Experimental autoimmune encephalomyelitis (EAE) has been studied for several decades as a model of autoimmune central nervous system disease, particularly human demyelinating multiple sclerosis (MS) [1-4]. The general features of EAE in animal models do not fulfill all characteristics of human MS although the neuropathologic features of EAE lesions occasionally display patterns similar to those of human MS [3, 4]. Thus, EAE is an alternative model for human MS because some of its features match those of human MS.

The advantages and disadvantages of animal models of EAE have been well reviewed and depend on pathology, T-cell phenotype, and the production of pro- and anti-inflammatory molecules [2, 3, 5, 6]. EAE may be induced in susceptible animals, including mice and rats, by immunization with brain tissue-specific antigens, including proteolipid protein (PLP), myelin basic protein (MBP), and myelin oligodendrocyte glycoprotein (MOG). Even though the autoimmune mechanism of EAE is similar in mice and rats, the pattern of EAE pathogenesis in mice [7] is slightly different from that in Lewis rats. Even among rat models, the pathology of Lewis rat EAE is distinct from that of Dark Agouti rats, which shows a recurrent pattern [8-10]. Thus, this review will limit the discussion to Lewis rat EAE.

After immunization of susceptible Lewis rats with brain tissue antigens plus complete Freund's adjuvant (CFA), the animals develop hind limb paralysis clinically [11]. At that time, neuropathological lesions are found mainly in the spinal cord and brain stem, but only rarely in the cerebrum. EAE lesions are characterized by edema, infiltration of mononuclear cells, and gliosis [12], but rarely demyelination. Animals afflicted with EAE experience spontaneous recovery from paralysis. Thereafter, the recovered animals are resistant to re-induction of EAE, showing signs of immunization. Instead of active induction of EAE upon immunization with myelin antigens, EAE can be passively induced by the transfer of encephalitogenic T cells; this is known as passive EAE [13-15].

Even though rat EAE does not show typical demyelination in target organs, the rat EAE model has generated considerable information regarding T-cell immunology and neurobiology, because in Lewis rats EAE is induced uniformly.

The present review will discuss the general features of active monophasic EAE in Lewis rats and highlights a novel view of the dual phenomenon of neuroinflammation and neuroprotection, with a particular emphasis on the macrophage phenotypes present during the course of EAE.

Acute monophasic EAE in Lewis rats is induced by immunization with whole homogenate of spinal cord tissue from guinea pigs with CFA [16], although rat MBP is also encephalitogenic [13, 17, 18]. Homogenate of guinea pig spinal cord tissue has been known to be more encephalitogenic than that of rats. Spinal cord homogenate contains fewer encephalitogenic antigens than does purified antigen, including MBP, PLP, and MOG. Thus, purified MBP, PLP, and MOG antigens, or synthetic peptides corresponding to encephalitogenic epitopes, have been used in studies of rat EAE. Of these, MBP is most commonly used as the immunogen for rat active EAE, and so will be the main antigen discussed in this review. The encephalitogenicity of myelin proteins and the general features of EAE in Lewis rats are well-reviewed elsewhere [13].

Apoptosis has been suggested to be involved in the recovery from rat EAE. This is because many T cells are apoptotic at the peak stage of EAE, and few T cells are found in the spinal cord at the recovery stage of EAE [14, 46, 47]. Furthermore, it has been shown that microglia induce T cell apoptosis in rat EAE [48, 49]. This is a plausible scenario, as inflammation in the CNS will be ameliorated by the loss of causative cells that secrete pro-inflammatory cytokines.

Regulatory T cells (T-reg) have been implicated in EAE remission [44, 50-52]. Recently, it was reported that increased numbers of Th17 and T-reg lymphocytes were associated with EAE remission in the rat [50, 51]. There is agreement that regulatory T cells secrete anti-inflammatory cytokines, which counteract pro-inflammatory cytokines.

Morphologically, what are presumed to be regulatory T cells have been found in the subarachnoid space (SAS) of rat acute EAE [12, 53, 54]. Two different phenotypes of CD4⁺ T cells, either CD45RC⁺ or CD45RC^-, have been detected in the SAS at the early stage of EAE [12]. In the SAS, although both cell types infiltrated at the same time in the early stage of EAE, encephalitogenic T cells, which were CD45RC^- (OX22-negative), preferentially infiltrated the parenchyma, followed by CD45RC⁺CD4⁺ T cells at the peak stage. These findings suggest that regulatory T cells infiltrate the EAE lesion, with encephalitogenic T cells in the SAS at the induction stage of rat EAE, and compete with the encephalitogenic T cells, consequently contributing to recovery from EAE via the secretion of anti-inflammatory mediators.

In rat tissues, various types of macrophages have been studied using a series of ED-specific monoclonal antibodies [55, 56]. Two macrophage phenotypes have been identified: pro-inflammatory, classically activated macrophages (M1 phenotype); and immunomodulatory, alternatively activated macrophages (M2 phenotype) [57].

Macrophages are regarded as an important cell type at the induction stage of rat EAE [32, 33], because blocking of macrophages ameliorated the condition. In addition, increased numbers of macrophages have been associated with the severity of rat acute EAE [22], and macrophages are associated with increased expression of pro-inflammatory mediators, including TNF-alpha and inducible nitric oxide synthase (iNOS) [22]. Macrophages that express TNF-alpha and/or iNOS are classified as classically activated macrophages, or M1 phenotype cells (Fig. 1).

