Several histopathological studies have examined the frequency and distribution of MCs in both CNS and peripheral lymphoid organs during the course of EAE in different animal species. Some work reported a decrease in the number of MCs in dura mater [72], velum interposition [72] and thalamus [34,73] during the acute phase of EAE in Lewis rats, while other described no change [74] or a three-fold rise in thalamic MCs [75]. Conversely, most of the studies were concordant in describing an increase of the percentage of degranulated MCs in the brain of EAE rats [34,74,75], thus proposing that MCs might be involved in the effector phase of the disease. In the CNS of naïve mice, MCs have been identified in perivascular areas of leptomeninges, hippocampus, habenula and thalamus [76,77]. During the course of mouse EAE, no degranulated MCs or MCs infiltrating acute lesions were detected in either WBB6F₁ or C57BL/6 strains [77,78]. In marmoset EAE, MC activation was increased in areas of demyelination in the diencephalon [79]. Indeed, in this model MCs were located in perivascular areas and displayed ultrastructural evidence of intragranular activation (indicating the release of selective mediators), but not degranulation [79].

The histological evaluation of peripheral lymphoid organs during the induction phase of EAE in C57BL/6 mice, revealed a greater number of MCs within the T-cell-rich perifollicular areas of DLN, with a certain degree of MC-clustering [78], and the presence of activated MCs establishing tight spatial interactions with Th17 cells and regulatory T cells [47]. These findings again evoked the occurrence of a potential MC-mediated modulation of Treg and Th17 cells immune functions [47].

The first studies attempting to clarify the role of MCs in EAE sought to modulate disease development by treatment with pharmacological agents able to block or induce MC degranulation. In 1989, Dietsch et al., showed that the incidence of passive EAE in Lewis rats was drastically abated if animals were treated intraperitoneally just before disease onset with proxicromil, a MC-stabilizer derivative of cromolyn [80]. They also reported that reserpine, a pharmacological agent inhibiting MC release of vasoactive amines, was effective in reducing the incidence of both active and passive EAE in Lewis rats [80]. However, a few years later, Levi-Schaffer and co-workers demonstrated that another derivative of cromolyn, nedocromil, was just efficacious in slightly delaying EAE onset in rats, and if administered at the time of disease induction, thus suggesting that MCs were only partially involved in the priming phase of disease, and perhaps dispensable for the effector phase [73]. Nevertheless, another group showed that intracisternal but not intraperitoneal administration of compound 48/80, which triggers MC degranulation, also reduced EAE severity in rats, underscoring the possibility of MC contribution to disease development in the CNS [72].

Since these first initial works, the involvement of MCs in EAE appeared somehow ambiguous. Considering that all these pharmacological agents are not MC specific being active also on other cell types [81,82], no direct conclusions could be drawn on the role of MCs in EAE by these pharmacological studies.

In recent years, a significant amount of work has attempted to assess MC involvement in EAE by using mouse strains harbouring spontaneous inactivating mutations of c-kit gene (or, in C57BL/6-Kit^W-sh/W-sh mice, a mutation that reduces c-kit expression [see below]) and consequently displaying severe MC deficiency [83,84]. The availability of c-kit mutant MC-deficient mouse models has provided the opportunity of applying an apparently more specific experimental approach to study MCs in EAE. However, data obtained with these mice appear often discordant and/or contradictory, and have depicted an equivocal and conflicting scenario about the exact impact of MCs in CNS autoimmunity.

For several years the most commonly used model for studying MCs has been the Kit^W/W-v strain on WBB6F₁ background [25,26]. WBB6F₁-Kit^W/W-v mice bear two mutated alleles at the White spotting (W) locus on chromosome 5, which corresponds to c-kit gene. The W mutation is a G to A point mutation at a splice donor site leading to exon skipping and production of a truncated c-kit, which lacks the transmembrane domain and is not expressed on the cell membrane [85]. The W-v mutation is a C to T point mutation (resulting in the change Thr660Met) in the c-kit tyrosine kinase domain that considerably reduces receptor signalling [86,87]. Kit^W/W-v mice display profound MC-deficiency, but also some other c-kit-dependent abnormalities, such as defective melanogenesis, sterility, anemia, deficiency of interstitial cells of Cajal (ICCs) and neutropenia [88].

