Absence of system xc− on immune cells invading the central nervous system alleviates experimental autoimmune encephalitis

Background Multiple sclerosis (MS) is an autoimmune demyelinating disease that affects the central nervous system (CNS), leading to neurodegeneration and chronic disability. Accumulating evidence points to a key role for neuroinflammation, oxidative stress, and excitotoxicity in this degenerative process. System xc − or the cystine/glutamate antiporter could tie these pathological mechanisms together: its activity is enhanced by reactive oxygen species and inflammatory stimuli, and its enhancement might lead to the release of toxic amounts of glutamate, thereby triggering excitotoxicity and neurodegeneration. Methods Semi-quantitative Western blotting served to study protein expression of xCT, the specific subunit of system xc −, as well as of regulators of xCT transcription, in the normal appearing white matter (NAWM) of MS patients and in the CNS and spleen of mice exposed to experimental autoimmune encephalomyelitis (EAE), an accepted mouse model of MS. We next compared the clinical course of the EAE disease, the extent of demyelination, the infiltration of immune cells and microglial activation in xCT-knockout (xCT−/−) mice and irradiated mice reconstituted in xCT−/− bone marrow (BM), to their proper wild type (xCT+/+) controls. Results xCT protein expression levels were upregulated in the NAWM of MS patients and in the brain, spinal cord, and spleen of EAE mice. The pathways involved in this upregulation in NAWM of MS patients remain unresolved. Compared to xCT+/+ mice, xCT−/− mice were equally susceptible to EAE, whereas mice transplanted with xCT−/− BM, and as such only exhibiting loss of xCT in their immune cells, were less susceptible to EAE. In none of the above-described conditions, demyelination, microglial activation, or infiltration of immune cells were affected. Conclusions Our findings demonstrate enhancement of xCT protein expression in MS pathology and suggest that system xc − on immune cells invading the CNS participates to EAE. Since a total loss of system xc − had no net beneficial effects, these results have important implications for targeting system xc − for treatment of MS.


Background
Multiple sclerosis (MS) is an autoimmune disease that affects about 2.5 million people worldwide. Myelin, the insulating sheath surrounding nerve cells, is attacked and impaired by infiltrating T cells and macrophages, ultimately leading to disturbed neural signaling and transmission by axons [1,2]. It is believed that an imbalance in glutamate homeostasis together with oxidative stress contributes to neurodegeneration [3][4][5][6]. In addition, neuroinflammation is undeniably involved in the pathogenesis of chronic neurodegeneration in MS [7].
The cystine/glutamate antiporter or system x c − is a sodium-independent glutamate transporter that imports extracellular cystine in exchange for intracellular glutamate [31]. System x c − is composed of the subunit 4F2 heavy chain and the specific light chain subunit xCT [32] and is a major source of extracellular glutamate in several rodent brain regions such as the striatum and hippocampus [33][34][35]. Cysteine, the reduced form of the imported cystine, is the rate-limiting amino acid necessary for the synthesis of the major antioxidant of the CNS, glutathione (GSH). Expression and activity of system x c − are enhanced by reactive oxygen species (ROS) and inflammatory stimuli [36][37][38] and, besides GSH production, this increase can potentially lead to glutamate toxicity and neurodegeneration [39,40].
