Excess Salt Intake Activates IL-21–Dominant Autoimmune Diabetogenesis via a Salt-Regulated Ste20-Related Proline/Alanine-Rich Kinase in CD4 T Cells Free

Although monozygotic twin studies elucidate that environmental factors contribute to initiation of autoimmune diseases in individuals with predisposing genetic risks (1), possible mechanisms by which a diet triggers autoimmunity or modulates disease severity remain largely unexplored. In addition to increasing the risk of cardiovascular disease and hypertension, overconsumption of salt from processed foods beyond the recommended amounts has also been implicated as a risk factor in the development of autoimmune diseases. It was noted that incubation of human mononuclear cells under hyperosmotic conditions can induce the release of proinflammatory cytokines (2), and the clinical use of hypertonic saline for plasma expansion enhances immune functions (3,4). Further investigations demonstrated that the mechanism underlying enhanced T-cell responses triggered by hyperosmolar sodium chloride (NaCl) concentrations is regulated by activation of the p38/MAPK cascade and the transcription factor nuclear factor of activated T cells 5 (NFAT5) osmotic stress response pathway (5,–,7). Considering that the tissue sodium content is highly compartmentalized under pathophysiological conditions, dietary salt intake may affect tissue sodium accumulation and immune function (8).

Even though a direct relationship between salt and multiple sclerosis (MS) has not been demonstrated, a higher amount of sodium intake is associated with increased clinical and radiological disease activity in patients with MS (9). Moreover, based on previous findings from investigations of the effects of elevated NaCl levels on human T cells and murine experimental autoimmune encephalomyelitis (EAE; an animal model of MS) (10,11), elevated NaCl concentrations promoted the differentiation of murine and human T helper (Th) 17 cells, which display a highly proinflammatory phenotype characterized by an upregulation of proinflammatory cytokines such as granulocyte macrophage colony-stimulating factor, tumor necrosis factor (TNF)-α, and interleukin (IL)-2, and activated the osmotic stress pathway, including p38/MAPK, NFAT5, and serum/glucocorticoid-regulated kinase 1 (SGK1) (10). In particular, a high-salt diet (HSD) induced SGK1 expression for stabilizing Th17-cell phenotype in an IL-23R–dependent manner and accelerated the development of EAE (11). Furthermore, several studies demonstrated that an HSD impaired suppressive function of T regulatory (Treg) cells by skewing Treg cells into a dysfunctional Th1-like phenotype (12–14), resembling a proinflammatory phenotype in patients with autoimmune diseases (13,14).

These results suggested a possible mechanism by which salt may potentiate autoimmunity and could be replicated in other autoimmune conditions such as autoimmune colitis (15). Nevertheless, in a mouse model of spontaneous autoimmune peripheral polyneuropathy, a knockout of CD86 in nonobese diabetic mice (CD86^−/− NOD), an HSD delayed onset of clinical symptoms and ameliorated disease course in these mice (16). However, a beneficial or detrimental effect of the high salt conditions solely based on the NOD mouse strain used in this study or in type 1 diabetes is unclear.

Patients with type 1 diabetes are more susceptible to hypertension than are individuals with no diabetes (ND), and this is partly mediated by enhanced blood pressure sensitivity to high salt intake associated with insulin resistance (17) and accompanied by absence of skin macrophage influx and lack of dilation of lymphatic vasculature (18). While children with type 1 diabetes in the U.S., Europe, and Australia demonstrate low adherence to dietary guidelines and overconsume higher-calorie and saturated fats than control study participants (19,–,21), it is important to highlight that a total of 80% of Australian children with type 1 diabetes and control participants are overconsuming sodium (2,837 ± 848 mg/day) above the recommended levels (22). To date, little information exists on the impact of salt intake on type 1 diabetes, and an HSD may be the reason for the increased incidence of type 1 diabetes in recent decades.

