Increased salt exposure affects both lymphoid and myeloid effector functions, influencing innate‐associated disease but not T‐cell‐associated autoimmunity

Summary

High salt consumption has since long been associated with elevated blood pressure and cardiovascular disease. Recently, mouse studies suggested that a high dietary salt intake exacerbates the clinical manifestations of autoimmunity. Using naïve cells ex vivo after pre‐exposure of mice to high salt intake, we showed that increased salt exposure affects the viability and effector functions of immune cells. CD4⁺ T‐cells evidenced a pro‐inflammatory phenotype characterized by increased secretion of IFN γ and IL‐17A, when exposed to high salt concentrations in vitro. Interestingly, this phenotype was associated with osmotic pressure, as replacing salt for d‐mannitol resulted in similar observations. However, high salt intake did not alter the development of T‐cell‐dependent autoimmunity. Instead, recruitment of peritoneal macrophages was increased in mice pre‐exposed to high salt concentrations. These cells had an increased production of both TNF α and IL‐10, suggesting that salt stimulates expansion and differentiation of different subsets of macrophages. Moreover, mice pre‐exposed to high salt intake developed exacerbated symptoms of colitis, when induced by dextran sulphate sodium. The aggravated colitis in salt‐exposed animals was associated with a higher frequency of CD4⁺ T‐cells and CD11b⁺ CD64⁺ macrophages producing TNF α. These phenotypes correlated with elevated titres of faecal IgA and higher lymphocytic cellularity in the colon, mesenteric lymph nodes and spleen. In conclusion, we report here that high salt intake affects both lymphoid and myeloid cells ex vivo. However, the effects of high salt intake in vivo seem less pronounced in terms of CD4⁺ T‐cell responses, whereas macrophage‐dependent pathologies are significantly influenced.

Introduction

Sodium chloride, generally referred to as salt, has been used for centuries as a food preservative. The intake of salt has risen over the years, particularly due to the increased consumption of processed food.1 Although salt contributes to many relevant processes of our cellular biology, a high intake of salt can be deleterious. Elevated salt consumption contributes directly to high blood pressure in various animal species and in humans. Moreover, a high‐salt diet (HSD) has found to be associated with elevated risks of cardiovascular disease, kidney disease and even insulin resistance.1, 2, 3, 4 Furthermore, high dietary salt intake has been associated with a higher risk to develop gastric cancer, through synergic action with Helicobacter pylori infection.5

Autoimmune diseases, such as rheumatoid arthritis (RA), multiple sclerosis (MS) and inflammatory bowel disease (IBD) are complex heterogeneous diseases characterized by chronic inflammation in which the immune response is modified by genetic and environmental factors. Epidemiological studies indicate that the prevalence of autoimmunity has risen in Western countries over the last decades.6 The reason for this increase has been attributed partially to the hygiene hypothesis.7 However, there is also rising evidence that dietary factors contribute to the pathogenesis of several autoimmune diseases.1, 8, 9, 10, 11 In this regard, recent studies in the mouse have shown that animals fed a HSD are likely to develop more severe autoimmune manifestations.12, 13, 14, 15, 16

The connection between salt and the immune system can be noticed by the hyperosmolality of the lymphoid microenvironment.17 It is therefore likely to assume that osmotic changes caused by high‐salt intake will have consequences at the level of immune cell activation and consequently in the building of immune responses. The mammalian adaptive osmotic stress response is based on the activity of the nuclear factor of activated T‐cells 5 (NFAT5), which indirectly increases the intracellular concentrations of osmolytes through genetic regulation.17 The presence of salt, in particular sodium, has been shown to activate distinct proteins and consequently modulate immune responses.12, 13

In the present study, we investigated whether a moderate increase of salt exposure affects the effector functions of naïve lymphoid and myeloid cells. Also, we addressed whether these effects occur due to osmotic pressure or other unrelated mechanisms. Furthermore, we assessed the development of distinct autoimmune mouse models upon exposure to increased salt intake. Our data suggest that exposure to moderate sodium chloride concentrations drives lymphoid and myeloid cells to a more pro‐inflammatory phenotype. Whereas the increase of salt exposure exacerbates the development of acute colitis, it did not alter the development of experimental autoimmune encephalomyelitis (EAE) or collagen‐induced arthritis (CIA), two mouse models of MS and RA, respectively.

Materials and methods

Animals

All animal experiments were conducted with male C57BL/10.Q mice (hereafter referred to as BQ), unless described otherwise. BQ mice were bred in the mouse facility of the division of Medical Inflammation Research (Karolinska Institutet, Stockholm, Sweden) under specific‐pathogen‐free (SPF) conditions and used for experiments at 10–14 weeks of age. C57BL/6J mice (B6) were purchased from The Jackson Laboratories and kept under similar SPF conditions (Boston, MA). HCQ318 T‐cell receptor (TCR) transgenic mice recognizing the galactosylated form of the immunodominant T‐cell epitope of type II collagen19 were used to assess antigen‐specific CD4⁺ T‐cell reactivity. Mice were housed in ventilated cages with soft bedding material and tissue paper as environmental enrichment and stress reducer. Three–five mice were housed together per cage, and fed a standard rodent chow and given water (or salt solution) ad libitum. All the experiments followed the ARRIVE recommendations.20 The local ethics committee approved all animal experiments (Stockholms Norra Djurförsöksetiska Nämnd, Stockholm, Sweden). In vivo and ex vivo experiments were performed under the ethical permits numbers N490/12 and N35/16 for arthritis, N83/13 for encephalomyelitis, and N181/13 for colitis. The colitis experiments have also been conducted and approved by IACUC and COMS of Harvard Medical School (Boston, MA). Anaesthesia of animals was accomplished by isoflurane inhalation, whereas the animals were killed using CO₂.