Due to its potential relevance in neurological disorders [40], system x c − has recently been studied in MS; nevertheless, the results are still controversial. Pampliega et al. showed enhanced xCT mRNA expression levels in leukocytes and postmortem optic nerve samples of MS patients compared to healthy controls. Additionally, increased xCT protein and mRNA expression was found in the spinal cord of rats exposed to EAE as compared to control rats [41]. Moreover, using positron emission tomography imaging with (4S)-4-(3-(18)F-fluoropropyl)-L-glutamate, a novel radiotracer to assess system x c − activity, it was shown that its function was enhanced in the same model [42]. In C3H/HeSnJSlc7a11 sut/sut (sut/ sut) mice that are deficient for system x c − , proteolipid protein (PLP)-induced EAE was attenuated compared to their wild type littermates. The latter effect could be replicated by pharmacological inhibition of system x c − with sulfasalazine (SAS) and (S)-4-carboxyphenylglycine [43]. Whereas the above-described studies suggest that enhancement of system x c − contributes to the pathogenesis of MS, a recent report of Morales Pantoja et al. showed diminished xCT mRNA and protein levels in the mouse spinal cord during the course of myelin oligodendrocyte glycoprotein (MOG)  peptide-induced EAE [44]. Moreover, Soria et al. demonstrated that pharmacological inhibition of system x c − with SAS-induced myelin damage and speculated that the cystine/glutamate antiporter is a key factor for the maintenance of oligodendrocyte homeostasis [45]. We here sought to shed light on this controversial subject and clarify the potential of system x c − as a drug target in the treatment of MS, by investigating xCT protein expression levels in NAWM of MS patients as well as the CNS and spleen of EAE-induced mice and by exploring whether EAE development is affected by xCT deletion using xCT knockout (xCT −/− ) mice with a C57BL/6J background [46] or bone marrow (BM) chimeras thereof.
We describe for the first time increased xCT protein expression in the NAWM of MS patients compared to controls without neurological disease. Moreover, we show that this xCT upregulation is independent of activating transcription factor 4 (ATF4), a transcription factor known to initiate transcription of xCT [38,47,48]. In agreement with our human data, xCT protein expression was augmented in the CNS as well as the spleen of EAE mice. xCT −/− mice were equally susceptible to EAE than their wild-type (xCT +/+ ) littermates. However, the lack of xCT exclusively in immune cells attenuated EAE compared to control mice. We therefore hypothesize that system x c − activity on immune cells invading the CNS might be involved in EAE-and MS-dependent neurodegeneration.

MS patients and controls
Freshly frozen brain tissue of MS patients and geographically matched controls has been collected from donors for or from whom a written informed consent for a brain autopsy and the use of the material and clinical information for research purposes had been obtained by the Human Brain and Spinal Fluid Resource Center in Los Angeles (USA). The frozen tissue was stored at −80°C until use. The tissue was sectioned in 5 and 50 μm sections. NAWM was collected from the 50-μm brain slices and used to extract total RNA or protein. Adjacent 5 μm cryosections were assessed by a neuropathologist to confirm the localization of lesions and NAWM.

Animals
All mice used in this study were 10-15-week-old females and were housed and maintained in accordance with national guidelines on animal experimentation. Experiments were approved by the ethical committee for animal experimentation of the Faculty of Medicine and Pharmacy of the Vrije Universiteit Brussel, of the Faculty of Sciences of Ghent University or of the University of Mainz. C57BL/6J mice were obtained from Charles River Laboratories (France). The results are presented in accordance with the ARRIVE guidelines for reporting experiments involving animals [49].
xCT −/− mice and their xCT +/+ littermates have a C57BL/6J background and are high-generation descendants of the strain described previously [46]. Despite the absence of xCT, these mice are healthy and fertile [46]. They have unaffected hippocampal and striatal GSH levels and do not show any signs of increased oxidative stress or increased susceptibility to oxidative stress-induced damage in the brain. Deletion of xCT does result in decreased extracellular glutamate concentrations in the hippocampus and striatum [33,34]. Genotypes were confirmed by PCR amplification of ear DNA using REDExtract-N-Amp Tissue PCR Kit (Sigma-Aldrich, USA) and the following primers: As for the BM transplanted mice, conventional chimeras were generated as described previously [50,51], with the following modifications: recipients were C57BL/6J mice and to obtain conventional BM chimera, donor cells were derived from xCT −/− mice and xCT +/+ littermate controls. In brief, recipient animals were sublethally irradiated with 1100 cGy (split dose). Donor animals were sacrificed by cervical dislocation. BM cells were isolated by flushing of the femur and tibia, resuspended and MACS®-depleted of CD90 + T cells. Recipients were reconstituted with 8-16 × 10 6 donor BM cells 8 h after irradiation. Mice were kept on 0.01% Enrofloxacin (Baytril®, Bayer Health Care) in drinking water for 4 weeks. Engraftment took place over 8 weeks of recovery.

