Salt Accelerates Allograft Rejection through Serum- and Glucocorticoid-Regulated Kinase-1–Dependent Inhibition of Regulatory T Cells

Abstract

A high-salt diet (HSD) in humans is linked to a number of complications, including hypertension and cardiovascular events. Whether a HSD affects the immune response in transplantation is unknown. Using a murine transplantation model, we investigated the effect of NaCl on the alloimmune response in vitro and in vivo. Incremental NaCl concentrations in vitro augmented T cell proliferation in the settings of both polyclonal and allospecific stimulation. Feeding a HSD to C57BL/6 wild-type recipients of bm12 allografts led to accelerated cardiac allograft rejection, despite similar mean BP and serum sodium levels in HSD and normal salt diet (NSD) groups. The accelerated rejection was associated with a reduction in the proportion of CD4⁺Foxp3⁺ regulatory T cells (Tregs) and a significant decrease in Treg proliferation, leading to an increased ratio of antigen-experienced CD4⁺ T cells to Tregs in mice recipients of a HSD compared with mice recipients of a NSD. Because serum- and glucocorticoid-regulated kinase-1 (SGK1) has been proposed as a potential target of salt in immune cells, we fed a HSD to CD4^CreSGK1^fl/fl B6-transplanted recipients and observed abrogation of the deleterious effect of a HSD in the absence of SGK1 on CD4⁺ cells. In summary, we show that NaCl negatively affects the regulatory balance of T cells in transplantation and precipitates rejection in an SGK1-dependent manner.

The discovery of salt is considered a fundamental milestone for humanity; throughout history, salt extraction, possession, and intake were a reflection of the prosperity and wealth of a nation. Virtual absence of dietary salt, as seen in the Yanomami tribes living in the Amazon rainforests, is associated with zero incidence of hypertension,¹ whereas a typical Western salt-rich diet has long been associated with hypertension and consequent cardiovascular morbidities.² Surprisingly, however, it has recently been shown that salt intake also affects the immune system: using a murine experimental autoimmune encephalomyelitis (EAE) model, mice fed a high-salt diet (HSD) had exacerbated disease mediated by the induction of pathogenic T_H17 T cells.^3–5 Furthermore, HSD has also been shown to promote lymph capillary network hyperplasia and increased skin infiltration by mononuclear phagocytic cells in mice and rats.^6,7

In transplantation, advances in immunologic screening combined with developments in immunosuppression achieved in the last several decades have resulted in significant improvements in early graft survival; the rates of late allograft loss, however, remain unacceptably high.⁸ There have been only isolated reports examining the correlation of salt intake with hypertension after transplantation,^9,10 whereas there are no reports of the potential contribution of salt intake to the incidence of rejection and allograft survival. It is, therefore, unknown whether dietary salt consumption could affect the alloimmune response and long-term allograft survival. In this report, we sought to examine the effect of salt on alloimmunity. We report, for the first time, that salt augments in vitro allospecific T cell proliferation, whereas in a mouse model of solid organ transplantation, feeding mice an HSD results in accelerated allograft rejection caused by disturbance of the regulatory balance of T cells in vivo.

We first examined the effect of higher salt concentration ([NaCl]) on T cell proliferation in vitro using cultures of naive murine splenocytes incubated in medium enriched with incremental concentrations of NaCl (ranging from 0 to 40 mM) in the presence of aCD3 and aCD28 (2 μg/ml) (Figure 1A). We observed that increasing [NaCl] from 0 to 40 mM resulted in a significant increase in proliferation, which was measured by thymidine incorporation, from 23,674±2063 to 49,801±2423 counts per minute (P<0.001). To further investigate this observation in an alloimmune milieu (Figure 1B), we primed reporter C57Bl/6 Foxp3.GFP mice (B6 wild type [WT]) a priori with intraperitoneal injections of BALB/c splenocytes; 2 weeks later, we isolated sensitized CD4⁺Foxp3⁻ cells from these mice and cultured them with irradiated BALB/c CD3⁻ splenocytes in the presence of incremental [NaCl]. Indeed, increasing [NaCl] from 0 to 40 mM again resulted in a significant increase in T cell proliferation from 7355±565.5 to 18,588±1635 counts per minute (P=0.004). This increase in proliferation was specific to NaCl. Adding urea to culture medium at 80 mM concentration resulted in decreased cell proliferation compared with standard medium as measured by CFSE-negative populations (77.1%±0.3% versus 79.2%±0.3%, respectively; P=0.01), whereas the addition of equiosmolar NaCl concentration (40 mM) resulted in increased cell proliferation compared with standard medium (85.2%±0.6% versus 79.2%±0.3%, respectively; P<0.001).

