Exacerbation of autoimmune neuroinflammation by dietary sodium is genetically controlled and sex specific

Abstract

Multiple sclerosis (MS) is a debilitating autoimmune neuroinflammatory disease influenced by genetics and the environment. MS incidence in female subjects has approximately tripled in the last century, suggesting a sex-specific environmental influence. Recent animal and human studies have implicated dietary sodium as a risk factor in MS, whereby high sodium augmented the generation of T helper (T_h) 17 cells and exacerbated experimental autoimmune encephalomyelitis (EAE), the principal model of MS. However, whether dietary sodium interacts with sex or genetics remains unknown. Here, we show that high dietary sodium exacerbates EAE in a strain- and sex-specific fashion. In C57BL6/J mice, exposure to a high-salt diet exacerbated disease in both sexes, while in SJL/JCrHsd mice, it did so only in females. In further support of a genetic component, we found that sodium failed to modify EAE course in C57BL6/J mice carrying a 129/Sv-derived interval on chromosome 17. Furthermore, we found that the high-sodium diet did not augment T_h17 or T_h1 responses, but it did result in increased blood–brain barrier permeability and brain pathology. Our results demonstrate that the effects of dietary sodium on autoimmune neuroinflammation are sex specific, genetically controlled, and CNS mediated.—Krementsov, D. N., Case, L. K., Hickey, W. F., Teuscher, C. Exacerbation of autoimmune neuroinflammation by dietary sodium is genetically controlled and sex specific.

Multiple sclerosis (MS) is a multifactorial neuroinflammatory disease that represents the most common neurologic disorder affecting young adults. This disease is thought to be initiated by T helper (T_h) 1 and T_h17 cells reactive to self-antigens expressed in the CNS, which become aberrantly activated, traffic to the CNS, and initiate an inflammatory cascade through the secretion of proinflammatory cytokines and recruitment of other immune cells. This immune attack results in focal lesions of demyelination, astrogliosis, and neuronal damage, over time causing the neurologic deficits that are the hallmark of MS.

Classic genetic linkage and genome-wide association studies have identified a clear-cut yet modest genetic component to MS, whereby the major histocompatibility complex (MHC) class II locus, together with other minor loci, govern susceptibility to disease (1). A large portion of MS risk is therefore attributed to environmental influences, or gene–environment interactions. This notion is further supported by the observation that MS incidence has roughly tripled in female subjects, but not male subjects, over the past century, suggesting the existence of an environmental risk factor influencing MS that has appeared within this time frame (2). Several putative risk factors have been identified in epidemiologic studies, including low exposure to sunlight, low vitamin D₃ levels, smoking, and exposure to Epstein-Barr virus (3).

Another putative recently identified risk factor for MS is dietary sodium, a factor that has long been linked to other pathologies, such as cardiovascular disease. Two studies have shown that addition of a modest amount of sodium to culture media in vitro can enhance generation of mouse and human T_h17 cells and that exposure of mice to a high-salt diet results in exacerbated experimental autoimmune encephalomyelitis (EAE) and enhanced generation of effector T_h17 cells (4, 5). Importantly, a subsequent epidemiologic study demonstrated that increased dietary sodium intake (as predicted by urinary sodium levels) was associated with an exacerbated progression of MS symptoms and MRI activity (6). Many important questions remain. Although the effect of sodium in vitro is clear, it is not evident how exposure to excess dietary sodium influences immune cells in peripheral lymphoid organs and/or in the CNS. Moreover, MS epidemiology suggests that environmental risk factors act in sex-specific and genetically controlled fashion, and whether this is the case for dietary sodium remains unknown.

Here, we show that exposure to high dietary sodium exacerbates EAE in 2 genetically divergent strains of mice in a sex-dependent fashion. Moreover, in our model, this phenomenon occurs in the absence of augmented T_h1/T_h17 responses, but correlates with enhanced breakdown of the blood-brain barrier (BBB) and more severe brain pathology. These findings support dietary sodium as an MS risk factor but suggest that the underlying mechanisms and potential contribution to the etiopathogenesis of disease are more complex than previously appreciated.

MATERIALS AND METHODS

Mice

Wild-type C57BL6/J mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA), and SJL/JCrHsd mice were obtained from Harlan Industries (Indianapolis, IN, USA). Mice were rested at the animal facility at the University of Vermont for at least 2 wk before any experimentation. Lysm-Cre mice (B6.129P2-Lyz2tm1(cre)Ifo/J) (7), Cd11c-Cre mice (B6.Cg-Tg(Itgax-cre)1-1Reiz/J) (8), and p38α floxed mice (Mapk14^tm1.2Otsu) (9) have been described previously and were obtained from The Jackson Laboratory or RIKEN BioResource Center (Tsukuba, Japan). Lck-Cre mice (B6.Cg-Tg(Lck-cre)1^Cwi) (10) were obtained from Taconic Farms (Germantown, NY, USA). All conditional knockout (CKO) mice were bred and housed in the vivarium at University of Vermont for 2–3 yr. The experimental procedures used in this study were approved by the Animal Care and Use Committee of the University of Vermont.