Induction and assessment of EAE
EAE was induced in C57BL/6J, xCT −/− , and xCT +/+ mice as well as BM-transplanted mice using the Hooke Laboratories kit EK-2110 (Hooke Laboratories, Inc., USA) and according to the manufacturer's instructions. Sham-treated mice received an injection of adjuvant components that was identical for the study group with omission of specific antigens (Hooke Laboratories kit CK-7110, Inc., USA). Briefly, mice were immunized by subcutaneous injection of a (MOG)  emulsion in complete Freund's adjuvant, followed by an intraperitoneal injection with pertussis toxin in phosphate-buffered saline (PBS) immediately after immunization as well as 24 h later. Starting from day 7 post-immunization, animals were daily evaluated for clinical signs of disease, according to the following scoring system: 0, no clinical/ obvious changes; 0.5, the tip of the tail is limp; 1, limp tail; 1.5, limp tail and hind leg inhibition; 2, limp tail and weakness of the hind legs; 2.5, limp tail and dragging of hind legs; 3, limp tail and complete paralysis of the hind legs; 3.5, limp tail and complete paralysis of the hind legs in addition to unable to right itself when placed on its side; 4, limp tail, complete hind leg and partial front leg paralysis; 4.5, complete hind and partial front leg paralysis, no movement around the cage, mouse is not alert; and 5, moribund or dead. According to the ethical guidelines for animal experimentation, mice with a clinical score of 4 or higher for two consecutive days were excluded from the experiments. A researcher blinded to the genotype performed all scorings and area under the curve (AUC), and highest clinical score reached were analyzed and compared between genotypes.
C57BL/6J mice were sacrificed 25 days after EAE induction by cervical dislocation. The brain, spinal cord, and spleen were collected, snap frozen, and stored at −80°C until use. xCT +/+ and xCT −/− mice were sacrificed with an overdose of pentobarbital (Nembutal®; Ceva Sante Animale, Brussels, Belgium) 25 days after EAE induction, followed by transcardial perfusion with saline and collection of the spinal cord for further histologic examination. BM-transplanted mice were sacrificed 15 days following EAE induction and transcardially perfused with saline (histological analysis) or ice-cold PBS (FACS analysis).

Histology
Mouse tissue was postfixed in 4% paraformaldehyde in 0.1 M PBS overnight, transferred to a 10% sucrose solution in 0.1 M PBS for 24 h and preserved in a 25% sucrose solution in 0.1 M PBS at 4°C. A 15-mm section of the lumbosacral spinal cord was placed in Tissue Tek OCT compound (Sakura, The Netherlands) and frozen on dry ice. Next, the tissue was sliced on a cryostat in 20-μm sections that were mounted on Superfrost Ultra Plus slides (Thermo Scientific, Waltham, MA, USA) and stored at −80°C until further analysis. To determine the extent of demyelination, luxol fast blue (LFB) staining was performed. Slides were immersed in 0.1% LFB solution (0.5 g LFB, 500 ml 99% ethyl alcohol, 2.5 ml glacial acetic acid) for 16 h at 56°C. Differentiation was performed in 0.05% lithium carbonate solution. Afterwards, the slides were gradually dehydrated in increasing concentrations of ethyl alcohol followed by xylene and coverslipped with DPX mounting medium (VWR, Radnor, PA, USA). Histological quantification was done by defining a demyelination score: 0, normal white matter; 1, rare foci; 2, a few areas of demyelination; and 3, large areas of demyelination.
The amount of invading cells was analyzed by cresyl violet staining using standard staining procedures. The sections were scored for infiltration: 0, no infiltrating cells; 1, few infiltrating cells; 2, numerous infiltrating cells; and 3, widespread infiltration. Scoring was performed blinded.