In vivo effects of salt

For the in vivo salt intake studies, mice were divided into two groups and supplied with normal drinking water or 1% (w/v) NaCl in water for 2–3 weeks prior to disease induction. Water intake was observed weekly in the different groups and calculated as average intake per mouse per day. To determine urine osmolality, urine samples were taken periodically and urine‐specific gravity (USG) was measured using urine dipsticks (Macherey‐Nagel, Düren, Germany).

Three weeks after 1% NaCl exposure, mice were euthanised peritoneal lavage was collected, and spleens were harvested. Peritoneal cells were counted and evaluated in terms of cytokine production, as described below. Splenocytes were stimulated with 50 ng/ml phorbol 12‐myristate 13‐acetate (PMA), 250 ng/ml ionomycin and 10 μg/ml brefeldin A (BFA) for 4 hr at 37° in a 5% CO₂ incubator. IL‐2, IL‐17A and TNFα production were assessed on live CD4⁺ T‐cells using flow cytometry.

Collagen antibody‐induced arthritis (CAIA)

Production of collagen type II (CII)‐specific B‐cell clones has been previously described.21, 22 On day 0, mice were injected i.v. with 4 mg mAb cocktail containing M2139 binding the J1 epitope of CII; CIIC1 binding the C1 epitope; CIIC2 binding the D3 epitope; and the UL1 clone binding the U1 epitope. On day 5, mice received a boost of 25 μg of lipopolysaccharide (LPS) from Escherichia coli, administered i.p. to enhance the severity and incidence of arthritis. Clinical scoring was performed daily based on the number of inflamed joints. Each red and swollen finger or knuckle was scored 1 point, while an inflamed wrist or ankle was scored 5 points, resulting in a maximum of 15 points per limb, and 60 points per animal. The animals were regularly monitored for weight loss. Mice were allowed to move freely, and killed when having a total arthritis score ≥ 50 points.

CIA

Collagen type II was isolated from the rat Swarm chondrosarcoma, prepared by limited pepsin digestion and dissolved at 5 mg/ml in 0·1 m acetic acid. On day 0, mice were immunized intradermally at the base of the tail with 100 μl of an emulsion of 100 μg of CII in complete Freund's adjuvant, containing a final concentration of 25 μg of Mycobacterium smegmatis (CFA; Difco, Detroit, MI). On day 35, mice received a booster dose of 50 μl emulsion containing 50 μg of CII emulsified in incomplete Freund's adjuvant (IFA; Difco). After the first signs of arthritis, mice were blindly scored three times per week as described above, and monitored for weight changes. Mice were killed when they had a total arthritis score ≥ 50 scoring points.

Anti‐CII antibody titres

All animals used for CIA were bled on days 21, 35 and 50 after immunization through the submandibular vein, and the obtained sera were used in ELISA for the quantification of anti‐CII antibody levels. Ninety‐six‐well Maxisorp plates (NUNC, Roskilde, Denmark) were coated with CII in phosphate‐buffered saline (PBS; 10 μg/ml), followed by blockage of non‐specific binding sites with 2% bovine serum albumin (BSA). Serum samples were plated in duplicates at a predetermined dilution together with a standard curve derived from a sample of pooled sera positive for anti‐CII antibodies. Total levels of anti‐CII immunoglobulins (Igs), IgG1 and IgG2b isotypes were determined using horseradish peroxidase (HRP)‐conjugated anti‐kappa monoclonal antibody (clone 181.1; Southern Biotech, Birmingham, AL), goat anti‐mouse IgG1 or IgG2b antibodies (Southern Biotech), respectively. The enzymatic reaction was achieved using ABTS solution as substrate (Roche Diagnostics, Germany), and colorimetric absorbance was measured at 405 nm (Synergy‐2; BioTek Instruments). Anti‐CII antibody titres were plotted as arbitrary units based on the standard curves.

EAE

At day 0, mice were immunized intradermally at the base of the tail with 100 μl of an emulsion containing 50 μg of myelin oligodendrocyte glycoprotein peptide sequence 75–96 (MOG_75–96; Schafer‐N, Copenhagen, Denmark) in CFA. On the same day and 2 days later, mice were given 400 ng of Bordetella pertussis toxin i.p. (Sigma‐Aldrich, St Louis, MO). Mice were monitored daily for weight loss and clinical signs of EAE. Disease scoring was performed daily using a 0–8 scale: 0 = normal; 1 = tail weakness; 2 = tail paralysis, normal gait; 3 = tail paralysis, low back and mild waddle; 4 = tail paralysis, severe waddle, less sure footing; 5 = tail paralysis and paralysis of one limb; 6 = tail paralysis and paralysis of a pair of limbs; 7 = tetraparesis; 8 = premorbid or deceased. Mice were allowed to move freely, and were killed when presenting a score ≥ 7 points.