Special diets

Dietary regimens were based on previously published experiments (4, 11). Starting at 6 wk of age, mice were randomly assigned to either a low-sodium diet (catalog E15430-24; Sniff, Soest, Germany) and regular tap water, or a high-sodium (4% NaCl) diet (catalog E15431-34; Ssniff) with 1% NaCl drinking water. The 2 diets were of identical composition, with the exception of NaCl content. Mice were maintained on these diets for 3 to 4 wk before EAE induction and were continued on these diets until the end of the experiment.

EAE induction and scoring

EAE was induced essentially as described previously (12). In brief, mice were injected subcutaneously with an emulsion containing 100 μg MOG_35-55 peptide (MEVGWYRSPFSRVVHLYRNGK; New England Peptide, Gardner, MA, USA) (B6 mice) or 100 μg PLP_135-151 (SJL mice) emulsified in complete Freund adjuvant (CFA) (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 200 μg Mycobacterium tuberculosis H37Ra (Difco Laboratories, Detroit, MI, USA) in the posterior right and left flanks. One week later, all mice received an identical injection at 2 sites on the right and left flank anterior of the initial injection sites (2 × MOG_35-55/CFA or 2 × PLP_135-151/CFA). Mice were scored daily starting at day 10 after injection, as previously described (12). Briefly, the clinical scores were as follows: 1, floppy tail; 2, floppy tail and weakened hind limbs; 3, hind limb paralysis; 4, hind limb paralysis and incontinence; 5, quadriplegia or death. EAE scoring was not performed in a blinded fashion as a result of the necessity of labeling the cages with the appropriate diet for the vivarium staff. Weights were monitored at the indicated time intervals and expressed as a percentage of starting weight (at day 0) for each mouse.

Cytokine production quantification

For the detection of cytokines in the cell culture supernatants, ELISAs were performed as described previously (12), using the primary capture mAbs: anti-IFN-γ, anti-IL-17A, anti-TNF-α, and anti-IL-6 and their corresponding biotinylated detection mAbs (Biolegend, San Diego, CA, USA). Other ELISA reagents included horseradish peroxidase–conjugated avidin D (Vector Laboratories, Burlingame, CA, USA), 3,3′,5,5′-tetramethylbenzidine (TMB) microwell peroxidase substrate, and stop solution (Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA). Recombinant mouse IFN-γ, IL-17A, and granulocyte macrophage colony-stimulating factor (GM-CSF; Biolegend) were used as standards.

For analysis of antigen-specific cytokine production by lymphocytes from mice immunized with 2 × MOG_35-55/CFA or 2 × PLP_135-151/CFA, spleen, draining (axillary, brachial, and inguinal) lymph nodes, and mesenteric lymph nodes were collected on day 10 after immunization, single-cell suspensions were prepared at 2 × 10⁶ cells/ml in RPMI 1640 medium with 5% fetal bovine serum, and stimulated with 50 μg/ml MOG_35-55 or 10 μg/ml PLP_135-151. Cell culture supernatants were collected at 72 h, and cytokine levels were measured by ELISA as described above.

Flow cytometry

For intracellular cytokine staining ex vivo, mice were immunized with 2 × MOG_35-55/CFA or 2 × PLP_135-151/CFA; spleen, draining lymph nodes, and mesenteric lymph nodes were collected on day 10 after immunization; and cells were stimulated with 5 ng/ml phorbol 12-myristate 13-acetate (PMA), 250 ng/ml ionomycin (Sigma-Aldrich), and Golgi Plug reagent (BD Biosciences, San Jose, CA, USA) for 4 h. Cells were then stained with the UV-Blue Live/Dead fixable stain (Invitrogen, Carlsbad, CA, USA) and then surface stained for the following markers: CD11b, CD4, CD8, T-cell receptor (TCR)γδ, and TCRβ. Cells were then fixed with 1% paraformaldehyde (Sigma-Aldrich), permeabilized with buffer containing 0.2% saponin, and stained with anti-IL-17A, anti-IFN-γ, and anti-GM-CSF (Biolegend).

For surface marker analysis and forkhead box P3 (Foxp3) staining, unstimulated isolated cells were stained directly ex vivo with the UV-Blue Live/Dead fixable stain and then surface labeled for different combinations of the markers CD25, CD11b, CD11c, MHCII, CD4, CD8, and TCRβ (Biolegend), then fixed using the Foxp3 fixation/permeabilization buffer (eBioscience, San Diego, CA, USA), followed by intracellular staining for Foxp3. All antibodies used for flow cytometry were directly conjugated to fluorophores and obtained commercially (Biolegend; BD Biosciences; eBioscience).

Cells were analyzed with an LSR II cytometer (BD Biosciences). Compensation was calculated using appropriate single-color controls. Data were analyzed by FlowJo software (Tree Star, Inc., Ashland, OR, USA).

