Neuropeptides and Microglial Activation in Inflammation, Pain, and Neurodegenerative Diseases

Microglial cells are responsible for immune surveillance within the CNS. They respond to noxious stimuli by releasing inflammatory mediators and mounting an effective inflammatory response. This is followed by release of anti-inflammatory mediators and resolution of the inflammatory response. Alterations to this delicate process may lead to tissue damage, neuroinflammation, and neurodegeneration. Chronic pain, such as inflammatory or neuropathic pain, is accompanied by neuroimmune activation, and the role of glial cells in the initiation and maintenance of chronic pain has been the subject of increasing research over the last two decades. Neuropeptides are small amino acidic molecules with the ability to regulate neuronal activity and thereby affect various functions such as thermoregulation, reproductive behavior, food and water intake, and circadian rhythms. Neuropeptides can also affect inflammatory responses and pain sensitivity by modulating the activity of glial cells. The last decade has witnessed growing interest in the study of microglial activation and its modulation by neuropeptides in the hope of developing new therapeutics for treating neurodegenerative diseases and chronic pain. This review summarizes the current literature on the way in which several neuropeptides modulate microglial activity and response to tissue damage and how this modulation may affect pain sensitivity.

Within the CNS, the proopiomelanocortin (POMC) gene is mainly expressed in neurons from the arcuate nucleus (ARC) of the hypothalamus which project into the paraventricular nucleus, lateral hypothalamus, and other regions of the brain such as amygdala, cortex, hippocampus, medulla, mesencephalon, and spinal cord [ 19 ]. Through the action of tissue-specific prohormone convertases, it is posttranslationally processed to yield adrenocorticotropic hormone (ACTH), α -melanocyte-stimulating hormone ( α -MSH), β -MSH, and γ -MSH, collectively known as melanocortins, and the endogenous opioid β -endorphin ( β -END). The melanocortin system is involved in the regulation of a great variety of functions such as food intake and energy expenditure, sexual behavior, exocrine gland secretion, fever control, pigmentation, learning, and memory [ 20 , 21 ]. This system is also tightly linked to control of inflammation, both centrally and peripherally, by exerting anti-inflammatory actions on cells of the immune system including lymphocytes, monocytes, and macrophages, as well as nonimmune cells such as melanocytes, keratinocytes, and adipocytes, among others [ 22 ]. Melanocortins may also regulate macrophage differentiation into tissue-specific cells such as osteoclasts, Kupffer cells, and microglia [ 23 ], further highlighting the role of these neuropeptides as modulators of immunity. Despite the long recognized anti-inflammatory properties of melanocortins, little is known about their direct influence on microglial cells, which are the main component of innate immunity in the CNS. In addition, most of the studies on the influence of POMC products on microglia used cell lines instead of primary microglia and focused on the effect of ACTH and α -MSH, along with several synthetic analogs of these peptides, whereas the effects of γ -MSH, β -MSH, and β -END have been studied less. Almost two decades ago, α -MSH was demonstrated to inhibit brain production of TNF- α in a mouse model of brain inflammation caused by bacterial lipopolysaccharide (LPS) injection [ 24 ]. The authors suggested that α -MSH was acting through the activation of MC1R present on astrocytes and microglia, the main producers of TNF- α within the brain. A year later, ACTH, α -MSH, and α -MSH COOH-terminal tripeptide ( α -MSH 11–13 ) were shown to inhibit LPS + Interferon- (IFN-) γ -induced TNF- α , IL-6, and NO release in N9 microglial cells, most likely by a mechanism involving cAMP production [ 25 ]. The authors also demonstrated that microglial cells release α -MSH upon stimulation with LPS + IFN- γ , thereby creating an autocrine anti-inflammatory loop. This finding placed microglial cells in the group of nonneuronal immune cells expressing POMC mRNA and producing α -MSH as an autoregulatory factor, similar to what had been previously demonstrated in RAW 264.7 murine macrophages [ 26 ] and in human THP-1 myelomonocytic cells [ 27 ]. Subsequently, α -MSH and α -MSH 11–13 were also demonstrated to reduce the release of TNF- α and NO and the expression of TNF- α and inducible nitric oxide synthase (iNOS) induced by costimulation of N9 cells with Amyloid- β (A β ) and IFN- γ [ 28 ], further supporting its role as an anti-inflammatory agent. Since then, melanocortins have proven to be anti-inflammatory and neuroprotective in vivo in several experimental models of nervous tissue damage, including brain ischemia [ 29 – 31 ] and traumatic brain injury, where neuroprotection exerted by α -MSH (11–13) was associated with decreased microglial activation [ 32 ], suggesting a contribution of local immunomodulation to the overall neuroprotective effect. Melanocortins exert their actions through MCRs of which five subtypes have been described up to now (MCR1 to MCR5). These are G protein-coupled receptors (GPCRs) with seven transmembrane domains, all positively coupled with adenylate cyclase (AC) [ 33 ]. Nonetheless, cAMP-independent pathways have also been shown to mediate MCR signaling (reviewed in [ 34 ]). In the brain, the most abundant MCR subtypes are MC3R and MC4R [ 35 , 36 ]. Expression of MC1R has been detected in RAW 264.7 cells [ 26 ], human monocytes [ 37 ], and murine macrophages [ 38 ]. MC3R expression was found in murine and rat macrophages [ 39 , 40 ] and in RAW 264.7 cells [ 41 ] and expression of MC1R, MC3R, and MC5R was detected in THP-1 cells [ 27 , 42 ]. Human monocytes express mRNA for MC1R, MC2R, MC3R, and MC5R [ 43 ]. MC4R, however, does not appear to be expressed in mouse macrophages [ 39 ] and RAW 264.7 cells [ 44 ]. On the other hand, MCR expression in microglial cells has not been thoroughly examined. Lindberg et al. showed that the human microglial cell line CHME-3 expresses MC1, MC3, MC4, and MC5 receptors and that α -MSH increases basal IL-6 secretion in these cells [ 45 ]. The authors suggested that this effect is mediated by either MC3R or MC5R, since agouti, an endogenous antagonist of the melanocortin system which has greater affinity

The Vasoactive Intestinal Peptide (VIP) is a 28-amino acid polypeptide discovered in 1970 in intestinal extracts, able to induce vasodilation [ 112 ]. VIP is an immunoregulatory neuropeptide produced by neurons, endocrine, and immune cells, widely distributed in the CNS and PNS, and also found in heart, lung, thyroid, kidney, urinary and gastrointestinal tracts, genital organs, and the immune system [ 113 ]. VIP is structurally related to the secretin/glucagon family of peptides and hormones, sharing 70% sequence identity with the neuropeptide pituitary adenylate cyclase-activating polypeptide (PACAP) [ 114 ]. The precursor molecule pre-pro-VIP is processed into mature VIP and the related peptide C-terminal methionine amide (PHM) in humans and peptide with N-terminal histidine and C-terminal isoleucine amide (PHI) in other mammalian species [ 115 ]. PACAP was discovered nearly two decades later as a 38-amino acid hypothalamic neuropeptide that stimulated AC activity and increased cAMP levels in pituitary cells [ 116 ]. The sequence of PACAP-38 contains an internal cleavage site giving rise to a 27-amino acid fragment corresponding to the N-terminal portion of PACAP-38, termed PACAP-27, which is also capable of inducing AC activity in pituitary cells [ 117 ]. In the CNS, the predominant form is PACAP-38, expressed mainly in the hypothalamus but also widely distributed in several extrahypothalamic regions. In the periphery, PACAP-38 is also the predominant form and has been found in numerous organs and tissues including the immune system [ 117 ]. Following the discovery of PACAP, it was found that VIP and PACAP act as endogenous ligands for the same receptors, which are three heterotrimeric GPCRs: PAC1, VPAC1 (or VIPR1), and VPAC2 (or VIPR2), each with a distinct expression pattern. PAC1 has much greater affinity for PACAP than for VIP, whereas VPAC1 and VPAC2 can bind PACAP and VIP with comparable affinity [ 118 ]. Expression of PAC1 is highest in the CNS, particularly in the thalamus, hypothalamus, hippocampus, olfactory bulb, and cerebellum. VPAC1 is also widely distributed in the CNS, particularly in the cerebral cortex and hippocampus, and in several peripheral organs [ 118 ]. VPAC2 expression has been detected in the thalamus, suprachiasmatic nucleus, hippocampus, brainstem, and dorsal root ganglia, and in some