Dextran sulphate sodium (DSS)‐induced colitis

Prior to colitis induction, mice were provided with normal drinking water or a 1% NaCl water solution for 3 weeks. Fresh faecal samples were collected weekly for faecal IgA quantification. After 3 weeks, acute colitis was induced by administration of 5% (w/v) DSS (40 000 g/mol; AppliChem GmbH, Darmstadt, Germany) in drinking water for 7 days. Drinking water was then replaced with normal or 1% salty water for an additional 7 days. Mice were monitored daily and a disease activity index score was assessed. The activity index consists of a combined score that includes weight loss, stool consistency, and bleeding, with a maximum score of 12 per animal (Table S1;Ref 23). On day 14, mice were killed, and mesenteric lymph nodes and spleens were collected for cellular analysis. For histology of the colons, proximal colon samples were collected from naïve BQ mice or at day 14 after DSS‐colitis induction. Tissues were fixed in 4% (w/v) paraformaldehyde, dehydrated, mounted in paraffin, sectioned longitudinally, and stained with haematoxylin and eosin. Colonic tissue was scored using a scoring system adapted from Erben et al.24 Individual scores were given for the following histological features: loss of lining epithelium; crypt damage; loss of goblet cells; and infiltration of inflammatory cells. The scoring system is presented in detail in Table S2. The histological score corresponds to the sum of the four individual scores, resulting in a maximum score of 12 per section. Each histological score was calculated as the average of three scoring areas of 1 mm². Histological images were acquired through a Zeiss Axioplan HBO50 microscope using an Olympus SC30 camera. cellsens entry 1.8 (Olympus Life Science, Hamburg, Germany) was used as software.

Colonic lymphoid and myeloid cell isolation and analysis

Colons from B6 mice were harvested 10 days after DSS‐colitis induction. Tissues were immediately cleaned from faecal content by smooth flushing of the colon with 5 ml of RPMI containing 10% fetal calf serum (FCS) and penicillin/streptomycin using a syringe and 18G blunt needle. Colons were then cut into small pieces and digested in four rounds of 0·5 mg/ml collagenase type II (Worthington, Biochemical, Lakewood, NJ) and 125 μg/ml of DNase I (Roche) in RPMI containing 5% FCS. The resulting cell suspensions were then centrifuged in 40%–75% Percoll gradient. The resulting cellular interphase was then assayed. Cells were cultured for 5 hr at 37° in a 5% CO₂ incubator in the presence of 10 μg/ml BFA. The frequency of immune cell populations and intracellular cytokine production was assessed by flow cytometry.

IgA quantification

A 100 mg/ml suspension was made by vortexing faecal pellets in PBS containing protease inhibitors (Roche, Basel Switzerland). Samples were centrifuged at 10 000 g for 10 min, and the resulting supernatant was stored at −20° until analysed. ELISA plates coated with 2 μg/ml rat anti‐mouse IgA in PBS (clone RMA‐1; Biolegend, San Diego, CA) were incubated overnight at 4°. Wells were blocked with 2% BSA solution in PBS for 1 hr at room temperature. After washing, samples were added in duplicate at a pre‐determined dilution followed by a 1 : 3 dilution series and incubated for 2 hr at room temperature. A standard curve of purified mouse IgA (500 ng/ml followed by 1 : 3 dilutions; BD Biosciences, San Jose, CA) was used in every plate. Anti‐mouse‐kappa‐HRP‐conjugated (clone 181.1; Southern Biotech, Birmingham, AL) secondary antibody was added to the wells and incubated for 1 hr. ABTS solution (Roche, Switzerland) was used as enzymatic substrate, and absorbance was read at 405 nm.

Lymphocyte cultures

Spleens from naïve BQ mice were minced and filtered through a 40‐μm mesh (BD Biosciences, San Jose, CA). Splenocytes were cultured in Dulbecco's modified Eagle's minimal essential medium (DMEM) supplemented with 5% heat‐inactivated fetal bovine serum, 10 mm HEPES buffer, 50 U/ml penicillin and 50 μg/ml streptomycin (P/S), and 50 μm β‐mercaptoethanol [complete DMEM (cDMEM)] in a sterile 96‐well U‐bottom plate (NUNC) at a concentration of 10⁶ cells/well. Cells were stimulated with 1 μg/ml of anti‐CD3 and anti‐CD28 mAb (anti‐CD3/28; BD Pharmingen), and incubated for the described time length at 37° in a 5% CO₂ incubator. When applicable, NaCl or d‐mannitol (Sigma‐Aldrich, St Louis, MO) was added to the cultures. T‐cell viability and activation were assessed using flow cytometry and ELISA after 48‐ and 72‐hr cultures. IL‐2 secretion was measured in the cell culture supernatant by ELISA at 48 hr, whereas IFN‐γ and IL‐17A levels were determined at 72 hr. Anti‐IL‐2 clones JES6‐IA12 and 5H4‐biotin (homemade), anti‐IFNγ clones R46‐A2 and AN18‐biotin (homemade), and anti‐IL‐17A clones TC11‐15H10.1 and TC11‐8H4‐biotin (Biolegend) were used. The read‐out was performed with streptavidin‐Eu (Delfia, PerkinElmer, Turku Finland). Cytokine secretion was calculated as Eu‐counts per frequency of live CD4⁺ T‐cells. Purified CD4⁺ T‐cell cultures were done as described above after enrichment with untouched CD4 Dynabeads (Invitrogen, Carlsbad, CA).