Real-time PCR
Total RNA was isolated from scrapings of the 50 μm slices of the human tissue using the TRIzol® (Life Technologies, USA) extraction method. cDNA was generated by the Taqman reverse transcription kit (Life Technologies, USA), and SYBR®Green Real-Time-PCR Master Mix (Life Technologies, USA) was used to perform realtime PCR. xCT RNA expression was analyzed using primer sets designed with the Primer Express 3 software (Applied Biosystems®, Life Technologies, USA): xCT upstream primer 5′-TGATTCATGTCCGCAAGCA-3′ and downstream primer 5′-TGTCGAGGTCTCCAGAGAAG AG-3′ and the reference gene GAPDH upstream primer 5′-TGCACCACCAACTGCTTAGC-3′ and downstream primer 5′-GGCATGGACTGTGGTCATGAG-3′. All primer sets had efficiencies between 98 and 100% over 5 log changes and amplified a single band as determined by melting curve analysis.
Protein concentrations were quantified using the NanoDrop Spectrophotometer ND-1000 (Thermo Scientific, Waltham, MA, USA). Equal concentrations of proteins were loaded on the gel. Following separation of proteins by SDS-polyacrylamide gel electrophoresis (PAGE; 4-12% Bis-Tris precast gel, Bio-Rad Laboratories, USA) under reducing conditions, proteins were transferred to a polyvinylidene fluoride membrane (Immun-Blot PVDF Membrane, Bio-Rad Laboratories, USA). After 1-h incubation with 5% membrane blocking agent (RPN2125V; GE Healthcare, UK), blots were incubated overnight at 4°C with an immunoaffinity purified rabbit polyclonal antibody to xCT diluted in membrane blocking agent (NB300-318, lot number G3; Novus Biologicals, USA; 1/2000 for NAWM and mouse spleen; 1/ 5000 for the mouse brain and spinal cord) [52]. Next, the membranes were incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit immunoglobulin G antiserum (1/4000 for NAWM; 1/15000 for mouse CNS homogenates and spleen; 30 min; Dako, Denmark). Immunoreactive bands were visualized using enhanced chemiluminescence (ECL Select, RPN2235; GE Healthcare, UK) and ImageQuant™ LAS 4000 biomolecular imager (GE Healthcare) and analyzed by ImageJ software (National Institute of Health, Bethesda, MD, USA) to obtain the optical densities (OD). OD of the immunoreactive bands was normalized to the OD of the entire lane after a total protein stain (SERVA purple; SERVA Electrophoresis GmbH, Germany). An arbitrarily chosen control sample was set as a reference (100%), and the OD of all other samples was calculated relative to this reference. As molecular weight standard, Novex Sharp Protein Standard (LC5800; Novex by Life Technologies, USA) was used. The specificity of the xCT antibody was confirmed on the knockout tissue [52].

FACS analysis
For immune cell isolation from the CNS, lethally anesthetized animals were transcardially perfused with icecold PBS. The brain and spinal cord were isolated, cut into small pieces, and diluted in Iscove's modified Dulbecco's medium (IMDM; Life Technologies, USA) substituted with 5 mg/50 μL collagenase/clostridiopeptidase (Sigma-Aldrich, USA), 1 mg/ml collagenase/dispase (Roche, Germany), and 1000 U/50 μL DNAse (Roche, Germany). After incubation for 30 min at 37°C in the water bath, the CNS tissue was put through a mesh (100 μm) and mononuclear cells were separated by conventional 40/70 Percoll centrifugation. FACS analysis of surface markers was performed directly after isolation. Cytokine analysis was preceded by plate-bound anti-CD3/anti-CD28 stimulation for 4 h (brefeldin A was added after 2 h). CNS cells were pre-gated on lymphocytes and monocytes, followed by selection of CD45 high cells. Antibodies used for FACS surface and intracellular stainings are as follows:

Statistical analysis
Statistical analysis was performed using GraphPad Prism 6.0 software. Data are expressed as mean ± standard error of the mean (SEM). Significance was assessed by Mann-Whitney U test. Differences between groups were rated significant at p < 0.05. Correlations between xCT protein and ATF4/p-NF-κB(p65)/NF-κB(p65)/iNOS protein expression levels were computed using Pearson's test.