peripheral tissues such as smooth muscle of the cardiovascular, gastrointestinal, and reproductive systems [ 118 ]. These receptors are coupled primarily with Gs and activate AC and PKA, although VPAC/PAC1 activation has also been reported to activate or inhibit several other signaling pathways [ 119 , 120 ]. VIP and PACAP proved to be neuroprotective in various models of nervous tissue damage [ 121 ]. Since microglial cells are the main source of inflammatory mediators in the CNS, it has been suggested that VIP and PACAP might be exerting their protective effects in part by acting directly on microglial cells as microglia-deactivating factors. Expression of VPAC1 and PAC1 (but not VPAC2) was detected in rat microglia, where these receptors mediate VIP and PACAP inhibition of LPS-induced TNF- α production in a cAMP-dependent manner [ 60 ]. Similarly, Delgado and colleagues found expression of VPAC1 and PAC1, but not VPAC2, in murine microglia, where VIP and PACAP inhibit LPS-stimulated production of the CXC chemokines macrophage-inflammatory protein- (MIP-) 2 and KC, and of the CC chemokines MIP-1 α , MIP-1 β , macrophage chemoattractant peptide (MCP)-1, and RANTES, and inhibit chemotactic activity of peripheral leukocytes [ 61 ]. VIP and PACAP also prevent the production of LPS-induced TNF- α , IL-1 β , IL-6, and NO in murine microglia [ 122 ]. These effects are attained through activation of VPAC1, induction of cAMP production, and inhibition of NF- κ B [ 61 , 122 ]. The mechanism for VIP and PACAP inhibition of NF- κ B in activated microglial cells involves a reduction in p65 binding to the coactivator CREB-binding protein (CBP) and parallel stimulation of CREB phosphorylation, thereby promoting CREB/CBP complex formation instead of p65/CBP interaction [ 123 ] and ultimately leading to reduced transcription of classic proinflammatory NF- κ B-regulated genes. VIP inhibits cyclooxygenase- (COX-) 2 expression and PGE 2 production in LPS- and LPS + IFN- γ -treated murine microglial cells [ 124 ]. Activation of VPAC1-cAMP signaling by VIP and PACAP was also found to inhibit the MEKK1/MEK4/JNK pathway and subsequent binding of the transcription factor AP-1 in LPS-stimulated microglial cells [ 125 ]. PACAP was also shown to prevent LPS-induced BV-2 microglial activation by inducing cAMP production and inhibiting p38 MAPK signaling pathway [ 126 ]. This neuropeptide also attenuated microglial hypoxia-induced TNF- α , iNOS, and NO production and p38 MAPK activation and prevented neuronal death caused by microglial neurotoxicity [ 127 ]. In IFN- γ -treated microglial cells, activation of VPAC1 by VIP an

Cortistatin (CST) is a cyclic peptide initially isolated from rat brain and named after its predominantly cortical expression and its ability to depress neuronal electrical activity [ 156 ]. Its precursor preprocortistatin (preproCST) is a 112-amino acid protein that may suffer proteolytic cleavage at multiple sites, yielding products of different lengths. CST belongs to the same family as SST; the 14-amino acid rat CST isoform and the 17-amino acid human isoform both share 11 amino acids with SST-14, including the motif responsible for SST receptor interaction, and are thereby able to bind all five known SST receptors in vitro with similar affinities to that of SST, acting as receptor agonists and inhibiting cAMP accumulation [ 157 ]. Although some of the biological actions of CST seem to be mediated by SST receptors, several effects of CST in the CNS are distinct from those of SST, such as induction of slow-wave sleep and reduction of locomotor activity, suggesting the existence of alternative signaling pathways for CST [ 157 ]. In support of this notion, other GPCRs have been shown to bind CST with selective affinity over SST, such as the ghrelin/growth hormone-secretagogue receptor (GHSR) which binds CST with similar affinity compared to ghrelin [ 158 ], and the human orphan receptor Mas-related G-protein-coupled receptor member X2 (MrgX2) [ 159 ] which also binds proadrenomedullin and related peptides [ 160 ]. More recently, truncated functional forms of sst5 have been described which show different signaling profiles in response to CST or SST [ 161 , 162 ]. Expression of preproCST mRNA in the CNS appears to be restricted to the cortex and hippocampus [ 156 ]. In addition to its role as regulator of sleep rhythm and locomotor activity, evidence strongly suggests