To assess the effect of salt on antigen‐specific T‐cell activation, 10⁵ purified CD4⁺ T‐cells (Dynabeads; Invitrogen) from HCQ3 (CD45.2⁺) TCR transgenic mice were co‐cultured with 5 × 10⁵ splenocytes (CD45.1⁺), in the presence of 1 μg/ml anti‐CD28 and 10 μg/ml galactopyranosyl‐hydroxylysine CII 260–270 peptide (CII_260–270; Gal‐Hyl at position 264), for 6 hr at 37° and 5% CO₂. When applicable, NaCl or d‐mannitol was added to the cultures. Intracellular CD40L expression on live CD45.2⁺ CD4⁺ T‐cells was assessed using flow cytometry.

Myeloid cell cultures

Peritoneal lavage was obtained by injection of ice‐cold PBS into the peritoneal cavity of naïve mice, followed by gentle massage of the abdomen and aspiration of the resulting fluid. Peritoneal cells were cultured in cDMEM at 5 × 10⁵ cells/well and stimulated with 1 μg/ml of LPS with or without additional NaCl. After 20–24 hr, cell viability was determined and the supernatant was used to assess secretion of TNFα, IL‐6 and IL‐10. Cytokine levels were detected using antibodies from the respective clones MP6‐XT22 and MP6‐XT3‐biotin, MP5‐20F3 and MP5‐32C11‐biotin, and JES5‐2A5 and JES5‐16E3‐biotin (BD). The read‐out was performed using streptavidin‐Eu (Delfia). Cytokine secretion was calculated as Eu‐counts per frequency of live CD11b⁺ peritoneal cells.

FACS antibodies

Flow cytometry analysis was performed in LSR‐II instrument (BD, San Jose, CA). The following antibodies were purchased from Biolegend or BD Biosciences: anti‐TCRβ (H57‐597), ‐CD4 (H129.19), ‐CD8α (53‐6.7), ‐CD44 (IM7), ‐CD45 (30‐F11), ‐CD62L (MEL‐14), ‐CD64 (X54‐5/7.1), ‐CD69 (H1.2F3), ‐CD169 (3D6.112), ‐Ly6C (HK1.4), ‐CD11c (N418), ‐IL‐17A (TC11‐18H10.1), ‐IL‐2 (JES6‐5H4), ‐IL‐10 (JES5‐16E3), ‐IFNγ (XMG1.2), anti‐TNFα (MP6‐XT22), anti‐CD40L (MR1), anti‐CD45.1 (A20), anti‐CD45.2 (104), anti‐CD11b (M1/70), anti‐F4/80 (BM8). Viability stain was achieved with LIVE/DEAD fixable violet or near‐IR dead cell kit (Invitrogen).

Statistical analysis

All mice and mice‐derived samples used in the experiments were included in the statistical calculations. Arthritis, EAE and colitis severity were analysed at each time point using the Mann–Whitney U‐test, while a Fisher's exact test was used to calculate the significance of disease incidence. For all the in vitro assays, a non‐normal distribution of the data was considered, and a Mann–Whitney U‐test was used for comparison of two different groups. All the statistical calculations were performed using graphpad prism 7.0a (GraphPad Software, San Diego, CA). Significance was considered when P < 0·05 for a 95% confidence interval.

Results

Increased salt intake does not affect urine osmolality or intestinal IgA excretion

According to the World Health Organization, the concentration of NaCl in drinking water varies between 0·3 mm and 4·3 mm (i.e. 18–250 mg/l NaCl).25 In 2003, the Environmental Protection Agency of the United States reported that the median levels of sodium detected in drinking water were generally below 30 mg/ml.26 Having these values as reference, we decided to determine whether exposure to high salt intake through drinking water could affect the general body homeostasis, and whether immune responses could be significantly altered under such conditions. Hence, we provided mice with drinking water containing 1% NaCl (10 g/l), which considerably increased the salt intake – > 300 times higher when considering 30 mg/l as the average concentration. During the course of 3 weeks, fresh urine and stool samples were collected (Fig. 1a). Despite the significantly higher water intake by mice in the 1% NaCl group (Fig. 1b), the USG (a measure for urine osmolality) was not affected (Fig. 1c). One other biological parameter that indicates homeostatic changes is the titre of faecal IgA. During the same period of time, we did not observe any significant changes in faecal IgA that could have been induced by a higher intake of salt (Fig. 1d).

