Multiple sclerosis-induced neuropathic pain: pharmacological management and pathophysiological insights from rodent EAE models

Dysaesthetic extremity pain is characterised by burning, tingling or aching dysaesthesia predominantly in the legs that may be unilateral or bilateral and is often worse at night (Truini et al. 2011; O’Connor et al. 2008). In patients with MS, dysaesthetic extremity pain is the most commonly reported type of neuropathic pain, having a life-time prevalence of 12–28 % and it is very challenging to treat (Truini et al. 2011; Nurmikko et al. 2010). Clinically, patients with primary progressive or progressive-relapsing MS are more likely to suffer from dysaesthetic CNP, whereas patients with relapsing-remitting disease are less affected (Boneschi et al. 2008; Nurmikko et al. 2010).

Trigeminal neuralgia (TN) is defined by the International Association for the Study of Pain (IASP) as “sudden, usually unilateral, severe, brief, stabbing recurrent episodes of pain in the distribution of one or more branches of the trigeminal nerve” (Merskey and Bogduk 1994). TN is further divided into two types, classical and symptomatic, as termed by The International Headache Society (Headache Classification Subcommittee of the International Headache 2004). Classic TN has no cause other than neurovascular contact between the trigeminal nerve root and a blood vessel (Zakrzewska and McMillan 2011). However, neuralgia is called symptomatic if it occurs as a consequence of MS or a tumour (benign or malignant) or if it is due to structural abnormalities such as arteriovenous malformations (Zakrzewska and McMillan 2011). The prevalence of trigeminal neuralgia in patients with MS ranges from 2 to 6.3 % (Putzki et al. 2009; Solaro et al. 2004; Osterberg et al. 2005). Although classic TN and MS-associated TN are both characterised by pain, more frequent sensory deficits and occasional bilateral pain that occur in 20 and 7 % of patients, respectively, are useful criteria for differentiating MS-associated TN from classical neuralgia (Truini et al. 2011; Cruccu et al. 2008).

L’hermitte’s phenomenon is described as a transient, paroxysmal electric shock-like discharge, usually evoked by neck flexion which is felt in the back of the neck and spreads to the lower limbs. Although this phenomenon is not an exclusive symptom of MS, it is frequently reported by patients with MS (Nurmikko et al. 2010) with a prevalence ranging from 9 to 41 % (Solaro et al. 2004; Al-Araji and Oger 2005). Although this symptom is typically self-limiting with spontaneous resolution after some weeks, its frequency and intensity may be troublesome in some patients (Nurmikko et al. 2010).

Although TCAs are recommended as the drugs of choice for management of CNP associated with MS (Truini et al. 2011; Solaro and Uccelli 2011), these agents have many adverse effects such as drowsiness, dry mouth, constipation, hypotension and urinary retention (Mir and Taylor 1997). In a randomised, single-blind clinical trial, the TCA, nortriptyline, was equi-effective with self-applied transcutaneous electrical nerve stimulation for reducing MS-related pain (Table 1). However, nortriptyline produced an array of typical adverse effects associated with TCA administration (Chitsaz et al. 2009). Overall efficacy and optimal dosing regimens of TCAs for the relief of MS-neuropathic pain are lacking as large placebo-controlled RCTs have not been undertaken (Solaro and Uccelli 2011). The mode of action of TCAs and SNRIs for relief of neuropathic pain is attributed to their inhibition of presynaptic reuptake of the biogenic amines, serotonin and noradrenaline, that augment descending inhibitory signalling from the brain to the spinal cord (Coluzzi and Mattia 2005). As selective serotonin reuptake inhibitors (SSRIs) are ineffective for the relief of neuropathic pain (Coluzzi and Mattia 2005), noradrenaline reuptake inhibition is assumed to be the mechanism underpinning the analgesic efficacy of TCAs and SNRIs for relief of peripheral neuropathic pain (Coluzzi and Mattia 2005). However, a recent randomised double-blind, placebo-controlled clinical trial of the SNRI, duloxetine, in patients with spinal cord injury-induced CNP, showed that although duloxetine significantly improved dynamic and cold allodynia relative to placebo, pain intensity was not significantly reduced above placebo (Vranken et al. 2011). Hence, the extent to which SNRIs have efficacy for the relief of MS-induced CNP is unclear.