CNS-infiltrating mononuclear cell isolation

Animals were perfused with PBS, and brains and spinal cords were removed. A single cell suspension was obtained and passed through a 70 µm strainer. Mononuclear cells were obtained by Percoll gradient (37%/70%) centrifugation and collected from the interphase. For intracellular cytokine analysis, cells were washed and stimulated with 50 μg/ml of MOG_35-55 for 4 h in the presence of brefeldin A (Golgi Plug reagent, BD Biosciences). Cells were labeled with the UV-Blue Live/Dead fixable stain (Invitrogen) followed by surface staining (CD11b, CD11c, MHCII, CD4, CD8, TCRγδ, and TCRβ). Afterward, cells were fixed, permeabilized, and stained for intracellular IL-17A, IFN-γ, and GM-CSF as described above.

Genome-wide single nucleotide polymorphism genotyping

Genetic background analysis was performed commercially by DartMouse (Lebanon, New Hampshire, USA). In brief, the Illumina GoldenGate Genotyping Assay was used to examine single nucleotide polymorphisms (SNPs) throughout the mouse genome. Each experimental sample was interrogated for over 1400 SNPs spaced throughout the mouse genome with an average density of 2 cM.

Bioinformatic analysis of genetic control of differential gene expression

The ImmGen database (http://immgen.org/) was used to identify genes differentially expressed between the sodium-unresponsive strain (129Sv) and the 2 sodium-responsive strains (B6 and SJL). The genetic variation feature of the online data browser was used utilizing previously published data sets (13), which profiled gene expression in CD4 T cells and neutrophils in 40 different inbred strains of mice. Using the strain comparison feature, we compared the strains described above (averaging expression in B6 and SJL to compare against 129Sv), using the default settings to generate lists of differentially expressed genes and associated relative fold change in expression. These lists were further filtered to include genes that showed a 3-fold or greater change in expression (up or down), then filtered by chromosomal location for those genes that fell within the coordinates of the 129Sv-derived locus on chromosome 17 (3000–50,000 kb).

Determination of electrolyte content in serum

Sera were collected at day 21 after induction of EAE. Concentrations of Na⁺, Cl⁻, and K⁺ were determined at the department of Clinical Chemistry at the University of Vermont Medical Center.

BBB permeability measurements

On day 21 after induction of EAE, 50 µg/g body weight of FITC-conjugated bovine serum albumin (BSA; Sigma-Aldrich) was injected i.v. in a volume of 200 µl PBS. Four hours later, plasma and cerebrospinal fluid (CSF) were collected. For collection of CSF, mice were terminally anesthetized with ketamine, and 2 to 10 µl CSF was collected from the cisterna magna as previously described (14). Plasma and CSF were diluted into 100 µl PBS, and FITC fluorescence was measured with a microplate fluorescence reader (BioTek Instruments, Winooski, VT, USA). The BBB permeability index was calculated by dividing the total fluorescence of the CSF by the total fluorescence of the plasma for each animal.

CNS histopathology

On day 21 after induction of EAE, animals were euthanized. The skull and vertebral column were removed and diffusion fixed in 10% formalin. Brain and spinal cord were extracted from calvaria and vertebral columns, respectively, and embedded in paraffin, sectioned (coronal and longitudinal for brain and spinal cord, respectively), and stained with hematoxylin and eosin (H&E). Histopathologic evaluation was performed in a blinded fashion. Representative areas of the brain and the spinal cord, including brain stem, cerebrum, cerebellum, and the cervical, thoracic, and lumbar segments of the spinal cord, were selected for histopathologic evaluation. The type and severity of the EAE lesions were evaluated in each animal and scored according to a semiquantitative scale as previously described (15). The following components of the lesions were assessed: 1) severity of the lesion as represented by each component of the histopathologic assessment; 2) extent and degree of myelin loss (demyelination was evident on H&E-stained tissue by the pallor of the parenchyma in areas of inflammation, the presence of naked axonal segments in the vicinity, and axonal spheroids not covered with a myelin sheath) and tissue injury (swollen axon sheaths, swollen axons, and reactive gliosis); 3) severity of the acute inflammatory response (predominantly neutrophils); and 4) severity of the chronic inflammatory response (lymphocytes/macrophages). A score was assigned separately to the entire brain and spinal cord for each lesion characteristic on the basis of a subjective scale ranging from 0 to 5. Scoring was as follows: 0, no lesions; 1, minimal lesions; 2, mild lesions; 3, moderate lesions; 4, marked lesions; and 5, severe lesions.

Statistical analyses

All statistical analyses were performed by GraphPad Prism 6 software (GraphPad Software, La Jolla, CA, USA). The significance of differences in cytokine production and flow cytometry data was determined by 2-way ANOVA and post hoc comparisons. The significance of differences observed in clinical course of EAE was determined by the Friedman 2-way ANOVA by ranks test (for overall effect of treatment) (16–19). If an overall effect of treatment was detected, post hoc analysis using Fisher’s least significant difference test was used to determine differences at individual time points. EAE data from replicate experiments were analyzed by heterogeneity testing. No significant experiment-to-experiment variation was observed, and therefore the data were pooled accordingly.

