C-polymodal nociceptors without loss of any sensory neurons¹; however, the idea that inflammation may play a role in neuropathic pain was excluded by categorizing these data as part of inflammatory pain, rather than as related to neuropathic pain. The acute changes in the endoneural microenvironment interestingly are positively related to the development of hyperalgesia. These changes include the recruitment of neutrophils and macrophages, followed by macrophage activation, glial cell activation, and cytokine release.2, 3 In fact, glial cells and Schwann cells may also be essential for the inflammatory process because they are capable of releasing mediators such as cytokines and acting as antigen-presenting cells after nerve injury. These accumulated data directly implicate tumor necrosis factor alpha, interleukin 1, and interleukin 6 in acute and chronic hyperalgesia and in allodynia.4, 5, 6, 7 Thus, the concept that inflammation is involved in neuropathic pain mechanisms seemed reasonable and was gradually taken for granted. However, no animal models have been employed to verify the idea that inflammation may attribute to the development of neuropathic pain. By contrast, the release of anti-inflammatory cytokines and opioid peptides has also been observed in recruited inflammatory lineages in neuropathic pain models, which may reduce behavioral hypersensitivity.8, 9 Thus, during nerve injury, the recruitment of circulating polymorphonuclear cells, monocytes, and opioid peptide-containing lymphocytes may indeed attenuate behavioral hypersensitivity while augmenting the process of inflammation in neuropathic pain models. These observed analgesic effects of recruited inflammatory leukocytes can be blocked by antiopioid peptide antibodies¹⁰ or by inhibiting their trafficking to the injured nerve.¹¹ Until 2011, when our group used immunocompromised mice models, the role of recruited inflammatory lineages in neuropathic pain models remained controversial.¹² To observe the difference in immune responses and hypersensitivity, we first used normal mice and immunocompromised mice with acute mechanical nerve injury. Our findings demonstrated attenuated inflammatory responses and behavioral hypersensitivity in the immunocompromised mice. This suggests that attenuated inflammation impedes the development of hypersensitivity. The severity of acute inflammation contributes to approximately 70% behavioral hypersensitivity shortly after an injury, but reduces hypersensitivity later; this reflects the influence of immune responses on the functional alteration of the injured nerve.¹² We furthermore tested the converse hypothesis by enhancing the recruitment of inflammatory lineages to the injured nerve, which was achieved by the continuous exogenous administration of recombinant mouse granulocyte-colony-stimulating factor (rmG-CSF).¹³ Increased inflammatory cytokines and the content of opioid peptides have been observed in mice receiving rmG-CSF; however, exacerbated behavioral hypersensitivity, but not attenuation, has also been recorded in these mice and thus indicates that augmented inflammation facilitates the development of hypersensitivity.¹³ Therefore, our group was the first to establish legally a direct connection between inflammation and neuropathic pain in animal models. Our data also suggested that endogenous opioid peptides released from recruited inflammatory lineage are an innate negative feedback system for the nociceptive inflammatory process.

3. Cytokine and chemokine networks in neuropathic pain

Accumulated evidence suggests that proinflammatory cytokines and chemokines play a part in the pathogenesis of neuropathic pain.¹⁴ The expression of proinflammatory cytokines such as IL-1β, TNFα, and IL-6 (which are upregulated in peripheral nerve injury) elicits the inflammatory process and subsequently elicits neuropathic pain.¹⁵ Macrophages and Schwann cells primarily produce IL-1β during neuroinflammation and are responsible for the recruitment and activation of other inflammatory lineages. Evidence also supports that IL-1β is capable of directly sensitizing nociceptors such as the transient receptor potential cation channel, subfamily V member 1 (TRPV1),¹⁶ which is a heat-sensitive and chemosensitive cation channel located in primary afferent neurons. Further evidence demonstrates that a simulated neuropathic pain response can be recorded by applying exogenous IL-6 and TNFα locally to the peripheral nerve.¹⁷ These results are in line with the data obtained from neuropathic pain models that used genetically modified IL-1β knockout mice and IL-6 knockout mice in which injured nerve-related autotomy, mechanoallodynia, spontaneous ectopic neuronal activity, and adrenergic sprouting in sensory ganglia had been attenuated.18, 19