The efficacy of anticonvulsants including lamotrigine (Cianchetti et al. 1999; Breuer et al. 2007; Silver et al. 2007), levetiracetam (Rossi et al. 2009; Falah et al. 2012) and topiramate (D’Aleo et al. 2001) for the relief of MS-associated persistent or paroxysmal (non-trigeminal) neuropathic pain has been assessed in several small clinical studies (Table 1). In each of these clinical trials, patients reported either incomplete pain relief or limited tolerability (Table 1). Open-label investigation of the efficacy of gabapentin in 25 patients with MS-associated neuropathic pain showed that although patients reported moderate to excellent pain relief (Houtchens et al. 1997), five patients discontinued treatment due to adverse effects including somnolence, insomnia and dyspepsia.

Since discovery of the endogenous cannabinoid system and identification of CB1 and CB2 receptors that are expressed predominantly in the CNS and on peripheral immune cells, respectively, cannabinoids have been investigated for efficacy in neuropathic pain (Rice 2008). Although cannabinoid efficacy for relief of MS-related CNP has been demonstrated (Rog et al. 2005; Rog et al. 2007; Svendsen et al. 2004; Svendsen et al. 2005a), there were many treatment-related adverse effects including dizziness, dry mouth, headache, tiredness or muscle weakness (Rog et al. 2007; Svendsen et al. 2004). Additionally, concern on the risks of acute psychosis and cannabis misuse associated with its therapeutic use (Rice 2008) has restricted these agents to second- or third-line medications for the treatment of MS-associated neuropathic pain (Dworkin et al. 2007; Dworkin et al. 2010; Attal et al. 2010).

In work by others, intravenous morphine was efficacious for the relief of MS-neuropathic pain in a few patients at high doses, consistent with its generally poor efficacy for the relief of other types of neuropathic pain (Kalman et al. 2002). In two small studies, intrathecal baclofen alone (Herman et al. 1992) or in combination with morphine (Sadiq and Poopatana 2007) reportedly alleviated MS-related neuropathic pain. However, baclofen analgesia was short-lived and may have been confounded by concurrent motor impairment (Herman et al. 1992).

Carbamazepine is the most effective first-line treatment for MS-associated TN, in a manner similar to classic TN, despite its poor tolerability profile (Solaro and Uccelli 2011). The frequency and severity of carbamazepine side effects often leads to treatment discontinuation (Hooge and Redekop 1995; Espir and Millac 1970) as these side effects including leg muscle weakness and micturition problems can mimic MS relapse (Solaro et al. 2005). Oxcarbazepine, the keto derivative of carbamazepine effectively alleviated TN in nine of twelve patients, but with better tolerability than carbamazepine (Solaro et al. 2007). By combining lower doses of carbamazepine with gabapentin, complete relief of MS-associated TN was achieved with very good tolerability (Solaro et al. 2000). Although there are no randomised, double-blind, placebo-controlled clinical studies of either gabapentin or lamotrigine for the relief of MS-associated TN, several small open-label studies have been published (Table 1). Their findings suggest that gabapentin alone (Solaro et al. 1998), lamotrigine alone (Lunardi et al. 1997; Leandri et al. 2000) or both in combination (Solaro et al. 2000) may have efficacy. Likewise, most patients reported at least partial relief of MS-associated TN in small open-label clinical studies of misoprostol (Reder and Arnason 1995; DMKG study group 2003) and topiramate (Zvartau-Hind et al. 2000; D’Aleo et al. 2001).

Anticonvulsants such as carbamazepine, oxcarbazepine and gabapentin are recommended as drug treatments for relentless pain due to L’hermitte’s sign (Truini et al. 2011). Although lignocaine infusion or to lesser extent oral mexiletine alleviated the paroxysmal pain of L’hermitte’s sign in small open-label clinical studies (Sakurai and Kanazawa 1999), neither drug was efficacious for relief of persistent painful symptoms associated with MS (Sakurai and Kanazawa 1999).

From the foregoing, it is clear that there is an urgent unmet medical need for new drug treatments that are highly efficacious and have good tolerability profiles for the relief of MS-associated neuropathic pain. This unmet medical need is driving preclinical research on the pathobiology of MS-associated neuropathic pain as a means for identifying novel “druggable” targets for use in pain therapeutics drug discovery programs.