RESULTS

Exposure to high dietary sodium exacerbates chronic and relapsing EAE

In order to confirm the effect of dietary sodium on EAE, we utilized 2 different genetic models of disease: chronic EAE in B6 mice and relapsing-remitting EAE in SJL mice. Six-wk-old male and female mice of both strains were exposed to either a control low-sodium diet or a high-salt diet for 3 to 4 wk, followed by induction of EAE using the 2 × MOG_35-55/CFA (B6) or PLP_135-151/CFA (SJL) immunization protocols. These EAE induction protocols avoid the use of pertussis toxin (PTX) as an ancillary adjuvant, a particularly important point because we and others have demonstrated that PTX can override many genetic checkpoints in EAE (20–22). Moreover, this protocol results in less severe disease compared to the conventional protocol including PTX, and thus is more suited to studying the effects of MS environmental risk factors on EAE susceptibility and severity, particularly those that exacerbate disease.

Immunized B6 mice exposed to the control diet developed typical chronic EAE accompanied by modest weight loss, which was more severe in males than in females (Fig. 1), consistent with our previous results using this immunization protocol (23). Exposure to the high-salt diet resulted in exacerbated clinical signs and significantly higher weight loss in both males and females, which was particularly pronounced at the peak of acute disease (Fig. 1).

Figure 1.

Dietary sodium exacerbates EAE in B6 mice. EAE was induced in male (A and C) and female (B and D) B6 mice using 2 × MOG_35-55/CFA. Mean daily clinical scores (A and B) and body weight change (C and D) are shown. Starting body weights on day 0 were set as 100% for each mouse. Data represent 2 independent experiments, pooled. Data were analyzed by 2-way ANOVA. A significant effect of diet on EAE clinical course and weight change was observed in females and males (P < 0.0001). For individual time points, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. Sample sizes were as follows: for EAE scores, n = 16; for body weight, n = 8.

Immunized SJL mice exposed to the control diet exhibited an acute episode of clinical signs, followed by remission and a relapse (Fig. 2). The acute episode was accompanied by pronounced weight loss. Unlike what was seen in B6 mice, exposure to high dietary sodium did not affect EAE or weight loss in male SJL mice (Fig. 2A, C). However, both parameters were exacerbated in females, particularly during the acute peak of disease (Fig. 2B, D). Of note, we monitored weights in B6 and SJL mice during the sodium-loading period before EAE induction (data not shown), and we did not observe any effects of sodium that correlated with weight loss after EAE induction, suggesting that the differences in the latter were due to differences in EAE severity. Taken together, these results demonstrate the existence of a sexually dimorphic EAE response to dietary sodium, which is genetically controlled.

Figure 2.

Sex-specific exacerbation of EAE by dietary sodium in SJL mice. EAE was induced in male (A and C) and female (B and D) SJL mice using 2 × PLP_135-151/CFA. Mean daily clinical scores (A and B) and body weight change (C and D) are shown. Starting body weights on day 0 were set as 100% for each mouse. Data represent 2 independent experiments, pooled. Data were analyzed as in Fig. 1. A significant effect of diet on EAE clinical course was observed in females (P < 0.0001) but not in males (P = 0.19). A significant effect of diet on body weight was observed in females (P < 0.0001) and in males (P = 0.004), although the latter did not show any significant differences at individual time points. For individual time points, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. Sample sizes were n = 16.

Exacerbation of EAE by dietary sodium is genetically controlled

Augmented induction of T_h17 cells by high sodium has been shown to be dependent on a signaling pathway involving serum glucocorticoid kinase 1 and p38α MAPK in T cells, although the in vivo cell-specific requirement for p38α for EAE exacerbation by dietary sodium has not been shown (4). Moreover, using cell type-specific genetic ablation of p38α, we have previously shown that p38α plays cell-specific roles in EAE independent of diet (24). Hence, we used the same approach to delineate the cell type–specific contribution of p38 MAPK signaling to sodium-mediated exacerbation of EAE. Several cohorts of mice homozygous for the floxed Mapk14 allele (p38α^fl/fl) and either positive for Cre (CKO), or Cre-negative littermates (wild-type controls) were subjected to the high-sodium diet paradigm or control diet, as above. Unexpectedly, we found that the high-sodium diet completely failed to exacerbate EAE in either male or female wild-type controls (Fig. 3A, B), thus making it impossible to determine the contribution of p38α to this phenotype.

Figure 3.

Dietary sodium does not influence EAE in B6 p38α^fl/fl mice. EAE was induced in male (A) and female (B) B6 p38α^fl/fl Cre-negative mice using 2 × MOG_35-55/CFA. Mean daily clinical scores are shown. Data represent 3 independent experiments, pooled. Data were analyzed as in Fig. 1. No significant effect of diet on EAE clinical course was observed in females (P = 0.1) and males (P = 0.7). Sample sizes were as follows: male control, n = 18; male high Na⁺, n = 16; female control, n = 14; female high Na⁺, n = 12. (C) Genome-wide SNP genotyping of p38α^fl/fl was carried out as described in Materials and Methods.