a role for CST in inflammation and immunomodulation. Expression of preproCST mRNA (but not of preproSST) was detected in human monocyte-derived macrophages and dendritic cells. In addition, CST was demonstrated to bind sst2, and both preproCST and sst2 are upregulated during macrophage differentiation and after LPS stimulation, suggesting CST might be an endogenous ligand for sst2 in the human immune system [ 163 , 164 ]. In peritoneal macrophages, CST prevented the LPS-induced production of TNF- α , IL-12, IL-1 β , IL-6, NO, MIP-2, and RANTES. Since these effects were only partially prevented by the sst antagonist cyclosomatostatin and were also partially blocked by a GHSR antagonist, evidence suggests CST is acting through sst-dependent and sst-independent pathways in these cells [ 165 , 166 ]. CST was also shown to dose-dependently inhibit basal and IL-1 β -induced PGE 2 release and to reduce IL-1 β -stimulated COX-2 mRNA expression in primary rat microglial cells [ 151 ]. Protective effects of CST have been demonstrated in vivo in experimental models of ulcerative colitis, arthritis, and endotoxemia, where the anti-inflammatory effects of CST were attributed mainly to its ability to deactivate resident and infiltrating macrophages, and other immune cells [ 165 – 167 ]. CST also showed beneficial effects in a model of meningoencephalitis caused by Klebsiella pneumoniae infection; it reduced white blood cell infiltration into the cerebrospinal fluid (CSF), attenuated clinical symptoms of illness, inhibited TNF- α , IL-1 β , and IL-6 brain expression and release into the CSF, and decreased neuronal cell death in the cortex and hippocampus [ 168 ]. Interestingly, the decrease in K. pneumoniae -induced proinflammatory cytokine production after CST treatment was also observed in vitro in neuron-glia cocultures, suggesting direct downregulation of glial activity may partially account for the anti-inflammatory effects of CST [ 168 ]. CST also exerted beneficial long lasting effects in models of chronic and relapsing-remitting EAE, where systemic treatment with the neuropeptide reduced incidence and severity of the disease [ 169 ]. CST treatment reduced inflammatory spinal cord infiltrates, decreased activation of autoreactive Th1/Th17 cells, and inhibited expression of proinflammatory mediators, while promoting regulatory T cell differentiation. The protective effects of CST were associated with its ability to promote the development of a glial neuroprotective phenotype, by inducing BDNF and activity-dependent neuroprotector protein release from neuron-glial cocultures, reducing astrocyte and microglial IL-6, TNF- α , and NO release, and preventing oxidative stress-induced oligodendrocyte death [ 169 ]. CST is expressed in interneurons from the spinal cord and in nociceptive neurons from the dorsal root ganglia, where it colocalizes mostly with peptidergic calcitonin gene-related peptide (CGRP)/substance P- (SP-) expressing nociceptors. It is considered to be an endogenous analgesic factor, as nociceptive responses to inflammatory pain are exacerbated in CST knockout mice, and administration of CST induces analgesia and decreases nocifensive behavi

CGRP and adrenomedullin (AM) are neuropeptides that belong to the CGRP/calcitonin peptide superfamily, which also includes calcitonin, amylin, and intermedin [ 198 – 202 ]. They are widely distributed in the CNS and PNS and both have potent vasodilator activity [ 203 ]. They exert their actions through specific receptors which are GPCRs formed by three components: the 7-transmembrane calcitonin receptor-like receptor (CRLR), the receptor component protein (RCP), and the receptor activity-modifying protein (RAMP) 1–3 [ 204 ]. Interaction of CRLR with RAMP1 forms the CGRP1 receptor, which has greater affinity for CGRP than for AM, whereas association with RAMP2 or RAMP3 gives rise to the AM preferring receptors 1 and 2, respectively. CGRP/AM receptor activation is known to stimulate AC and cAMP production, although other signaling pathways have been described such as PLC, intracellular calcium increase, and NO production [ 203 ]. Functional evidence indicates microglial cells express CGRP/AM receptors, as treatment with CGRP upregulates immediate-early c-fos gene expression and induces cAMP accumulation [ 67 , 68 ], and both CGRP and AM modulate expression of LPS-induced inflammatory mediators in cultured microglial