Intracellular ion homeostasis is essential for proper cell function in response to hypertonic stimuli like NaCl. The dynamic regulation of signaling pathways controls the activity of ion transport to maintain homeostasis (23). In response to hypertonicity and cell shrinkage, the members of the WNK (“with-no-lysine”) kinase family, which are mutated in Gordon’s hypertension syndrome (24), phosphorylate and activate oxidative stress–responsive kinase 1 (OSR1) and STE20/SPS1-related proline/alanine–rich kinase (SPAK) (25,26), encoded by STK39. Once activated, OSR1/SPAK kinases activate Na⁺/K⁺/2Cl⁻ cotransporter (NKCC) and Na⁺/Cl⁻ cotransporter (NCC), which are members of solute carrier 12A (SLC12A) family of cotransporters, in the context of hypertonicity. In addition, the phosphorylation states of OSR1/SPAK and NCC were increased by a low-salt diet and decreased by an HSD (27). A genome-wide association study identified STK39 as a hypertension susceptibility gene (28), and there is a novel association between salt sensitivity and STK39 genetic variation in the Korean population (29). On the other hand, SPAK expressed in CD4⁺ T cells has been implicated in the regulation of T-cell activation as a direct substrate of PKCθ in the pathway leading to AP-1 activation (30) and bound to CD46 to regulate IL-10 production (31). Overexpression of SPAK contributing to production of inflammatory cytokines can increase experimental colitis susceptibility (32), whereas SPAK deficiency attenuates intestinal inflammation (33). Additionally, a recently identified, novel, rare single nucleotide polymorphism within the first intron of STK39 has large effects on risk of type 1 diabetes (34). In vitro pharmacological inhibition of Stk39 activity using a specific STK39 inhibitor (closantel) (35) in primary murine T cells augmented effector responses through enhancement of IL-2 signaling and the secretion of IFN-γ (34).

Even though the association between a novel STK39 variant and type 1 diabetes risk and involvement of Stk39 in murine T-cell activation and effector functions have been reported, the potential link between salt sensitivity in patients with type 1 diabetes and the functional impact of Stk39 activity on T-cell pathogenicity is missing. In this study, we aimed to determine whether there is a direct association between high salt intake and type 1 diabetes, taking into account the potential role of SPAK in T cells as a central player linking high-salt-intake influences to type 1 diabetes pathophysiology.

NOD/Sytwu (K^d, D^b, I-A^g7, I-E^null) and NOD.Rag1^−/− mice were purchased from The Jackson Laboratory. Stk39^−/− (SPAK KO) and Stk39^F/F mice with a C57BL/6 genetic background (36) were backcrossed to the NOD strain for at least 10 generations. NOD/Stk39^F/F mice were bred with the Lck^Cre mice to generate T-cell–specific SPAK knockout (SPAK CKO) mice.

Mice were monitored for spontaneous diabetes by measuring urine glucose concentrations twice a week. Mice with two consecutive urine glucose measurements >500 mg/dL were diagnosed with diabetes using Chemstrips (Roche, Boehringer Mannheim). For histology, pancreases from 12-week-old, female NOD mice fed either a normal-salt diet (NSD) or an HSD were fixed with 10% formalin, and paraffin-embedded samples were analyzed by staining with hematoxylin and eosin. The severity of insulitis was scored by two investigators blindly, as previously described (37).

Data are presented as mean ± SD. Statistical analysis for group differences was performed using the unpaired Student t test or a one-way ANOVA followed by Fisher least significant difference test, and Kaplan-Meier analysis was performed using the log-rank test (GraphPad Prism, version 8.00; GraphPad Software). A P value <0.05 was considered statistically significant.

All the study data are provided in this article and the Supplementary Material. Full methods are available in the Supplementary Material.

Intracellular ion homeostasis is essential for regulating cell volume to maintain necessary cellular functions. In response to perturbations in intracellular ion content or extracellular osmolality, mediators of ion transport, including NKCC1 and KCC3 cation–Cl⁻ cotransporters, significantly affect regulation of cell volume through homeostasis of the intracellular Cl concentration (23). In a previous report, researchers demonstrated that altered expression of Cl⁻ transporters with downregulated NKCC1 and upregulated KCC3/KCC4 in peripheral blood mononuclear cells from patients with type 1 diabetes compared with ND individuals, as a result of the effects of GABA_A receptors (38).