xCT protein levels are increased in the NAWM of MS patients
Characteristics of MS patients and controls without neurologic disease that were included in this study are outlined in Table 1. There is a significant difference in age and postmortem delay between the group of controls and MS patients; however, there is no correlation between those two parameters and xCT protein levels or any other protein evaluated in these samples (results not shown).
In conclusion, although xCT expression levels were increased in NAWM of MS patients compared to controls, we could neither identify the pathway that leads to this upregulation nor correlate the changes in xCT expression with inflammatory markers.
xCT protein expression levels are increased in the CNS and spleen of C57BL/6J mice after EAE induction To corroborate our findings and clarify their functional significance, we first studied xCT protein expression levels in the brain, spinal cord, and spleen of EAEinduced C57BL/6J mice (n = 4) compared to shamtreated mice (n = 5) that only received adjuvant components. Clinical signs were established from day 7 after immunization with an increase of severity until day 17, followed by partial recovery (Fig. 3a). A significant increase in xCT protein expression at day 25 was observed in the brain (relative expression control mice vs. EAE mice: 100.00 ± 8.17 vs. 158.80 ± 22.76%; p = 0.016, Fig. 3b), spinal cord (relative expression control mice vs. EAE mice: 100.00 ± 8.55 vs. 139.50 ± 7.78%; p = 0.016, Fig. 3c), and spleen (relative expression control mice vs. EAE mice: 100.00 ± 14.31 vs. 178.50 ± 17.79%; p = 0.032, Fig. 3d) of EAE-induced mice compared to sham controls.

Discussion
The goal of this study was to further investigate the potential of system x c − as a drug target for the treatment of MS. Consistent with earlier reports [41][42][43], our data support that system x c − is involved in the pathophysiology of MS. Yet, our data show that the absence of system x c − on immune cells invading the CNS, rather than resident cells of the CNS, seems to ameliorate the clinical outcome of mice in the EAE model, and therefore suggest that enhancement of system x c − on these invading immune cells contributes to EAE disease.
Glutamate homeostasis plays an essential role in limiting neuronal toxicity because elevated extracellular glutamate levels can lead to excessive activation of ionotropic glutamate receptors and, as a result, excitotoxicity and neuronal death. Glutamate toxicity has been shown to be a key player in several neurological disorders [8,58]. System x c − releases glutamate in exchange for cystine and was shown to be the major source of extracellular glutamate in several brain regions [33][34][35]; enhancement of this transporter may therefore induce detrimental effects via release of toxic amounts of glutamate in the extracellular space. However, by providing cysteine [59] and stimulating GSH synthesis [45,60], system x c − fulfills a dual role Fig. 3 xCT protein expression is increased in the CNS and spleen of C57BL/6J mice after EAE induction. a Clinical score of C57BL/6J mice after EAE induction (n = 4). Sham-treated mice (n = 5) were included in this study to correct for adjuvant-induced responses and had a clinical score of 0 throughout the entire experiment. xCT protein expression levels in the brain (b), spinal cord (c), and spleen (d) were significantly increased in C57BL/6J mice after EAE induction compared to sham-treated animals as measured using semi-quantitative Western blotting (data are expressed as the mean ± SEM). Mann-Whitney U test, *p < 0.05. A representative xCT band for the brain (b), spinal cord (c), and spleen (d) of control and EAE mice is shown in the inset and its relevance in neurological disorders remains a matter of debate. Previous reports have shown that xCT mRNA expression is increased in postmortem optic nerve and peripheral blood cells of MS patients [41]. Accordingly, in our study, we report for the first time the increased xCT protein levels in the NAWM of MS patients. These increased levels are most likely the result of xCT upregulation on resident CNS cells rather than infiltrating immune cells given the absence of overt leukocyte infiltration in the NAWM [61]. In combination with the previously reported reduction in EAAT2 expression at the periphery of MS lesions [15,62], glutamate release via system x c − might lead to excitotoxicity, in accordance with elevated glutamate levels reported in NAWM of MS patients [11].