Because some chemokine receptors such as C-C chemokine receptor type 2 (CCR2) and C-C chemokine receptor type 5 (CCR5) are present in primary afferent neurons and secondary neurons of the spinal dorsal horn,²⁰ it is possible that chemokines produced by recruited inflammatory lineages may alter nociceptive transduction by activating chemokine receptors in the dorsal root ganglia (DRG). Therefore, it is not surprising that behavioral hypersensitivity could be elicited by the peripheral administration or spinal administration of CCL2, C-C chemokine receptor type 3 (CCL3), CCL5, or by the chemokine (C-X-C motif) ligand 12 (CXCL12).²¹ However, a more complex role of the chemokine-chemokine receptor in modulating neuropathic pain exists in the CXCL2-CXCR2 axis, which demonstrates bidirectional nociceptive modulation. This includes the analgesic effect observed by blocking CXCR2 with a selective antagonist and the analgesic effect observed by the peripheral administration of CXCL2 in animal models.²² These results indicate that chemokines may exert a complicated, reciprocally regulated process for nociceptive modulation. Further efforts are needed to elucidate their individual roles in the development of neuropathic pain mechanisms.

4. Chemokines and neuropathic pain: CX3CL1/CX3CR1

The chemokine (C-X3-C motif) ligand 1 (CX3CL1) was previously not considered part of the neuropathic pain-related chemokine network because of its constitutive expression as full length mRNA in the spinal cord and in the DRG neurons,23, 24, 25 and because the expression of CX3CL1 mRNA does not change significantly after nerve injury. It was later realized that CX3CL1 is characterized by two distinct forms—a membrane-bound form and a soluble form—that can be cleaved by metalloproteases.²⁶ The membrane-bound CX3CL1 is remarkably reduced in the DRG after spinal nerve ligation, which also induces the upregulation of CX3CR1 (the only receptor for CX3CL1) on the ipsilateral side of the spinal cord.²⁷ Because CX3CL1 is primarily expressed in neurons and CX3CR1 is substantially expressed in the microglia in the spinal cord dorsal horn,¹² it is plausible to postulate that a reciprocal cross-talk may exist between spinal neurons and microglial cells via this complementary distributed CX3CL1/CX3CR1. Later experiments identified significant chemotaxis activity of soluble CX3CL1 for the microglia,²⁶ which suggests CX3CL1-mediated microglia recruitment during spinal nerve ligation.

Several proteases may be involved in the cleavage of CX3CL128, 29; of great interest is the metalloproteases.²⁶ We have established the expression profiles in the central nervous system by using murine sciatic nerve ligation models.³⁰ Our data showed that matrix metalloproteinase 2 (MMP-2) and matrix metalloproteinase 9 (MMP-9)—but not matrix metalloproteinase 12 (MMP-12)—are expressed in the murine spinal cord. In neurons and microglial cells, MMP-9 is constitutively expressed and immediately upregulated after nerve injury. In glial-like cells, MMP-2 is expressed and gradually increases after nerve injury.³⁰ The upregulation of MMP-9 in DRG neurons in the early phase of nerve injury interestingly leads to microglial activation and the development of neuropathic pain.³¹ However, no evidence supports the existence of MMP-2-mediated CX3CL1 cleavage, except in hepatocytes in which CX3CL1 is primarily processed by MMP-2.³² Therefore, MMP-9 and MMP-2 may participate in the development of neuropathic pain by activating microglia in early-phase and late-phase nerve injury, respectively, primarily viacleavage of CX3CL1 from the spinal neuronal membrane in the DRG.

The involvement of neuronal-derived CX3CL1 in microglia activation is further supported by several in vivo studies, described later. As expected, CX3CR1 upregulation was observed in the spinal microglia after peripheral nerve injury.²⁷Minocycline can block thermal hyperalgesia in microglia that is induced [via the activation of p38 mitogen-activated protein kinase (p38 MAPK)] through the spinal administration of CX3CL1. Phosphorylated p38 MAPK is complementarily and substantially expressed in the spinal microglia. Intrathecal blockage of CX3CR1 by a specific neutralizing antibody can decrease p38 MAPK activation. These results altogether demonstrate a critical pathway for neuron-microglia communication in neuropathic pain, whereas soluble neuronal CX3CL1 binds to CX3CR1 on microglial cells and subsequently phosphorylates p38 MAPK and induces microglia activation. Activation of p38 MAPK would furthermore induce the synthesis of proinflammatory cytokines such as IL-1β, TNFα, and IL-6, which may lead to central sensitization at low concentrations via a distinct pathway.³³ Interleukin 1β exerts an intriguing dual function in enhancing excitatory synaptic transmissions and in reducing inhibitory synaptic transmissions in the dorsal horn neurons, and it results in central sensitization; by contrast, TNFα only enhances excitatory synaptic transmissions and IL-6 reduces inhibitory synaptic transmission at the same spinal cord level.³⁴ These results suggest the pivotal role of the cytokines in regulating synaptic plasticity and neuronal excitability.