To better understand the association of salt sensitivity with CD3⁺ T cells from type 1 diabetes, we used this clinical cohort to compare the expression signature of molecules and ion transporters downstream of salt-responsive pathway (Fig. 1A) in a published RNA-sequencing data set of CD3⁺ T cells from ND participants and patients with type 1 diabetes (Gene Expression Omnibus series GSE111876). Analysis of cation-chloride cotransporter-related gene expression did not reveal any major changes in gene expressions of WNKs, master regulators of the electroneutral cation-Cl cotransporters, and the SLC12A family, including the Na⁺ coupled Cl⁻ importers (NCC and NKCC1) and the K⁺-coupled Cl⁻ exporters (KCC1, KCC3, and KCC4) in patients with type 1 diabetes (Fig. 1B). We also observed no significant differences in gene expression levels of the osmotic stress pathway, including NFAT5, SGK1, SLC5A3, and p38/MAPK in type 1 diabetes patients (Fig. 1C). Nevertheless, our data revealed an upregulation of SPAK in T cells from patients with type 1 diabetes, suggesting that a potential role of SPAK in T cells is associated with autoimmune diabetes for triggering homeostatic counter-responses due to perturbations of intracellular ion content.

It has been demonstrated that the increased extracellular sodium concentration due to excessive salt intake affects T-cell function (10,11). To explore the effect of NaCl on autoimmune diabetogenesis in vivo, we investigated the impact of high salt intake on the development of diabetes in NOD mice compared with a regular diet (NSD). After 6 weeks on an HSD, starting from age 6 weeks, we observed an accelerated diabetic kinetic and more severe insulitis in 12-week-old, HSD-fed mice compared with that of the control group, without weight loss in both groups (Fig. 2A–C). Nevertheless, the absolute total cell numbers and populations of CD4⁺ T, CD8⁺ T, and γδ T cells; Treg cells; B cells; macrophages; and dendritic cells in spleens, pancreatic lymph nodes (PLNs), and pancreas-infiltrating immune cells were similar between NSD- and HSD-fed mice (Fig. 2D). Even the activation status and proliferation of CD4⁺ and CD8⁺ T cells were comparable between these two groups (Fig. 2E). Moreover, populations of activated B cells, germinal center B cells, autoreactive B cells, and plasmablasts were also similar between two groups (Supplementary Fig. 1).

While CD4⁺ and CD8⁺ T cells play a crucial role in autoimmune diabetogenesis, we further dissected the salt-induced pathogenic role of CD4⁺ and CD8⁺ T cells by adoptive transfer. We cotransferred CD4⁺ cells, isolated from either NSD- or HSD-fed mice, in combination with CD8⁺ T cells, sorted from either NSD- or HSD-fed mice (Supplementary Fig. 2), into NOD.Rag1^−/− mice to monitor their diabetogenic potential to transfer disease. Cotransfer of CD4⁺ and CD8⁺ T cells from the HSD group accelerated diabetogenesis compared with the NSD group (Fig. 2F, left), supporting a salt-triggered T-cell–dependent pathogenicity. Of note, the salt-induced diabetogenicity was specifically confined to CD4⁺ T cells, whereas only cotransfer of HSD CD4⁺ T cells, instead of NSD CD4⁺ T cells, in combination with NSD CD8⁺ T cells accelerated diabetes progression (Fig. 2F, middle and right), suggesting that a salt-accelerated diabetic kinetic is initiated in a CD4⁺ T-cell–autonomous manner.