In humans, MS-associated neuropathic pain develops secondary to demyelination and plaque formation in the brain and spinal cord (Nurmikko et al. 2010). In EAE-rats induced with spinal cord homogenate, extensive CNS demyelination induces sensory abnormalities as well as hindlimb weakness and paralysis (Pender 1986b). These clinical signs are underpinned by hyperexcitability of demyelinated axons, ectopic impulses, ephatic activity and conduction block (Pender 1986b). In the CNS of MOG-induced EAE-rodents (Table 3), there is widespread demyelination (Storch et al. 1998) together with anti-MOG antibodies and MOG-reactive T-cells (Iglesias et al. 2001) effectively mimicking neuropathological changes in the CNS of patients with MS (Storch et al. 1998; Iglesias et al. 2001). However, altered nociceptive behaviour in MBP-induced EAE-rats is underpinned, at least in part, by demyelination of primary afferent Aδ-fibres in the sacral and coccygeal dorsal root ganglia (DRGs), dorsal roots and dorsal root entry zones (Pender 1986a). Using electrophysiological methods, changes in the functional properties of peripheral nerve fibres in PLP-induced EAE-mice have also been characterised (Lu et al. 2012). Hence, as MS-associated neuropathic pain in humans is thought to develop secondary to CNS neuropathological changes, MOG-induced EAE-rodent models more closely mimic the human condition.

The high prevalence of axonal damage in the CNS of EAE-rodents appears to involve direct damage induced by cytotoxic T-cells and myelin autoantibodies (Hans 2010). In humans with MS and rodents with MOG-induced EAE, axonal damage is present not only in active lesions in the white matter, but also in the normally appearing white and grey matter (Kornek et al. 2000; Herrero-Herranz et al. 2008). As axonal damage contributes to ectopic firing of primary afferent sensory nerve fibres in rodent models of peripheral neuropathic pain (Costigan et al. 2009; Moalem and Tracey 2006), axonal damage may also underpin ephatic activity of CNS sensory neurons to induce central sensitisation in the spinal cord of EAE-rodent models as well as patients with MS-associated neuropathic pain.

A role for various T-cell subsets including CD4⁺, CD8⁺ and regulatory T-cells (T_reg) in the pathophysiology of MS that is recapitulated in EAE-rodent models is well established (Chitnis 2007; Huseby et al. 2012; Anderton and Liblau 2008). However, the extent to which various T-cell populations contribute to the aetiology of MS-associated neuropathic pain is largely unknown. Prophylactic treatment of EAE-mice with the immunosuppressant, rapamycin, not only attenuated the development of neuropathic pain behaviours, but also reversed clinical signs of EAE disease (Lisi et al. 2012). The mechanism appeared to involve inhibition of effector T-cells, augmentation of T_reg cells (Esposito et al. 2010) and/or inhibition of glial cell activation (Lisi et al. 2012). Hence, future research directed at elucidating the contribution of various T-cell subsets to the pathobiology of MS-induced neuropathic pain is warranted. Other innate immune cells that contribute to CNS neuroinflammatory processes in MS and in EAE-rodents include mast cells, macrophages, neutrophils, dendritic cells and natural killer cells (Gandhi et al. 2010). These may also contribute to the pathogenesis of MS-associated neuropathic pain (Moalem and Tracey 2006).

Pro-inflammatory cytokines, particularly TNF-α, IFN-ϒ, IL-1β, IL-6, IL-12 and IL-18, contribute to disruption of the integrity of the blood–brain barrier (Minagar and Alexander 2003) and to induction of CNS neuroinflammation in MS as well as in EAE-rodent models (Imitola et al. 2005). Their pro-inflammatory effects are offset by the actions of anti-inflammatory cytokines such as IL-10 (Imitola et al. 2005). Pro-inflammatory cytokines have a pathogenic role in the development of neuropathic pain (Zhang and An 2007). Conversely, augmented synthesis of IL-10 in the spinal cord of EAE-mice secondary to intrathecal administration of plasmid DNA encoding IL-10, suppressed glial cell activation and reversed mechanical hypersensitivity in the hindpaws, a hallmark symptom of neuropathic pain (Sloane et al. 2009). In EAE-rats, increased expression levels of TNF-α in the DRGs during peak disease with insignificant TNF-α mRNA expression in the spinal cord (Melanson et al. 2009) implicate anterograde transport of DRG-derived TNF-α to the spinal cord in the development of central sensitisation (Melanson et al. 2009). In support of this notion, thermal hypoalgesia in the tail and limbs of EAE-rats during disease onset was correlated significantly with inflammation induced by elevated expression of TNF-α in the DRGs with anterograde transport to the spinal cord (Begum et al. 2013). Although anti-TNF agents failed to attenuate MS disease progression in clinical studies, this may have been due at least in part to TNF receptor-1 gene variants in patients with MS (Gregory et al. 2012). Potential benefit for relief of MS-associated neuropathic pain by treatment with TNF-α antagonists such as the TNF receptor-Fc fusion protein (etanercept) and anti-TNF monoclonal antibodies (adalimumab and infliximab) remains to be investigated.