5. Chemokines and neuropathic pain: CCL2/CCR2

Monocyte chemoattractant protein 1 (MCP-1), also known as CCL2, is a major inflammatory chemokine and is specifically responsible for recruiting monocytes to the site of inflammation. It binds to CCR1, CCR2, and CCR4 to exert distinct biological functions; however, CCL2 has a much higher affinity for CCR2.³⁵Evidence supports that CCL2/CCR2 signaling is essential to the development of neuropathic pain.³⁶ The existence of CCL2 in DRG neurons is quite certain; however, the expression level is controversial. Most reports show a very low basal level,37, 38 but other reports show a normal CCL2 expression level in 40% of small and medium DRG neurons.³⁹ However, solid evidence supports that CCL2 is highly inducible in DRG neurons and satellite cells after peripheral nerve injury. Further studies using staining techniques show that CCL2 is coexpressed with substance P, calcitonin gene-related peptide (CGRP), and the capsaicin receptor (i.e., TRPV1).³⁹ In the DRG, it is also partially colocalized with CCR2.¹³ In situ hybridization demonstrates that CCR2 expression in neurons and in glial cells can be induced by nerve injury.⁴⁰ Thus, neuron-derived CCL2 may bind to CCR2 on neurons and glial cells to modulate pain transmission.13, 40

Whether this CCL2/CCR2 signaling is true for the spinal cord and for the higher central nervous system remains uncertain because the expression of CCL2 and CCR2 in the spinal cord and cerebrum is controversial. Recent information in the medical literature demonstrates that CCL2 is expressed in primary afferents and in astrocytes,⁴¹ as evidenced by the colocalization of CCL2 with glial fibrillary acidic protein, and can be released from activated astrocytes on peripheral nerve injury.⁴¹Other evidence shows that CCL2 is also expressed in cerebral astrocytes after different types of cerebral injury such as axonal transaction⁴² and focal cerebral ischemia.⁴³ There is similar debate for the location of CCR2 expression in the spinal cord and the higher level of central nerve system. The expression of CCR2 was first discovered in spinal microglial cells.²¹ Later reports further detected CCR2 in spinal astrocytes⁴⁴ and in neurons.⁴¹ However, more evidence suggests that CCR2 is constitutively expressed in the spinal cord neurons.⁴¹ This evidence includes the presence of green fluorescent protein (GFP) signals in the dorsal horn neurons of CCR2-GFP reporter mice, the detection of inducible CCR2 mRNA in the dorsal horn neurons and motor neurons by in situ hybridization,21, 41, 44 the rapid response of spontaneous excitatory postsynaptic currents (sEPSCs) and N-methyl-d-aspartate (NMDA) currents in the dorsal horn neurons with the local application of CCL2⁴¹ (but not CX3CL1²⁷) in electrophysiological studies. These findings suggest the constitutive expression of functional CCR2 in dorsal horn neurons.

The expression of CCL2/CCR2 is principally located in the DRG and the dorsal horn; however, it may be that CCL2 is transported from the DRG to the central terminal (as mentioned previously), to the substance P (SP), and to CGRP-positive primary afferents in the superficial dorsal horn.³⁹

The idea that CCL2 can be released from primary afferents in the dorsal horn in an active manner is further supported by the following reports. Important data reported by Thacker et al⁴⁵ in an ex vivo model using the isolated dorsal horn and the dorsal root of the spinal cord show that electrical stimulation can induce the rapid release of CCL2 in the dorsal root in neuropathic preparations, but not in sham preparations. Thus, CCL2 may also modulate the nociceptive signaling pathway. In addition, the spatial expression of CCL2 and the activation of microglia, induced by peripheral nerve injury, are very similar to each other.45, 46, 47 The activation of microglia and p38 MAPK activation can also be attenuated by the local application of a CCL2 antibody in the spinal cord of CCR2 knockout mice.21, 48Therefore, it may be that CCL2/CCR2 signaling, which can be activated by peripheral nerve injury, plays a pivotal role in spinal microglia activation and in the development of neuropathic pain.