Figure 1.

Increasing NaCl concentration in vitro augments polyclonal and allospecific T cell proliferation. (A) Proliferation of splenocytes after stimulation with aCD3 and aCD28 (2 μg/ml) in the presence of incremental concentrations of NaCl. (B) Proliferation of splenocytes after allospecific stimulation in the presence of incremental concentrations of NaCl. B6 Foxp3.GFP mice were sensitized with intraperitoneal BALB/c splenocytes injection; 2 weeks later, CD4⁺Foxp3⁻ cells were isolated by FACS sorting and cultured with magnetically isolated, irradiated BALB/c CD3⁻ splenocytes. Thymidine was added at 72 hours of culture, and incorporation was quantified 12 hours later. P values >0.05 are omitted. CPM, counts per minute. *P<0.05; **P<0.01; ***P<0.001;

In light of these encouraging in vitro findings, we then investigated the in vivo response using a mouse model of chronic rejection, in which bm12 hearts are transplanted into B6 recipients. In this MHC class II–mismatched cardiac transplant model, allografts typically survive >56 days, although they develop progressive vasculopathy.^11–14 The survival of allografts in this model is dependent on the presence of regulatory T cells (Tregs) that inhibit the expansion of the small clone size of allospecific effector T cells, which recognize the single mismatched MHC II molecule on donor grafts.^12,15,16 B6 recipients were fed either normal-salt diet (NSD) or HSD and allowed unrestricted access to free water. We found that feeding mice a HSD resulted in decreased allograft survival compared with NSD, with a median survival time (MST) of 48 days compared with >56 days, respectively (n=8–9 per group; P=0.01) (Figure 2A). Given the nature of the transplant (vascularized) and the intervention (HSD), we investigated the potential influence of hypertension: measurement of BP (Figure 2B) 25 days post-transplantation showed that mice fed a NSD or a HSD had similar mean arterial pressures (MAPs) of 65.35±2.5 and 65.76±4.4 mmHg, respectively (n=8 per group; P=0.93). We also evaluated the effect of diet on serum sodium and weight change (Figure 2, C and D): mice fed a NSD or a HSD had serum sodium of 149.5±3.7 and 150.4±2.4 mEq/L (n=6–7 per group; P=0.83) and percentage weight change of 7.449%±0.5% and −2.421%±0.6% (n=4 per group; P<0.001), respectively. To reconcile the in vitro effects of higher sodium concentration with the in vivo findings associated with HSD despite unchanged serum sodium, it is worth mentioning that previous work had shown that HSD results in increased interstitial fluid and lymphatic sodium concentration without effect on serum sodium concentration.⁷

Figure 2.

HSD accelerates cardiac allograft rejection independently of serum sodium or BP. (A) Kaplan–Meier curves of allograft survival (n=8–9 per group) of bm12 hearts transplanted into B6 recipients fed either a NSD or a HSD. (B) Tail cuff MAP measured at 25 days post-transplantation (n=8 per group; P=0.93). (C) Serum sodium of transplanted mice measured at 3 weeks post-transplantation (n=6 per group; P=0.80). (D) Percentage weight change at 3 weeks post-transplantation in relation to baseline weight (n=4 per group). Dietary modification was started 2 weeks before transplantation. *P<0.05; ***P<0.001.

