Increased expression of cystine/glutamate antiporter in multiple sclerosis

Peripheral blood for RNA expression studies was obtained from the Neurology Service of the Hospital of Basurto. Healthy controls matched by sex and age were recruited at the University of the Basque Country. Characteristics of peripheral blood samples are described in Table 1. All the MS patients used in this study suffered the relapsing form of the disease (R-MS), which included relapsing-remitting (RR) and secondary-progressive (SP) subtypes. Thirty five per cent of all the patients were under long term treatment with immune-suppressors. Among R-MS patients, 24 were experiencing a clinical relapse and were not being treated with steroids at the time of blood sampling. All subjects gave informed consent to participate in the study. The study was approved by the local ethical committee.

For RNA expression studies, postmortem optic nerve samples from 16 long-standing MS patients and 12 control subjects (who died from non-neurological diseases) were obtained at autopsy under the management of the Netherlands Brain Bank. These samples have been previously characterized ([21]; see Table 2). MS samples were classified according to clinical data together with macroscopic tissue analysis as normal appearing (NAON) or damaged optic nerves (DON).

For immunohistochemistry studies, additional postmortem spinal cord samples from MS and control subjects were obtained at autopsy from the Netherlands Brain Bank. Characteristics of spinal cord samples are described in Table 3. Frozen tissue was kept at -80°C until use.

For comparisons, all MS samples were matched with control samples for sex, age and post-mortem delay.

Blood was taken into heparin-coated tubes and peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation over Ficoll-Paque PLUS (GE Healthcare). For flow cytometry analysis PBMCs were plated at 1.5 × 10⁶ cells per ml in complete RPMI media with GlutaMax and 25 mM HEPES (Gibco), and supplemented with 4.5 g/l glucose, 1 mM sodium pyruvate, 10% heat-inactivated FBS and 1% penicillin-streptomycin. Monocytes were purified from isolated PBMCs by MACS using the Monocyte Isolation Kit II (Milteny Biotec) following manufacturer's instructions. Cells were plated at 0.5 × 10⁶ cells per ml in complete RPMI media for 72 h. Then, the day before glutamate release experiments media was changed to DMEM with 1% of heat-inactivated FBS and 1% of penicillin-streptomycin. Human U-937 monocytic cell line (ATCC) was cultured in suspension in complete RPMI media supplemented as before. Drugs were added 30 min before cell activation by LPS (1 μg/ml; Escherichia coli O11:B4; Sigma-Aldrich), and 48 h after, cells and culture supernatants were collected for experiments.

Chronic, relapsing EAE was induced in C57BL/6 mice by immunization with 300 μl of MOG(35-55) (200 μg; Sigma, Barcelona, Spain) in incomplete Freund's adjuvant supplemented with 8 mg/ml Mycobacterium tuberculosis H37Ra. Pertussis toxin (500 ng; Sigma) was injected on the day of immunization and again two days later. Neurological impairment was monitored as in acute EAE. Mice were sacrificed at 20 days after immunization, when maximal motor symptoms were detected (see Additional file 1, Figure S1).

Glutamate release from U-937 and human primary monocytes was monitored by an enzymatic assay based on the activity of the enzyme L-glutamate dehydrogenase (15 U/ml; Sigma-Aldrich) as previously described [13, 22]. Glutamate was oxidized by L-glutamate dehydrogenase in the presence of 1 mM NADP⁺ to α-ketoglutarate with the formation of NADPH and fluorescence emission at 430 nm (excitation light 335 nm). Release was quantified in reference to standard curves constructed with exogenous glutamate. In each assay, 20 μl of culture supernatant were used in a final volume of 300 μl.

Intracellular reduced glutathione (GSH) levels were determined using the QuantiChrom™ GSH Assay Kit (BioAssay Systems) according to the instructions of the manufacturer. Briefly, control and LPS-treated U-937 cells were harvested, rinsed in cold PBS and homogenized by sonication in a solution containing 50 mM phosphate (pH = 6) and 1 mM EDTA. Lysates were centrifuged at 10,000 g for 10 min at 4°C and supernatant collected for measurement of GSH. The method comprises a colorimetric reaction of GSH with 5,5'-dithiobis(2-nitro-benzoic acid), which yields a yellow product. Optical density was measured at 415 nm in a plate reader.