In preclinical studies, evaluation of the efficacy of specific monoclonal antibodies for a range of pro-inflammatory cytokines and chemokines implicated in the pathogenesis of MS-associated neuropathic pain is warranted. The chemokine, CCL2 (also called monocyte chemotactic protein-1, MCP-1), (Mahad and Ransohoff 2003) and the pleiotropic cytokine, osteopontin (Imitola et al. 2005) facilitate recruitment of T-cells and macrophages into the CNS of EAE-rodents. Hence, therapeutic strategies aimed at inhibition of CCL2 and/or osteopontin signalling have potential as treatments for MS-associated neuropathic pain.

During active demyelination in MS, there are elevated numbers of CXCR3- or CCR5-positive mononuclear cells in the cerebrospinal fluid (CSF) relative to the peripheral circulation (Sorensen et al. 1999). A role for chemokine signalling via the chemokine receptor, CXCR3, in MS-associated neuropathic pain is supported by observations of reduced neuropathic pain behaviour in EAE-mice null for CXCR3, that was accompanied by attenuated EAE disease progression (Schmitz et al. 2013). However, work by others found no benefit on EAE disease progression (Lalor and Segal 2013).

Considerable research on peripheral nerve injury-induced neuropathic pain implicates glial cell activation in the spinal cord associated with the development of central sensitisation and neuropathic pain (Zhuo et al. 2011; Scholz and Woolf 2007; Hald 2009). Activated glial cells in the spinal cord release pro-inflammatory cytokines, glutamate and nitric oxide that amplify CNS neuronal hyperexcitability in neuropathic pain (Zhuo et al. 2011; Scholz and Woolf 2007; Hald 2009).

In EAE-mice, reactive gliosis and neuroinflammation in the spinal cord developed concurrently with robust neuropathic pain behaviours in a manner similar to that for neuropathic pain of peripheral origin (Olechowski et al. 2009; Lu et al. 2012). The underlying mechanisms include release of pro-inflammatory cytokines (e.g. IL-Iβ, IL-6, TNF-α) and neurotoxic molecules such as nitric oxide and prostaglandins from activated glia in the CNS (Rasmussen et al. 2007; Ponomarev et al. 2005). Apparently, contradictory reports that activated astrocytes enhance (Wang et al. 2005) and suppress (Matsumoto et al. 1992) neuroinflammation in EAE-rodents, require further investigation (Miljkovic et al. 2011).

The neuropeptides, calcitonin gene-related peptide (CGRP) and substance P (Sub-P) are markers of small-diameter sensory neurons (Latremoliere and Woolf 2009). Hyperexcitability of these sensory neurons leads to the development of central sensitisation and neuropathic pain (Latremoliere and Woolf 2009). In EAE-mice, CGRP regulates IL-17 expression to promote Th17 cell-mediated CNS inflammation (Mikami et al. 2012). However, a role for CGRP in the aetiology of MS-neuropathic pain is discounted as spinal cord expression levels did not differ significantly between EAE and control mice (Lu et al. 2012; Olechowski et al. 2009). Although Sub-P and its NK-1 receptor are positively linked to CNS inflammation in EAE-mice (Reinke et al. 2006), a role for Sub-P signalling via the NK-1 receptor in the pathobiology of MS-neuropathic pain remains to be investigated.