As mentioned previously, CCL2 expression can be induced in astrocytes and CCR2 is constitutively expressed in spinal neurons. It is plausible to hypothesize that CCL2/CCR2 signaling may also serve as the bridge between astrocytes and neurons in the development of neuropathic pain. This theory can be supported by the aforementioned evidence that CCL2 increases the frequency and amplitude of sEPSCs and enhances the NMDA current and the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) inward current in the dorsal horn lamina II neurons.⁴¹ These results suggest CCL2 presynaptically49, 50 enhances glutamate release and postsynaptically⁴¹ enhances glutamate receptor function, and it has a postsynaptic effect on the NMDA current.⁴¹ Other evidence moreover suggests that CCL2 further negatively regulates the inhibitory gamma-aminobutyric acid (GABA)-induced current in spinal neurons independently.⁵¹ Further evidence that CCL2 is involved in neuropathic pain and has a direct action on spinal cord neurons is represented by behavior studies in which the spinal injection of CCL2 elicited heat hyperalgesia and phosphorylation of extracellular signal-regulated kinases (ERK), a nociceptive-specific signal and marker for central sensitization, in dorsal horn neurons.⁴¹ These data altogether support a direct and indirect effect of CCL2/CCR2 signaling in the development of neuropathic pain.

6. Chemokines and neuropathic pain: CCL3/CCR1, CCL4/CCR5, and CCL5/CCR5

In recent years, the roles of inflammatory chemokines have been examined in neuropathic pain models. The involvement of macrophage inflammatory protein-1 alpha (MIP-1α, CCL3) was examined in murine partial sciatic nerve ligation (PSNL) models.52, 53 The upregulation of CCL3 and CCR1, but not CCR5, was observed after PSNL, which paralleled the development of allodynia and thermal hyperalgesia. The expression of CCL3 after PSNL was identified primarily in macrophages and Schwann cells in the injured sciatic nerve.⁵³ The behavioral and biochemical findings could be attenuated by the administration of intrathecal CCL3 antibody and exaggerated by the administration of recombinant CCL3 protein.52, 53Macrophage inflammatory protein-1 beta (MIP-1β; also called CCL4) was similarly examined for its role in neuropathic pain by using murine PSNL models.⁵⁴ The expression of CCL4 after PSNL was also identified in macrophages and Schwann cell in the injured sciatic nerve. The upregulation of CCL4 and CCR5 after PSNL was observed, which was in line with the development of allodynia and thermal hyperalgesia. These pathological findings can be attenuated by the administration of either CCL4 antibody or CCR5 antagonist.⁵⁴

The C-C motif chemokine ligand 5 [CCL5; also known as regulated on activation, normally T-expressed, and presumably secreted (RANTES)] acts as a ligand for the chemokine receptors CCR1, CCR3, and CCR5 and directs the migration of monocytes/macrophages and T cells.55, 56 Accumulated evidence indicates that CCL5 modulates the inflammatory response in several pathological conditions. Aberrant regulation of CCL5 activity may result in disease, and it indeed has been implicated in the pathogenesis of several inflammatory disorders such as multiple sclerosis, human immunodeficiency virus dementia, and rheumatoid arthritis,⁵⁷and in animal models of arthritis, endotoxemia, and chronic colitis.58, 59 The authors of other reports conclude that reduced CCL5 production and inflammation may be a productive strategy for future therapy in clinical settings. In regard to the CCL5-related pathogenesis of painful neuropathies, previous reports of the intracisternal administration of CCL5 show that it increases the number of scratching bouts and the scratching duration that had been induced by a subcutaneous injection of formalin in the orofacial area.⁶⁰ It induced dose-dependent hyperalgesia when infused directly into the periaqueductal gray.⁶¹ Recent studies also confirm a role of CCL5 in nociception via the induction of pain, after the intradermal injection of CCL5 into the rat paw.⁶² In injured nerves, CCL5 has been identified, which implicates a potential role for this chemokine in neuropathology, and recent studies confirm the role of CCL5 in peripheral nociception.⁶³ These data suggest that peripheral CCL5 is implicated in the processing of pain information. However, direct evidence and the possible mechanism or mechanisms for the involvement of peripheral CCL5 in nociceptive processing in chronic neuropathic pain has been lacking.