For leukocyte isolation, blood samples from MS and controls were collected in K₃EDTA and kept not longer than 6 h at 4°C for avoiding changes in the RNA expression profile. Blood was centrifuged at 300 × g during 10 min and then, the plasma was removed. Red cells were lysed with a buffer containing 155 mM NH₄CL, 01.mM EDTA and 10 mM KHCO₃, pH 7.4. The blood-buffer solution was centrifuged at 300 × g during 10 min and the pellet was washed twice. The end resulting pellet of leukocytes was stored at -80°C until use.

Total RNA was isolated from U-937, rat optic nerve, and human optic nerve and leukocytes using TRIzol Reagent (Invitrogen) according to manufacturer's guidelines. Subsequently, 2 μg of RNA were retrotranscribed into cDNA using SuperScript III retrotranscriptase (200 U/μl; Invitrogen) and random hexamers as primers following manufacturer's instructions.

Gene expression levels were analyzed using qPCR. Primers were designed with PrimerExpress software (Applied Biosystems), and their sequences and gene accession numbers are provided in Additional file 2, Table S1. qPCR reactions were carried out as previously described [21] using SYBRGreen (Invitrogen) as the fluorescent reporter dye in an ABI PRISM 7000 Sequence Detection System instrument (Applied Biosystems). qPCR products were subjected to a dissociation protocol to ensure that a single product of the expected melting temperature was obtained. mRNA expression level of a given sample was calculated from a standard curve of stock cDNA obtained from the same tissue. Then, a normalization factor was calculated for each sample by GeNorm v.3.4 software based on the expression levels of 4 housekeeping genes for human samples (18s ribosomal RNA, HRPT-1, β-2 microglobulin and cyclophilin A), and 3 genes for rat samples (18s ribosomal RNA, cyclophilin A and GAPDH) [23]. Expression level of each sample was divided by this normalization factor. Data are expressed as fold change in expression compared to the mean of the matched controls. The relative abundance of the different transporters was determined by using the ΔCt method (see User Bulletin 2; Applied Biosystems).

Western blot analysis of cystine/glutamate antiporter expression was done in U-937 cell cultures and in rat and mice spinal cord homogenates by conventional SDS-PAGE polyacrylamide electrophoresis as described previously [24]. The system x_c ^- was revealed with antibodies against the catalytic subunit xCT (0.25 μg/ml; ab37815; AbCam). Densitometric analysis was performed using the NIH Image program (n = 3 in triplicate).

xCT expression in human blood monocytes was analyzed by flow cytometry. PBMCs were stained using an antibody to CD14 conjugated to PE (BD Pharmigen), and an antibody to the intracellular epitope of xCT (2.1 μg/ml; ab37815; AbCam) conjugated to FITC. IgG (Sigma) was used as isotype control for xCT antibody. Extracellular staining was done in PBS buffer containing BSA, and for intracellular staining saponin was added to the previous buffer. Acquisition was performed in BSA with PBS buffer, using a Coulter EPICS Elite flow cytometer. Data were analyzed using the WinMDI v 2.9 software. Density plots were used for gating CD14⁺ cells in PBMC populations, and then histograms for xCT-FITC were opened. The geometric mean of the fluorescence was used for statistical analysis.

xCT protein expression in rat and human spinal cord was examined by double immunofluorescence. Animals were deeply anesthetised with chloral hydrate (500 mg/kg, i.p.) and transcardially perfused with 0.1 M sodium phosphate buffer (PB), pH 7.4, followed by 4% paraformaldehyde (PFA) in the same buffer. After extracting spinal cord, tissue was postfixed 2 h in 4% PFA in PB and subsequently dehydrated in PBS containing 20% sucrose. Tissue was included in Tissue-tek resin (Sakura) in a Frigocut 2800 E cryostat (Leica) and frozen. Serial sections of 10 μm from both rat and human spinal cords were incubated with monoclonal antibodies against CD68 (1:50; DakoCytomation) for human monocyte-macrophage staining, and OX-42 (1:50; Serotec) for rat macrophage-microglia staining. Next, samples were incubated with anti-xCT antibody (5 μg/ml; ab37815; AbCam). Appropriate fluorescent secondary antibodies (Invitrogen) were used in each case. In human samples, Hoechst 33258 labeling (5 μg/ml, Sigma) was used for cell nucleus staining. Control staining was done in the absence of primary antibodies and by preincubating with peptide (ab93079; Abcam) and gave no labeling (Additional file 3, Figure S2). The same results were obtained with antibodies to xCT from Novus (NB300-317) and the staining with this antibody also disappears after preincubation with Novus peptide (NB300-317PEP) (data not shown). Images were taken by Olympus Fluoview FV500 and Leica LCS SP2 AOBs confocal microscopes.