In patients with MS, bradykinin signalling via the B₁ receptor has been proposed as an index for disease activity (Prat et al. 2005). In EAE-mice, bradykinin signalling via both B1 and B2 receptors is implicated in the pathogenesis of MS-neuropathic pain. Specifically in EAE-mice, mechanical hypersensitivity in the hindpaws was alleviated by treatment with a B₁ receptor antagonist and abolished by genetic deletion of the B₁ receptor, whereas treatment with a B₂ receptor antagonist alleviated heat hyperalgesia in the hindpaws (Dutra et al. 2013). In EAE-mice, augmented bradykinin signalling was associated with increased expression levels of cyclo-oxygenase-2 (COX-2) and nitric oxide synthase-2 (NOS-2) (Dutra et al. 2013). The net result was upregulated endogenous production of prostaglandins and nitric oxide, respectively, as well as increased expression levels of the pro-inflammatory cytokines, IFN-ϒ and IL-17 (Dutra et al. 2013).

Increased phosphorylation of cyclic AMP (cAMP) response element-binding protein (CREB) by protein kinase-ϒ (PKC-ϒ) (Mao et al. 2007) that is co-localised with macrophages and astrocytes in the spinal cord of EAE-rats (Kim et al. 2007) is implicated in neuroinflammation in MS (Kim et al. 2007). In EAE-mice with severe hindlimb paralysis (chronic disability phase), there was a significant decrease in PKC-ϒ expression in the corticospinal tract (CST), whereas levels were normalised during the remission phase (limp or paralysed tail only) (Lieu et al. 2013). However, as spinal dorsal horn expression levels of PKC-ϒ did not differ significantly in EAE versus control mice, PKC-ϒ is unlikely to contribute significantly to the pathobiology of MS-neuropathic pain (Lieu et al. 2013).

Central sensitisation and long-lasting changes in synaptic plasticity in the dorsal horn of the spinal cord are mediated by phosphorylation of multiple receptors and ion channels by other kinases including p38 mitogen-activated protein kinase (MAPK) and p44/p42 MAPK (also known as ERK1/ERK2), protein kinase A (PKA), protein kinase G (PKG) and calmodulin protein kinase II (CaMK II) (Latremoliere and Woolf 2009). However, pharmacological manipulation of these pathways and their downstream effects on transcription factors such as AP-1 proteins, c-fos, and c-Jun in the aetiology of MS-associated neuropathic pain, remains for future investigation.

Neurotrophins including nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF), signalling via the tropomyosin-related kinase A receptor (TrkA) and tropomyosin-related kinase B receptor (TrkB), respectively, are implicated in the pathogenesis of peripheral neuropathic pain conditions (Siniscalco et al. 2011). Although neurotrophins also regulate neuronal growth and myelination in the CNS (Kalinowska-Lyszczarz and Losy 2012), it is unclear whether they have a protective or detrimental effect in MS-associated neuropathic pain. In a demyelinating EAE-rat model of MS, altered thermal nociception and motor deficits were improved by chronic treatment with the adrenocorticotrophic hormone analogue, [ACTH]_4–9, suggesting that neurotrophic therapy may have benefit for relief of MS-associated neuropathic pain (Duckers et al. 1996).

A role for aberrant Wnt (wingless-type protein-1) signalling in the pathogenesis of MS-associated neuropathic pain is supported by observations of upregulated expression of the Wnt ligands, Wnt3a, Wnt5a and their downstream effector, β-catenin in the spinal dorsal horn in EAE-mice, with pain behaviours attenuated by treatment with the Wnt5a antagonist, Box5, or the β-catenin inhibitor, indomethacin (Yuan et al. 2012).

Apart from development of hindpaw hypersensitivity to applied mechanical and noxious heat stimuli in EAE-mice, most studies also report hyponociception (Table 3) particularly during disease onset (Pender 1986a; Aicher et al. 2004; Sloane et al. 2009) and at peak disease (Olechowski et al. 2009; Thibault et al. 2011). However, the validity of reports of hyponociception in EAE-mice has been questioned due to the potentially confounding effects of concurrent motor deficits (Table 3). Investigation of this issue in EAE-mice after intraplantar injection of formalin into one hindpaw showed that hyponociception appears to be underpinned by dysregulated glutamatergic signalling in the spinal cord (Olechowski et al. 2010). This was characterised by the downregulation of glutamate transporters resulting in increased synaptic levels of glutamate sufficient to activate presynaptic inhibitory group II mGluRs with the net result being attenuation of pain behaviours (Olechowski et al. 2010). In other work, thermal hypoalgesia in EAE-rats in the absence of motor deficits was correlated significantly with the extent of neuroinflammation and axonal swelling (Begum et al. 2013).