Several transcription factors that regulate xCT expression have been identified. Recently, the PI3K/Akt/ GSK3β/eIF2α/ATF4 pathway was shown to be responsible for increased xCT expression in the hippocampus of patients suffering from temporal lobe epilepsy [53]. ATF4 binds to the amino acid response element in the xCT promoter region to induce xCT transcription. Moreover, the phosphorylation of eIF2α in response to cellular stress situations has been associated with other white matter disorders such as childhood ataxia with central hypomyelination/vanishing white matter disease syndromes, a collection of autosomal recessive neurological leukodystrophies that are characterized by white matter hypomyelination [63]. Elevated levels of phosphorylated eIF2α are also found in CNS oligodendrocytes of EAE mice compared to control animals [64]. In Fig. 4 xCT −/− and xCT +/+ mice are equally susceptible for EAE development. Clinical score of xCT +/+ (n = 6) and xCT −/− mice (n = 7) after EAE induction (a). No significant differences in EAE severity were found between the two genotypes comparing AUC (b) and highest score reached (c). On day 25, the spinal cord was collected, sliced, and stained for demyelination (d, e), infiltration (f), and microglial activation (g) and subsequently scored (n = 3-4). Scoring was performed blinded. Data are expressed as mean ± SEM, Mann-Whitney U test. Representative images of LFB (d), MBP (e), cresyl violet (f), and Iba1 (g) staining in the lumbar section of EAE-induced xCT +/+ and xCT −/− mice are shown below each graph (scale bar, 100 μm) the present study, however, we could not correlate increased xCT protein expression levels in NAWM of MS patients to ATF4 alterations.
Many inflammatory stimuli, including the cytokine interleukin-1β [54] and tumor necrosis factor α [36] as well as bacterial lipopolysaccharide (LPS), which are associated with the activation of the transcription factor NF-kB [37,55,65] have been shown to upregulate xCT [36,37,54]. A classical NF-kB downstream target in the context of inflammation is iNOS [66]. Although some evidence suggests that xCT is not a direct downstream target of NF-kB [37], NF-kB and its downstream targets might represent surrogate parameters for the still unknown mechanism of inflammatory upregulation of xCT. Moreover, several studies provide evidences that NF-κB is involved in MS and EAE [67]. In line with these observations and in accordance with previous studies [67][68][69], we observed a significant upregulation of iNOS in NAWM of MS patients. Yet, despite the increased iNOS expression levels, no changes could be identified in either the expression of NF-κB or its phosphorylation, suggesting that iNOS expression is regulated by other signaling pathways, independent of NF-κB [70]. However, we cannot exclude that nuclear translocation of NF-κB, an essential step for activation of target genes, is the main pathway of iNOS induction in MS. Moreover, phosphorylated proteins might be subjected to dephosphorylation during the postmortem period. Still, as iNOS levels did not correlate with xCT protein levels, the pathways that regulate the expression of iNOS and xCT seem to be different in MS NAWM. As such, the transcription factor/pathway that upregulates xCT in NAWM of MS patients remains to be determined. One possibility would be Nrf2, a transcription factor that binds to an antioxidant response element in the promoter region of xCT [40]. However, it has been described that Nrf2 expression is increased around the active lesion edge in MS patients, decreasing towards the NAWM [71], and therefore, we speculate that this pathway is most likely not involved.