Therefore, our group first investigated the potential of the selective CCL5 receptor antagonist methionylated RANTES (Met-RANTES), administered via peritoneal administration, to modulate the recruitment of inflammatory cells at injured sites and to attenuate nociceptive responses in a mouse neuropathic pain model.⁶⁴Nociceptive sensitization, immune cell infiltration, multiple cytokine secretion, and opioid peptide expression in damaged nerves were studied. Our results indicated that Met-RANTES-treated mice had less behavioral hypersensitivity after partial sciatic nerve ligation. The infiltration of macrophages, the secretion of proinflammatory cytokine proteins (e.g., TNFα, IL-1β, IL-6, and IFNγ), and the expression of enkephalin, β-endorphin, and dynorphin mRNA in damaged nerves after partial sciatic nerve ligation were significantly decreased in Met-RANTES-treated mice.⁶⁴ These results suggest that CCL5 is capable of regulating the microenvironment that controls behavioral hypersensitivity at the level of the peripheral injured site in a murine model of chronic neuropathic pain.

We further tested the hypothesis that the lack of CCL5 would modulate the recruitment of inflammatory cells to painful, inflamed sites and could attenuate pain in a murine model of chronic neuropathic pain.⁶⁵ Our results indicated that CCL5 knockout mice had less behavioral hypersensitivity after PSNL. The infiltration of macrophages and the secretion of proinflammatory cytokines (e.g., TNFα, IL-1β, IL-6, and IFNγ) in damaged nerves after PSNL were significantly decreased in the CCL5 knockout mice. By contrast, several anti-inflammatory cytokine proteins (i.e., IL-4 and IL-10) were significantly increased in the CCL5 knockout animals, and the expression of enkephalin, β-endorphin, and dynorphin was significantly lower in knockout mice than in wild-type control mice. These results represent the first evidence that CCL5 is capable of regulating the pathway that controls hyperalgesia at the level of the peripheral injured site in a murine chronic neuropathic pain model.⁶⁵ We demonstrated that a lack of CCL5 modulated cell infiltration and the proinflammatory milieu within the injured nerve. Attenuated behavioral hypersensitivity in CCL5 knockout mice observed in the study could have resulted from decreased macrophage infiltration, mobilization, and functional ability at injured sites.

Recent evidence implies that CCR5 may play an essential role in the development of neuropathic pain, but there is still no report that provides direct evidence of CCR5 in the development of neuropathic pain.

Our group has studied the role of CCR5 in neuropathic pain by using CCR5 knockout mice and bone marrow transplant chimera. Our results demonstrated that CCR5 could be critical for T lymphocyte function during inflammation and during the development of neuropathic pain (manuscript in preparation). Further studies are needed to explore the details of the mechanism of CCR5 in modulating neuropathic pain transmission.

7. Chemokines and neuropathic pain: leptin and STAT3 signal transduction

Chemokines participate in the development of neuropathic pain, which involves various inflammatory mediators in the peripheral and central nerve system. A rising feature is the adipokines (which are adipocyte-derived mediators), some of which are also classified as chemokines. One adipokine, leptin, was recently found to be a novel inflammatory molecule and directly involved in the development of neuropathic pain. When a peripheral nerve is injured, leptin is highly expressed and released in the surrounding adipose tissue, which subsequently upregulates the expression of CCL3, CCL4, and CCL5 in macrophages via activating the signal transducer and activator of transcription 3 (STAT3) pathway. The STAT3 signal pathway interestingly can be activated by nerve injury in glial cells residing in the spinal dorsal horn, which lacks adipose tissue and certainly lacks leptin. The central activation of STAT3 in glial cells can be elicited by the release of IL-6 and can subsequently enhance neuropathic pain. These findings altogether suggest that the STAT3 signal pathway is a novel key target that may lead to novel therapeutic strategies for future drug development because of its involvement in peripheral and central cytokine/chemokine network-mediated regulation of neuroinflammation and related neuropathic pain mechanisms.

8. Conclusion

The importance of chemokines in neuroinflammation and neuropathic pain is twofold. First, inflammatory chemokines participate in the recruitment of leukocyte infiltration and they are relevant to the process of neuroinflammation. Second, chemokines and chemokine receptors may be directly involved in the alteration of neuronal plasticity, the development of nociception and hyperalgesia, the modulation of spinal neuronal activity, microglia and astrocyte communication, and individual activities in peripheral nerve injury. Therefore, the results obtained from chemokine and chemokine receptor studies may clarify the future development of target drugs for intractable neuropathic pain.