Next, we sought to understand why sodium failed to exacerbate EAE in the wild-type controls. Importantly, the conditionally targeted Mapk14 allele was generated using 129Sv embryonic stem cells and subsequently backcrossed onto the B6 background (9). However, Mapk14 lies on the proximal end of chromosome 17, a region that has been associated with suppression of recombination under certain settings (25, 26). We performed genome-wide SNP genotyping, which revealed that approximately half of chromosome 17 (flanking the Mapk14 locus) in the p38α CKO mice was indeed derived from 129Sv (Fig. 3C), suggesting that this locus renders B6 p38α^fl/fl mice insensitive to the effects of sodium. Importantly, both B6 and 129 mice possess the H2^b haplotype at the H2 locus, which is also located within this region (27), suggesting that polymorphism at H2 is not responsible for this phenotype. Using mouse genome informatics (MGI), we undertook an analysis of this locus, searching for polymorphisms between B6 and 129Sv that are predicted to result in protein coding sequence changes and thus would be predicted to affect protein function. Several interesting candidate genes with known functions in sodium homeostasis emerged: Slc9a3r2 (solute carrier family 9 [sodium/hydrogen exchanger], member 3 regulator 2), Slc22a7 (solute carrier family 22 [organic anion transporter], member 7), and Sik1 (salt-inducible kinase 1). Furthermore, using the ImmGen database (13), we set out to find genes located within the chromosome 17 locus that were differentially expressed between the sodium-responsive strains (B6 and SJL) strains and the sodium-unresponsive strain (129Sv). This analysis identified 83 differentially expressed transcripts, of which only 3 genes were within the chromosome 17 locus—Zfp760, Dnahc8, and, remarkably, Slc9a3r2—suggesting that the latter gene also exhibits an expression level polymorphism in addition to the change in its coding sequence. Taken together, these findings further support the concept that genetic control of responsiveness to dietary sodium influences EAE susceptibility, and identify polymorphism or polymorphisms in Slc9a3r2 as a potential candidate gene.

High dietary sodium does not augment encephalitogenic T cell responses

Previous studies suggested that the primary mechanism behind exacerbation of EAE by dietary sodium is through enhancement of T_h17 differentiation, as NaCl treatment of T_h17 cells in vitro stimulated production of IL-17, and a high-sodium diet was associated with augmented T_h17 responses in vivo during EAE (4, 5). Although the in vitro effects of sodium on T cells are clear, it is less evident how high-salt in the diet can augment the generation of encephalitogenic T cells in lymphoid tissues distal to the intestine, i.e., the lymph nodes draining the immunization sites, where this response is primarily generated in active EAE. To address this question, we measured the effects of a high-sodium diet on CNS antigen-specific responses in B6 and SJL mice at day 10 after EAE induction in peripheral lymphoid organs (spleen and draining lymph nodes) using ex vivo peptide restimulation and ELISA. Surprisingly, we did not observe significant sodium-dependent upregulation of either IL-17, or IFN-γ and GM-CSF, the 2 other signature cytokines of encephalitogenic T cells; in fact, in some cases production of these cytokines was decreased by the high-salt diet (Figs. 4 and 5).

Figure 4.

Dietary sodium does not augment peripheral encephalitogenic T-cell recall responses in B6 mice. B6 mice were immunized using 2 × MOG_35-55/CFA. On day 10 after immunization, cells were isolated from the draining lymph nodes (A–C) and spleen (D–F) and then restimulated with 50 µg/ml MOG_35-55 for 72 h. Supernatants were assayed for the production of the indicated cytokines by ELISA. Data were analyzed by 2-way ANOVA and post hoc comparisons; n = 8 for each group.

Figure 5.

Dietary sodium does not augment peripheral encephalitogenic T-cell recall responses in SJL mice. SJL mice were immunized using 2 × PLP_135-151/CFA. On day 10 after immunization, cells were isolated from the draining lymph nodes (A–C) and spleen (D–F), then restimulated with 10 µg/ml PLP_135-151 for 72 h. Supernatants were assayed for the production of the indicated cytokines by ELISA. Data were analyzed by 2-way ANOVA and post hoc comparisons; n = 8 for each group.