Figure 1.

Intake of 1% salty water does not disturb kidney function or intestinal barrier. (a) Mice were exposed to normal drinking water or water containing 1% NaCl for 3 weeks. Urine and stool samples were collected weekly. (b) The weekly water intake was registered for four individual cages per group and calculated as an average water intake per mouse per day (n = 12 mice per group). (c) Urine‐specific gravity (USG) was measured weekly in all mice. (d) Total titres of IgA excreted in faeces. Titres are presented as μg/g of stool. Data show mean ± SEM. A Mann–Whitney U‐test was used, and data were considered significant when P < 0·05 for a 95% confidence interval. ***P < 0·0001.

T‐cell effector functions are affected by exposure to high salt concentrations

Prevailing hyper‐osmolality of lymphoid organs in relation to blood suggests that significant changes in salt concentrations may affect the homeostasis of these lymphoid structures, and consequently the immune responses there established.17, 27 In this regard, several studies have shown that exposure to NaCl potentiates polarization towards Th17 cells.12, 13, 28 Using ex vivo cell cultures from naïve mice pre‐exposed to normal or 1% NaCl drinking water for 3 weeks, we assessed how these conditions could influence the effector functions of CD4⁺ T‐cells, namely in terms of cytokine production profiles. CD4⁺ T‐cells originating from mice exposed to higher salt intake showed a significant increase in IL‐2, TNFα and IL‐17A production (Fig. 2a), after PMA stimulation. The levels of IFNγ production were similar between the groups (data not shown).

Figure 2.

In vitro pro‐inflammatory CD4⁺ T‐cell effector functions and reduced viability after salt exposure are associated with osmotic pressure. (a) Mice were exposed to standard drinking water or 1% NaCl in water for 3 weeks prior to organ harvesting. Splenocytes were stimulated with phorbol 12‐myristate 13‐acetate (PMA)/ionomycin for 4 hr, and IL‐2, TNF α and IL‐17A production was quantified in CD4⁺ T‐cells. Mean fluorescence intensity (MFI) is shown. Naïve splenocytes (n = 5) were stimulated with anti‐CD3/CD28 mAb for 48 hr (b) or 72 hr (c), in the presence or absence of external NaCl or d‐mannitol in the culture medium. Cell viability was assessed by flow cytometry, whereas IL‐2 (b), IFN γ and IL‐17A (c) production were measured by ELISA in the culture supernatant and represented as units per frequency of live CD4⁺ T‐cells. (d) Naïve CD4⁺ T‐cells from HCQ3 mice in a RAG‐deficient background were co‐cultured with naïve wild‐type splenocytes and stimulated with anti‐CD28 mAb and cognate antigen (Gal‐Hyl₂₆₄ peptide) for 6 hr. CD40L upregulation by HCQ3 T‐cells was assessed by flow cytometry. Data show mean ± SEM, and are representative of two (a and d) or six (b and c) independent experiments with a similar number of mice. A Mann–Whitney U‐test was used, and data were considered significant when P < 0·05 for a 95% confidence interval. *P < 0·05; **P < 0·01.

To determine whether the effect of salt on CD4⁺ T‐cell effector function could be solely explained by its osmotic properties, we compared in vitro T‐cell responses in the presence of NaCl with those of d‐mannitol at identical osmotic concentrations. Despite the elevated cell death induced by salt and d‐mannitol (Fig. 2b,c), the relative production of IL‐2, IFNγ and IL‐17A was significantly increased in cell cultures with high salt or d‐mannitol concentrations. Although pointing in the same direction as a previous publication14 regarding the effect of salt in potentiating IL‐17A secretion, it was curious to observe that d‐mannitol could also increase the production of pro‐inflammatory cytokines. This pro‐inflammatory skewing occurred albeit at a lower extent than NaCl. Together these data show that the increased effector function of CD4⁺ T‐cells is at least partially mediated by osmotic effects.

To assess the effect of NaCl on antigen‐specific T‐cell responses, naïve TCR transgenic CD4⁺ T‐cells were cultured with naïve wild‐type splenocytes in the presence of cognate antigen. In this setting, high concentrations of NaCl and d‐mannitol significantly reduced T‐cell reactivity, as measured by CD40L expression (Fig. 2d).

A moderate increase of salt intake does not predispose for autoimmunity

Recent reports have linked a higher intake of salt to a predisposition of mice to develop severe autoimmunity, particularly EAE, a mouse model of MS.12, 13, 14, 15 Thus, we evaluated whether a moderate increase of salt intake could also affect the development of other T‐cell‐dependent autoimmune conditions, such as CIA, a mouse model for human RA. Here, we use the term ‘moderate intake’, referring to 1% NaCl in drinking water, as a distinction to the studies mentioned above where HSD alone (4% NaCl) or in combination with 1% salty water have been used. We observed that the development of arthritis in mice with a moderate salt intake was similar to that of mice under standard housing conditions. Both macroscopic arthritic manifestations as well as autoantibody titres were comparable between groups (Fig. 3a,b). Interestingly, and in contrary to what was previously described,12, 13, 14, 15 we did not observe any increased susceptibility to EAE after increased salt intake (Fig. 3c). Here, it is relevant to point out once again that mice were exposed to a salt concentration 3·5 times lower than in the previous studies.