To elucidate the mechanisms underlying the accelerated rejection observed with HSD, we immunophenotyped lymphocytes isolated from the allograft-draining lymph nodes (LNs) and spleens 25 days post-transplantation using flow cytometry (Figure 3). Analysis of the LNs of recipients fed a HSD revealed a significant reduction in the proportion of Tregs (Foxp3⁺ of CD4⁺ T cells) compared with those fed a NSD (5.36%±0.9% versus 15.47%±1.7%, respectively; P=0.02); furthermore, these Tregs proliferated significantly less than those isolated from NSD-fed mice, which was determined by expression of the intracellular marker Ki67 (14.45%±0.15% versus 20.17%±1.37%, respectively; P=0.05). There was a slight reduction in the proportion of splenic Tregs, albeit not statistically significant (7.69%±0.87% versus 9.77%±0.46% with NSD, respectively; P=0.10), whereas the proliferation of splenic Tregs (Ki67⁺ of CD4⁺Foxp3⁺ cells) was also reduced in mice fed a HSD compared with a NSD (23.65%±6.9% versus 44.97%±0.42%; P=0.03). The ratio of splenic CD4⁺CD44⁺/CD4⁺Foxp3⁺ cells (Teffmem/Treg) was increased in the HSD group compared with the NSD group (2.570±0.12 versus 1.547±0.15, respectively; P=0.02). Although salt has been shown to induce Th17 cells (CD4⁺IL-17⁺) expansion in an EAE model,⁴ spleen Th17 cells were extremely rare in our model and not significantly different between groups (0.22%±0.1% versus 0.08%±0.04%, respectively; P=0.28). In sum, the accelerated rejection seen in HSD-fed recipients was associated with a reduction in the proportion of Tregs, primarily in the draining LNs, along with a significant decrease in their proliferation, leading to an increase in the CD4⁺ Teffmem/Treg ratio compared with NSD-fed recipients. Given these findings, we investigated the extent of acute rejection and allograft Treg infiltration. Histologic examination of the grafts 25 days post-transplant revealed a similar degree of rejection in both groups as measured by the International Society of Heart and Lung Transplantation classification²⁷ (ISHLT-R) score (2.3±0.3 versus 2.4±0.2, respectively; n=5–6 per group; P=0.88); however, there was significantly decreased Foxp3⁺ cell infiltration in mice fed HSD compared with NSD (1.6±0.3 versus 10.1±1.6 cells per HPF, respectively; n=3 per group; P=0.02) as measured by immunohistochemistry.

Figure 3.

Transplant recipients fed with an HSD displayed both decreased proportion and proliferation of Tregs; bm12 hearts were transplanted into B6 mice fed with either an NSD or an HSD. Recipients were then euthanized at 25 days after transplantation; spleens and allograft-draining LNs were harvested and analyzed by flow cytometry. Representative dot plots and histograms of flow cytometry gating strategy on CD4⁺, Foxp3⁺, and Ki67⁺ subsets in (A) spleens and (B) LNs. Respective bar plots of percentages of Tregs (CD4⁺Foxp3⁺) and their proliferation (CD4⁺Foxp3⁺Ki67⁺) are included. n=3 per group. Data are representative of three independent experiments. *P<0.05.

Serum- and glucocorticoid-regulated kinase-1 (SGK1) is a signaling kinase that is induced by a variety of stimuli, including glucocorticoids, aldosterone, and hypertonicity.^17–19 SGK1 has been shown to play a major role in sodium homeostasis and is expressed in a wide array of cells, including those of the immune system. As an example, SGK1 activation was shown to promote Th17 cell polarization in an EAE model.^4,5 Furthermore, in an experimentally induced asthma model in mice, SGK1 activation favored T_H2 phenotype polarization, and SGK1 deletion in T cells was protective against asthma.²⁰ We hypothesized that the deleterious effects consequent to HSD consumption observed in our model were mediated by SGK1, and therefore, we sought to examine its potential role. Using mice with selective deletion of SGK1 in CD4⁺ cells (CD4^CreSGK1^fl/fl), we again examined the effect of incremental concentrations of salt on proliferation in vitro followed by investigation of the alloimmune response in vivo using our cardiac transplantation model.

In vitro, the salt-induced increased proliferation of BALB/c–primed WT B6 CD4⁺ T cells was abrogated when primed B6 CD4^CreSGK1^fl/fl CD4⁺ T cells were used instead (ANOVA P values of 0.01 versus 0.11, respectively) (Figure 4A). In vivo, the observed accelerated rejection of bm12 hearts transplanted into WT B6 mice fed an HSD was abrogated in HSD-fed B6 CD4^CreSGK1^fl/fl mice (MST>56 days). Furthermore, the proportion of splenic Tregs was also similar between HSD-fed CD4^CreSGK1^fl/fl mice and NSD-fed controls (10.83%±0.12% versus 9.770%±0.4676%; n=3 per group; P=0.09) (Figure 4B), resulting in a similar splenic Teffmem/Treg ratio between groups (1.773±0.08 versus 1.547±0.15; n=3 per group; P=0.28). The proportion of Tregs in allograft-draining LN did not significantly differ between HSD-fed B6 CD4^CreSGK1^fl/fl mice and NSD-fed controls (10.03%±3.6% versus 15.47%±1.7%, respectively; n=3 per group; P=0.22). On allograft histology, we observed a similar degree of acute rejection between WT mice fed NSD, WT mice fed HSD, and CD4^CreSGK1^fl/fl mice fed HSD (2.3±0.3 versus 2.4±0.2 versus 2.67±0.3, respectively; n=3–6 per group; ANOVA P=0.78); however, the HSD-induced reduction in allograft Tregs observed in WT mice was again abrogated in HSD-fed CD4^CreSGK1^fl/fl mice (1.61±0.3 versus 10.4±1.3 cells per HPF, respectively; n=3 per group; P<0.01) (Figure 4C), indicating a critical role of SGK1 in the acceleration of rejection by reducing Tregs in mice fed a HSD.