We also examined the effect of dietary sodium on the generation of a T_h1/T_h17 polyclonal response during EAE at the single cell level, using direct ex vivo restimulation of T cells with PMA/ionomycin, followed by intracellular cytokine staining and flow cytometry. Using this approach, we also failed to see sodium-induced augmented production of the cytokines associated with T_h1/T_h17 cells in either draining lymph nodes or spleens of either SJL or B6 mice (Fig. 6A, B, D, E). We also failed to see a significant effect of diet on the numbers of Foxp3⁺ T-regulatory (T_reg) cells, MHC class II expression by antigen-presenting cells, or the overall composition of major immune cell subtypes in lymphoid organs (Fig. 7; data not shown). Because the organ that initially encounters high sodium from the diet is the intestine, we examined whether the diet had any effects in mesenteric lymph nodes. Contrary to published results (5), we again did not find any significant effects of dietary sodium on the expression of IL-17, IFN-γ, GM-CSF (Fig. 6C, F), or Foxp3 (Fig. 7).

Figure 6.

Dietary sodium does not augment generation of peripheral encephalitogenic T-cell responses. Male (M) and female (F) B6 mice (A–C) were immunized using 2 × MOG_35-55/CFA. SJL mice (D–F) were immunized using 2 × PLP_135-151/CFA. On day 10 after immunization, cells were isolated from the draining lymph nodes (A and D), spleen (B and E), or mesenteric lymph nodes (C and F), then stimulated for 4 h ex vivo with PMA and ionomycin in the presence of brefeldin A. Cells were then surface stained and fixed, followed by permeabilization, staining for cytokines, and flow cytometry analysis. Cells were gated on live TCRβ⁺ and CD4⁺ cells, and the percentages of cells positive for IL-17, IFN-γ, and GM-CSF were calculated. Data were analyzed by 2-way ANOVA and post hoc comparisons for effect of diet; n = 8 for each group. Representative flow cytometry plots of B6 male spleen cells are shown (G).

Figure 7.

Dietary sodium does influence generation of T_reg cells. Male (M) and female (F) B6 mice (A) were immunized using 2 × MOG_35-55/CFA. SJL mice (B) were immunized using 2 × PLP_135-151/CFA. On day 10, postimmunization cells were isolated for the indicated organs (dLN, draining lymph nodes; SPL, spleen; mLN, mesenteric lymph node), surface stained, fixed, and permeabilized, followed by intracellular staining for Foxp3 and flow cytometry analysis. Cells were gated on live TCRβ⁺ and CD4⁺ cells, and percentages of cells positive for Foxp3 were calculated. Representative flow cytometry plots of B6 male spleen cells are shown (C). Data were analyzed by 2-way ANOVA and post hoc comparisons for effect of diet; n = 8 for each group.

Last, we examined whether dietary sodium augmented the presence of encephalitogenic T cells in the CNS during the peak of disease in female B6 mice (the sex/strain combination that exhibited the most pronounced effect of sodium on clinical disease). No significant differences were observed in the total numbers of infiltrating cells isolated from the CNS (Fig. 8A). Moreover, no significant difference was observed in percentages of cells positive for IL-17, IFN-γ, and GM-CSF (Fig. 8B). Taken together, these results show that dietary sodium can exacerbate EAE in the absence of augmented T_h1 or T_h17 generation.

Figure 8.

Dietary sodium does not augment encephalitogenic T-cell responses in the CNS. Female B6 mice were immunized using 2 × MOG_35-55/CFA. On day 21 after immunization, cells were isolated from the draining lymph nodes (dLN) or the CNS using a Percoll gradient (as described in Materials and Methods) and counted (A). CNS cells were stimulated for 4 h ex vivo with 50 µg/ml MOG_35-55 in the presence of brefeldin A, then surface stained and fixed, followed by permeabilization, staining for cytokines, and flow cytometry analysis (B). Cells were gated on live TCRβ⁺ and CD4⁺ cells, and the percentages of cells positive for IL-17, IFN-γ, and GM-CSF were calculated. Data were analyzed by 2-way ANOVA and post hoc comparisons for effect of diet; n = 10 for each group. Representative flow cytometry plots are shown (C).

High dietary sodium causes elevated serum sodium, increased BBB permeability, and more severe brain pathology

These results led us to speculate that exacerbation of EAE by the high-salt diet in our model is not dependent on direct effects of sodium on immune cells. It has been shown that a high-sodium diet can result in accumulation of sodium in peripheral tissues, such as the skin (11). Thus, we tested whether the high-sodium diet affected serum sodium levels, and we found that indeed the serum sodium levels were elevated (Fig. 9A). Serum chloride levels were also significantly elevated, as expected, while potassium levels were unchanged (Fig. 9B, C). One of the hallmarks of MS and EAE is the loss of BBB integrity (28). Intriguingly, initial BBB breakdown and immune cell entry is thought to be initiated at the sensory circumventricular organs (CVO) (29, 30). The sensory CVO are unique organs in the brain that carry out many functional roles, with a major one being the sensing and regulation of electrolyte content and osmolarity of plasma (31, 32). Thus, we hypothesized that elevated systemic sodium may be influencing CVO early during EAE, and in doing so, it may contribute to enhanced BBB permeability. To test this hypothesis, we measured BBB permeability during peak EAE using intravenous injection of FITC-labeled BSA and subsequent monitoring of its leakage across the BBB into the CSF. We found that high dietary sodium resulted in significantly greater BBB permeability (Fig. 9D).