Figure 3.

Increased salt intake does not affect the development of T‐cell‐dependent arthritis and encephalomyelitis mouse models. Mice received standard drinking water or 1% NaCl in water starting 2 weeks prior to immunization. (a) Mice were immunized with rat collagen type II (CII), and monitored for weight changes and development of arthritis. (b) Serum samples from immunized mice were collected at days 21, 35 and 50 post‐immunization and assessed for titres of anti‐CII antibodies: total Ig, IgG1 and IgG2b. (c) Mice were immunized with myelin oligodendrocyte glycoprotein (MOG) _75–96 peptide and monitored for weight changes and development of experimental autoimmune encephalomyelitis (EAE). The values within brackets indicate the number of mice that developed disease out of the total number of animals in the experimental group. Data show mean ± SEM, and are representative of three (a,b) and two (c) independent experiments with similar numbers of mice. A Mann–Whitney U‐test was used, and data were considered significant when P < 0·05 for a 95% confidence interval. n.s., not significant.

In summary, despite the increased pro‐inflammatory T‐cell effector functions induced by high salt concentrations in vitro (Fig. 2), the in vivo exposure to higher NaCl concentrations did not affect the manifestations of T‐cell‐dependent autoimmunity (Fig. 3).

Macrophage effector functions and survival are affected by increased salt exposure

Apart from affecting the activation status of T‐cells and their downstream effector functions, elevated levels of dietary salt have also been shown to influence pro‐inflammatory macrophages.17, 27 Moreover, the different effector functions induced by high salt exposure to splenocytes (Fig. 2) and purified CD4⁺ T‐cells (Fig. S1) suggested that the influence of salt on T‐cells may, in part, be mediated by other salt‐sensitive immune cells. To determine whether a moderate increase in salt intake would have an effect on primary macrophages at steady‐state, we assessed the number of cells in the peritoneal cavity of mice exposed to different salt concentrations. The number of live F4/80⁺ CD11b⁺ cells was significantly increased in the peritoneum of mice exposed to 1% NaCl in drinking water (Fig. 4a). However, no significant differences in terms of activation markers (MHCI‐II, CD80, CD86) were observed between both groups (Fig. S2).

Figure 4.

Increased salt intake influences peritoneal macrophage cellularity and their effector functions. (a) Mice were exposed to standard drinking water or 1% NaCl in water for 3 weeks prior to peritoneal lavage. Absolute cell numbers of peritoneal macrophages (F4/80⁺/CD11b⁺), 3 weeks after salt exposure. (b) Naïve peritoneal cells were stimulated with lipopolysaccharide (LPS) in the absence or presence of external 40 mm NaCl for 20 hr (n = 6). The frequency of live CD11b⁺ cells was assessed by flow cytometry. (c) Cytokine production profile produced by cells described in (b). Cytokine levels are represented as units per frequency of live CD11b⁺ cells. Data show mean ± SEM, and are representative of two independent experiments with similar numbers of mice. A Mann–Whitney U‐test was used, and data were considered significant when P < 0·05 for a 95% confidence interval. *P < 0·05, **P < 0·01, ***P < 0·001.

We have further evaluated the viability and effector functions of peritoneal cells ex vivo, after exposure to high salt concentrations. An increase of 40 mm NaCl in the culture medium significantly decreased the viability of CD11b⁺ cells (Fig. 4b), whereas no viable F4/80⁺ cells could be detected (Fig. S2). Moreover, exposure of peritoneal cells to a high NaCl concentration significantly decreased the production of IL‐6, whereas TNFα and IL‐10 levels were significantly increased (Fig. 4c).

Macrophages are known to play a crucial role in the development of antibody‐mediated autoimmunity.29, 30 Because a moderate increase of NaCl intake resulted in increased numbers of macrophages (Fig. 4a) as well as elevated TNFα production (Fig. 4c), we assessed whether the initial stages of antibody‐mediated arthritis could be influenced by exposure to higher salt intake. During the initial phase of CAIA, which is predominantly influenced by neutrophil and macrophage infiltration of the affected joints,29 mice exposed to higher salt levels manifested less arthritic symptoms (Fig. S3). This was true for both severity as well as prevalence of the disease. However, after LPS stimulation (day 5) and consequent increase in the involvement of macrophages in the pathology, we did not observe any macroscopic differences of arthritis. Despite these unexpected results, mice exposed to salty water showed significantly more weight loss in comparison to control mice, which is in contrast to what we observed during CIA (Fig. 3a).