After EAE induction in C57BL/6J mice, we also confirmed increased xCT protein expression in the brain, spinal cord, and spleen compared to controls, in line with the observations in the spinal cord of EAE rats [41,42]. However, Morales Pantoja et al. recently reported diminished xCT protein levels in EAE mice 21 days post-immunization; nevertheless, no differences were observed 30 days post-immunization. Since we collected our samples 25 days after EAE induction, this contradictory result may be due to the different time-point evaluated, suggesting interesting dynamics in xCT protein expression during disease course [44]. Another important reason for this contradictory finding may be the use of non-specific anti-xCT antibodies [52]: although most of the commercially available anti-xCT antibodies label a protein with a molecular weight of 55 kDa, we recently showed that xCT protein migrates at 37 kDa in SDS-PAGE and confirmed the specificity of this signal on xCT −/− tissue [72]. System x c − can fulfill a dual role in the CNS. Induction of xCT might result in an overload of glutamate into the extracellular space, supporting the hypothesis that enhancement of system x c − during EAE disease, together with dysfunction of EAATs [18], might result in glutamate dysregulation followed by excitotoxicity, and hence axonal degeneration. On the other hand, system x c − imports cystine in the cells, and thereby promotes the maintenance and synthesis of GSH, the most important antioxidant in the brain and essential to prevent oxidative stress and cell damage [34,39,73]. Several research groups demonstrated that reduction of oxidative stress ameliorates EAE [74][75][76], and GSH plays an important role in this process. Moreover, GSH levels are decreased in the spinal cord of EAE mice during the acute phase of the disease [77].
Surprisingly, genetic loss of xCT in C57BL/6J mice did not affect the clinical outcome after EAE induction. Nonetheless, and in contrast to our results, recent data provided by Evonuk et al. show a resistance to EAE development in sut/sut mice, which harbor a spontaneous large deletion that comprises the last exon of the slc7a11 gene which encodes xCT and leads to functional system x c − deficiency [43,78]. Yet, discrepancies in EAE outcome could be due to two dissimilarities, i.e., different EAE induction protocols and the genetic background of the sut/sut mice as compared with our xCT −/− mice. Sut/sut mice arose spontaneously in the C3H/HeDiSn colony, whereas xCT −/− mice have a defined C57BL/6J background. Moreover, the mutation in sut/sut mice is not restricted to deletion of the last exon of Slc7a11 but extends from intron 11 through exon 12 resulting in a truncated mRNA transcript for xCT and covers a large fragment of 480 kb [78,79]. Sut/sut mice develop brain atrophy starting from week 15 after birth [80], a phenomenon that is not observed in xCT −/− mice where no anatomical changes are found in the brain [34]. As such, results obtained in sut/sut mice are not always in line with the data from xCT −/− mice [34,79,80]. Also, SAS was shown to attenuate EAE disease and to reduce demyelination [43]. Here, we should keep in mind that although SAS is known to inhibit system x c − [81], it also exerts anti-inflammatory actions [82], which can obviously bias the result. Yet, the authors confirmed these results with (S)-4-carboxyphenylglycine (CPG), a group I Fig. 6 Immune cell infiltration into the CNS of irradiated mice transplanted with BM originating from xCT +/+ or xCT −/− mice is not different between the two groups after EAE induction. Proportion of immune cells relative to the total amount of isolated cells (a) and the proportion of CD4 + , CD11b + , CD4 + IL-17 + , CD4 + IFN-γ + , and CD4 + FoxP3 + cells within the immune cell population (b) in the CNS after EAE induction in mice that are expressing (xCT +/+ BM mice) or lacking xCT in their immune cells (xCT −/− BM mice). Data shown are means ± SEM; Mann-Whitney U test metabotropic receptor antagonist that also inhibits system x c − [83], supporting the idea that the observed effect is most probably related to system x c − inhibition. Finally, it is important to mention that sut/sut mice were immunized with PLP to induce EAE, a model that is closely related to relapsing-remitting MS, whereas our xCT −/− mice were EAE-induced with MOG, which serves as a more chronic model. Therefore, discrepancies in the effects of xCT deficiency on EAE severity might arise from differences in the experimental EAE paradigm, background strain or the genetic modification that leads to xCT deficiency.