We have previously demonstrated that activation of U-937 monocytes induces glutamate release [13]. So, we now analyzed the mechanism of glutamate release in U-937 cells as well as in peripheral blood monocytes in vitro. LPS-activated U-937 monocytes (LPS 1 μg/ml, 48 h) showed a significant increase in glutamate levels (n = 4; p < 0.001; Figure 1A), as previously described [13]. Incubating cells with the non competitive blocker of glutamate transporters DL-threo-beta-benzyloxyaspartate (TBOA; 100 μM) or with dihydrokainate (DHK; 1 mM), a selective inhibitor of EAAT2 glutamate transporter, did not block LPS-induced glutamate release by monocytes (n = 4; Figure 1A), excluding any role of glutamate transporters in such glutamate release. In contrast, incubating cells with aminoadipic acid (AAA, 1 mM), an inhibitor of cystine/glutamate antiporter [25, 26], significantly reduced glutamate release by LPS-activated monocytes (n = 4; p = 0.02; Figure 1A), indicating that the x_c ^- system is the main glutamate release mechanism in monocytes. Accordingly, intracellular glutathione levels were increased after LPS treatment (n = 3; p = 0.03; Figure 1B), as a result of an increased function of xCT antiporter.

Finally, 6-diazo-5-oxo-L-norleucine (DON; 1 mM), an inhibitor of glutaminase enzyme, also inhibited significantly glutamate release by monocytes (n = 4; p = 0.02; Figure 1A). Glutaminase is the key enzyme to replenish glutamate pool in neuronal presynaptic terminals. Similarly, it is possible that blocking glutaminase in monocytes depletes the pool of glutamate that is released after their activation with LPS.

Because the system x_c ^- is modulated by oxidative stress [27], we next examined whether LPS modulated system x_c ^- expression in monocytes. System x_c ^- is a heteromeric protein complex consisting on a catalytic L chain (xCT) and a regulatory H chain (4F2hc) [27]. In control patients, the levels of xCT in CD14⁺ monocytes are much higher than the levels of the major glutamate transporters EAAT1 and EAAT2 (Figure 1C). We then analyzed the expression of xCT in U-937 cells after their activation with LPS (1 μg/ml, 48 h) by quantitative PCR (qPCR) and Western blot. LPS activation (1 μg/ml) induced an increase in xCT mRNA expression (n = 5; p = 0.014; Figure 1D). Accordingly, Western blot analysis revealed a two-fold increase in xCT protein levels (n = 3; p = 0.01; Figure 1E). Upregulation of xCT expression is not secondary to glutamate homeostasis alteration since the incubation with glutamate (100 μM to 1 mM; 48 h) did not change xCT mRNA expression (n = 3; Figure 1F), suggesting that extracellular glutamate increase is correlative, not causative, of xCT upregulation. In addition, the expression of other enzymes involved in glutamate metabolism, such as glutamine synthetase, was not change after LPS treatment (Figure 1D). Similar to LPS, activation with TNFα also induced an increase in xCT mRNA expression (Figure 1F). Altogether, these results show that LPS activation enhances xCT expression and function in U-937 monocyte cell line.

We then examined if the aforementioned features are also present in blood monocytes freshly isolated by MACS. After isolation, cells were maintained in culture for 96 h and subsequently activated with LPS (1 μg/ml, 48 h). Similarly to U-937 cell line, monocytes showed an increase of 33.5% in glutamate levels after their activation with LPS (p = 0.04; Figure 2A), which was blocked in the presence of AAA (n = 3; p = 0.01; Figure 2A). To confirm changes in xCT expression in LPS-activated human monocytes we carried out flow cytometry analysis. LPS-treated monocytes, not lymphocytes, showed a change in the granularity, a parameter indicative of the activated state of monocytes (Figure 2B left-up). The analysis of xCT expression was performed by gating CD14⁺ monocytes. Staining cells with xCT antibodies showed a clear shift versus the IgG isotype control, indicating that the staining with this antibody was specific (Figure 2B left-bottom). Similar to the results obtained in the U-937 cell line, there was an increase in xCT expression in CD14⁺ cells after their activation with LPS (n = 5; p = 0.0098; Figure 2B right). Thus, LPS activation also enhances xCT expression and function in peripheral blood monocytes.