DSS‐induced colitis is significantly affected by moderate salt intake

The primary tissue/organ exposed to dietary salt intake, and therefore likely to be greatly affected by it, is the gastrointestinal tract. The DSS‐induced murine colitis model is a very well‐studied animal model for IBD.31 As with many autoimmune diseases, the development and severity of IBD is dependent on genetic and environmental factors. During IBD, the epithelial barrier in the intestinal mucosa gets impaired, leading to translocation of microbial products from the gut and increased IgA excretion as a result of its release from the lamina propria. Measurement of faecal IgA is therefore a read‐out used to assess the permeability and compromised structure of the intestinal epithelial barrier. Increased salt intake did not affect the total titres of faecal IgA (Figs 1d and 5b). However, this increased salt intake amplified the disease activity index (DAI; see Fig. S4 and Table S1 for description of the scoring) and disease prevalence when mice were exposed to DSS (Fig. 5c). Two days after DSS experience, all mice pre‐exposed to salt had developed colitis, whereas control animals developed the disease significantly later and in a less severe manner (Fig. 5c). Mice exposed to salty water evidenced a greater weight loss after exposure to DSS (Fig. S4). The frequency of diarrheic stool and stool bleeding were likewise increased in mice exposed to higher salt intake (Fig. S4). Additionally, the elevated titres of faecal IgA (Fig. 5b) and increased cellular counts in spleen and mesenteric lymph nodes (Fig. 5d) confirmed the compromised status of the intestinal barrier in these mice. Histology of the proximal colons from naïve and DSS‐colitis animals under different water regimens emphasized once again the aggravated inflammatory phenotypes induced by salt exposure (Fig. 5e,f; Table S2). Moreover, moderate salt‐exposed mice also showed an elevated colonic cellularity relative to the colon length (Fig. 5g). Here, a significantly higher frequency of CD4⁺ T‐cells and CD11b⁺ CD64⁺ macrophages displayed a pro‐inflammatory phenotype indicated by high TNFα production (Fig. 5h,i). Interestingly, the above‐described parameters of colitis pathology and its associated cellular phenotypes were identical between normal water‐ and d‐mannitol‐exposed mice (Fig. S5). In summary, these data suggest that a higher oral exposure to NaCl has the potential of increasing and sustaining intestine‐associated diseases. Moreover, the inflammatory effects of salt occur via a pathway that is independent of its osmotic properties.

Figure 5.

Exposure to high salt intake exacerbates the severity of dextran sulphate sodium (DSS)‐induced colitis. (a) Mice were pre‐exposed to 1% NaCl in water or standard drinking water for 3 weeks, followed by 5% DSS for 1 week, and again 1% NaCl in water or standard drinking water for another 1 week. (b) Faecal IgA titres from colitis mice described in (a) compared with naïve mice under identical salty water regiments. (c) Colitis disease symptoms were monitored daily and presented as disease activity index (DAI) score (see also Fig. S4 and Table S1) and prevalence. (d) Absolute cell numbers in spleen and mesenteric lymph nodes of mice 14 days after DSS exposure. (e) Haematoxylin and eosin stain showing cellular infiltration of proximal colons from naïve mice or mice under DSS‐induced colitis (day 14) pre‐exposed to 1% NaCl or normal water. The scale bar represents 200 μm. (f) Histology score based on (e) (see also Table S2). Three of six naïve samples, and four of eight colitis normal water and four of seven colitis 1% NaCl water colon samples were used to calculate the histology score. (g) Colon length and absolute cell numbers isolated from the respective colons of B6 mice, 10 days after DSS exposure as described in (a) (see also Fig. S4). Frequency of isolated colonic (h) CD4⁺ T‐cells and (i) CD11b⁺ CD64⁺ macrophages and their respective TNF α production, from samples shown in (g). Values in brackets indicate the number of mice in each experimental group. Data show mean ± SEM. Fisher's exact test was used to calculate statistics regarding disease prevalence, whereas a Mann–Whitney U‐test was used to compare samples at individual time points. Data were considered significant when P < 0·05 for a 95% confidence interval. *P < 0·05, **P < 0·01, ***P < 0·001.

Discussion

Many of the earlier mouse studies implicating dietary salt intake and the development of more severe pathologies used a ‘HSD’. Rodent HSD contains 4% (w/w) salt, in contrast to 0·4% (w/w) of a standard diet. The intake of a 4% salt diet is a dramatic increase in dietary salt consumption, even for humans. In fact, most of the processed foods contain no more than 1%–1·5% of sodium.1 Therefore, we performed our studies comparing mice on a standard diet and mice supplemented with 1% NaCl in drinking water. In the present study, we show that exposure of mice to moderate salt concentrations influenced the effector functions of naïve T lymphocytes and myeloid cells, in particular macrophages, with pathological consequences during the development of inflammatory diseases. Nevertheless, and in contrast to what has been previously suggested,12, 13, 14, 15 the association between salt and exacerbated pathogenesis did not result exclusively from a skewed pro‐inflammatory T‐cell activation. Instead, the exacerbation of disease symptoms was also associated with myeloid cells, especially in the gastrointestinal tract.