In contrast to the observations described above, Soria et al. recently showed that chronic administration of SAS causes a decrease in myelin levels together with a significant reduction in rotarod performance after 28 days of treatment [45]. Despite the fact that the essential role of the cystine/glutamate antiporter for oligodentrocyte survival in vitro is clear cut, as well as for other cell types such as neurons [39], it is commonly accepted that alternative mechanisms are able to provide cyst(e)ine to cells and therefore counterbalance a GSH shortage in vivo. In our hands, chronic administration of 320 mg/kg once daily or 160 mg/kg twice daily of SAS did not induce any signs of motor deficits in xCT +/+ nor xCT −/− mice after 28 days of treatment (unpublished observations). Yet, since information such as sex and strain of the mice used in the experiment were not reported in the aforementioned study, we cannot exclude that these discrepancies are related to these essential factors.
xCT mRNA is upregulated in peritoneal macrophages after injection with LPS [36] and MOG peptide (unpublished results), in line with the upregulation of xCT protein we observed in the spleen of EAE mice. Furthermore, leukocytes from MS patients display higher xCT mRNA levels [41]. These findings suggest that system x c − might be enhanced on immune cells in MS and after EAE induction, whereby immune cells invade the spinal cord, resulting in increased glutamate release and ensuing damage. Indeed, it is known that the blood-brain barrier (BBB) in MS disease is disrupted and that a larger activity of ROS and inflammatory cytokines promotes the migration of immune cells into the CNS to trigger inflammation and, consequently, MS lesions [84]. To test this hypothesis, we transplanted BM of xCT −/− mice into irradiated mice to create animals expressing xCT throughout their entire body, except for their immune cells. This setup demonstrated a general protective outcome for mice lacking xCT in immune cells, suggesting that system x c − on immune cells plays an important role in the degree of EAE disease and possibly, in human disease as well. The differences in clinical score that we observed in BM-transplanted mice are not related to differences in invasion of immune cells into the CNS, since the number of infiltrating cells was similar in both groups of mice, as was seen with the infiltration score as well as with the FACS analysis. It is noteworthy to mention that in the study of Evonuk and colleagues, a decrease in the number of infiltrating CD4 + , IFN-γ + , IL-17A + , and FoxP3 + T cells was noticed in the spinal cord of EAE mice after SAS treatment, suggesting a modulating role in immune cell infiltration for this compound [43]. However, although the authors also looked at the effects of CPG as well as genetic loss of xCT on the clinical outcome after EAE induction, they only examined the effect of SAS, and not CPG or genetic deletion of xCT, on immune cell infiltration. As such, the possible effect of the anti-inflammatory action of SAS on the BBB permeability and/or the infiltration of immune cells cannot be entirely excluded and might explain the discordance with our present study. Reduced xCT expression levels in the spinal cord of SAS-treated EAE mice compared to PBStreated EAE mice [43] support indeed the idea that the anti-inflammatory actions of SAS are not to be neglected in this model.
So far, information about xCT expression on immune cells is quite limited, and the exact function of system x c − in these cells needs further investigation. Cystine transport in human T cells was barely detected [85][86][87][88]. On the other hand, cystine/glutamate exchange in dendritic cells [88] and macrophages [89,90] has been shown to mediate T cell activation and GSH synthesis, respectively. Moreover, increased glutamate levels released via system x c − can affect T cells function, acting on AMPA receptors, known to be expressed by activated T cells [91] and increased in MS patients [92]. Treatment with NBQX, an AMPA receptor antagonist, ameliorates clinical EAE symptoms in both mice [93] and rats [94] without affecting lesion size or CNS inflammation, in line with our observations in xCT −/− BM mice. Since xCT levels are increased in infiltrating dendritic cells and macrophages in EAE lesions [41] and glutamate released by dendritic cells is known to interact with T cells [91], we can speculate that the absence of xCT and the resulting missing interaction between toxic glutamate levels and AMPA receptors present on T cells might affect EAE, ameliorating the course of the disease without affecting lesion size and degree of CNS inflammation.

Conclusions
Our findings confirm the involvement of system x c − in MS although additional research is needed. Indeed, in contrast to other neurological disorders where loss of system x c − is clearly beneficial, this is less straightforward in MS due to concurring effects in the CNS and in the immune system. We hypothesize that glutamate release via system x c − on CNS-invading immune cells participates in excitotoxicity and contributes to MS pathology and that preventing the induction of xCT in inflammatory cells rather than inhibiting system x c − pharmacologically might be a promising strategy to treat inflammatory CNS demyelination.