Figure 4.

Mice lacking SGK1 in CD4⁺ T cells seem to be protected against both the in vitro NaCl-induced increased proliferation and the in vivo effects of HSD in transplantation. (A) WT or CD4^CreSGK1^fl/fl B6 mice were sensitized with intraperitoneal BALB/c splenocytes injections; 2 weeks later, CD4⁺ T cells were magnetically isolated and cultured for 72 hours with irradiated BALB/c CD3⁻ cells in the presence of incremental concentrations of NaCl. Proliferation was measured by thymidine incorporation (n=6 per group). (B) Spleen Teffmem/Treg ratios (CD4⁺CD44⁺/CD4⁺Foxp3⁺) and Tregs in spleens, draining LNs, and allografts at 25 days after transplantation from WT B6 mice placed on NSD, WT mice on HSD, and CD4^CreSGK1^fl/fl B6 mice on HSD (n=3 per group). Data are representative of three independent experiments. (C) Photomicrographs of high-power fields (×400) of immunohistochemistry staining of Foxp3⁺ cells in bm12 cardiac allografts 25 days after transplantation in WT B6 mice placed on NSD, WT B6 mice placed on HSD, and CD4^CreSGK1^fl/fl B6 mice placed on HSD. (D) Effect of salt on Tregs. Green arrows represent activation, whereas the red arrow represents inhibitory signals. (E) Phosphorylated (inactivated) FoxO1 and FoxO3a transcription factors measured in control and CD4^CreSGK1^fl/fl Tregs cultured with and without additional NaCl (+0 and +40 mM, respectively); BioHeat map generated by flow cytometry illustrates fold change in the mean fluorescent intensity (MFI) of anti-phosphorylated FoxO1/FoxO3a. CPM, counts per minute. *P<0.05; **P<0.01.

Two transcription factors, FoxO1 and FoxO3a, play an important role in regulating Foxp3 expression in Tregs,^21,22 and their deletion in mice results in a fatal multifocal inflammatory disorder caused by a defect in Tregs.²³ Interestingly, SGK1 is known to inactivate both FoxO1 and FoxO3a by phosphorylating them and promoting their sequestration in the cytoplasm.^4,24,25 Therefore, we hypothesized that salt may inhibit FoxO1/3a by activation of SGK1, leading ultimately to the reduction in Tregs that we observed thus far (Figure 4D). To test this hypothesis, we measured phosphorylated FoxO1 and FoxO3a by flow cytometry in WT and CD4^CreSGK1^fl/fl Tregs cultured with and without additional NaCl. Indeed, increasing [NaCl] in culture medium resulted in increased FoxO1/FoxO3a phosphorylation in an SGK1-dependent fashion (Figure 4E), indicating a direct correlation between salt, SGK1 activation, FoxO1/FoxO3a phosphorylation, and ultimately, Treg inhibition.

In summary, we report, for the first time, immunologic effects of salt consumption in a mouse model of solid organ transplantation. First, we showed increased proliferation of primed T cells after exposure to higher [NaCl]. Second, we observed that HSD-fed mice displayed accelerated allograft rejection in the absence of elevated BP or alteration of serum sodium concentration. Third, we found that, in transplanted mice fed HSD, there was a shift in the Teffmem/Treg balance and that HSD was associated with both decreased proliferation of splenic and graft-draining LN Tregs and significantly fewer Tregs infiltrating the allograft. Fourth, we examined SGK1 as a potential mediator of our observations: using a conditional knockout model in which CD4⁺ T cells lack SGK1, we found that the increased proliferation, accelerated rejection, skewing of Teffmem/Treg balance, and decreased allograft Tregs observed in HSD-fed WT recipients were abrogated by the lack of SGK1. Our findings constitute a proof of concept of a deleterious, immune-modifying role of salt in accelerating rejection in transplantation, complementing previous evidence of salt-induced autoimmunity. Given the endemic nature of salt consumption, additional investigation of its effect in human transplantation is warranted.

Concise Methods

Mice

WT C57BL/6 (B6), bm12, and BALB/c mice were purchased from The Jackson Laboratory. B6 CD4^CreSGK1^fl/fl mice were a gift from the Kuchroo Laboratory. All mice were 8–12 weeks of age, and they were harbored and used in accordance with Harvard Medical School and National Institutes of Health guidelines.