The group of M. Brown, first in 2000, reported that MC-deficient WBB6F₁-Kit^W/W-v mice developed MOG_35–55-induced chronic EAE with a significantly lower incidence and milder severity than controls. Engraftment of Kit^W/W-v mice with bone marrow-derived, in vitro cultured MCs (BMMCs) before EAE induction, restored disease susceptibility to levels of wild-type mice, thus clearly indicating a detrimental role of MCs in this model [77]. Activation of BMMCs through the Fc receptor common γ-chain (shared by FcγRI, FcγRIII and FcγRI) or through FcɛRIII was essential to promote EAE in this model [89]. Further studies by the same group have outlined that MCs influenced disease development by acting both in peripheral lymphoid organs [16] and the CNS [90]. Indeed, MCs were proposed to be necessary for the establishment of an optimal encephalitogenic Th1 cell response in both lymph nodes and spleen [17]. Adoptive transfer of myelin-activated T cells also resulted in less severe EAE in Kit^W/W-v mice compared to controls [17]. In the CNS, meningeal MCs were suggested to contribute to the breach of the BBB occurring in EAE, by favouring the recruitment of neutrophils into the CNS parenchyma through secretion of TNF [90]. Recently, Brown and co-workers have shown that SJL/J-Kit^W/W-v mice displayed attenuated PLP_139–151-induced EAE, thus indicating that MCs may promote also the relapsing-remitting model of MS [91].

Although these results obtained in mouse models of MC deficiency have highlighted an important contribution of MCs to the development of EAE, other studies have recently questioned these data, showing that EAE in WBB6F₁-Kit^W/W-v developed with no significant difference or even with higher severity compared with wild-type littermates [57,78,92]. The reasons for these discrepant results are still to be understood. However, in the original work by Secor et al., showing that Kit^W/W-v mice were protected from EAE [77], the protocol of disease induction was much stronger than those generally used to induce EAE, and consisted of 300 μg of MOG_35–55 emulsified in 500 μg of M. tuberculosis in CFA (injected on days 0 and day 7 post-immunization) and 500 ng of PTX (on days 0 and 2 p.i.). In the two papers reporting full susceptibility of Kit^W/W-v mice to EAE, the disease was elicited by administration of lower amounts of peptide and adjuvants (i.e., 200 μg of MOG_35–55 in 550 or 800 μg of M. tuberculosis in CFA (on day 0) and 200 ng of PTX (on days 0 and 2 p.i.) [57,92]. We have tried somehow to reconcile the divergent EAE outcomes obtained in different works by proposing that EAE expression in Kit^W/W-v model was “tunable” according to the doses of peptide and adjuvants used to elicit EAE. Indeed, we have demonstrated that Kit^W/W-v mice developed milder EAE than controls only when immunized with a “strong” protocol of immunization (i.e., similar to the one used by Secor et al. [77]) [78]. Conversely, when a low/normal protocol of immunization was used (i.e., 100 μg of MOG_35–55 in 200 μg of CFA on day 0 and 200 ng of PTX on days 0 and 2 p.i.) EAE was actually slightly exacerbated in Kit^W/W-v mice [78]. Indeed, the reliance on the specific experimental setting observed in this strain is common to animal models of asthma, contact hypersensitivity and bacterial infection, where the induction protocol can drastically affect the importance of MC’s contributions to the disease model under investigation [93–95]. It can be hypothesised that diverse experimental conditions/protocols for disease elicitation may result in different pathological mechanisms, which might impact on the same mutation in alternative ways. However, a more recent report from Brown’s group has described reduced EAE severity in Kit^W/W-v mice even upon the application of a relatively mild immunization protocol (100 μg of MOG_35–55 in 5 mg/mL CFA and 250 ng of PTX) [90], rendering the interpretation of such discrepancies unclear. Based on the results of Brown et al., it seems that different protocols of EAE induction appear not to be the only factor involved in the divergent results obtained by different groups with Kit^W/W-v mice.