Collectively, research to date on MS-associated neuropathic pain using EAE-rodent models implicates complex neuroimmune mechanisms in the pathobiology of this condition. Multiple targets on CNS neurons and glial cells have been identified for potential use in novel analgesics discovery programs.

Optimal pharmacological management of CNP in patients with MS is a significant challenge for clinicians as there are very few randomised, double-blind, placebo-controlled clinical trials of analgesics or adjuvant drug treatments for alleviation of this condition. Most published clinical studies are limited by their open-label design, small patient numbers and relatively short duration (Table 1). Patients with MS suffer from multifaceted pain including neuropathic (e.g. ongoing extremities pain, trigeminal neuralgia or L’hermitte’s sign), nociceptive (e.g. pain arising from musculoskeletal problems) or mixed neuropathic, nociceptive pain (e.g. tonic painful spasms or spasticity) that involve distinct pathophysiological mechanisms (Truini et al. 2013). Although there are a few randomised double-blind placebo-controlled clinical studies of cannabis and low-dose naltrexone for the relief of MS-associated pain, the specific type of MS pain assessed in these patients was not specified (Wade et al. 2004; Sharafaddinzadeh et al. 2010).

Utilisation of standardised criteria that distinguish between the various pain components experienced by patients with MS and their respective responsiveness to analgesic/adjuvant drug treatments is needed to improve the conduct, reporting and value of such studies. In addition, changes to the design of clinical studies aimed at assessing the efficacy of disease-modifying treatments for not only reducing the rate of disease progression (relapses) in patients with MS, but also to include concurrent assessment of the effect of such treatments on the various components of MS-associated pain would be invaluable.

Apart from the paucity of RCTs assessing the efficacy of analgesic and adjuvant drug treatments for alleviation of MS-induced neuropathic pain, it is only relatively recently that research directed at elucidating the neurobiology of MS-induced neuropathic pain using EAE-rodent models has been undertaken.

A significant limitation of most studies using EAE-rodents in MS-associated neuropathic pain research is the use of CFA comprising heat-killed mycobacterium and paraffin oil, as the adjuvant to induce EAE disease (Gold et al. 2006; Stromnes 2006) (Table 3). This is problematic because administration of CFA alone increases expression of the pro-inflammatory mediator, COX-2 in the spinal cord and brainstem of control mice to match the levels in EAE-mice (de Lago et al. 2012). In addition, intraplantar injection of CFA in rats produced pain behaviours that correlated with elevated pro-inflammatory cytokine levels and glial cell activation in the CNS (Raghavendra et al. 2004). Hence, caution is required in the interpretation of pain behaviours in EAE-rodents when CFA is used as the adjuvant in the immunisation protocol. Replacement of CFA by adjuvants that do not themselves induce marked CNS neuroinflammation is a critical step in optimising EAE disease induction protocols in rodents for use in studies on the pathobiology of MS-neuropathic pain.

In EAE-rodents, the clinical disease course in most studies designed to gain insight into the mechanistic basis of MS-associated neuropathic pain includes periods of paralysis and motor impairment which preclude unambiguous assessment of behavioural pain endpoints at those times (Table 3). Thus, reliable assessment of pain behaviours is restricted to time points before motor impairment is evident. Hence, development of an optimised EAE-rodent model that exhibits robust neuropathic pain behaviours with a prolonged relapsing-remitting clinical disease time course, but without confounding motor impairment is needed. Such a model would facilitate a detailed investigation of the mechanisms underpinning the development and maintenance of neuropathic pain in a longitudinal fashion over the relapsing-remitting EAE disease course.

Myelin oligodendrocyte glycoprotein is a glycoprotein important to the structural integrity of the myelin sheath of neurons in the CNS (Baumann and Pham-Dinh 2001) and it is widely used as a powerful encephalitogen in EAE-rodents (Wekerle and Kurschus 2006; Mendel et al. 1995). MOG induces widespread demyelination and other pathological features in the CNS of EAE-rodents in a manner similar to that observed in patients with MS (Storch et al. 1998; Iglesias et al. 2001). Although EAE may be induced with either MBP or PLP, both of these proteins are found in the peripheral nervous system as well as the CNS. Hence, MOG is the preferred antigen as MS-associated neuropathic pain is regarded as being of central origin.