Transplantation

Vascularized heart grafts were placed in an intra-abdominal location using microsurgical techniques as described by Corry et al.²⁶ Graft function was assessed by palpation of the heartbeat; rejection was determined by complete cessation of palpable heartbeat and confirmed by direct visualization after laparotomy. Graft survival is shown as the MST in days. Mechanistic experiments with transplanted mice were performed at the time point in which rejection started to occur in the HSD-treated group (approximately day 25 post-transplant).

Diet Modification

Mice were fed either standard chow containing 0.28% Na (NSD) or chow containing 3.15% Na (HSD; 1810179; Pharmaserv Inc.) beginning 2 weeks before transplantation and continued thereafter. Mice were allowed continuous access to free water.

Serum Sodium and BP Measurements

Sodium was measured by the Roche Cobas c501 Module sodium ion–specific electrode on mice sera after at least 3 weeks on HSD or NSD. BP was measured by tail transmission photoplethysmography by the Visitech System BP-2000 Series II, and MAP was calculated as follows: MAP=[(2×diastolic)+systolic]/3.

In Vitro Culture Media

Standard culture medium was enriched to attain additional NaCl concentrations between 0 and 40 mM by adding NaCl (Sigma Life Science). Osmotic control was done with urea-enriched culture media (0–80 mM; Sigma Life Science).

In Vitro Cell Cultures and Proliferation Assay

For polyclonal proliferation, naïve B6 splenocytes were cultured with aCD3 and aCD28 (final well concentration of 2 μg/ml) and increments of NaCl-enriched culture media for 72 hours. For allospecific proliferation, WT B6, B6 FoxP3.GFP, or B6 CD4^CreSGK1^fl/fl mice were sensitized with an intraperitoneal injection of 15 million BALB/c splenocytes and euthanized at least 2 weeks thereafter. CD4⁺FoxP3.GFP⁻ T cells were isolated by FACS sorting and cocultured in culture media enriched by incremental concentrations of NaCl with irradiated CD3⁻–depleted BALB/c splenocytes. Proliferation was measured 72 hours later by quantification of thymidine incorporation. For osmotic control, cells were stained with 10 μM CFSE in the presence of additional salt (0–40 mM) or urea (0–80 mM), and proliferation was measured by CFSE dilution by flow cytometry 72 hours later.

Flow Cytometry

Transplanted mice spleens and draining LNs were harvested, and single-cell suspensions were prepared. Cells were stained with fluorochrome-labeled mAbs against CD4, CD8, B220, CD62 ligand (CD62L), CD44, CD25, Ki67, and Foxp3. Intracellular staining for Foxp3 and Ki67 was performed after permeabilization of the cells using the eBioscience Foxp3 Fixation/Permeabilization Solution. Flow cytometry was performed using a BD FACSCanto II Cytometer and analyzed using FlowJo software.

Phosphorylated FoxO1 and FoxO3a Measurement

Tregs were isolated using the EasySep Mouse CD4⁺CD25⁺ Regulatory T Cell Isolation Kit (Stem Cell). Cells were cultured with aCD3 and aCD28 (final well concentration of 2 μg/ml) for 10 or 60 min with and without additional NaCl. Cells were then fixed with BD Cytofix (BD Biosciences) and permeabilized with Perm Buffer III (BD Biosciences); then, they were stained with a primary anti-phospho–FoxO1/3a^T24/T32 antibody (Cell Signaling Technology) and a secondary anti-rabbit IgG Alexa 488 antibody (Molecular Probes). Flow cytometry was performed using a BD FACSCanto II Cytometer and analyzed using FlowJo software. BioHeat maps were generated with the web-based software Cytobank (www.cytobank.org).

Histopathology

Cardiac graft samples from transplanted mice were harvested from both NSD- and HSD-fed groups at 25 days post-transplantation. Grafts were then fixed in 10% formalin, embedded in paraffin, transversely sectioned, and stained with H&E stain. Using the revised ISHLT-R,²⁷ a blinded transplant pathologist graded the degree of rejection (0–3). Immunohistochemistry for Foxp3⁺ cells was performed on formalin-fixed, paraffin-embedded allografts sections; cells were counted in at least 10 high-power fields per sample, and individual sample counts were averaged thereafter.

Statistical Analyses

Statistical analyses were performed using Prism 5.0b of Graphpad software. Kaplan–Meier curves were used to generate allograft survival curves, and log-rank test was used to compare median survival. Two-way unpaired t test and ANOVA were used to compare datasets. P values <0.05 were considered statistically significant.

Disclosures

None.