Figure 9.

High dietary sodium enhances BBB breakdown. Female B6 mice were immunized using 2 × MOG_35-55/CFA. On day 21 after immunization, concentration of serum sodium (A), chloride (B), and potassium (C) was determined. (D) FITC-BSA was injected intravenously, followed by collection of CSF and plasma 4 h later. The BBB index was calculated by dividing the total FITC fluorescence in CSF by that of plasma. Data were analyzed by 2-tailed Student’s t test; n = 10.

On the basis of these data, we hypothesized that increased clinical signs of EAE induced by dietary sodium were a result of enhanced BBB breakdown, dysregulated sodium homeostasis, and subsequent CNS pathology. To further test this hypothesis, we examined histopathologic parameters at peak of clinical signs of EAE. We found that dietary sodium did not affect pathology in the spinal cord (Fig. 10A). However, in the brain, pathology was more severe in mice exposed to the high-sodium diet (Fig. 10B). Signs of axonal damage in the brain were also detected more frequently in the high-sodium group (3 of 10 mice) compared to the control group (1 of 10), although this did not reach statistical significance. Taken together, these results demonstrate that high dietary sodium results in more severe EAE pathology in the brain, likely as a result of enhanced BBB breakdown and dysregulated sodium homeostasis.

Figure 10.

High dietary sodium results in more severe brain pathology. Female B6 mice were immunized using 2 × MOG_35-55/CFA. On day 21, animals were euthanized, and spinal cords (A) and brains (B) were removed, fixed in formalin, extracted, and sectioned for H&E staining. Histopathologic evaluation was performed as described in Materials and Methods. Data were analyzed by 2-way ANOVA and post hoc comparisons; n = 10.

DISCUSSION

Since the publication of the 2 EAE studies reporting a potential association between sodium and autoimmunity, an epidemiologic study has indeed found a positive correlation between increased dietary sodium intake and exacerbation of disease course in MS (6). Overall, our findings support the potential role of dietary sodium in MS progression and shed new light on the mechanisms behind this phenomenon. We find that this response is 1) genetically controlled, 2) biased to female sex, and 3) can be independent of the effects of sodium on immune cells. This has important implications for MS. The strikingly increasing incidence of MS in women is suggestive of an environmental factor that has changed rapidly in the last 50 to 80 yr. Sodium intake has potentially increased within the same time frame as a result of increased consumption of processed foods (33–35), making this a plausible contributor to increasing incidence of MS in women. Importantly, the increase in the incidence of MS in females associated with environmental risk factors over the past century is primarily a function of relapsing–remitting disease, which is best modeled in SJL mice—the strain in which we find that sodium exacerbates disease in a female-specific fashion. Furthermore, the finding that this phenotype is genetically controlled suggests that the response to this new environmental stimulus may vary quite significantly between genetically diverse individuals and across populations. The findings that the EAE response to dietary sodium is genetically controlled are also in line with the well-documented genetic control of salt sensitivity of blood pressure in humans and in rat models (36, 37). Whether the 2 phenomena share a similar genetic architecture remains to be investigated.

Farez and colleagues reported that sodium intake was strongly correlated with MS progression in 2 small patient cohorts (6). Although their cohorts included both males and females, and although the model was adjusted for the effects of sex on disease progression, both sexes were analyzed together for the effect of sodium. Hence, it is unclear whether there is any sex bias in sodium’s influence on MS, but based on our results, this may be dependent on the individual’s genetic background and disease course. Interestingly, Farez and colleagues also reported a male bias toward higher sodium intake (6), which has been reported previously (38). This may help explain in part why males tend to exhibit more severe and rapid progression of MS.

Two recent reports have shown that the addition of a modest amount (40 mM) of NaCl to cell culture media can profoundly enhance the differentiation of mouse T_h17 cells (4, 5). These studies then extrapolated these results to link high dietary sodium intake to exacerbated EAE symptoms and augmented generation of T_h17 cells in vivo. Kuchroo and colleagues (5) reported that a high-salt diet augmented the numbers of T_h17 cells in the lamina propria of naive mice and in the mesenteric lymph nodes and CNS, but not the spleen, of MOG_35-55-immunized B6 mice. Similarly, Hafler and colleagues (4) reported increased IL-17 expression in the CNS and the spleen of MOG_35-55-immunized B6 mice. In contrast, in our model, we failed to see a positive effect of dietary sodium on T_h17 differentiation in MOG_35-55-immunized B6 or PLP_135-151-immunized SJL mice in the draining lymph nodes or spleen, where the initial encephalitogenic T_h response is generated; in the mesenteric lymph nodes, a major lymphoid organ in the gut; and in the CNS, the target organ in EAE. Nonetheless, in agreement with the previous studies, we found that clinical signs of EAE were exacerbated by the high-sodium diet, suggesting that additional effects aside from those of sodium on T_h17 cells also contribute to clinical disease progression. An important difference between the 2 aforementioned studies and ours is that our experimental design does not use PTX as part of the EAE induction protocol. Although we cannot exclude the notion that dietary sodium can influence CNS autoimmunity via T_h17 cells in the presence of PTX, our results suggest that this is not required in our model of CNS autoimmunity. Moreover, it is unclear how dietary sodium could influence T_h17 generation. We did observe a modest increase in sodium concentration in serum (∼5 mM), which is lower than the reported concentrations affecting T_h17 generation in vitro. It has been reported that a high-sodium diet can result in accumulation of sodium in tissues distal to the intestine, such as the skin (11). Thus, it is possible that a similar phenomenon occurs in lymphoid organs or even in the CNS, but this possibility has not been directly addressed.