cells [ 69 ]. Cumulative evidence suggests a strong protective anti-inflammatory role for CGRP and AM in several models of diseases, although some proinflammatory effects have also been described for both neuropeptides [ 205 , 206 ]. Treatment with AM has shown protective effects in a variety of experimental models of disease such as arterial and pulmonary hypertension, heart failure, septic shock, and ischemia-reperfusion injury [ 207 ]. AM expression is induced after hypoxia in various cell types of the brain, including neurons, astrocytes, endothelial cells, and microglia [ 208 – 210 ], suggesting a role for the neuropeptide in modulating hypoxic damage in the CNS. In addition, several studies have demonstrated a potent antioxidant role for AM in different cell types, including macrophages [ 211 – 214 ]. In a transient focal ischemia model using AM knockout heterozygous (AM (+/−) ) mice, the authors showed increased iNOS expression in microglial cells from AM (+/−) mice compared to wild type mice [ 215 ]. Furthermore, supernatants of microglial cells exposed to OGD effectively protected neurons against OGD-induced death, and this effect was prevented by an AM receptor antagonist [ 209 ], suggesting a neuroprotective role for microglia-derived AM under hypoxic/ischemic stress. CGRP and AM also inhibit LPS-induced TNF- α , IL-6, and NO release in rat cultured microglia and prevent LPS-induced expression of MIP-1 α and MCP-1 in microglia/astrocyte cocultures. The anti-inflammatory effect appears to be stimulus-specific since neither CGRP nor AM inhibits IL-6 and NO release induced by a cytokine mix composed of TNF- α , IFN- γ , and IL-1 β [ 69 ]. The ability of neuropeptides to inhibit cytokine and chemokine release in an inflammatory context is of particular relevance in neuroinflammatory diseases such as MS, in which local release of these factors promotes CNS infiltration of leukocytes and autoreactive T cells, which in turn perpetuate and amplify the inflammatory reaction [ 216 ]. Concordantly, in a murine model of chronic-progressive EAE, treatment with AM reduced inflammatory infiltration and demyelination in the CNS, partly by impairing activation of autoreactive Th1/Th17 cells and increasing the number of Th2 cells and regulatory T cells and also by preventing oxidative stress-induced oligodendrocyte death and inhibiting expression of astroglial and microglial-derived proinflammatory mediators such as IL-6, IL-12, TNF- α , and NO [ 217 ]. In another murine model of chronic EAE, infusion of CGRP also decreased clinical signs of disease and prevented microglial activation, evidenced by a reduced proportion of amoeboid Iba1 + cells [ 218 ]. Altogether, evidence suggests that modulation of microglial activation by CGRP/AM signaling may account, at least in part, for the protective effects of these neuropeptides in neuroinflammatory diseases. CGRP has been strongly linked to increased pain sensitivity in the spinal cord. Binding of CGRP to CGRP1 receptors in the rat spinal cord produces hyperalgesia through a mechanism involving activation of PKA and PKC signaling pathways [ 219 ]. Injection of CGRP stimulates activation of spinal cord microglia in a model of temporomandibular joint disorder, evidenced by increased OX-42 immunostaining [ 220 ]. Administration of the CGRP antagonist (CGRP 8–37 ) attenuates mechanical hypersensitivity and reduces microgliosis in a rat model of collagen-induced arthritis-induced hypersensitivity [ 221 ]. Furthermore, CGRP was demonstrated to contribute to the development of tolerance to morphine-induced analgesia partly through induction of p38 MAPK phosphorylation, upregulation of IL-6, and activation of the NF- κ B signaling pathway in spinal cord microglial cel