This controversial picture has been further complicated by data produced on a more recently tested c-kit mutant MC-deficient strain, the C57BL/6-Kit^W-sh/W-sh mouse. The W-sash (W-sh) mutation consists of an inversion mutation upstream from the c-kit promoter, covering approximately 3Mb and including 27 genes. The 3′ end of this inversion breaks a regulatory locus that controls c-kit expression specifically in MCs, whereas the 5′ breakpoint is localized between exons 5 and 6 of corin gene, which as a result is disrupted [96,97]. Kit^W-sh/W-sh mice exhibit severe MC-deficiency, lack melanocytes and ICCs, but they are not anaemic nor sterile, unlike the Kit^W/W-v animals [88]. Nevertheless, they are affected by some other hematopoietic alterations such as splenomegaly with expanded myeloid populations, and an increased number of circulating neutrophils, platelets [97] and basophils [94].

Although in a first report Kit^W-sh/W-sh mice were described to develop milder EAE compared to control mice [98], we and others have independently shown in subsequent work that, surprisingly, EAE in C57BL/6-Kit^W-sh/W-sh mice was exacerbated compared to that in control mice with MCs, with an earlier disease onset and a more severe progression compared to sibling controls [78,99]. Kit^W-sh/W-sh mice also displayed more severe EAE under different conditions of immunization [78]. Bennett et al. reported no significant clinical difference in EAE between Kit^W-sh/W-sh and Kit^+/+ mice [92]. Nevertheless, all of these studies were concordant in describing a more pro-inflammatory profile of autoreactive T cells in peripheral lymphoid organs of Kit^W-sh/W-sh animals. Indeed, myelin-specific T cells from MC-deficient mice exhibited an increased proliferative response to MOG_35–55[78,92,99], enhanced secretion of Th1/17 cytokines such as IFN-γ, IL-6 and IL-17A, and a decreased production of Th2 or suppressor cytokines, such as IL-4, IL-5 and IL-10 [78,99]. Higher clinical severity was also associated with a reduction of Treg frequencies in the spleen [78] or the CNS [99] of Kit^W-sh/W-sh mice.

Mast cell knock-in studies have also been conducted to verify the contribution of MC to the EAE output observed in Kit^W-sh/W-sh mice. In our work, intravenous transplantation of BMMCs in Kit^W-sh/W-sh mice 6–8 weeks before EAE induction (in line with common procedures for performing MC-knock-in experiments [20,100]) was not effective in restoring EAE severity to wild-type mice levels [78]. In this setting of engraftment, MCs engrafted only partially the priming sites, (i.e., the inguinal and axillary lymph nodes) but not the CNS, as also observed in previous work [88,100]. Nonetheless, in these conditions we could verify the rescue of some MC-related biological functions, such as normal percentages of Treg and granulocytes in lymph nodes and spleens, respectively, of the MC-engrafted mice. However, this engraftment setting did not allow us to evaluate the contribution of MCs to EAE development into the CNS. By using an “alternative” MC-engraftment experiment, Li et al. showed reversion of increased EAE severity of Kit^W-sh/W-sh mice [99]. Indeed, they injected BMMCs into Kit^W-sh/W-sh animals just before EAE onset. BMMCs-transplanted Kit^W-sh/W-sh exhibited EAE severity, frequency of Treg in the CNS and peripheral myelin-specific immune response comparable to those observed in wild-type mice. Remarkably, in this MC-knock-in setting, MCs were found also in the CNS [99]. It is possible that the enhanced immune cell-infiltration in the CNS of Kit^W-sh/W-sh mice [78,99] might have been the effect of an exacerbated peripheral activation of immune cells but also of their increased trafficking (and/or re-activation) through the BBB into the CNS occurring in absence of MCs. Of note, passive transfer of myelin-activated T cells resulted in earlier EAE onset in Kit^W-sh/W-sh mice compared to controls [78], again suggesting a possible impact of Kit^W-sh/W-sh mutation and/or MCs directly in the effector phase of the disease occurring in the CNS.

Taken together, the data obtained in the Kit^W-sh/W-sh model indicated that MCs might be dispensable for, or even limit, the establishment of anti-myelin T cell responses in both peripheral lymphoid organs and the CNS, regardless the conditions of immunization. Interestingly, we have shown that histidine decarboxylase (HDC)^−/− mice, carrying histamine deficiency but also MC paucity, develop EAE with earlier onset and extensive granulocytic infiltration of the CNS [101]. This phenotype resembles somehow the clinical and histological outcome of MC-deficient Kit^W-sh/W-sh mice, which bear just about one-third of wild-type histamine levels in the brain [102]. In this regard, it should be considered that in the context of CNS autoimmunity, histamine has been demonstrated to reduce BBB permeability by stimulating histamine receptor 3 (H3R) on brain presynaptic neurons and histamine receptor 1 (H1R) on brain endothelial cells [103,104]. In addition, histamine has been shown to reduce the firm arrest of encephalitogenic T cells to the inflamed brain circulation in an in vivo model of early EAE inflammation [105]. Thus, it could be speculated that reduced levels of histamine of Kit^W-sh/W-sh mice might contribute, in part, to EAE phenotype and autoreactive T cell responses observed in these mice.