To identify signature molecules potentially contributing to salt-triggered distinct diabetogenicity of CD4⁺ T cells, CD4⁺ T cells were isolated from spleens of NSD- or HSD-fed mice, and gene expression was analyzed by bulk RNA sequencing (RNA-seq) to determine differentially expressed genes (DEGs) between these two CD4⁺ T-cell subsets. RNA-seq analysis revealed a distinct gene expression signature of top significant genes upon HSD treatment (Fig. 3A). Gene ontology (GO) analysis of the DEGs identified molecular functions that were enriched, including those associated with the structural constituent of ribosome, heat shock protein binding, and unfolded protein binding, reflecting the activation of heat shock response under salt stress whereby chaperone proteins prevent cellular stress response by protein folding and refolding (Fig. 3B). In line with this GO analysis, upregulated DEGs included those encoding the heat shock response–related proteins (namely, Hsph1, Hspa1b, Hsp90ab1, and Stip1). In addition, upregulated DEGs also included genes encoding regulators of osmotic stress pathway (Naft5, Sgk1, and Slc5a3) and mediators of ion transport important for intracellular ion homeostasis in response to osmotic stress (Wnk4, Stk39, and Slc12a4). In agreement with previous reports (39,– 41), the expression pattern of CD4⁺ T cells after HSD treatment was strongly indicative of a diabetogenic phenotype, with the upregulation of key proinflammatory Th1- and IL-21–associated genes (Ifng, Il21, Tbx21, Bcl6, and Maf) and downregulation of Th17-related cytokine genes (Il17a and Il17f) (Fig. 3C). Consistent with RNA-seq analysis, quantitative PCR analysis showed that the mRNA levels of Stk39, Sgk1, Ifng, Bcl6, and Maf were significantly increased and levels of Il17a were decreased in HSD-treated CD4⁺ T cells (Fig. 3D). Meanwhile, immunoblot analysis of CD4⁺ T cells from the HSD group confirmed there was only significantly increased protein expression of SGK1 and SPAK (Fig. 3E) but not expression levels of phosphorylation of p38/MAPK (Supplementary Fig. 3), suggesting that salt activates the SGK1- and SPAK-dependent pathway without being accompanied by activation of the p38/MAPK cascades.

To determine salt-triggered cytokine production by CD4⁺ T cells, we analyzed the frequency of cytokine-producing CD4⁺ T cells as well as Treg cells in peripheral lymphoid organs and the pancreases from NSD- and HSD-fed mice. Although an HSD has been reported to induce IFN-γ–secreting Treg cells with associated losses of function (12–14), we demonstrated here that HSD neither altered IFN-γ/IL-10 balance of Treg cells (Supplementary Fig. 4A) nor inhibited the suppressive capacity of Treg cells from HSD-fed NOD mice (Supplementary Fig. 4B and C). Importantly, only the IL-21–producing CD4⁺ T-cell population was increased preferentially in spleens from HSD-fed mice (Fig. 3F), indicating IL-21 is a signature cytokine of the salt-induced pathogenic phenotype in CD4⁺ T cells.

To examine a direct effect of increased NaCl intake on CD4⁺ T cells, which, therefore, would represent a risk factor for autoimmune diabetes, we investigated the effect of NaCl on in vitro–cultured CD4⁺ T cells from NOD, B6, and BALB/c mice. Notwithstanding increasing the extracellular salt concentration significantly increased Stk39 levels and decreased levels of Il17a and Il2, only observed in B6 CD4⁺ T cells (Supplementary Fig. 5A), stimulation under increased NaCl concentrations markedly induced expression of Il21 in NOD, B6, and BALB/c CD4⁺ T cells (Fig. 3G and Supplementary Fig. 5A and B), suggesting that a high-salt effect on Il21 expression appeared to be similar among mouse strains. Taken together, in contrast to the SGK-induced Th17-cell activation in dietary NaCl intake, our data show a downregulation of Il17a levels in vitro and in vivo, whereas there was no increase in IL-17A production in any of the NOD mice fed an HSD, implying that this NaCl-induced inflammatory program in NOD CD4⁺ T cells involves both SPAK and IL-21, rather than SGK-1 and IL-17A.

The Stk39 gene in mice located on chromosome 2 is not presented in the published insulin-dependent diabetes regions (42) and, to our knowledge, not previously associated with autoimmunity. To further determine whether SPAK is involved in the pathogenesis of autoimmune diabetes, we monitored the diabetes development in SPAK KO (Stk39^−/−) NOD mice. We observed no abnormality of T-cell development in SPAK KO mice (Supplementary Fig. 6) and confirmed the lack of SPAK protein in CD4⁺ T cells from SPAK KO mice by Western blotting (Fig. 4A). The incidence of spontaneous diabetes in female SPAK KO mice was similar to that in female wild-type NOD mice (Fig. 4B).