Homeostasis of sodium and other electrolytes in the body is tightly regulated by the renin–angiotensin–aldosterone system. This system is activated by low blood pressure to stimulate increased sodium absorption in the kidney, which in turn increases body fluid volume and blood pressure (39). In this regard, angiotensin signaling has been shown by several groups to have proinflammatory effects in EAE, whereby its blockade suppresses clinical signs of disease and immune cell activation (40–42). However, a high-salt diet typically has the opposite effect, inactivating the renin–angiotensin–aldosterone system to decrease sodium absorption (39), and thus would be predicted to suppress any EAE-promoting functions of this system, which is not consistent with our results. Another important caveat is that the high-salt diet regimen used in our study does not induce elevated blood pressure in B6 mice (4, 43).

What are other potential non-immune-mediated mechanisms by which dietary sodium can exacerbate EAE and MS? The sensory CVO—which include the area postrema, the subfornical organ, and the organum vasculosum of the lamina terminalis—are specialized regions of the brain devoid of a tight BBB and are viewed as the portals of entry for immune cells into the CNS and the sites of initial BBB breakdown in EAE (29, 30, 44). Importantly, these organs are critical for sensing of physiologic levels of sodium in plasma and in response regulating its homeostasis via several complex signaling pathways (31, 32). We observed that the high-salt diet caused elevated system sodium levels, which we hypothesize exerted detrimental effects on the CVO, leading to accelerated BBB breakdown and CNS pathology. Our finding that the brain, rather than the spinal cord, exhibited more severe EAE pathology in response to the high-salt diet further supports this model. Moreover, several recent studies have reported a striking dysregulation of sodium homeostasis in MS lesions in the brain (45, 46). Higher sodium concentrations were observed in MS lesions compared to unaffected areas or control healthy brains, likely reflecting neuroaxonal pathophysiology, and this sodium accumulation correlated with more severe disease progression. Although it is unclear whether dietary sodium can influence sodium retention in the brain, we speculate that it is possible that in our model elevated systemic sodium levels and that a weakened BBB could lead to sodium influx into the CNS and subsequent retention of sodium in the brain, aggravating ongoing neuronal pathophysiology. Last, we also note that we cannot rule out the possibility that enhanced BBB breakdown and brain pathology in mice exposed to a high-sodium diet is the result of enhanced inflammation in the brain, rather than the other way around.

Another potentially important variable is the influence of the high-sodium diet on the gut microbiome and vice versa. It is likely that the high-sodium diet profoundly changes the gut microbiome, which is well known to influence EAE progression (47). However, these effects appear to be mostly immune mediated (e.g., modulation of T_reg numbers), and we failed to observe any immune phenotypes associated with EAE in correlation with exposure to high sodium. We cannot, however, rule out any effects of the microbiome on CNS function and development, which are also well documented (48). Moreover, we note that some of the mouse strains used in our experiments that exhibited differential responses to sodium are from different vendors or colonies and thus are likely harbor divergent microbiomes, which could in turn influence their response to sodium. Further studies are needed to address these possibilities.

It is important to note that the EAE exacerbation due to increased dietary sodium (Figs. 1 and 2) was relatively modest, although clear cut. Moreover, the high-sodium regimen used in our study uses supraphysiologic levels compared to sodium intake levels in humans, although the exposure time (loading) is relatively short (3 wk), to try to approximate a lifelong consumption of sodium in a human. Overall, this suggests it is likely that variation in dietary sodium intake in humans would have a relatively modest effect on MS severity or incidence, and that sodium is just one of the many environmental factors influencing this complex disease.

Glossary

B6

C57BL6/J mice

BBB

blood–brain barrier

BSA

bovine serum albumin

CFA

complete Freund adjuvant

CKO

conditional knockout

CSF

cerebrospinal fluid

CVO

circumventricular organ

EAE

experimental autoimmune encephalomyelitis

Foxp3

forkhead box P3

GM-CSF

granulocyte macrophage colony-stimulating factor

H&E

hematoxylin and eosin

MHC

major histocompatibility complex

MS

multiple sclerosis

PTX

pertussis toxin

SJL

SJL/JCrHsd mice

SNP

single nucleotide polymorphism

TCR

T-cell receptor

T_h

T helper

T_reg

T-regulatory