Collectively, the works discussed above depict an ambiguous scenario about the involvement of MCs in the pathogenesis of EAE. Indeed, in c-kit mutant models, a certain variability in disease outcome in the same MC-deficient strain as well as in Kit^W/W-vversus Kit^W-sh/W-sh mice has been observed. On the whole, the expression of EAE in the Kit^W/W-v model appears to a certain extent to be affected by the immunization conditions, while in Kit^W-sh/W-sh mice, developing similar or more severe EAE compared to wild type mice, there was a trend toward an exacerbated anti-myelin pro-inflammatory T cell response. Different results obtained in the same MC-deficient model may reflect MCs plasticity and their “tunable” response to different kind or amount of stimuli, but may also depend on mouse housing conditions, gut micro flora composition or other reasons that still need to be elucidated. Divergences between Kit^W/W-v and Kit^W-sh/W-sh models may depend on genetic background or may be the result of different and complex hematopoietic alterations of these mice. Indeed, MC-engraftment via intravenous route has been shown not to recapitulate in MC-deficient mice the distribution and amount of MCs observed in wild-type mice [20]. Even though in some cases MC-engraftment was sufficient to recover some biological responses to wild-type levels [78], it cannot be ruled out that MCs may play an aberrant and not-physiologic role in the context of severe immune alterations, such as the neutropenic or neutrophilic status of Kit^W/W-v and Kit^W-sh/W-sh mice, respectively. In this regard, studies on models of antibody-mediated arthritis have provided a straight example of how granulocytes abnormalities of Kit mutant strains might impact the development of immune responses. Initially Kit^W/W-v mice were shown to be resistant to arthritis induced by injection of antibodies (Abs) to glucose 6-phosphate isomerase and MC-secreted IL-1β was proposed to promote joint inflammation in this model [54,55]. A subsequent study demonstrated that, surprisingly, Kit^W-sh/W-sh but not Kit^W/W-v mice were fully susceptible to arthritis induced by Abs to type II collagen and LPS [56]. Also, by depleting Gr1⁺ immune cells, authors had shown that granulocytes, rather than MCs, were playing a major role of in the pathogenic mechanisms driving tissue damage in this model, suggesting that neutropenic status of Kit^W/W-v mice was actually responsible for their resistance to disease [56].

A valuable effort to elucidate MC contribution to EAE pathogenesis has been recently made by Feyerabend et al.[57], who evaluated EAE in a novel mouse model of MC-deficiency, independent of c-kit abnormalities. In this strain, the insertion of a Cre-recombinase into the mast cell carboxypeptidase A3 (Cpa3) locus by targeted recombination resulted in selective deletion of MC-lineage, due to the genotoxic effect of sustained Cre expression [57]. Analysis of MC frequency in heterozygous Cpa3^Cre/+ mice (on a C57BL/6 background) revealed ablation of both mucosal and connective-tissue MC populations in the intestine, skin and peritoneal cavity, and a partial reduction of splenic basophils [57]. Cpa3^Cre/+ mice developed EAE with clinical severity, CNS infiltration by immune cells and peripheral anti-MOG_35–55 T cell response comparable to wild-type mice. Of note, Kit^W/W-v mice developed EAE with the same clinical severity of Cpa3^Cre/+ and Cpa3^+/+ mice, even though they were not compared to Kit^+/+ control littermates [57]. Collectively, data obtained in MC-deficient Cpa3^Cre/+ model have indicated that MCs are neither promoting nor dampening CNS autoimmune response occurring in EAE and play a redundant role in the clinical expression of disease. Interestingly, this study also demonstrated that Cpa3^Cre/+ mice developed full antibody-mediated arthritis with no significant difference compared to Cpa3^+/+ mice, suggesting that MCs are dispensable for development of arthritis and that c-kit mutation rather MC deficiency was responsible for disease resistance in Kit^W/W-v strain [57].