Since glutamate levels are reported to be altered in MS disease [12, 13, 28] and monocyte infiltration and microglial activation are MS hallmarks [29], we hypothesized that the expression of system x_c ^- could be altered in MS disease. To test this hypothesis, we analyzed the expression of xCT in the spinal cord of rats after induction of acute EAE, an experimental model of MS which reproduces the inflammatory and immune component of the disease.

Rats immunized with myelin basic protein showed maximal motor deficits at around 14 days postimmunization (Figure 3A). At that stage, expression of mRNA encoding xCT was more abundant in spinal cord samples from rats with EAE than in control animals (n = 6; p = 0.013; Figure 3B left). Accordingly, xCT protein levels were significantly increased in EAE rats (n = 5; p < 0.05; Figure 3B right). We also analysed the expression of xCT in the spinal cord by double immunofluorescence using antibodies to OX-42, a marker of microglia lineage. No immunolabeling was detected after preincubation xCT antibodies with the corresponding peptide (Additional file 3, Figure S2), demonstrating the specificity of the immunofluorescence signal. xCT expression was dramatically increased in the spinal cord of rats with EAE. The expression was particularly increased in meninges in EAE animals, showing clear signs of inflammation (asterisk in Figure 3C top right). There was also a clear increase of xCT in OX42⁺ infiltrating cells, which were organized in clusters around vessels (Figure 3C bottom). In addition, activated microglial cells in EAE rats showed a massive upregulation of xCT (Figure 3D). Activated microglia was identified by their characteristic ameboid shape in contrast with the ramified appearance of resting microglia in control animals (arrows in Figure 3D). Increases in xCT expression at the level of RNA and protein were also detected in the chronic EAE mice (n = 5 for RNA and protein; p < 0.05; Figure 3E), a model that reproduces the chronic and progressive phase of MS.

We next checked whether system x_c ^- expression is similarly altered in MS disease. We analyzed the expression of xCT in peripheral blood and in postmortem optic nerve samples from MS patients by quantitative PCR. xCT mRNA levels showed a 2.5 average fold increase in isolated leukocytes from MS patients (n = 52) compared to controls (n = 25; p = 0.003; Figure 4A). In turn, the increase in xCT levels was higher in R-MS patients having a clinical relapse (n = 24; p = 0.016; Figure 4A) than during remission (n = 28; p = 0.055; Figure 4A). Elevated xCT expression was also observed in CNS samples. Thus, the levels of xCT mRNA in optic nerve (ON) from MS patients showed a two fold increase compared to matched controls (n = 16; p = 0.027; Figure 4B). The increase in xCT mRNA expression was significantly higher in damaged optic nerves showing macroscopic plaques, atrophy and/or optic neuritis. Moreover, levels of xCT mRNA showed a high correlation with CD8 mRNA (Pearson r = 0.83; n = 16; p = 0.0013, Figure 1C), a marker of infiltration and lymphocyte activity, suggesting that lesion activity correlates with xCT expression. Interestingly, the levels of xCT expression in the ON of MS patients correlated positively with the levels of EAAT2 reported previously [21] (Pearson r = 0.81; n = 10; p = 0.005; Figure 1D). Because an enhanced xCT expression is associated to an increase in extracellular glutamate in monocytes in vitro, the increase in glutamate transporter expression could be an adaptive response to counteract the excess of glutamate driven by xCT in MS.

Finally, we performed double immunofluorescence analysis of xCT expression in the spinal cord of MS patients and controls. Overall, we observed a higher expression of xCT in the CNS of MS patients than in controls. In particular, infiltrating CD68⁺ macrophages revealed higher xCT expression in MS patients than in controls (Figure 5). Thus, CD68⁺ cells located in blood vessels (Figure 5A) or nearby, and usually forming clusters, showed a clear xCT overexpression in MS patients (Figure 5B). These results are consistent with the increase of xCT expression in EAE animals, and suggest that altered xCT expression may have a pathophysiological role in MS disease.