In previous rat EAE studies, macrophages/activated microglia, but not unstimulated microglia in normal rat CNS tissue, were shown to be positive for ED1, while ED2⁺ macrophages were localized mainly in perivascular lesions [55], suggesting that phenotypic differences exist in rat EAE lesions.

Neuropathologically, it is evident that hematogenous macrophages (presumably of the classically activated M1 phenotype) preferentially infiltrate with autoimmune T cells in rat acute EAE [12]. At the EAE induction stage, microglia proliferate in response to T cell infiltration [12, 58]. Even though the number of microglia was increased in rat EAE lesions, the pathological lesions were diminished. This finding suggested that some cellular factors, in addition to regulatory T cells, contribute to EAE modulation. In support of this, activated microglia and/or macrophages [49, 59] were suggested as candidate cell types involved in EAE remission because microglia are activated by expression of CD4-modulated rat EAE [60], and amelioration of clinical EAE was achieved by administration of M2 activated monocytes [61]. Furthermore, even though microglia are known to release potentially cytotoxic molecules, such as pro-inflammatory cytokines, reactive oxygen intermediates, and proteinases [62], their role may be either beneficial or detrimental, depending on the neuropathological conditions [62, 63]. Thus, it is suggested that either activated microglial cells or M2 macrophages, or both, are involved in EAE remission in acute monophasic rat EAE. In a recent study, the presence of M2 macrophages was further evaluated in rat active EAE [11]. In brief, M1 macrophages were predominant at the early stage of EAE, while the proportion of M2 macrophages overwhelmed that of M1 cells at the peak stage and remained high during the recovery stage. Phenotypic differentiation from M1 to M2 macrophages, or activation of microglial cells to the M2 phenotype at the peak stage of EAE remains an area of speculation. The functional role of M2 macrophages was confirmed because type II monocytes have been shown to modulate T cell-mediated mouse EAE through the differentiation of naïve T (Th0) cells into Th2 and CD4⁺CD25⁺FoxP3⁺ regulatory T cells [64]. In either case, the increased numbers of M2 macrophages, through the secretion of anti-inflammatory mediators such as activin A [11], may play a role in recovery from rat EAE. To summarize, the data have shown that this bias toward M2 macrophage activation is at least associated with the amelioration of rat active EAE.

Many molecules have been found in ED1⁺ macrophages in rat active EAE, including osteopontin [65], erythropoietin [66], heat-shock protein 27 [67], nitric oxide synthase [46], and arginase-1 [11]. Some, including erythropoietin [68] and osteopontin [69], have been shown to protect neurons in neurodegenerative disease models, even though the role of osteopontin remains unclear in autoimmune disease models [70]. Thus, a particular macrophage population in rat EAE seems to be a "supplier" of neuroprotective molecules.

The relationship between iNOS and arginase-1, a competitive enzyme for iNOS, will be further discussed (Fig. 2). Upon activation in EAE, macrophages in the lesions are immunopositive for iNOS, which generates nitric oxide through L-arginine. Physiological levels of nitric oxide are beneficial, but can be transiently harmful in rat EAE, when autoimmune T cells infiltrate at the time of EAE induction [71]. However, nitric oxide is also regarded as an EAE-resistant molecule [6, 72], suggesting that the timing of inhibition of iNOS is associated with disease suppression. It is believed that iNOS activity is quickly exhausted in macrophages once activated. Arginase, a competitive enzyme of iNOS, then substitutes for the catalysis of L-arginine, resulting in less production of nitric oxide in the CNS, during recovery-stage EAE [11]. Since some macrophages express both iNOS and arginase, and the expression of iNOS in macrophages is inversely related to that of arginase, it is possible that the phenotypic changes from hematogenous M1 macrophages to M2 macrophages may occur within the same cell in rat EAE. However, it remains possible that phenotypically differentiated M2 macrophages infiltrated alongside hematogenous M1 macrophages. Another possibility is that M2 macrophages originated from proliferation of microglial cells, because microglia were shown to proliferate in rat EAE [14, 58], and alternative activation of microglial cells occurred via IL-4 in a mouse EAE model [73]. A similar result was found in a rat spinal cord injury model in which the majority of macrophages were found to be proliferating [74]. An inverse relationship between iNOS and arginase [75] was identified in a model of spinal cord injury in rats. Taken together, data from the two models of spinal cord inflammation suggest that M2 macrophages may originate from either hematogenous monocytes (via phenotypic differentiation) or activated microglial cells. Both sources of M2 macrophages have been associated with modulation of EAE through the secretion of immunomodulatory molecules, including activin A, in the rat active model. The neuroprotective capacity of macrophages in rat EAE lesions represents a distinct story, but is also beneficial for remission of rat EAE.

Active monophasic EAE in Lewis rats has been extensively studied in the past few decades as a model of human autoimmune disease. Many studies have demonstrated the immunological nature of autoimmune T cells and bystander cells during the initial stage of autoimmune inflammation in the target organ, the spinal cord. Since this model lacks demyelination of CNS tissues, the research has focused mainly on T cell immunology. In the field of T cell immunology, rat EAE has served as a good model for the study of pro- and anti-inflammatory cytokines, depending on the disease status, and will in future be utilized for pilot studies of anti-inflammatory drugs. Furthermore, this review has emphasized the involvement of alternatively activated M2 macrophages in remission from EAE and pointed out their importance in recovery from rat EAE, which suggests a promising strategy for the treatment of autoimmune diseases. Since EAE lesions were protected from further autoimmune attack, and numerous neuroprotective mediators were upregulated at both the peak and recovery stages of EAE, rat EAE is a useful model for the study of both neuroprotection and neuroinflammation.