To further exclude any bystander effect of SPAK-deficient cells and study the function of SPAK specifically in T cells, we generated T-cell–specific SPAK knockout (CKO) NOD mice (Stk39^F/FLck^Cre) in which SPAK was deleted in T cells (Fig. 4C), enabling us to analyze the critical role of SPAK for T-cell pathogenicity during diabetes progression. SPAK CKO mice had significantly reduced diabetic incidence and severity of insulitis (Fig. 4D and E). Meanwhile, the IL-21–producing CD4⁺ T cell population from spleens of 12-week-old SPAK CKO mice was significantly reduced, whereas the population of IFN-γ–producing CD4⁺ T cells was unaffected compared with those from NOD mice (Fig. 4F). When we stimulated the naïve splenic SPAK CKO CD4⁺ T cells upon anti-CD3/CD28 activation for 48 h, the SPAK-deficient T cells had decreased expression levels of Il21 and upregulated Il17 levels compared with control T cells, whereas no notable changes were observed in mRNA levels of Ifng, Il2, Il4, and Il10 (Fig. 4G). Thus, our data indicate that SPAK in CD4⁺ T cells preferentially facilitates the production of the proinflammatory cytokine IL-21, contributing to diabetogenesis.

Considering SPAK has been reported to act as a mediator for sodium homeostasis and that T-cell–specific SPAK deficiency ameliorates the progression of autoimmune diabetes, we hypothesized that excess salt intake may affect the T-cell diabetogenicity through SPAK. To test this, we fed an NSD or HSD to control littermates (Stk39^F/F) and SPAK CKO NOD mice. After 6 weeks on an HSD, control mice had an earlier onset of diabetes and accelerated diabetic kinetics than NSD-fed littermates, and this acceleration of diabetogenesis was abolished in SPAK CKO mice (Fig. 5A), indicating that SPAK is involved in this HSD-initiated inflammatory pathway. Consistent with our data demonstrated above, we observed an increase in the IL-21–but not IFN-γ–producing CD4⁺ T-cell population from spleens of control mice fed an HSD compared with those from NSD-fed littermates. Of note, the frequency of IL-21–producing CD4⁺ T cells from control mice fed the HSD was significantly higher than in that from HSD-fed SPAK CKO mice, whereas there were no notable changes in the IFN-γ–, IL-10–, and IL-2–producing CD4⁺ T-cell populations (Fig. 5B). Moreover, in contrast to Il2, we detected augmented Il21 mRNA expression in the splenic CD4⁺ T cells of HSD-fed control mice compared with those from HSD-fed SPAK CKO mice (Fig. 5C).

Our data presented here indicate that high sodium intake potentiates IL-21–producing CD4⁺ T-cell generation in a SPAK-dependent manner to accelerate progression of autoimmune diabetes. To understand better the molecular role of high-salt–induced, SPAK-mediated IL-21 production in CD4⁺ T cells, we analyzed changes in gene expression of CD4⁺ T cells from HSD-fed control littermates or SPAK CKO NOD mice by RNA-seq analysis. In line with our data presented above, high NaCl concentration induced the expression of Stk39, Slc12a2 (encoding NKCC1), and Slc12a3 (encoding NCC) in control T cells. In addition, HSD-treated control CD4⁺ T cells had a proinflammatory phenotype with highly upregulated Il21 and Tnf expression, whereas key signatures of Th17 cells including Il17a, Il17f, and Rorc were downregulated. Notably, HSD treatment also upregulated Maf expression, a central player for inducing IL-21 production (43), in control T cells. Importantly, this high-salt-triggered, IL-21–predominant phenotype was attenuated in HSD-treated SPAK CKO T cells as proinflammatory Th1- and IL-21–related genes (Il21, Ifng, Tnf, Eomes, Tbx21, and Maf) were downregulated. Meanwhile, decreased Bcl6 and increased Prdm1 expression, with opposing effects on IL-21 production (44,45), in high-salt-treated SPAK-deficient T cells corelated with Il21 downregulation (Fig. 5D). Together, our results illustrate a critical role of SPAK in high-salt–induced, IL-21–driven diabetogenic phenotype of CD4⁺ T cells.