Recent reports have discussed how a high salt exposure affects the immune system and, in particular, how HSD influences the development of encephalomyelitis. In these studies,13, 14, 15 CNS tissues isolated from EAE‐diseased animals showed increased levels of cellular infiltrates, when mice were exposed to HSD. However, despite this observation and the fact that NaCl potentiates the stability of in vitro polarized Th17 cells,12, 13 exposure of mice to HSD did not translate into a dramatic increase in disease severity (encephalomyelitis). Here, using 1% NaCl in water as a way to expose mice to higher salt concentrations, we showed that a higher salt intake skews CD4⁺ T‐cells to a more pro‐inflammatory effector profile, at steady‐state both in vitro and ex vivo. The effect was partially mediated through osmotic mechanisms, as an equal osmolality of d‐mannitol resulted in similar phenotypes (Fig. 2). Curiously, when we looked at antigen‐specific CD4⁺ T‐cell responses at steady‐state, we observed that exposure to high salt concentrations resulted in a reduced frequency of T‐cell activation. However, the development of T‐cell‐mediated autoimmune phenotypes was similar between mice under standard housing conditions and mice exposed to 1% NaCl (CIA and EAE; Fig. 3), despite the elevated T‐cell effector functions observed in vitro. The discrepancy in observations between our study and others12, 13, 14, 15 may result from the lower exposure to salt in our study, but it may also reside on the strength of the generated T‐cell responses. Whereas our mice expressed a MHC‐II H‐2^q molecule, mice used in the previous studies expressed H‐2^b. The strength and duration of interactions between TCR and peptide‐MHC complexes have previously been shown to play an important role in the resulting differentiation and activation of T‐cells (reviewed in Refs 32, 33). These differences in signalling could possibly revert any effects induced by salt exposure.

Apart from influencing T‐cell differentiation and polarization, exposure to high salt concentrations have also been implicated in the modulation of myeloid cells, particularly in skewing of bone marrow macrophages to a more pro‐inflammatory phenotype.15 Nevertheless, and as previously reported,15 exposure of naïve macrophages to high salt concentrations did not alter the surface levels of co‐stimulatory markers (MHC‐II, CD80, CD86). We did not assess whether the same was true under pathological conditions. At steady‐state, we observed that mice exposed to high salt concentrations showed a significantly greater number of peritoneal macrophages. These elevated numbers could be explained by an increased recruitment of new cells into the peritoneal cavity after high dietary salt exposure, as salt was also shown to increase cell death of macrophages (Fig. 4b). Ex vivo, macrophages produced higher levels of not only the pro‐inflammatory cytokine TNFα, but also the anti‐inflammatory cytokine IL‐10. This observation suggests that salt exposure does not promote the expansion of one single macrophage population, but rather a mixture of distinct populations (e.g. M1 and M2). In this study, we did not explore that possibility in detail.

Recently, two studies have shown that mice exposed to HSD developed exacerbated experimental colitis.16, 34 Here, by exposing mice to 1% NaCl, we were able to reproduce these findings. Mice pre‐exposed to 1% NaCl developed severe colitis phenotypes, such as increased weight loss, faecal bleeding and reduced stool consistency (Fig. 5, and Figs S4 and S5). Disruption of the intestinal barrier in these mice was evidenced by the increased titres of faecal IgA as well as lymphoid cellularity in the colon, spleen and mesenteric lymph nodes. Hence, it is not completely unforseen that high salt intake has been associated with a higher risk of gastrointestinal cancers.5 Moreover, in agreement with our in vitro observations (Figs 2 and 4), colonic CD4⁺ T‐cells and CD11b⁺ CD64⁺ macrophages from colitis mice pre‐exposed to high salt concentrations showed a capacity for strong TNFα production (Fig. 5h,i). Immune therapy against TNFα is recommended in glucocorticoid‐resistant patients with moderate to severe Crohn's disease. The same applies to patients suffering from ulcerative colitis,35 which clearly indicates a deleterious effect of TNFα in the pathogenesis of IBD.

In summary, our data confirms that exposure to higher salt concentrations in vitro affects the differentiation of naïve T‐cells, with a preferential skewing into pro‐inflammatory phenotypes. Nevertheless, when exposing mice to reasonably increased amounts of salt, the consequences of high salt intake in the development of T‐cell‐dependent autoimmunity were almost negligible. In contrast, myeloid cells responded to high salt concentrations in very distinct ways. Both pro‐ and anti‐inflammatory cytokine responses were observed, ex vivo. In vivo, mice exposed to moderate salt concentrations developed exacerbated colitis symptoms. These were associated with high levels of TNFα‐producing CD4⁺ T‐cells and macrophages. Our study indicates that the in vitro phenotypes observed in naïve lymphoid/myeloid cells exposed to high salt concentrations and the respective cell‐associated disease models in vivo cannot be directly correlated.

Disclosures

The authors declare that they have no competing interests.

Supporting information

Figure S1. Effect of high salt concentrations on purified CD4⁺ T‐cells.

Figure S2. Salt does not affect expression of activation markers but decreases viability of F4/80⁺ CD11b⁺ cells.

Figure S3. Increased salt intake decreases onset of antibody‐mediated effector‐phase of arthritis.

Figure S4. Exposure to higher salt intake exacerbates several colitis phenotypes.

Figure S5. Increased colitis from high salt exposure does not occur via osmotic mechanisms.

Table S1. Disease activity index score in DSS‐colitis

Table S2. Colonic histology score