Ghrelin is a 28-amino acid peptide originally isolated from rat stomach as an endogenous ligand for the GHSR with the ability to stimulate growth hormone (GH) release from the pituitary [ 251 ]. It is expressed at high levels in the stomach and is also produced in the ARC of the hypothalamus [ 251 ]. Ghrelin mRNA expression in the stomach and circulating plasma levels increase after fasting and decrease after refeeding [ 252 ]. It acts as an orexigenic peptide, antagonizing the effects of leptin on food intake through activation of the hypothalamic NPY/Y 1 R pathway [ 253 ]. The ghrelin receptor GHSR has two isoforms: GHSR1 α and GHSR1 β [ 254 ]. Ghrelin acylation in Ser3 is required for hormonal activity and for ghrelin to activate its cognate receptor GHSR1 α [ 255 ], which is expressed in many tissues including pituitary gland and hypothalamus [ 256 ]. Nonetheless, unacylated ghrelin is found in circulation at greater concentrations than ghrelin, suggesting a relevant physiological role [ 255 ]. Ghrelin can act directly on hypothalamic NPY and AgRP neurons, increasing food intake and body weight gain through a GHSR-dependent, GH-independent pathway [ 257 ]. However, certain effects of ghrelin have been observed in tissues where only the GHSR1 β isoform is expressed, leading to the postulation of the existence of novel nonspecific ghrelin receptors called ghrelin receptor-like receptors and of specific unacylated ghrelin receptors. Finally, GHSR has been shown to form heterodimers with other GPCRs such as sst5, MC3R, and the dopamine receptors D1 and D2. Since receptor heterodimerization can alter G protein coupling as well as ligand potency, these findings add complexity to the pathways involved in ghrelin signaling [ 254 ]. Ghrelin has been demonstrated to exert neuroprotective effects in several models of neurodegenerative and inflammatory diseases, in which its protective effects have been associated, at least in part, with its ability to reduce microglial activation. In experimental SCI, microglial p38 MAPK activation followed by pronerve growth factor (proNGF) release is known to mediate oligodendrocyte death [ 258 ]. Ghrelin has been shown to promote functional recovery after SCI, partly by inhibiting microglial p38 MAPK activation and proNGF release in the spinal cord and by preventing apoptotic cell death of neurons and oligodendrocytes through a GHSR1 α -dependent pathway [ 259 ]. The inhibitory effects of ghrelin on microglial activation have also been observed in vitro using BV-2 microglial cells. In this assay system, ghrelin prevented LPS-induced BV-2 p38 MAPK and JNK activation and attenuated proNGF and ROS production [ 260 ]. However, while expression of GHSR1 α has been detected in spinal cord neurons and oligodendrocytes, it has not been detected in microglial cells or astrocytes by immunohistochemistry [ 259 ] or in BV-2 cells by RT-PCR and western blot [ 260 ], suggesting that the mechanism underlying the inhibitory effects of ghrelin on microglial activation may be either indirect or through a GHSR1 α -independent pathway. In an in vitro assay system of OGD followed by reoxygenation (OGD/RO), modeling SCI, endothelial cell-derived MMP-3 was shown to mediate BV-2 microglial p38 MAPK activation and proNGF release. In this study, supernatant of OGD/RO endothelial cells transfected with MMP-3 siRNA failed to induce BV-2 microglial activation compared to supernatant from cells transfected with control siRNA, and microglial activation was also attenuated in MMP-3 knockout mice. Moreover, addition of ghrelin to OGD/RO endothelial cells inhibited MMP-3 production in a GHSR1 α -dependent fashion [ 261 ], providing a possible mechanism for ghrelin-induced inhibition of microglial activation. In an in vivo model of kainic acid- (KA-) induced hippocampal neurodegeneration, systemic administration of ghrelin prevented KA-induced neuronal cell death, attenuated microglial and astroglial activation, and reduced TNF- α , IL-1 β , and COX-2 expression in the hippocampus. Furthermore, ghrelin inhibited KA-induced MMP-3 expression in hippocampal neurons, and the protective effects of ghrelin were exerted through a GHSR1 α -dependent pathway [ 262 ]. Once again, evidence supports a link between ghrelin-induced modulation of microglial activation and inhibition of MMP-3 expression. Ghrelin also attenuated motoneuron loss in organotypic rat spinal cord cultures exposed to threohydroxyaspartate (THA), a model of excitotoxic motoneuron degeneration, and prevented spinal cord microglia activation and expression of IL-1 β and TNF- α , suggesting a possible therapeutic role for ghrelin in amyotrophic lateral sclerosis [ 263 ]. Ghrelin has also shown neuroprotective effects in animal models of PD. In a murine model of MPTP-induced dopaminergic neuron loss, systemic ghrelin administration improved dopaminergic neuron survival, prevented the loss of striatal dopaminergic fibers, and reduced MPTP-induced microglial

The authors declare that there is no conflict of interests regarding the publication of this paper.