To investigate whether the pharmacological inhibition of SPAK abrogates IL-21 production, we analyzed the effect of closantel, a specific SPAK inhibitor (35,46), on cytokine responses in NOD CD4⁺ T cells in vitro. As expected, closantel treatment under increased concentrations markedly reduced expression of IL-21 in CD4⁺ T cells, as determined by quantitative PCR with reverse transcription (Fig. 6A) and by flow cytometry (Fig. 6B), but did not induce apoptosis (Supplementary Fig. 7). In contrast to IL-21 production, the frequency of IL-17–producing CD4⁺ T cells was increased upon closantel treatment, whereas the IFN-γ–, IL-4–, or IL-10–producing CD4⁺ T-cell populations were not affected (Fig. 6B)—findings that are in line with mRNA expression patterns displayed in SPAK-deficient CD4⁺ T cells (Fig. 4G). These results indicate that increased closantel concentrations specifically prevent the generation of an IL-21–dominant phenotype.

Given that in vitro administration of closantel inhibited IL-21 expression in CD4⁺ T cells, we monitored its in vivo potential to alleviate autoimmune diabetogenesis by treating NOD.Rag1^−/− mice with 10 mg/kg/day closantel every other day for 2 weeks starting from day 0 after adoptive cotransfer of purified CD4⁺ and CD8⁺ T cells. Of note, closantel treatment ameliorated diabetes development compared with mice that received control solvent (Fig. 6C). In summary, the data presented here indicate that pharmacological inhibition of SPAK has the potential to decrease the risk of salt-promoting autoimmunity as the HSD-induced, IL-21–predominant phenotype is dependent on SPAK.

Our study demonstrates for the first time, to our knowledge, that an HSD accelerates and aggravates diabetogenesis in NOD mice when compared with NSD group through a SPAK-dependent CD4⁺ T-cell–autonomous mechanism by which an IL-21–predominant phenotype potentiates T-cell pathogenicity. Most importantly, the pharmacological inhibition of SPAK and its specific effect on downregulation of HSD-triggered IL-21 seem to be promising therapeutic targets in excess salt intake–driven autoimmune diabetes progression.

Accumulating evidence has revealed that high salt levels can modulate the differentiation, activation, and function of T cells. NaCl induced Th17-cell differentiation via the SGK-1–Th17 axis in the pioneering work of Kleinewietfeld et al. (10) and Wu et al. (11) in humans and mice, respectively, contributing to aggravated neuropathology in the EAE model, because NaCl-triggered SGK-1 is critical for regulating IL-23R expression to stabilize the Th17 phenotype (11). On the contrary, high NaCl concentrations impair Treg functions by promoting a proinflammatory Th1 phenotype via SGK-1 in a xenogeneic graft-versus-host disease model and in adoptive transfer models of experimental colitis (12). Instead of IFN-γ expression, increased inflammatory Th17-like differentiation in murine Treg cells upon NaCl treatment was observed in a T-cell transfer-induced intestinal inflammation model (47), suggesting salt preconditions a proinflammatory conversion of Treg cells through either IFN-γ or IL-17 induction. Furthermore, Matthias et al. (48) showed that NaCl upregulates the Th2 signature cytokines IL-4 and IL-13 and abrogates IFN-γ production in human and murine cells, aligning with the Th2 bias for the pathogenesis of atopic dermatitis (48). Inconsistent with previous studies, our results reveal that a markedly reduced expression of Il17 was observed, whereas SGK-1 expression was upregulated, in HSD-fed NOD CD4⁺ T cells, implying a differential effect of NaCl on the proinflammatory cytokine production under the Th1 milieu independent of SGK-1. Moreover, even though our RNA-seq analysis of an increase in Il4 and Il13 mRNA levels in HSD-fed NOD CD4⁺ T cells is consistent with findings from a previous study that NaCl also acts as a potent inducer of Th2 cells (48), Th2 cells have only minimal effects on salt-triggered accelerated diabetogenesis in NOD mice. Together, the discrepancies of dietary effects of NaCl on Th-cell differentiation indicate that the cytokine context determines the overall outcome of NaCl effects on Th-cell effector functions in the context of a specific genetic background predisposing to autoimmunity, suggesting an involvement of more complex immune networks whereby salt exerts its effects on human autoimmune disease.

Previous work has established a central role of SPAK in the WNKs-OSR1/SPAK-CCCs cascade for NaCl homeostasis to regulate cell volume (23) and salt-sensitive hypertension (25). The importance of SPAK is also implicated in T cells in that its interaction with PKCθ is involved in the activation of AP-1 critical in the production of IL-2 and cell proliferation, but its physiological function in T cells has not been explored yet. In the present study, we illustrated a novel link between salt-regulated SPAK and T-cell pathogenicity through enhancing SPAK-mediated IL-21 expression in NOD mice, revealing a dual function of SPAK. Although a previous report demonstrated that pharmacological inhibition of Stk39 activity enhances the secretion of IFN-γ and IL-2 and CD25 expression level of murine T cells (34), our results indicate that SPAK-deficient NOD T cells exhibit a downregulation of an IL-21–producing T-cell population accompanied by similar expression levels of IFN-γ, IL-2, and CD25, implying a potent inhibitory effect of SPAK on T-cell pathogenicity in the context of autoimmune diabetes. In line with our results observed in SPAK CKO NOD mice, our in vitro experiments provide evidence that closantel treatment specifically suppresses IL-21 expression in NOD T cells but has no impact on IFN-γ and IL-2 secretion. However, the underlying mechanisms whereby SPAK facilitates IL-21 expression are not yet defined.

In our RNA-seq analysis, we observed a decrease in Maf expression in combination with an increase in Prdm1 expression, whereas an HSD-induced IL-21–predominant phenotype was abolished in HSD-fed SPAK CKO mice, implying that SPAK is involved in the activation of transcription factors critical in the regulation of IL-21 production (43–45). Of interest, unique β-cell antigen–responsive naïve CD4⁺ T cells, which were characterized by a high expression of pre–TH1/TH17/T follicular helper cell–like gene expression profile (CCR6, IL21, TBX21, TNF, RORC, EGR2, TGFB1, and ICOS), from children who developed β-cell autoimmunity were found in infancy, suggesting that the immune system is preprimed in infancy (49). In agreement with this report, HSD-triggered IL-21 phenotype in NOD CD4⁺ T cells is possibly a consequence of genetic or environmental priming leading to a latent immunosusceptible phase before emergence of autoreactive memory CD4⁺ T cells, inferring that the SPAK–IL-21 axis may be predictive of and preventive for future β-cell autoimmunity.

Collectively, our findings provide an explanation of how high dietary salt intake may regulate CD4⁺ T-cell diabetogenicity in type 1 diabetes pathophysiology, in particular, taking account of a higher sensitivity to salt and a predisposition to autoimmune diabetes, and point to additional possible applications of SPAK modulation between the crosstalk of cytokines with the excess salt milieu implicated for therapeutic strategies.

Acknowledgments. The authors thank the Laboratory Animal Center of National Defense Medical Center for animal care and Jian-Liang Chou and Yu-Chuan Huang (Instrument Center, National Defense Medical Center) for technical support and help in the analysis of the RNA-sequencing data.

Funding. This work was funded by Ministry of Science and Technology (grants MOST109-2320-B-400-018-MY3 and MOST110-2320-B-400-011-MY3), the National Science and Technology Council (grant NSTC112-2320-B-400-026-MY3), and Tri-Service General Hospital (grants TSGH-C02-112029 and VTA112-T-1-1).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. J.-J.C., S.-H.F., and H.-K.S. wrote the manuscript. M.-W.C., C.-Y.H., S.-H.F., and H.-K.S. conceived the hypothesis and designed the experiments. J.-J.C., M.-W.C., C.-Y.H., Y.-W.L., J.-L.D., S.-Y.T., and S.-H.F. conducted the experiments and performed the data analyses. S.-S.Y., S.-H.L., and B.L.-J.Y. provided material support and contributed to discussion. All authors contributed relevant text to the results and discussion and critically reviewed the manuscript. S.-H.F. and H.-K.S. are the guarantors of this work, as such, had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis.